Life Cycle Assessment and Environmental Efficiency of

Concrete Materials

by

Karina E. Seto

A thesis submitted in conformity with the requirements

for the degree of Master of Applied Science
Civil Engineering
University of Toronto

© Copyright by Karina E. Seto 2015

 ProQuest Number: 1605403

All rights reserved

INFORMATION TO ALL USERS

The quality of this reproduction is dependent upon the quality of the copy submitted.

In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if material had to be removed,
a note will indicate the deletion.

ProQuest 1605403

Published by ProQuest LLC (2015). Copyright of the Dissertation is held by the Author.

All rights reserved.

This work is protected against unauthorized copying under Title 17, United States Code
Microform Edition © ProQuest LLC.

ProQuest LLC.
789 East Eisenhower Parkway
P.O. Box 1346
Ann Arbor, MI 48106 - 1346
 ii

Life Cycle Assessment and Environmental Efficiency of Concrete

Materials

Karina E. Seto

Master of Applied Science

Civil Engineering
University of Toronto

2015

Abstract

Many materials have been proposed as ‘green’ alternatives to conventional concrete constituents.

A life cycle assessment (LCA) approach is adopted to model the processes included in the life

cycle of concrete. Five LCA software packages are evaluated to select a tool for use in this work.

The sensitivity of LCA results to the composition of Ontario’s electricity grid is analyzed.

Allocation of environmental impact to by-product materials, namely fly ash, is investigated. Six

functional units representing various levels of complexity and incorporating strength and

durability performance are developed. Finally, a methodology for calculating Environmental

Efficiency Indicators (EEIs) for concrete is presented. EEIs are calculated for eight mix designs

containing varying amounts of alternative concrete materials. The results show that

supplementary cementitious materials and limestone cement improve EEI. The EEI of

photocatalytic cement concrete is improved if SCMs are incorporated. Recycled aggregate

concrete has a lower EEI compared to conventional concrete.

 iii

Acknowledgments
I would like to thank Professor D. K. Panesar and Professor C. J. Churchill for their roles as
advisors and mentors. Professor Panesar, thank you for guiding me, challenging me, and
inspiring me during my time at the University of Toronto. Professor Churchill, thank you for
providing your support and insights throughout the past two years, as well as for being an
important part of my undergraduate experience at McMaster University. You have both
broadened my concept of what engineering can be, and what engineers can do.

I would like to gratefully acknowledge the contributions of the Ministry of Transportation of

Ontario (MTO), which supported this research through the MTO HIIFP (2013-2015) program. In
particular, I would like to thank Ms. H. Schell and Mr. D. Rhead of the MTO for providing
valuable insights throughout this research.

I am so thankful for the support of my family and friends. To my parents, thank you both for
gently setting such high standards through your own accomplishment, and for having such
unconditional belief in my ability to attain them. To my brother, Matthew, thank you for
supporting and encouraging me, and for being my link to home as we both started new lives in
the city. Naghmeh, thank you for the tea, lunches, and walks that made my graduate experience
so enjoyable. Finally, Binh – thank you for being by my side to encourage me and challenge me.
You continue to inspire me to do more and be more and I could not have done it without you.
 iv

Table of Contents
Acknowledgments.......................................................................................................................... iii

Table of Contents ........................................................................................................................... iv

List of Tables ................................................................................................................................. vi

List of Figures ............................................................................................................................... vii

List of Appendices ....................................................................................................................... viii

1 Introduction ................................................................................................................................ 1

2 Impetus and Scope of Research ................................................................................................. 2

3 Background ................................................................................................................................ 4

3.1 Alternative Concrete Constituents ...................................................................................... 4

3.2 Fundamentals of LCA ......................................................................................................... 5

3.2.1 ISO Standards and the Fundamentals of LCA ........................................................ 5

3.2.2 Stages of an LCA Study .......................................................................................... 6

3.2.3 Important Topics from LCA Literature .................................................................. 8

3.3 Literature on LCA of Concrete Materials ......................................................................... 14

4 LCA Methodology ................................................................................................................... 18

4.1 Goal and Scope Definition ................................................................................................ 18

4.1.1 Goal ....................................................................................................................... 18

4.1.2 Concrete Mix Designs........................................................................................... 18

4.1.3 Delineation of the System Boundary .................................................................... 19

4.1.4 Selection of Functional Unit ................................................................................. 20

4.2 Life Cycle Inventory ......................................................................................................... 22

4.2.1 Processes Included in the Life Cycle of Conventional Concrete .......................... 22

4.2.2 Processes Included in the Life Cycles of Alternative Concrete Constituents ....... 25

4.3 Life Cycle Impact Assessment.......................................................................................... 27

 v

4.3.1 LCIA Method ........................................................................................................ 27

4.3.2 Impact Categories ................................................................................................. 28

4.4 Interpretation ..................................................................................................................... 28

4.4.1 Green Indicator Methodology ............................................................................... 29

4.4.2 Environmental Efficiency Indicator Methodology ............................................... 30

5 Summary of Draft Manuscripts ................................................................................................ 31

5.1 Draft Manuscript 1: LCA software package selection for LCA of cement-based
materials ............................................................................................................................ 31

5.2 Draft Manuscript 2: Modelling of Ontario’s electricity grid for LCA of cement-based
materials ............................................................................................................................ 31

5.3 Draft Manuscript 3: Sensitivity of LCA of fly ash concrete to fly ash allocation
procedures ......................................................................................................................... 32

5.4 Draft Manuscript 4: Impact of the selection of functional unit on the life cycle
assessment of green concrete ............................................................................................ 33

5.5 Draft Manuscript 5: Life cycle assessment and development of environmental

efficiency indicators for alternative concrete constituents................................................ 33

6 Conclusions .............................................................................................................................. 35

7 Suggestions for Future Work ................................................................................................... 40

8 References ................................................................................................................................ 41
 vi

List of Tables
Table 1 Environmental and performance properties of alternative concrete constituents. ............. 4

Table 2 Key tasks to be completed during an LCA. ....................................................................... 7

Table 3 Examples of Data Quality Indicators (DQIs) (Coulon et al., 1997) .................................. 9

Table 4 Data quality pedigree matrix (reproduced from (Weidema & Wesnaes, 1996)) ............. 10

Table 5 Literature review of concrete LCI and LCA studies. ....................................................... 15

Table 6 Mix designs ...................................................................................................................... 19

Table 7 Hardened properties ......................................................................................................... 21

Table 8 Relevant factors for the calculation of transportation energy consumption and emissions
(Athena, 2005) .............................................................................................................................. 24

Table 9 Worst-case MTO transportation scenarios for concrete constituents .............................. 24

Table 10 Impact Categories .......................................................................................................... 28

 vii

List of Figures
Figure 1 LCI building methodology ............................................................................................. 12

Figure 2 System boundary for LCA of 100GU concrete .............................................................. 20

Figure 3 Ontario's electricity supply mix for 2014 (IESO, 2015)................................................. 23

 viii

List of Appendices
Appendix A Criteria for the evaluation of life cycle assessment software packages and life
cycle inventory data with application to concrete

Appendix B Modelling and sensitivity of Ontario’s electricity grid mix for the life cycle
assessment of cement-based materials

Appendix C Influence of fly ash allocation methodologies in life cycle assessment of cement-
based materials

Appendix D Impact of the selection of functional unit on the life cycle assessment of green
concrete

Appendix E Life cycle assessment and development of environmental efficiency indicators for
concrete constituents
 1

1 Introduction
This section summarizes the research presented in this thesis. The impetus and scope of
research are described in Section 2. Section 3 presents an overview of concepts that form
the basis for this work, including i) alternative concrete constituents, ii) the fundamentals
of life cycle assessment (LCA), and iii) previous LCA studies of cement and concrete
materials. In Section 4, the methodology used for LCA and for the development of
environmental efficiency indicators (EEIs) is presented. Five manuscripts are
summarized in Section 5 and are included in the appendices, including:

 Criteria for the evaluation of life cycle assessment software packages and life
cycle inventory data with application to concrete

 Modelling and sensitivity of Ontario’s electricity grid mix for the life cycle
assessment of cement-based materials

 Influence of fly ash allocation methodologies in life cycle assessment of cement-

based materials

 Impact of the selection of functional unit on the life cycle assessment of green
concrete

 Life cycle assessment and development of environmental efficiency indicators for

concrete constituents

Key conclusions are presented in Section 6, and potential future research is discussed in
Section 7.
 2

2 Impetus and Scope of Research

The production of cement-based products is responsible for a significant portion of global
CO2 emissions. In particular 6-7% of global CO2 emissions (Shi, Jimenez, & Palomo,
2011) can be attributed to the production of cement, which requires 3 GJ of energy per
ton of clinker (Van den Heede & De Belie, 2012) and is the component of concrete that
has the highest environmental impact. Over the past two decades efforts have been made
to ‘green’ the concrete industry, whether by improving process efficiencies, selecting
more sustainable materials, or developing more efficient designs. These efforts have
resulted in the proliferation of concrete mix designs that are claimed to be ‘greener’ or
more sustainable than traditional concrete mix designs with general use cement and
virgin aggregate.

Many of these ‘green’ concrete mix designs involve the replacement of a portion of
cement with a supplementary cementitious material (SCM). Commonly, SCMs are
recycled waste materials that are generated from various industrial processes such as slag
from the steel industry, fly ash from coal burning and silica fume from silicon production.
Incorporating SCMs into concrete has several potential environmental benefits, including
reducing the demand for virgin raw materials and providing a more functional end-of-life
for waste materials than landfilling (Siddique & Khan, 2011). General use limestone
cement has also been put forward as an alternative cementitious material to reduce
cement content and improve environmental performance. Concrete incorporating
recycled aggregate is also considered by many to be ‘sustainable’ as it incorporates a
waste material and reduces the demand for virgin aggregate. Photocatalytic concrete
containing titanium dioxide is said to clean the air of harmful pollutants by catalyzing
reactions.

Although alternative concrete constituents can potentially improve environmental

performance, it is important to assess these materials in a rigorous way over their entire
life cycle such that benefits and burdens are neither excluded nor exaggerated. Life cycle
assessment (LCA) enables analysis of all activities that occur during a product’s life
cycle, including raw materials extraction, transportation, production, use, maintenance
 3

and end of life. As the life cycles of concrete materials present a broad scope of potential
impacts, a LCA perspective is useful for assessing the environmental performance of
these materials.

As the field of LCA develops more and more, an increasing number of software packages
developed specifically for conducting LCAs have become available. These software
packages enable quick LCA calculations even when systems and databases are very
large, as they are for a typical LCA study. LCA practitioners must be able to evaluate
these products critically in order to determine which product is suitable for their study.
This research, therefore, required the development of a methodology for evaluating
software packages using preliminary LCA data and assumptions.

After the evaluation and selection process, LCA parameters are further refined to reflect
the context of Ministry of Transportation of Ontario (MTO) infrastructure in Ontario in
2015. Life cycle inventory (LCI) data is collected, assessed, and built into models that are
used to calculate the life cycle impacts of each type of concrete. The results of these
calculations are then interpreted based on the relative environmental performance of each
type of concrete, the functional requirements of concrete products, and the objectives and
standards of MTO through the development of Environmental Efficiency Indicators
(EEIs).
 4

3 Background
3.1 Alternative Concrete Constituents
Six alternative concrete constituents are included in this study: i) ground granulated blast
furnace slag (SL), ii) fly ash (FA), iii) silica fume (SF), iv) general use limestone cement
(GUL), v) recycled aggregate concrete (RAC), and vi) photocatalytic concrete (PCAT).
As the potential environmental impacts and benefits associated with these materials are
realized throughout the product life cycle, LCA is an appropriate tool to assess the
potential environmental impacts of using these materials. Detailed information about
each type of concrete can be found in the appendix to Draft Manuscript 1, included as
Appendix A to this thesis. Table 1 summarizes the environmental and performance
properties of the six alternative concrete constituents. The column ‘MTO Usage’ includes
comments based on discussions with MTO and a review of MTO standards.

Table 1 Environmental and performance properties of alternative concrete constituents.

Concrete Description Performance in Concrete MTO Usage*
Material
SL By-product of steel Strength: generally similar strength development to Most commonly used- 95%
industry PC concrete (Kosmatka et al., 2011) of all SCMs
Direct replacement Durability: generally improved compared to PC Sourced from Hamilton,
for cement concrete (Xu et al., 2008); however, potential for Sault Ste. Marie and
scaling is a concern in exposures with deicing salt Nanticoke
application, with elevated cement replacement levels MTO limits replacement to
Environmental: reduction in cement use and use of a 25% for structural concrete
waste material exposed to the environment
(i.e. where there are
concerns with scaling) and
higher levels (up to 50%)
are permitted for other
applications (e.g. for
drainage products, sulphate-
resistant foundations)
FA By-product of Strength: may improve strength at recommended Used in less than 5% of
coal-fired power replacement rates (Marsh, 2003) MTO projects
stations Durability: good durability and effective mitigation MTO limits replacement to
Direct replacement of alkali silica reaction (ASR) (Marsh, 2003); 10% of cementing materials
for cement however, potential for scaling is a concern in for environmentally
exposures with deicing salt application, with exposed conventional
elevated cement replacement levels concrete and 25% for high
Environmental: replaces cement content, repurposes performance concrete
waste; beneficial even with large transportation
distances (O'Brien et al., 2009)
 5

(cont’d) Table 1 Environmental and performance properties of alternative concrete constituents.

Concrete Description Performance in Concrete MTO Usage*
Material
SF By-product from Strength: improved as it is a very fine pozzolanic Used in less than 5% of
the manufacturer material (Khan & Siddique, 2011) MTO projects
of silicon or Durability: improved resistance to abrasion, freeze Typically used in the form
ferrosilicon thaw, chloride penetration, sulfate attack, and ASR of a pre-blended cement
Typically added in due to low permeability (Khan & Siddique, 2011) that contains up to 8% SF
blended cement Environmental: replaces cement content, repurposes and superplasticizer must be
waste, improves durability used

GUL Produced by i) Strength: if ground finer than cement, generally has Research and development
replacing some similar strength development (Tennis et al., 2011) projects are underway. In
clinker with Durability: sufficient durability with adequate SCMs the field, GUL concrete is
limestone, or ii) (CSA >40%) (Ramezanianpour & Hooton, 2013) used in less than 5% of all
blending cement MTO projects.
and limestone Environmental: replaces cement content
during batching
RAC Replacement of Strength/durability properties highly dependent on Research and development
virgin aggregate quality of aggregate, which is highly variable projects are underway. In
with construction depending on the source of the waste material the field, RAC is typically
and demolition 5-10% more cement recommended for strength made with aggregates from
waste development (Butler et al., 2013) crushed returned concrete
and is used as unshrinkable
Durability: performance uncertain (Marinkovic et fill.
al., 2010).
Environmental: replaces virgin aggregate,
repurposes waste products; however, processing can
be energy intensive; transportation is key variable
(Knoeri et al., 2013).
PCAT Titanium dioxide Strength: little effect on strength in typical Research and development
photocatalyst application (Dolatabadi, 2013) projects are underway. In
Typical Durability: adequate durability in typical application the field, PCAT use by
application: TiO2 (Hassan, 2010). MTO is limited to trials.
blended into Environmental: titanium dioxide catalyzes a reaction
cement (~5%), that converts air pollutants (ex. NOx, SOx) to less
mixed into damaging forms
concrete and
applied as surface
layer
* Comments on MTO usage are sourced from a December 4 2014 meeting with D. Rhead and H. Schell, as well as an
MTO document titled Specifier's Perspective- Supplementary Cementing Materials on MTO Contracts (Konecny,
2005)

3.2 Fundamentals of LCA

3.2.1 ISO Standards and the Fundamentals of LCA
Prior to the development of ISO standards for LCA in the late 1990s, LCA suffered from
a lack of consensus on its applications, how it should be conducted, and how results
could be considered valid (Marsmann, 1997). A critique by Ayres (1995) highlights the
main issues as being inconsistent and unverifiable data, a lack of common units and mass
 6

balances, and the discrediting of LCA because of its improper application. The ISO
14040 and 14044 standards provide for the standardization of LCA at the international
level, establishing principles and framework (ISO, 2006a) and requirements and
guidelines (ISO, 2006b) respectively. The ISO standards are widely accepted and present
a fairly straightforward framework for assessing the life cycle environmental impacts of
activities and processes. As described in Section 3.2.3, however, there are nonetheless
several important topics in the LCA literature that LCA practitioners continue to debate.

3.2.2 Stages of an LCA Study

ISO 14040 defines four stages of an LCA study as Goal and Scope Definition, Life Cycle
Inventory (LCI), Life Cycle Impact Assessment (LCIA), and Interpretation (ISO, 2006a).

Table 2 describes some key tasks that need to be completed in conducting an LCA.
Decisions made during these key tasks can significantly influence the results and must be
carefully considered and well documented.

The decisions made for the specific concrete LCA included in this study are documented
in Section 4, as well as in the attached draft manuscripts which form the appendices of
this thesis. Selected important topics from LCA literature are also presented in Section
3.2.3.
 7

Table 2 Key tasks to be completed during an LCA.

Stage Key Tasks
Goal and Define functional unit
Scope  Functional unit is the basis for all analysis and allows for comparisons to be made based on
Definition functional equivalency
 Should reflect the desired function of the product/process being studied
 There is limited research on functional units that reflect more than one function
Establish system boundary
 Determines which life cycle processes are included or excluded
 Often decisions are based on a predetermined cut-off value for estimated impacts (ex. if a
process contributes <1% it is not included). This approach is only possible if estimated
impacts are known from previous work.
 Can be based on geography, time, production stages, etc.
Life Cycle Develop a methodology for data collection
Inventory  Establish data quality objectives based on the objectives of the study
(LCI)
 Identify sources of data (literature, industry, government, etc.)
 Create methodology for managing poor quality or missing data
 Compile data for product, process or activity being analyzed
Foundational step for any assessment of environmental impacts
Every input or output of materials or energy included within the system boundary must be identified
and quantified
Collecting adequate high quality data is the primary objective of this stage
Life Cycle Overall objective: evaluate the potential human health and environmental impacts of the inputs and
Impact outputs identified in the inventory analysis
Assessment Define impact categories
(LCIA)
 Can include a range of environmental and health impacts such as global warming, water use,
and resource depletion
Choose impact assessment methods (problem-oriented vs. damage-oriented)
 Problem-oriented methods: environmental flows are simply grouped based on their
influence on certain environmental problems
o Reliable results with lower uncertainties (than damage-oriented)
o Lack information about cause-effect chain of environmental impact
 Damage-oriented methods: carry the calculations further down the chain to the endpoint or
actual damage
o Results are generally less reliable, as determining what and how much damage
will be caused by every environmental flow require assumptions and leads to
higher uncertainty
o Results are easy to interpret
Classify and characterize data
 Classification is the assigning of inputs and outputs to relevant impact categories
 Characterization factors are science-based conversion factors that allow different amounts of
various chemicals to be shown on an equal scale (SAIC, 2006).
Interpretation Identify significant issues
Evaluate results for completeness, sensitivity and consistency
May include peer review (required if results are asserted in a public forum)
 8

3.2.3 Important Topics from LCA Literature

3.2.3.1 Functional Unit

Establishing a functional unit is a critical preliminary step in any LCA. This unit is the
basis for all analysis and allows for interpretation of the results based on functionality,
and comparison between multiple products or processes based on functional equivalency.
The types of concrete included in this study are very diverse and might be used for a
variety of applications. The environmental impact of concrete could feasibly be measured
in terms of a volume or mass of concrete (i.e. functional unit is volume (1 m3) or mass (1
kg)). The results of this LCA, however, would not be as meaningful as if the impacts
were measured in terms of key functional characteristics such as strength or durability. A
more complex functional unit would allow for comparisons between diverse concretes,
based on this functional equivalency.

A review of the literature on cement and concrete LCAs (as compiled in Table 5)
revealed that despite reference to a ‘functional’ unit, many of these LCAs use units of
analysis that don’t correspond to a performance characteristic of concrete. Often, physical
characteristics such as mass or volume are used (as in Brown et al., 2014 and Lee & Park,
2005). This occurs most often in studies that use simple functional units (based on one
variable). For those studies that incorporate a performance characteristic, compressive
strength (typically 28-day) is the most consistently applied functional unit (as in Athena,
2005 and Prusinski et al., 2004). Finally, many functional units are tied to a specific type
of structural element and/or a specific environmental exposure (as in Kawai et al., 2005
and Knoeri et al., 2013).

Researchers have proposed various methods of calculating functional units that capture
more than one intended function. De Schepper et al. (2014) proposed the use of a
functional unit calculated as the volume of concrete required to achieve 1 MPa of
strength and 1 year of service life. This research is based on the work of Damineli et al.
(2010), who propose two indices that can be used as a functional unit: the binder intensity
 9

(binder materials (kg/m3) required divided by a performance requirement (e.g.

compressive strength)) and the carbon dioxide intensity (CO2 emissions divided by
performance requirement). The strength of these two approaches is the ability to compare
materials with different concrete mix designs and different efficiencies of production.

The influence of the choice of functional unit on LCA results and EEIs is explored
further in Section 4 and in Draft Manuscript 4, included in this thesis as Appendix D.

3.2.3.2 Data Collection Methodology and Data Quality

The creation of a Life Cycle Inventory (LCI) is a foundational step for any further
assessment of environmental impacts. Generally, every input or output of materials or
energy included within the system boundaries needs to be identified and quantified.
There are several factors that can influence the quality of an LCI, with the primary
concern being the quality of the data itself. Sources of data can be diverse, and may
originate from governments, academic institutions, industry organizations, or
corporations, amongst others. This data may represent a wide variety of geographical
regions or time periods. Some of this data may also be proprietary. The quality of the
LCI used for an LCA study can greatly influence the results, and data quality has both
quantitative and qualitative aspects as shown Table 3.

Table 3 Examples of Data Quality Indicators (DQIs) (Coulon et al., 1997)

Type of DQI Examples of DQIs
Quantitative Accuracy, bias, completeness, distribution, homogeneity, precision, uncertainty
Qualitative Accessibility, applicability, comparability, consistency, derived models, identification of anomalies,
peer review, representativeness, reproducibility, stability, transparency

The collection of high quality data is a particularly important and challenging stage of
LCA. There has been extensive research on how to define, measure, and improve data
quality. Data collection methods have been reviewed and three concepts were identified
as particularly important for this task:
1. Data Quality Metrics

Data quality is of critical importance to building an LCI, which is the foundation

of any LCA. Data quality is multidimensional and not necessarily quantitative.
Many LCAs and LCIs use the semi-quantitative matrix developed by Weidema &
 10

Wesnaes (1996) as shown in Table 4. Further research (De Barba Junior et al.,
2014) established data quality indicator (DQI) thresholds for excellent (<10) and
very good (<12) data quality, where the best score (cumulative of five indicators
where 1 is the strongest score and 5 is the weakest score) is 5 and the worst score
is 25. Note that ‘unknown’ conditions merit the worst scores. This highlights the
difficulty of working with datasets that are poorly documented.

Table 4 Data quality pedigree matrix (reproduced from (Weidema & Wesnaes, 1996))
Indicator 1 (Strongest) 2 3 4 5 (Weakest)
Reliability Verified data based Verified data Non-verified data Qualified estimate Non-qualified
on measurements partly based on partly based on (e.g. by industrial estimate
assumptions or assumptions expert)
non-verified
measured data
Completeness Representative Representative Representative Representative Representative-
data from a data from a smaller data from an data from a smaller ness unknown or
sufficient sample number of sites but adequate number # of sites/shorter incomplete data
of sites over an for adequate of sites but from periods or from a smaller # of
adequate period to periods shorter periods incomplete data sites and/or from
even out normal from adequate shorter periods
fluctuations number of sites
Temporal < 3 years < 6 years < 10 years < 15 years Age unknown or >
correlation difference to year difference difference difference 15 years difference
of study
Geographical Data from area Average data from Data from area Data from area Data from
correlation under study larger area in with similar with slightly unknown area or
which the study production similar production area with different
area is included conditions conditions production
conditions
Further Data from Data from Data from Data on related Data on related
technological enterprises, processes and processes and processes or processes or
correlation processes & materials under materials under materials but same materials but
materials under study but from study but from technology different
study different different technology
enterprises technology

2. Dominance
A review of the literature shows that within LCA, considering which processes
‘dominate’, or have the largest influence on the results, can be useful in building
an efficient LCI and assessing uncertainties. For example, Osses de Eicker et al.
(2010) studied the applicability of non-local LCI data for use in LCA. They
calculated the contribution of various processes to their impact assessment results,
using local and non-local LCIs. They argued that the consistent dominance of
some processes was one key indicator in determining that non-local LCI data can
be used in LCA. Weckenmann & Schwan (2001) suggest applying dominance to
 11

impact assessment results to determine which processes should be targeted for

improvement.

3. Data Fusion
A common issue in building an LCI is that databases are often incomplete.
Compounding this problem is the issue that often, environmental flows are not
recorded using a consistent methodology or with consistent units. At times, it
becomes necessary to combine several datasets. Gavankar & Suh (2014) have
proposed statistical methods for combining data in four different situations: two
data points, two un-nested intervals, multiple values from various sources, and
information as linguistic propositions. De Saxce et al. (2014) also suggest using
principal component analysis (PCA) and hierarchical clustering analysis in order
to reduce multicollinearity and group similar variables.

These data collection methods were applied to create a data collection methodology that
is specific to this study and its objectives. The data quality pedigree matrix shown as
Table 4 was applied to evaluate data quality and assess potential LCI datasets for this
research. Although data fusion methods are useful for filling in data ‘gaps’, they are time
consuming and may be difficult to scale up for large datasets. As a result, other
approaches will be used first before data fusion methods are considered (see Figure 1).
Drawing upon these concepts, a methodology for building an LCI has been developed
and is shown as Figure 1, where ‘decisions’ are shown as diamonds. This methodology
aims to build an LCI at the process level.
 12

Figure 1 LCI building methodology

3.2.3.3 Weighting and Single Indicators

Whenever multiple values are combined to create a single indicator value, some form of
weighting must be applied. There is no consensus on weighting in LCA, and there should
not be: the debate is valuable and critical to the interpretation of results. There is no way
to definitively determine that, for example, in all cases global warming potential should
trump all other forms of environmental impact. Similarly, there is no way to definitely
determine that all environmental impacts must be treated as equally important- which is
the underlying decision enacted when no weighting is applied. Rather, LCA practitioners
need to be aware of the context that forms the scope of their research, and the context in
which their results may be interpreted. As noted by Bengtsson & Steen (2000, p. 101),
“weighting is not meant to deliver the final verdict about the environmental performance
[…] it is meant to give an additional input into the process”. This quotation highlights
two important points about weighting in LCA: 1) weighting should not be used to
 13

selectively determine the ‘final verdict’, as this makes the results vulnerable to bias, and
2) testing the model using a variety of weighting schemes that favour one impact
category or another can provide valuable information about the sensitivity of the model,
and the robustness of the results. For example, if it is found that results do not change
significantly when many different weighting schemes are considered, then the results can
be said to be robust and independent of the priorities of different stakeholders. If results
do change significantly, then assessing various weighting schemes can provide insight for
various stakeholders about how the results may be interpreted from different
perspectives. In this work, the issue of weighting is addressed through sensitivity
analysis, as described in Section 4 and in the draft manuscripts that form the appendices.

3.2.3.4 Allocation
An interesting problem in LCA is the treatment of multifunctional processes; that is, what
is the appropriate way to model processes that produce more than one product? Coal
combustion is an example of a multifunctional process as it is a single process that
produces both electricity and FA. That is, for every kilogram of coal that is combusted, a
certain amount of both electricity and FA are produced, and these products cannot be
produced independently of one another when burning coal. The difficulty for LCA
practitioners is determining which environmental costs (from coal burning) are
attributable to the production of electricity, and which are attributable to the production
of FA, if any. This process of distributing environmental impacts is called ‘allocation’.
The results of this allocation can then be used to assess the environmental performance of
FA further in its life cycle- for example, when it is incorporated into concrete. To
perform an LCA of concrete containing FA, therefore, it is necessary to determine how to
resolve the problem of allocation for FA production. The sensitivity of the results of this
work to various allocation methodologies is described in Draft Manuscript 3, which is
Appendix C of this thesis.
 14

3.3 Literature on LCA of Concrete Materials

Known published concrete LCA studies have been reviewed, and are summarized in
Table 5. Some key outcomes of this literature review with regards to scope, functional
unit, and LCA results in previous published concrete LCA studies are presented:

Scope

 Often, the scope of concrete material LCAs is limited to ‘cradle-to-gate’ analysis

(i.e. as in Athena (2005), Huntzinger & Eatmon (2009) and Nisbet et al. (2002)).
This means that the scope is limited to the production of either cement or a
specific concrete product. Less commonly, LCAs may be ‘cradle-to-grave’, which
includes all life stages, or ‘cradle-to-cradle’, which includes the recycling of all
materials at the end of life of a product.
 There has not been a full LCA that has focused on the Canadian context, or more
specifically on the province of Ontario.
 Typically the upstream profiles of admixtures and supplementary cementing
materials (SCMs) are not included within the scope of concrete LCAs (Nisbet et
al. (2002), Damineli et al. (2010)).

Functional unit

 Many studies use physical characteristics of concrete such as volume or mass (ex.
1 m3 or 1 kg) as the functional unit (Athena (2005), Huntzinger & Eatmon
(2009)). This neglects important functional requirements for concrete, such as
strength and durability.

LCA results

 The energy required in these systems (both electricity grid and other fuels) is
often a key contributor to environmental impact (Huntzinger & Eatmon (2009).
 The production of cement (and in particular pyroprocessing) is typically the
process that contributes the most to the overall environmental impact of cement
and concrete products (Shi, Jimenez, & Palomo, 2011)
 15

Table 5 Literature review of concrete LCI and LCA studies.

Source LCA Scope System Boundary Functional Unit Geographical Alternative Concrete
Context Constituents
Anderson & Silman, LCI only Cradle to grave; no Steel or concrete building United States Recycled concrete
2009 EIO-LCA maintenance

Athena, 2005 LCI only Cradle to gate 1 m3 of 15, 20 or 30 MPa ready mixed concrete; precast Canada Fly ash, slag, silica
double ‘T’ beam, precast hollow deck, standard concrete fume
blocks, cement mortar
Brown et al., 2014 LCI only Cradle to gate for 1 tonne of cement Canada -
cement manufacture
Celik et al., 2015 LCA Cradle to gate 1m3 of ready mixed SCC concrete United States SCC with fly ash and
limestone
Chen et al., 2010a LCA Cradle to gate for 1 kg of cement Europe -
cement manufacture
Chen et al., 2010b LCA Cradle to gate for Binding equivalent value of fly ash, slag, and CEM 1 Europe Slag and fly ash
cement manufacture
Churchill & Panesar, LCI and Cradle to grave Noise barrier system Ontario Slag and
2013 costs only photocatalytic
Collins, 2010 LCA Cradle to cradle Concrete bridge (life of 100 years, then crushed and Recycled concrete,
recycled in bridge) slag, fly ash
Crossin, 2012 LCA Cradle to gate + end 1 m3 of AS 1379:2007 40 MPa concrete for 50 years Australia Proprietary blends;
of life (no use or slag, fly ash, recycled
maintenance) aggregate, silica fume
De Schepper et al., LCA Cradle to cradle Total amount of concrete necessary to deliver 1 MPa of Belgium Recycled aggregate,
2014 strength and one year of service life concrete reused for
cement production
 16

(cont’d) Table 5 Literature review of concrete LCI and LCA studies.

Source LCA Scope System Boundary Functional Unit Geographical Green Concrete Types
Context
Garcia-Segura et al., LCI for Cradle to grave (+ Reinforced concrete building column (3x0.3x0.3m) Spain Fly ash, slag
2014 GHG only post-demolition) during its lifetime
Hassan, 2009 LCA Cradle to gate for 5 mm thick surface mix of ultrafine TiO2 anatase, United States Photocatalytic
TiO2 (concrete not cement, filler, and water (0.1:1:3:0.4) applied on
included) pavement surface
Huntzinger & Eatmon, LCA Cradle to gate for 1 ton of cement x 4 types: traditional, blended (natural United States Natural pozzolans
2009 cement manufacture pozzolans), cement + sequestration in waste materials, (volcanic ash, rice
cement + recycled CKD husk ash, fly ash)
Jonsson et al., 1998 LCA Cradle to grave 1 m2 of floor area for seven frame cases (including in-situ Sweden/ Nordic -
cast and precast concrete frames) nations
Kawai et al., 2005 LCA Cradle to gate Prestressed concrete bridge, overflow dike of dam, Japan Slag
retaining wall, secondary lining in tunnel
Knoeri et al., 2013 LCA Cradle to gate 1 m3 of a specific strength class of concrete Switzerland Recycled concrete
Lean, indoor and outdoor structural concrete
Lee & Park, 2005 LCA Cradle to gate for 4 FUs: 1 kg of cement clinker, Portland cement, slag South Korea Slag
cement and slag powder, silicate fertilizer
powder
Li et al., 2015 LCA Cradle to gate for 2 FUs: 1 t of portland cement, 1 t of Portland cement with China and Japan -
cement 42.5 MPa of strength grade
Marinkovic et al., 2010 LCA Cradle to gate for 1 m3 of ready-mixed natural aggregate concrete and Serbia Recycled aggregate
concrete including recycled aggregate concrete concrete
transport
Nisbet et al., 2002 LCI only Cradle to gate 3 FUs: 1 m3 of ready mixed concrete, 100 concrete United States Fly ash, slag, silica
masonry units, 1 m3 of precast concrete fume
 17

(cont’d) Table 5 Literature review of concrete LCI and LCA studies.

Source LCA Scope System Boundary Functional Unit Geographical Green Concrete Types
Context
Prusinski et al., 2004 LCI only Cradle to gate 5 FUs: 20 and 35 MPa ready mixed concrete, 50 and 70 United States Slag
MPa precast concrete, concrete block mix
Tosic et al., 2015 LCA Cradle to gate 1 m3 of ready mixed natural and recycled aggregate Serbia Recycled aggregate
concrete
Wang et al., 2013 LCA Cradle to gate + 4047 m2 of permeable pavement (sized to store runoff Northeastern Permeable pavement
maintenance and associated with 2.5 cm of rainfall generated over 1 acre United States
stormwater runoff watershed)
and treatment Other FUs include other types of runoff systems designed
for the same capacity
Xing et al., 2008 LCI only Cradle to grave 1 m2 building area Shanghai, China -
Steel and concrete office structures
Yang et al., 2014 LCA Cradle to gate for Binder intensity and CO2 intensity concepts (Damineli et South Korea Slag, fly ash, silica
concrete al., 2010) in terms of the unit strength (1MPa) of concrete fume
 18

4 LCA Methodology
4.1 Goal and Scope Definition
4.1.1 Goal
The objective of this work is to use the principles of LCA to develop a methodology for
calculating EEIs for concrete mix designs that capture strength and durability performance as
well as environmental impact. This methodology is applied in a case study where eight mix
designs containing alternative concrete constituents are critically compared based on their
environmental and functional performance.

4.1.2 Concrete Mix Designs

The case study presented in this thesis includes concrete mix designs that are used by MTO.
MTO utilizes concrete that contains SCMs (SL, FA, and SF) and GUL, which replaces GU, in
transportation infrastructure. Data from the literature was also used to model two more novel
concrete constituents that are of interest to MTO and are currently the subject of trials: PCAT
(Dolatabadi, 2013) and RAC (Prasada Rao & Sindhu Desai, 2014). Eight mix designs are shown
in Table 6. The mix design in the first column, designated 100GU, is taken as the base case as it
is a conventional concrete mix design with GU, virgin aggregate, and no alternative concrete
constituents. There are two types of cement used in these mixtures: GU and GUL with a
limestone content of approximately 12%.
 19

Table 6 Mix designs

Mix Design ID
*j = - 1 2 3 4 5 6 7 8
Quantity 100GU* GU- GUL- GU-8SF- GU-10FA GUL- GU- GU-25SL- GU-
(kg/m3) * 25SL 25SL 25SL 10FA 5PCAT 5PCAT 100RAC
Water 180 155 150 140 128 147 180 180 283
GU Cement 429 300 - 276 316 - 407 305 565
GUL Cement - - 300 - - 319 - - -
SL - 53 54 100 - - - 107 -
FA - - - - 35 35 - - -
SF - - - 24 - - - - -
TiO2 - - - - - - 22 16 -
Coarse agg. 900 1064 1071 1000*** 1037 1036 900 900
Recycled - - - - - - - - 1955
coarse agg.
Fine agg. 875 779 781 800*** 777 771 869 876 966
Air entraining 52 - - - - - 52 52 -
admixture
(mL/kg
cement matl.)
* Index values (j) corresponding to Eq. 1 to Eq. 4
** Base case scenario (Dolatabadi, 2013)
*** Aggregate quantities are estimated based on typical MTO proportions.

4.1.3 Delineation of the System Boundary

Figure 2 presents the cradle-to-grave system boundary for 100GU, which is the base case mix
design for this work, and includes GU, fine and coarse aggregate, and water. The processes for
conventional concrete represent the baseline system boundary (i.e. the system boundary for each
of the other mix designs is created through modification to this baseline). Each process
represented in Figure 2 is comprised of input and output data such as energy use, raw material
use, and emissions to air, land, and water, that correspond to the activities in the concrete life
cycle described as follows. The Electricity Grid Mix process, which is shown in Figure 2 as
encompassing the other life stages due to its important role in many processes, is modelled based
on Ontario’s 2014 electricity grid mix and includes nuclear, hydroelectric, natural gas, and
alternative means of energy production. The Water Treatment process is the extraction and
processing of water. Cement Production is the extraction and transportation of raw materials, the
manufacturing of cement including blending, grinding, and pyroprocessing, and cement
transportation. Aggregate Production includes the extraction, processing, and transport of fine
 20

and coarse aggregates. Concrete Plant Operations includes batching and mixing activities, and
transport of concrete to the site where it will be placed. Service Life includes maintenance, and
for PCAT, pollutant abatement. End of Life is the in-place rubblizing of concrete. Further
information about each of these processes is provided in Section 4.2.1 of this thesis. This system
boundary is modified to model alternative concrete constituents as described in Section 4.2.2.

Figure 2 System boundary for LCA of 100GU concrete

4.1.4 Selection of Functional Unit

The purpose of the functional unit is to provide a basis for the quantification of all inputs and
outputs (i.e. a reference point for which data is collected), and to allow for comparison of LCA
results based on equivalent functional performance of different processes or products. Most
commonly concrete is specified based on strength and durability requirements that vary
depending on the application and on the type of materials included in the mix designs and the
geographic location (as it directly relates to exposure conditions). As MTO typically specifies
concrete used for transportation infrastructure according to compressive strength (in MPa) and
durability (typically measured using the rapid chloride permeability test (RCPT)) requirements,
these two properties are included in the functional unit. The experimental results for compressive
strength and RCPT corresponding to the mix designs shown in Table 6 are shown in Table 7.
 21

Table 7 Hardened properties

Mix Design ID
Hardened 100GU* GU- GUL- GU-8SF- GU-10FA GUL- GU- GU-25SL- GU-
Property 25SL 25SL 25SL 10FA 5PCAT 5PCAT 100RAC
28 day f’c 48.5 36.0 34.7 56.5 51.2 50.6 42.3 52.2 27.7
(MPa)
28 day 3186 2135 1969 421 3614 4024 5436 2002 5759
RCPT (C)
* Base case scenario

Equation 1 describes how the functional unit for this study is calculated.

Eq. 1

where j corresponds to each alternative mix design, as described in Table 6.

The strength value for the base material is on the denominator because, in general, a higher
compressive strength is a more desirable performance characteristic for structural concrete.
Concrete typically provides compressive strength capacity in structural systems and using higher
strength concrete has the potential to both improve functional performance and reduce the
amount of concrete needed to obtain the same performance (De Schepper et al., 2014).
Therefore, if a mix design containing alternative concrete constituents has a higher compressive
strength, this should raise the EEI. Under the same logic, for durability, the value for the base
material is on the numerator; this is because a higher RCPT value indicates a concrete that is
more permeable and therefore more vulnerable to the ingress of harmful chemicals such as
sulphates or chlorides. A higher RCPT value, therefore, represents a lower durability
performance for concrete. Therefore, if a mix design containing alternative concrete constituents
has a lower RCPT value compared to the base case, this should raise the EEI. As a result, the
functional unit, and the resulting EEI, has its highest value when the strength value is higher and
the RCPT value is lower. This methodology for calculating the functional unit corresponds with
the scenario that stronger and more durable concrete is better for the environment, also as
described in Van den Heede & De Belie (2012).
 22

4.2 Life Cycle Inventory

The Life Cycle Inventory (LCI) for each type of concrete, representing the inputs and outputs
throughout the product’s life cycle, was compiled based on a baseline LCI for conventional
concrete containing GU cement and virgin aggregate (represented by mix 100GU), which was
then modified to model each constituent. The data collection methodology described in Section
3.2.3.2 was applied. Based on preliminary LCA results (described in Draft Manuscript 1,
Appendix A), it was determined that dominant processes for LCA of concrete related to cement
production, electricity production, and transportation; as a result, achieving high geographical,
temporal, and technological correlation for these processes was a priority. Data fusion methods
were not required to build a high quality LCI for this study.

4.2.1 Processes Included in the Life Cycle of Conventional Concrete

4.2.1.1 Cement Production Process
A report released by the Athena Sustainable Materials Institute (2005) contains detailed LCI data
for the production of cement. The raw material, energy use, atmospheric emissions, liquid
effluent, and water usage data presented in this report was used to create the Cement Production
process in all of the LCA models. This dataset provides an overview of the Canadian Portland
cement industry, and is based on data from specific regions of the industry. The data for the
Central Region of Canada (Manitoba and Ontario) was used for this research, as it is most
relevant to the jurisdiction of the MTO.

4.2.1.2 Electricity Grid Mix Process

The electricity grid mix process is based on the Canadian electricity grid mix process from the
GaBi Extension database XVII: Full US, which includes biogas, biomass, hard coal, heavy fuel
oil, hydropower, natural gas, nuclear, photovoltaics, wind and waste-to-energy. To increase the
geographical and temporal correlation to the scope of this work, the ratio of the individual types
of generation was modified to reflect the supply mix of Ontario in 2014 (shown in Figure 3).
Draft Manuscript 3 describes the methodology used to evaluate alternate electricity supply mix
scenario (2010 and 2025 (predicted)).
 23

Figure 3 Ontario's electricity supply mix for 2014 (IESO, 2015).

4.2.1.3 Water Treatment Process

Racoviceanu et al. (2007) used the economic input-output life-cycle assessment model and the
GHGenius model to quantify the energy use and greenhouse gas emissions for the City of
Toronto municipal water treatment system. Several processes were considered including
chemical manufacturing, chemical transportation, and water treatment facility operation.
Although the scope of the study was somewhat limited, given that only energy use and
greenhouse gas emissions were considered, a review of the National Pollutant Release Inventory
(NPRI) has shown that there are no emissions recorded for the four Toronto water treatment
plants (which are: R.C. Harris, R.L. Clark, F.J. Horgan, and Island). This suggests that none of
the 363 NPRI substances (including NOx, SO2 or particulate matter) are released (Environment
Canada, 2013).

4.2.1.4 Aggregate Processes

There are four aggregate processes developed for the models used in this research: i) fine
aggregate production, ii) coarse aggregate production, iii) fine aggregate transportation, and iv)
coarse aggregate transportation. To build each of these processes, data for energy requirements,
raw material requirements, atmospheric emissions and liquid effluents were collected for the
extraction, processing, and transportation of the fine and coarse aggregates (Athena, 2005).
Based on a previous LCI (Athena, 2005), it is assumed that 1.18 MJ/t-km are consumed for road
transportation by diesel-powered truck. Emissions factors for road transportation by truck are
 24

taken from Athena (2005) and are shown in Table 8. Worst-case transportation scenarios for all
materials are shown in Table 9, where the worst-case transportation scenario corresponds to the
route that has the highest environmental impact. Transportation distances are from the MTO’s
Designated Source for Materials database (MTO, 2014) and discussions with MTO (D. Rhead &
H. Schell, personal communication, Aug. 14 2015). For this work, where all the transportation
processes use the same mode (truck), the route with the longest distance indicates the worst-case
transportation scenario.

Table 8 Relevant factors for the calculation of transportation energy consumption and emissions (Athena, 2005)
Energy Emissions Factors [kg/GJ]
CO2 SO2 NOx VOC CH4 CO
70.7 0.102 0.807 0.0869 0.0217 0.443

Table 9 Worst-case MTO transportation scenarios for concrete constituents

Material Route Total Distance (km) Transportation Mode
SL Hamilton to Picton, Picton to Toronto 498 Truck
FA Pleasant Prairie, WI to Toronto 937 Truck
SF Niagara Falls, NY to St Basile, QC, St Basile, QC to Toronto 1623 Truck
TiO2 Essexville, MI to Toronto 487 Truck
Coarse agg. Typical route 200 Truck
Fine agg. Typical route 200 Truck

4.2.1.5 Concrete Production Process

The concrete production process represents the stage in the life cycle where the various concrete
‘ingredients’, including water, GU, fine aggregate and coarse aggregate, are combined in order to
create concrete. The mix proportions are entered into the parameterized model according to the
mix designs described in Table 6. In addition to the raw material requirements, the batching and
mixing processes require energy; Prusinski et al. (2004) calculated a value of 247 MJ for the
production of a cubic metre of concrete (this value is consistent even with the addition of SCMs).

4.2.1.6 Concrete Maintenance Process

Concrete maintenance activities are highly dependent on the specific concrete application and
environmental conditions that the concrete is exposed to over its life. In terms of life cycle
assessment, however, they can be represented as additional quantities of materials and energy
that are required over the entire life of the concrete. As operational and cost considerations are
 25

not within the scope of this work, it is appropriate to combine all maintenance activities into a
single activity. In order to estimate these impacts, the concept of a Maintenance Factor (MF) was
used, where a MF value of 20% means that over the entire life of the concrete, 20% more of the
materials and energy required for initial placement must be input in order to maintain the
concrete. This value was selected in consultation with MTO (D. Rhead & H. Schell, personal
communication, Dec. 4 2014) and is applied to all of the mix designs included in this study.
Although detailed information about maintenance processes could inform MFs that are specific
to each given mix design, these data were unavailable for this work and therefore the 20%
approximation was applied.

4.2.1.7 End-of-Life Process

‘End of life’ is the final stage in a cradle-to-grave LCA. Potential disposal options for concrete
products include recycling, landfill, and reuse. Discussions with the MTO (D. Rhead & H.
Schell, personal communication, Dec. 4 2014) revealed that MTO infrastructure is typically
crushed in place for reuse as fill. This activity, as with all on-site activities, is assumed to be
powered by diesel. The energy required to crush one tonne of concrete is estimated as 34 MJ
(Wilburn & Goonan, 1998). The emissions resulting from the use of diesel were estimated using
the emission factors shown in Table 8.

4.2.2 Processes Included in the Life Cycles of Alternative Concrete

Constituents

4.2.2.1 Slag Processing and Transportation Processes

SL, a by-product of the steel industry, requires processing before it can be incorporated into
concrete. Energy and emissions for this processing are quantified by Marceau et al. (2006).
Upstream energy, raw materials, and emissions are not included in this process as SL is
generated regardless of whether it is incorporated into concrete, disposed of in a landfill, or any
other usage. Using SL in concrete is environmentally beneficial, and so not including any
upstream processes is conservative. This is an assumption commonly used in concrete LCAs
(CANMET & Radian Canada, 1993; Yang et al., 2014). SL transportation is modeled based on
SL sources and modes of transportation that are typical for the MTO, where emissions factors are
shown in Table 8 and the worst-case transportation scenario is shown in Table 9.
 26

4.2.2.2 Fly Ash Transportation Process

FA is a by-product of coal burning. It is a fine substance that does not require further processing
before it is incorporated into concrete. As with SL, upstream energy, raw materials, and
emissions are not included in this process, which is a common conservative assumption (this
assumption is explored further in Draft Manuscript 3, included in this thesis as Appendix C). FA
transportation is modeled based on FA sources and modes of transportation that are typical for
the MTO, where emissions factors are shown in Table 8 and the worst-case transportation
scenario is shown in Table 9.

4.2.2.3 Silica Fume Transportation Process

SF is a by-product of silicon production. It is a very fine substance that does not require further
processing before it is incorporated into concrete. As with SL and FA, upstream energy, raw
materials, and emissions are not included in this process, which is a common conservative
assumption. SF transportation is modeled based on SF sources and modes of transportation that
are typical for the MTO, where emissions factors are shown in Table 8 and the worst-case
transportation scenario is shown in Table 9.

4.2.2.4 Limestone Process

For GUL production, transportation of raw materials is not included. This is because cement
production facilities are typically located at limestone quarries, and so there is no transportation
requirement for these materials (CANMET & Radian Canada, 1993). As a result, the only added
process for GUL production is for the grinding of raw materials (Athena, 2005).

4.2.2.5 Titanium Dioxide and Photocatalytic Processes

Three processes are required to model PCAT: titanium dioxide production, titanium dioxide
transportation, and PCAT pollutant abatement. Titanium dioxide production was modeled using
data from Ortiz (2003). The production process modeled was the chlorine process for TiO2
production, which is more commonly used in North America and is therefore most appropriate
for this context. The assumptions made by Ortiz (2003) to build the portion of the LCI that
pertains to transportation, such as the transportation modes and distances, are based on Spanish
production, and are not appropriate for this research. As a result, a new transportation process
was created based on the emissions factors shown in Table 8 and the distance shown in Table 9.
 27

A PCAT NOx abatement rate of 6 mg/h/m2 was selected (Churchill & Panesar, 2013), and this is
included in the model during as part of the Service Life phase, where a service life of fifty years
is assumed.

4.2.2.6 Recycled Aggregate Process

To model RAC production, a crushing process is required. This process was created using the
same data as that used for the end-of-life rubblization stage described previously, the energy
required to crush 1 tonne of concrete is estimated as 34 MJ (Wilburn & Goonan, 1998) and it is
assumed that diesel is used to power the equipment. Note that this value is converted to calculate
the energy for crushing 1 m3 of concrete based on the density of each mix design. Emissions
factors for diesel burning are from Athena (2005).

4.2.2.7 Admixture Production Process

Data for air entraining admixtures has been gathered from an environmental product declaration
prepared by the European Federation of Concrete Admixtures Associations (2010). The
environmental product declaration includes the raw materials, emissions to air, emissions to soil,
and emissions to water.

4.3 Life Cycle Impact Assessment

4.3.1 LCIA Method
In order to compile the LCI data and conduct the impact assessment, a software package was
selected based on a rigorous comparison of five software packages. The methodology for and
results of this comparative study are presented in Draft Manuscript 1, which is included with this
thesis as Appendix A. The software package with the highest score was GaBi 6 (PE
International, 2014). The main advantage of GaBi 6 is that it allows for the user to have a high
level of control over the system being modelled and the calculation methods used.

GaBi 6 is packaged with several impact assessment methods, including the International
Reference Life Cycle Data System (ILCD) method which was used for this study (European
Commission Joint Research Centre, 2011). The ILCD method is essentially the result of
extensive evaluation and consultation which has resulted in a set of recommendations for the best
methodology to use within each impact category (PE International, 2011).
 28

4.3.2 Impact Categories

Four impact categories: i) Acidification, ii) Global Warming Potential, iii) Resource Depletion
and iv) Water Depletion- were selected for their analysis based on MTO’s assessment of priority
(H. Schell and D. Rhead, personal communication, December 4 2015) and the level of
international consensus on the classification and characterization that has been reached for each
category, as described in Stranddorf & Schmidt (2005). Table 10 presents a description of each
impact category as well as the characterization factor used in the analysis. In LCA,
characterization factors are science-based factors that are used to convert LCI data to actual
environmental impacts (Scientific Applications International Corporation, 2006).

Table 10 Impact Categories

*i= Impact Category Description Characterization Factor
1 Release of protons and leaching out of anions from a Acidification Potential
Acidification
system (mole H+ ion equivalents)
2 Effect of increasing temperature in the atmosphere Global Warming Potential
Global Warming due to the reflection of radiation by greenhouse gases
back to the surface (kg CO2 equivalents)

3 Depletion of abiotic non-water resources (ex. minerals Resource Depletion Potential (kg of
Resource Depletion
or fossil fuels) antimony (Sb) equivalents)
4 Water use (reuse, degradation, or incorporation into Water Depletion Potential (volume of
Water Depletion
final product) water used (m3))
* Index values (i) corresponding to Eq. 2 and 3.

4.4 Interpretation
This section describes the methodology used to interpret the LCA results by calculating two
indicators: 1) a Green Indicator, which uses weighting to combine results from multiple impact
categories into a single value; and 2) an EEI, which combines the Green Indicator (GI) with the
functional unit (from Eq. 1) to provide a measure of both environmental and technical
performance. For both of these indicators, the results for the mix designs containing the
alternative concrete constituents are calculated relative to the results for the 100GU mix design,
which are normalized to 1. This allows for a quick determination of whether the mix designs
containing alternative concrete constituents offer an improvement over conventional concrete
(EEI and GI >1) or whether they have a higher environmental impact for a given functional
performance (EEI and GI <1).
 29

In the Equations 2 to 4 that follow, the index ‘i’ refers to the four impact categories included in
this study (acidification, global warming potential, resource depletion, water depletion) as
described in Table 10. The index ‘j’ references each of the eight mix designs included in this
study as described in Table 6. These equations therefore present a methodology for calculating
GIs and EEIs that incorporate LCA results from all impact categories and are specific to each
mix design.

4.4.1 Green Indicator Methodology

The life cycle impact assessment results were factored by dividing the results for 100GU, the
base case mix design, by those for the alternative materials as shown in Eq. 2. The results for
100GU are placed on the numerator, because a higher LCIA result (ex. higher acidification,
higher global warming potential) is less desirable, and a higher value for the green indicator
corresponds to a reduced environmental impact.

Eq. 2

where i= 1 to 4 and corresponds to impact category

and j = 1 to 8 and corresponds to alternative materials

These normalized results were then combined into a single green indicator score that combines
the acidification, global warming, resource depletion, and water depletion potentials, as shown in
Eq. 3. An initial weighting of 0.25:0.25:0.25:0.25 was used to combine these categories.
Alternative weightings also tested to evaluate the sensitivity of the results, as discussed in Draft
Manuscripts 4 and 5, appended as Appendix D and E respectively. A higher green indicator
indicates a concrete with a higher environmental performance (and a lower environmental
impact).

Eq. 3

where i= 1 to 4 and corresponds to impact category

and j = 1 to 8 and corresponds to alternative materials
 30

4.4.2 Environmental Efficiency Indicator Methodology

The environmental efficiency indicator is calculated as shown in Eq. 4. For every alternative
concrete material (j), a specific EEI can be calculated by multiplying together the weighted result
and the functional unit.
Eq. 4

where j = 1 to 8 and corresponds to alternative materials

 31

5 Summary of Draft Manuscripts

5.1 Draft Manuscript 1: LCA software package selection for LCA
of cement-based materials
Life cycle assessment (LCA) is defined as “the compilation and evaluation of the inputs, outputs
and potential environmental impacts of a product system throughout its life cycle” (ISO, 2006a).
Since ISO 14040 was established, LCA has developed and grown and at the same time LCA
software packages have proliferated and evolved. There is now a multitude of LCA software
packages that must be critically evaluated by users. With such a broad scope of and timeframe
for potential impacts, a LCA perspective is useful for assessing the environmental performance
of various cementing material types in Canada. Prior to conducting a comparative LCA study on
different concrete materials it is necessary to examine a variety of software packages for this
specific focus. The paper evaluates five LCA tools in the context of the LCA of seven mix
designs (conventional concrete, concrete with fly ash, slag, silica fume or limestone as cement
replacement, recycled aggregate concrete, and photocatalytic concrete). An evaluation
methodology is presented which prevents duplication of efforts by simplifying the scope of an
LCA while enabling rigorous comparison of software packages. Three key evaluation criteria
required to assess the quality of analysis are: adequate flexibility, sophistication and complexity
of analysis, and usefulness of outputs. The quality of life cycle inventory (LCI) data included in
each software package is also assessed for its reliability, completeness, and correlation to the
scope of LCA of concrete products in Canada. A questionnaire is developed for evaluating LCA
software packages, and is applied to five LCA tools. The result is the selection of a software
package, which will be used to complete a full LCA study.

5.2 Draft Manuscript 2: Modelling of Ontario’s electricity grid for

LCA of cement-based materials
Ontario’s industries, including the cement and concrete industry, rely on Ontario’s electricity
supply mix and grid to power many fundamental processes. Ontario’s electricity grid mix is
complex, with nuclear, gas/oil, hydroelectric, wind, biofuel, and solar generation installed
throughout the province, and it is also dynamic. From 2010 to 2015, the Ontario energy
landscape has changed dramatically, partly as a result of i) the government mandate to shut down
all coal burning power plants, ii) an increase in nuclear power capacity, and iii) the growing
 32

importance of renewable energy technologies. Furthermore, the electricity supply mix is

projected to continue to evolve based on policy directives from the Ontario government. Since
industry is so reliant on the electricity grid, an accurate model of the relevant electricity grid is
crucial when evaluating the environmental impact of industrial processes through Life Cycle
Assessment (LCA). In this study, this model is built using a Top-Down Approach where data for
the Canadian electricity grid is adjusted to represent Ontario. LCA results for Portland cement
(PC) concrete, concrete containing fly ash (FA), and concrete containing Portland limestone
cement (PLC) are presented. The sensitivity of the results to changes in the composition of the
electricity supply mix is tested using models for the 2010 and 2025 (predicted) electricity supply
mix.

5.3 Draft Manuscript 3: Sensitivity of LCA of fly ash concrete to

fly ash allocation procedures
Supplementary cementing materials (SCMs) are commonly used to improve the environmental,
economic and in some cases technical performance of concrete products. One such SCM is fly
ash, which is produced as a by-product during the combustion of coal. Life cycle assessment
(LCA), which is defined as “the compilation and evaluation of the inputs, outputs and potential
environmental impacts of a product system throughout its life cycle” (ISO, 2006a), is a useful
way to model the complex processes that are included in the life cycle of a concrete product. An
interesting problem in LCA is the treatment of multifunctional processes; that is, processes
which produce more than one product. This is the case for the process of coal combustion, which
produces both electricity and fly ash. To perform an LCA of fly ash concrete, it is necessary to
determine how to resolve the problem of allocation for fly ash production. This paper presents a
critical comparison of four scenarios for the allocation of the environmental impact of fly ash
production in an LCA of concrete products; namely i) Baseline scenario- no allocation, ii) Mass
Allocation, iii) Economic Allocation, and iv) Disposal Avoidance. The Baseline scenario has no
allocation procedure applied. For the Mass Allocation scenario, the environmental inputs and
outputs associated with coal burning are allocated to both electricity and fly ash based on their
relative mass. For the Economic Allocation scenario, the environmental inputs and outputs
associated with coal burning are allocated to both electricity and fly ash based on their relative
economic value. For the Disposal Avoidance scenario, the environmental benefit of not
 33

landfilling fly ash is accounted for. The objective of this paper is to assess which of these
allocation scenarios is most appropriate for use in an LCA of fly ash concrete.

5.4 Draft Manuscript 4: Impact of the selection of functional unit

on the life cycle assessment of green concrete
An important step in defining the scope of an LCA is to select the functional unit for the
analysis. A functional unit “defines the quantification of the identified functions (performance
characteristics) of the product” (ISO, 2006a). The challenge for practitioners is how to select a
functional unit when a product or process has multiple identified functions, and ultimately how
to determine the influence of such a choice on the results. Despite the fundamental importance of
the functional unit in LCA studies, the ISO standards that relate to LCA (ISO 14040 and ISO
14041) provide little guidance on how to resolve these challenges.

A brief review of cement and concrete LCAs shows that a variety of functional units are used as
the basis for analysis including measurements of volume, mass, strength, and durability. The
importance of the selection of an appropriate functional unit in the context of LCA of concrete
materials is explored by examining the sensitivity of LCA results generated based on five
functional units i) volume of concrete, ii) 28-day compressive strength, iii) 28-day RCP, iv)
binder intensity; defined as amount of cementing materials (kg) per unit of 28-day compressive
strength (MPa), v) combination of compressive strength and RCP, and vi) combination of binder
intensity and RCP. A methodology is presented for calculating Environmental Efficiency
Indicators (EEIs) which combine LCIA results with measurements of performance in order to
compare various concrete materials. The objective of this paper is to assess the sensitivity of
LCA results to the selection of different functional units in the context of an LCA of concrete
containing GU cement, GUL cement, slag, and silica fume in Ontario.

5.5 Draft Manuscript 5: Life cycle assessment and development

of environmental efficiency indicators for alternative
concrete constituents
For the past several decades, one strong motivation for the development and implementation of
alternative concrete mix designs other than conventional concrete has been an effort to reduce
both cost and negative environmental impacts. These efforts have resulted in the proliferation of
concrete mix designs in the literature and in the industry that are claimed to be ‘greener’ or more
 34

sustainable than traditional concrete mix designs. This paper includes concrete mix designs that
have historically been used by the MTO, such as concrete containing slag, fly ash, silica fume,
and limestone cement. Data from the literature was also used to model two other more novel
alternative concrete constituents that are currently of interest to MTO and are the subject of
trials: photocatalytic concrete and recycled aggregate concrete.

Life cycle assessment (LCA), which is defined as “the compilation and evaluation of the inputs,
outputs and potential environmental impacts of a product system throughout its life cycle” (ISO,
2006a) is a useful way to environmentally model the complex processes that are included in the
life cycle of a concrete product, and rigorously assess different mix designs. At the same time, it
is critical that functional performance be included in any comparative analysis of concrete mix
designs. In this paper Environmental Efficiency Indicators (EEIs) are proposed as metrics for the
assessment of both environmental and functional performance. This paper presents a
methodology for LCA and the development of EEIs for alternative concrete constituents. The
objective of this paper is to critically compare six alternative concrete constituents based on their
environmental and functional performance.
 35

6 Conclusions
The Draft Manuscripts summarized in Section 5 yielded some important conclusions that are
relevant to researchers interested in concrete materials and/or in life cycle assessment. Many of
these conclusions informed the methodology presented in this thesis over the course of its
development.

The critical review of software packages in manuscript 1 revealed the following:

 It is critically important that LCA practitioners select a software package that is

appropriate for a specific research project.

 The ability to accurately model the chosen functional unit and system boundary is an
important criterion for the selection of a software package for use in LCA.

 By prioritizing the most important selection criteria, and making some preliminary
assumptions to simplify models, LCA practitioners can assess the advantages and
disadvantages of different tools without having to fully develop and optimize their
models in multiple software packages. This can enable a critical and rigorous comparison
without excessive and redundant duplication of efforts.

 The software package with the highest score is SP-C (GaBi 6), with 44 out of a possible
48 points. Its main advantage is that it allows for the user to have a high level of control
over the system being modelled and the calculation methods used.

The results of the sensitivity analysis with regards to the modelling of the electricity grid mix, as
presented in manuscript 2, showed that:

 For all three electricity models (2010, 2014, and 2025 (predicted)), PLC has the best
environmental performance of the three concrete mix designs evaluated, while
conventional concrete has the worst.

 The environmental impact of all three concrete materials, based on LCA results in four
impact categories (acidification, global warming potential, resource depletion, and water
 36

depletion), is lowest in 2014 and increases in 2025. This increase is primarily due to the
resources required to increase the proportion of photovoltaic energy in the supply mix.

 LCIA results are sensitive to changes in the composition of the electricity grid mix, even
when electricity is not the main product under study (i.e. electricity is one process in the
life cycle of concrete)

 Datasets must correlate well to the temporal, geographical, and technological scope of the
study.

 As a result, it was determined that the 2014 electricity grid mix has the highest
correlation to the scope of the study, and so this data will be used for the final LCA.
However, the model should be modified as Ontario’s energy profile continues to evolve.

The results of the sensitivity analysis with regards to the allocation of the environmental impacts
of fly ash production, as presented in manuscript 3, showed that:

 A review of the literature shows that allocation methods are not frequently applied when
fly ash is used in concrete because: i) there is lack of data, ii) fly ash is generated and
collected regardless of how it is used, and iii) the relative environmental impact of fly ash
processes are expected to be small compared to the overall concrete life cycle.

 Compared to the Baseline scenario, when the Disposal Avoidance Allocation scenario is
applied the change in environmental performance is very small (~0%). This aligns with
the common assumption that the relative environmental impact of fly ash processes is
expected to be small.

 Compared to the Baseline scenario, when the Economic Allocation and Mass Allocation
scenarios are applied the changes in environmental performance of the overall concrete
life cycle are more significant (~5% and ~10% respectively). This runs counter to the
common assumption that the relative environmental impact of fly ash processes in
concrete LCA is expected to be small.

 When different weightings of impact categories are considered, the Mass and Economic
Allocation scenarios are more sensitive than the other scenarios considered. Conversely,
 37

for the Baseline and Disposal Avoidance scenarios the results are generally stable even
when different weightings are applied. This indicates that the Mass and Economic
Allocation scenario leads to results that are sensitive to modelling decisions. This
suggests that LCA practitioners should assess the sensitivity of their results to alternative
weighting schemes when single indicators are used.

 Based on these results, and an assessment of the sensitivity of the results to changes in
allocation coefficients, it appears that no allocation scenario is more valid than the others
in application to an LCA of fly ash concrete. This is because all of the allocation
coefficients vary with changes in parameters that are highly variable, such as market
value of goods, exchange rates, raw material input quality, and production processes.
Therefore, the Baseline scenario will be selected for future work as this will reduce the
uncertainty of the results by eliminating the need for additional data sources and
modelling assumptions.

The results of the study of the impact of the selection of functional unit on LCA results, as
presented in manuscript 4, showed that:

 Varying the functional unit and using functional units that incorporate performance
characteristics can significantly influence the value of EEIs for each concrete mix design
relative to the value of EEIs with FU1 of 1 m3 (ex. the results for FU6 were 845% higher
than the results for FU1)). This is because of the significant variability in functional
performance for various mix designs depending on their constituents.

 Despite the variation, mix designs containing SL, GUL and SF have been shown to have
a consistently improved environmental efficiency over the base case (100GU) for all
functional units. This is because of the reduction in cement content due to cement
replacement, improvements in durability (all other mix designs compared to 100GU), and
improvement in strength (GU-8SF-25SL).

 In addition, the concrete mix design containing slag and silica fume (GU-8SF-25SL) has
been shown to generally have the highest environmental efficiency of the four mix
designs included in this analysis. This is likely because of its high functional performance
(highest compressive strength and lowest RCP). One exception is for FU1 (volume). This
 38

suggests that to accurately assess the environmental efficiency of concrete materials it is

important to select more sophisticated functional units, which incorporate performance
characteristics specific to the product under study (concrete) such as compressive
strength and durability as measured by RCP.

 The EEI results were assessed for their sensitivity to the choice of weighting scheme. It
was determined that the relative ranking of the materials in terms of environmental
performance did not change, even when weighting schemes were changed significantly.

Key findings from the LCA and EEI case study presented in manuscript 5 were that:

 The mix designs containing SCMs, limestone cement, or PCAT with cement replacement
(GU-25SL, GUL-25SL, GU-8SF-25SL, GU-10FA, GUL-10FA and GU-25SL-5PCAT)
have EEIs that are greater than 1 when normalized to the base case (100GU), indicating
an improved environmental efficiency compared to the base case. This is because in these
mix designs, there is significant (at least 10%) cement replacement. As cement is the
most environmentally intensive component of concrete, cement replacement is an
effective way to improve the life cycle environmental efficiency of concrete products.

 The mix design containing PCAT with no SCMs (GU-5PCAT) has an EEI that is less
than one when normalized to the base case (100GU), indicating a reduced environmental
efficiency compared to the base case. This is likely due to the relatively low functional
performance of these materials, and in particular the high RCPT values which indicate
poorer durability performance. In addition, there is little cement replacement in GU-
5PCAT, and PCAT cement is not a waste product so it does have upstream production
environmental impacts which are included in this analysis.

 The mix design containing recycled aggregate (GU-100RAC) has an EEI that is less than
one when normalized to the base case (100GU), indicating a reduced environmental
efficiency compared to the base case. This is likely due to the relatively low functional
performance of this material, and in particular the high RCPT value which indicate
poorer durability performance. Furthermore, a high replacement rate was used (100% of
coarse aggregate replaced by recycled aggregate) which reduced the functional
performance of concrete. This highlights the importance of incorporating functional
 39

performance into measures of environmental efficiency. In addition, there is no cement

replacement in GU-100RAC, and in fact recycled aggregate concrete typically requires
additional cement compared to an equivalent mix with virgin aggregate to achieve similar
strength performance (Butler et al., 2013).

 GU-25SL has an improved environmental efficiency (EEI>1) compared to 100GU, and

GU-25SL-5PCAT has an improved environmental efficiency compared to GU-5PCAT.
This suggests that SL improves environmental efficiency. Again, this is because cement
is environmentally intensive and its replacement can improve the environmental
efficiency of concrete mix designs.

 The only mix containing SF (GU-8SF-25SL) has an environmental efficiency far superior
to the other materials: EEI = 12.16 compared to the next highest value of EEI = 2.29
(GU-25SL-5PCAT) and the lowest value of EEI = 0.25 (GU-RA100). Although the
replacement rate is not as high for SF (8%) as for other SCMs, this material can
significantly influence strength and permeability and so this result is likely due to its high
strength and low permeability. In addition, the combined cement replacement rate (8%
SF and 25% SL) is the highest out of all the mix designs.

 Large changes in weighting generally have little influence on the EEI results (<2%),
except for mixtures containing PCAT, which unlike many of the other alternative
concrete constituents modelled, requires the additional extraction and processing of
virgin materials. These EEI results suggest that for mixtures containing PCAT, LCIA
results are more sensitive to modeling decisions and sensitivity analysis should be used to
explore the rigorousness of results and the implications for stakeholders and decision-
makers.

 The sensitivity of the results to the transportation distances was also explored by
generating results using best-case and worst-case transportation scenarios. Changes in
transportation distance up to approximately 1470 km have an overall influence on the
results for each given mix design that is <6%. This suggests that the transportation
processes do not have a large contribution to the overall LCIA results.
 40

7 Suggestions for Future Work

The LCA model and EEI methodology developed for this work can be applied to study the
relative environmental performance of diverse concrete mix designs.

 The model could be modified (by expanding the LCI) to include other alternative
concrete constituents as they are developed or as they become more commonly used by
MTO. This LCA model could also be expanded to include other types of concrete
products, such as pre-cast products. This would provide further insights into the
environmental efficiencies of these materials, while also increasing the scope and utility
of the model.

 The sensitivity of the results for other supplementary cementitious materials such as SL
and SF to the method of allocation should be assessed as data becomes available.

 Electricity and transportation LCI data should be updated as the energy climate in
Ontario and Canada continues to evolve over time. This will ensure that LCA results
generated by the model remain temporally appropriate for the scope of future studies.

 The LCI for photocatalytic concrete should be updated as more data becomes available
with regards to its ability to catalyze the reduction of non-NOx pollutants. As this is an
area of continued research, it is anticipated that future published results will inform a
more accurate LCI for PCAT, and subsequently more accurate LCA results.

 The results presented in this thesis could be combined with economic data, such as in a
life cycle costing (LCC) study which would incorporate the relative economic value of
the various alternative concrete constituents. This would provide additional insights for
the comparison of these materials and could result in the creation of an indicator that
incorporates environmental impact, functional performance, and cost.
 41

8 References

Anderson, J. E., & Silman, R. (2009, March). A life cycle inventory of structural engineering design strategies for

greenhouse gas reduction. Structural Engineering International, pp. 283-288.

Athena. (2005). Cement and Structural Concrete Products: Life Cycle Inventory Update #2. Ottawa: Athena

Sustainable Materials Institute.

Ayres, R. U. (1995). Life cycle analysis: A critique. Resources, Conservation, and Recycling, 14, 199-223.

Bengtsson, M., & Steen, B. (2000). Weighting in LCA- Approaches and applications. Environmental Progress,

19(2), 101-109.

Brown, D., Sadiq, R., & Hewage, K. (2014). An overview of air emission intensities and environmental performance

of grey cement manufacturing in Canada. Clean Technologies and Environmental Policy, 16, 1119-1131.

Butler, L., West, J., & Tighe, S. (2013). Effect of recycled concrete coarse aggregate from multiple sources on the

hardened properties of concrete with equivalent compressive strength. Construction and Building

Materials(47), 1292-1301.

CANMET & Radian Canada. (1993, October). Raw material balances, energy profiles and environmental unit factor

estimates: Cement and structural concrete products. Report. Ottawa, ON, Canada.

Celik, K., Meral, C., Gursel, A., Mehta, P., Horvath, A., & Monteiro, P. (2015). Mechanical properties, durability,

and life-cycle assessment of self-consolidating concrete mixtures made with blended portland cements

containing fly ash and limestone powder. Cement and Concrete Composites, 56, 59-72.

Chen, C., Habert, G., Bouzidi, Y., & Jullien, A. (2010a). Environmental impact of cement production: detail of the

different processes and cement plant variability evaluation. Journal of Cleaner Production, 18, 478-485.

Chen, C., Habert, G., Bouzidi, Y., Jullien, A., & Ventura, A. (2010b). LCA allocation procedure used as an

incitative method for waste recycling: An application to mineral additions in concrete. Resources,

Conservation and Recycling, 54, 1231-1240.

 42

Churchill, C., & Panesar, D. (2013). Life-cycle cost analysis of highway noise barriers designed with photocatalytic

cement. Structure and Infrastructure Engineering: Maintenance, Management, Life-Cycle Design and

Performance, 9(10), 983-998. doi:10.1080/15732479.2011.653574

Collins, F. (2010). Inclusion of carbonation during the life cycle of built and recycled concrete: influence on their

carbon footprint. International Journal of Life Cycle Assessment, 15, 549-556.

Coulon, R., Camobreco, V., Teulon, H., & Besnainou, J. (1997). Data quality and uncertainty in LCI. International

Journal of Life Cycle Assessment, 2(3), 178-182.

Crossin, E. (2012). Comparative life cycle assessment of concrete blends. Melbourne: Centre for Design RMIT

University.

Damineli, B. L., Kemeid, F. M., Aguiar, P. S., & John, V. M. (2010). Measuring the eco-efficiency of cement use.

Cement and Concrete Composites, 32, 555-562.

De Barba Junior, D. J., de Oliveira Gomes, J., & Bork, C. A. (2014). Reliability of the Sustainability Assessment.

21st CIRP Conference on Life Cycle Engineering, (pp. 361-366).

de Saxce, M., Rabenasolo, B., & Perwuelz, A. (2014). Assessment and improvement of the appropriateness of an

LCI data set on a system level- application to textile manufacturing. International Journal of Life Cycle

Assessment, 19, 950-961.

De Schepper, M., Van den Heede, P., Van Driessche, I., & De Beile, N. (2014). Life cycle assessment of completely

recyclable concrete. Materials(7), 6010-6027.

Dolatabadi, M. (2013). Properties and Performance of Photocatalytic Concrete. Toronto, ON, Canada: Library and

Archives Canada.

Environment Canada. (2013, July 11). Summary of National Pollutant Release Inventory Reporting Requirements.

Retrieved from Environment Canada- Pollution and Waste: https://www.ec.gc.ca/inrp-

npri/default.asp?lang=en&n=629573FE-1
 43

European Commission Joint Research Centre. (2011). International Reference Life Cycle Data System (ILCD)

Handbook. Instute for Environment and Sustainability. Luxemburg: Publications Office of the European

Union.

European Federation of Concrete Admixtures Associations. (2010). EFCA Publications. Retrieved January 10,

2015, from EFCA: http://www.efca.info/publications.html

Garcia-Segura, T., Yepes, V., & Alcala, J. (2014). Life cycle greenhouse gas emissions of blended cement concrete

including carbonation and durability. International Journal of Life Cycle Assessment(19), 3-12.

Gavankar, S., & Suh, S. (2014). Fusion of conflicting information for improving representativeness of data used in

LCAs. International Journal of Life Cycle Assessment, 18, 480-490.

Hassan, M. (2009). Life-cycle assessment of titanium dioxide coatings. Building a Sustainable Future –

Proceedings of the 2009 Construction Research Congress. 2, pp. 836-845. Seattle: American Society of

Civil Engineers.

Hassan, M. (2010). Evaluation of the durability of titanium dioxide photocatalyst coating for concrete pavement.

Journal of Construction and Building Material, 1456-1461.

Huntzinger, D., & Eatmon, T. (2009). A life-cycle assessment of Portland cement manufacturing: comparing the

traditional process with alternative technologies. Journal of Cleaner Production, 17, 668-675.

Independent Electricity System Operator (IESO). (2015). Supply Overview. Retrieved February 10, 2015, from

IESO: http://www.ieso.ca/Pages/Power-Data/Supply.aspx

ISO. (2006a). ISO 14040 International Standard. In: Environmental Management - Life Cycle Assessment -

Principles and Framework. Geneva, Switzerland: International Organisation for Standardisation.

ISO. (2006b). ISO 14044 International Standard. In: Environmental Management - Life Cycle Assessment -

Requirements and Guidelines. Geneva, Switzerland: International Organisation for Standardisation.

Jonsson, A., Bjorklund, T., & Tillman, A.-M. (1998). LCA of concrete and steel building frames. International

Journal of Life Cycle Assessment, 3(4), 216-224.

 44

Kawai, K., Sugiyama, T., Kobayashi, K., & Sano, S. (2005, October). Inventory data and case studies for

environmental performance evaluation of concrete structure construction. Journal of Advanced Concrete

Technology, 3(3), 435-456.

Khan, M. I., & Siddique, R. (2011). Utilization of silica fume in concrete: Review of durability properties.

Resources, Conservation and Recycling, 57, 30-35.

Knoeri, C., Sanye-Mengual, E., & Althaus, H.-J. (2013, June). Comparative LCA of recycled and conventional

concrete for structural applications. The International Journal of Life Cycle Assessment, 13(5), 909-918.

Konecny, J. (2005, March 11). Specifier's Perspective- Supplementary Cementing Materials on MTO Contracts.

CIRCA Collaborative Series. Toronto, ON, Canada: Ontario Ministry of Transportation. Retrieved January

8, 2015, from http://www.circainfo.ca/sem_050207.htm

Kosmatka, S. H., Kerkhoff, B., Hooton, R. D., & McGrath, R. J. (2011). Design and Control of Concrete Mixtures

(8th Edition ed.). Ottawa, ON: Cement Association of Canada.

Lee, K.-M., & Park, P.-J. (2005). Estimation of the environmental credit for the recycling of granulated blast furnace

slag based on LCA. Resources, Conservation and Recycling(44), 139-151.

Li, C., Cui, S., Nie, Z., Gong, X., Wang, Z., & Itsubo, N. (2015). The LCA of portland cement production in China.

International Journal of Life Cycle Assessment, 20, 117-127.

Marceau, M., Nisbet, M., & VanGeem, M. (2006). Life Cycle Inventory of Portland Cement Manufacture. Skokie,

IL, US: Portland Cement Association.

Marinkovic, S., Radonjanin, V., Malesev, M., & Ignjatovic, I. (2010). Comparative environmental assessment of

natural and recycled aggregate concrete. Waste Management, 30, 2255-2264.

Marsh, B. (2003, April). High-volume fly ash concrete. Concrete, 37(4), 54-55.

Marsmann, M. (1997). ISO 14040- The First Project. International Journal of Life Cycle Assessment, 2(3), 122-123.
 45

Ministry of Transportation of Ontario. (2014, December 1). Hydraulic Cements and Supplementary Cementing

Materials. Retrieved August 7, 2015, from Designated Sources for Material:

http://applications.roadauthority.com/mpl/mpl.asp?MPIShortName=MTO+DSM

Nisbet, M. A., Marceau, M. L., & VanGeem, M. G. (2002). Environmental Life Cycle Inventory of Portland Cement

Concrete. Portland Cement Association. Portland Cement Association.

O'Brien, K. R., Menache, J., & O'Moore, L. M. (2009). Impact of fly ash content and fly ash transportation distance

on embodied greenhouse gas emissions and water consumption in concrete. International Journal of Life

Cycle Assessment, 14, 621-629.

Ortiz, I. (2003). Life cycle assessment as a tool for green chemistry: application to kraft pulp industrial wastewater

treatment by different advanced oxidation processes. Barcelona: Institut de Cie`ncia i Tecnologia

Ambientals, The Universitat Auto` noma de Barcelona.

Osses de Eicker, M., Hischier, R., Kulay, L. A., Lehmann, M., Zah, R., & Hurni, H. (2010). The applicability of

non-local LCI data for LCA. Environmental Impact Assessment Review, 30, 192-199.

PE International. (2014). GaBi life cycle inventory data documentation. Retrieved October 10, 2014, from GaBi:

http://www.gabi-software.com/canada/support/gabi/gabi-database-2014-lci-documentation/

Prasada Rao, D., & Sindhu Desai, P. (2014). Experimental investigations of coarse aggregate recycled concrete.

International Journal of Advances in Engineering & Technology, 1522-1530.

Prusinski, J. R., Marceau, M. L., & VanGeem, M. G. (2004). Life Cycle Inventory of Slag Cement Concrete. Eighth

CANMET/ACI Eighth CANMET/ACI International Conference on Fly Ash, Silica Fume, Slag and Natural

Pozzolans in Concrete.

Racoviceanu, A., Karney, B., Kennedy, C., & Colombo, A. (2007, December). Life-Cycle Energy Use and

Greenhouse Gas Emissions Inventory for Water Treatment Systems. Journal of Infrastructure Systems,

13(4), 261-270.

Ramezanianpour, A. M., & Hooton, R. D. (2013). Sulfate resistance of Portland-limestone cements in combination

with supplementary cementitious materials. Materials and Structures, 46, 1061-1073.

 46

Scientific Applications International Corporation (SAIC). (2006, May). Life Cycle Assessment: Principles and

Practice. Retrieved September 20, 2013, from http://www.cs.ucsb.edu/~chong/290N-

W10/EPAonLCA2006.pdf

Shi, C., Jimenez, A. F., & Palomo, A. (2011). New cements for the 21st century: The pursuit of an alternative to

Portland cement. Cement and Concrete Research, 41, 750-763.

Siddique, R., & Khan, M. I. (2011). Supplementary Cementing Materials. New York: Springer.

Stranddorf, H., & Schmidt, A.L. (2005). Update on impact categories, normalisation and weighting in LCA. Danish

Ministry of the Environment Environmental Protection Agency.

Tennis, P., Thomas, M., & Weiss, W. (2011). State-of-the-Art Report on Use of Limestone in Cements at Levels of

up to 15%. Skokie, IL: Portland Cement Association.

Tosic, N., Marinkovic, S., Dasic, T., & Stanic, M. (2015). Multicriteria optimization of natural and recycled

aggregate concrete for structural use. Journal of Cleaner Production, 87, 766-776.

Van den Heede, P., & De Belie, N. (2012). Environmental impact and life cycle assessment (LCA) of traditional and

'green' concretes: Literature review and theoretical calculations. Cement & Concrete Composites, 34, 431-

442.

Wang, R., Eckelman, M., & Zimmerman, J. (2013). Consequential environmental and economic life cycle

assessment of green and gray stormwater infrastructures for combined sewer systems. Environmental

Science & Technology, 47, 11189-11198.

Weckenmann, A., & Schwan, A. (2001). Environmental Life Cycle Assessment with Support of Fuzzy-Sets.

International Journal of Life Cycle Assessment, 6(1), 13-18.

Weidema, B., & Wesnaes, M. (1996). Data quality management for life cycle inventories- an example of using data

quality indicators. Journal of Cleaner Production, 4(3-4), 167-174.

Wilburn, D., & Goonan, T. (1998). Aggregates from natural and recycled sources. US Geological Survey, US

Department of the Interior. Denver: US Geological Survey.

 47

Xing, S., Xu, Z., & Jun, G. (2008). Inventory analysis of LCA on steel- and concrete-construction office buildings.

Energy and Buildings(40), 1188-1193.

Xu, H., Provis, J. L., Deventer, J. S., & Krivenko, P. V. (2008, Mar/Apr). Characterization of aged slag concretes.

ACI Materials Journal, 105(2), 131-139.

Yang, K.-H., Jung, Y.-B., Cho, M.-S., & Tae, S.-U. (2014). Effect of supplementary cementitious materials on

reduction of CO2 emissions from concrete. Journal of Cleaner Production, 1-10.

doi:http://dx.doi.org/10.1016/j.jclepro.2014.03.018
APPENDIX A
CRITERIA FOR THE EVALUATION OF LIFE CYCLE ASSESSMENT

SOFTWARE PACKAGES AND LIFE CYCLE INVENTORY DATA WITH

APPLICATION TO CONCRETE

Karina E. Setoa, Daman K. Panesara*, Cameron J. Churchillb

a
Department of Civil Engineering, University of Toronto, 35 St. George St., Toronto, Ontario, Canada, M5S 1A4
b
Engineering and Society Program, McMaster University, 1280 Main St W, Hamilton , Ontario , Canada, L8S 4L7
*
Corresponding author. Email: d.panesar@utoronto.ca

Acknowledgement

The authors acknowledge the MTO HIIFP (2013-2014) program for support of this research. This research was

supported by a contribution from the Ministry of Transportation of Ontario. Opinions expressed in this report are

those of the authors and may not necessarily reflect the views and policies of the Ministry of Transportation of

Ontario.

48
ABSTRACT

Purpose: LCA software packages have proliferated and evolved as LCA has developed and grown. There is now a

multitude of LCA software packages that must be critically evaluated by users. Prior to conducting a comparative

LCA study on different concrete materials it is necessary to examine a variety of software packages for this specific

focus. The paper evaluates five LCA tools in the context of the LCA of seven mix designs (conventional concrete,

concrete with fly ash, slag, silica fume or limestone as cement replacement, recycled aggregate concrete, and

photocatalytic concrete). Methods: Three key evaluation criteria required to assess the quality of analysis are:

adequate flexibility, sophistication and complexity of analysis, and usefulness of outputs. The quality of life cycle

inventory (LCI) data included in each software package is also assessed for its reliability, completeness, and

correlation to the scope of LCA of concrete products in Canada. A questionnaire is developed for evaluating LCA

software packages, and is applied to five LCA tools. Results: The result is the selection of a software package for

the specific context of LCA of concrete products in Canada, which will be used to complete a full LCA study. The

software package with the highest score is Software Package C (SP-C), with 44 out of a possible 48 points. Its main

advantage is that it allows for the user to have a high level of control over the system being modelled and the

calculation methods used. Conclusions: This comparative study highlights the importance of selecting a software

package that is appropriate for a specific research project. The ability to accurately model the chosen functional unit

and system boundary is an important selection criterion. This study demonstrates a method to enable a critical and

rigorous comparison without excessive and redundant duplication of efforts.

KEYWORDS

Life cycle assessment, software, life cycle inventory, cement, concrete, questionnaire, evaluation criteria

1 INTRODUCTION

In today’s concrete construction industry there is a wide variety of types of concrete. For the past several decades,

one strong motivation for the development and implementation of concrete mix designs other than conventional

Portland cement concrete has been an effort to reduce cost and negative environmental impacts. Life cycle
49
assessment (LCA) is defined as “the compilation and evaluation of the inputs, outputs and potential environmental

impacts of a product system throughout its life cycle” (ISO 14040, 1997). The International Organization for

Standardization (ISO) has created a framework, ISO 14040, that serves as a guide to the four main stages of life

cycle assessment, namely, goal and scope definition, inventory analysis, impact assessment, and interpretation

(Scientific Applications International Corporation, 2006). The LCA approach enables analysis of all activities that

occur during a product’s life cycle, including raw materials extraction, transportation, production, use, maintenance

and end of life. With such a broad scope of potential impacts areas, an LCA perspective is regarded as a valuable

tool to assess the environmental performance of products, and processes.

To quantify the differences in environmental impact between various concrete mix designs, an LCA is a reliable

comparative approach. However, prior to conducting an LCA of a variety of concrete materials, it must be

recognized that as the field of LCA continues to develop, an increasing number of software packages designed

specifically for conducting LCAs have become available. Therefore, it is critical to evaluate these software packages

in order to determine the most appropriate product. This study proposes a methodology for evaluating five software

packages using preliminary LCA data of seven different concrete material types. The outcome of this evaluation

process is the selection of a software package to use for a full LCA study.

The objective of this paper is to conduct a critical and comparative review of five LCA software packages in context

with LCA evaluation of alternative concrete products, including concrete with slag, fly ash, silica fume or limestone

as cement replacement, concrete with recycled aggregates, and concrete with photocatalytic cement, in Canada.

2 BACKGROUND

2.1 Characteristics of LCA Software Packages

LCA software packages have evolved significantly in the last 20 years, from calculations using spreadsheets or

general mathematical modeling software, to highly functional applications developed specifically for LCA studies.

There is currently a multitude of products on the market, with different levels of functionality and specificity, and at

a variety of price points. Ciroth (2012) identifies four main characteristics of LCA software systems as platform,

pricing model, development model, and purpose, as shown in Table A1.

50
The platform used by a software package may influence both the user experience, and the ability to collaborate with

others. For example, a user may appreciate the convenience of being able to log in to their work from any device

with an internet browser, and so they may select a web tool. However, web tools cannot be used without an internet

connection, and they may be vulnerable to crashes or errors, making them potentially less reliable than desktop

tools. Also, while desktop tools store data on a local computer or network, for web tools the ability to export or save

data outside of the application may be critical for preserving results. For both types of platforms, the ability to export

data to different forms such as spreadsheets or text files may also be important for facilitating collaboration.

The pricing model of a software package can have ramifications beyond the financial expense incurred by the user.

More expensive commercial tools may offer value-added products and services, such as access to larger databases,

more powerful features, and more comprehensive technical support. Freeware tools have the immediate benefit of

easy and open access to powerful tools. Without the motivation of enticing or retaining paying customers, however,

freeware tools may be limited by smaller databases, less support, and less frequent updates leading to quicker

obsolescence.

Modern tools are generally developed using either an open- or closed-source model. In an open-source tool, a

special open-source license ensures that all users have access to the source code of the software, and can update and

modify the software. openLCA is an example of the open-source model. In closed-source models, these activities are

performed by the original creator of the software. SimaPro and GaBi are examples of the closed-source model. The

open-source development is dependent on the inputs from all of the collaborative users. As there is no single entity

in control of the final result, it could be argued that there is no accountability. Conversely, because there are many

experts using and contributing to the development of the tool, they share an interest in maintaining the integrity and

transparency of the product. Closed-source tools are more tightly controlled by whoever is creating or selling the

software. This means there may be more consistency and support for users, but there is more limited sharing of

information between experts and a slower development cycle.

The purpose of the software package is a key characteristic, as it defines its functionality and therefore its usefulness

for different LCA studies. A software package developed for general LCA is flexible and relevant for a project of

any size or scope. This is advantageous if a user intends to use the software package for many different types of

51
studies, or if a specialized software package is not available for a certain LCA topic. Specialized tools are often

developed to model a specific product or system (ex. buildings only). This creates a more streamlined user

experience, and these packages often include specialized databases that may be more useful than those found in a

general software package, since they are targeted towards a specific industry.

The capabilities of a software package cannot be fully assessed unless the evaluator actually uses the software.

Similarly, evaluation should be done within the context of a specific project, as the capabilities of the software

package must align with the parameters and objectives of the study. It is impractical, however, that a LCA

practitioner should fully develop models using multiple software packages solely for the purposes of comparison. In

this paper baseline LCA parameters and assumptions are established to perform a comparative study to evaluate five

software packages.

2.2 Cementing Materials

The potential environmental benefits of alternative concrete materials are diverse, including the reduction of cement

content, the incorporation of recycled or waste materials, and the oxidation of air pollutants. These environmental

benefits are realized at various points throughout the product life cycle. LCA is therefore an appropriate tool to study

the interplay between the strength, durability, and environmental impact of these materials throughout their life

cycle. The alternative concrete materials used in this study are ground granulated blast furnace slag (GGBFS), fly

ash (FA), silica fume (SF), Portland limestone cement (PLC), recycled aggregate concrete (RAC), and

photocatalytic cement (PCAT). Further information about each of these materials can be found in the appendix.

3 SCOPE OF LCA

Five software packages are included in this review, and they are identified as SP-A, SP-B, SP-C, SP-D and SP-E as

shown in Table A2. SP-A and SP-B, were developed for two areas of construction that use different concrete mixes-

highway and building elements- respectively. They are included in this review as they correlate well to Canada. SP-

D, which was partially developed in Canada, is also geographically relevant. SP-C and SP-E are included in this

review as they are two of the most widely used commercially available LCA software packages.

52
The five software packages will be assessed in the context of an LCA that aims to evaluate and compare the

environmental impact of conventional concrete with various concrete mix designs which are often described as

being more sustainable or ‘green’, including concrete with slag, fly ash, silica fume or limestone as cement

replacement, concrete with recycled aggregates, and concrete with photocatalytic cement.

The scope of any LCA study is defined by the selection of a functional unit and the establishment of the system

boundary. A functional unit is a parameter that enables analysis based on functional equivalency; 28-day strength of

a 1 m3 unit of concrete is a common functional requirement for concrete structural applications and is adopted

herein. The system boundary for the LCA, which summarizes the scope of the LCA, includes the entire life cycle

from raw material extraction, production, maintenance, use and end-of-life reuse.

Figure A1 presents the system boundary for conventional concrete, made of Portland cement, fine and coarse

aggregate, and water. The processes for conventional concrete represent the baseline system boundary. Each process

is comprised of input and output data such as energy use, raw material use, and emissions to air, land, and water,

that correspond to the activities described as follows. The Water Treatment process is the extraction and processing

of water. Cement Production is the extraction and transportation of raw materials, the manufacturing of cement

including blending, grinding, and pyroprocessing, and cement transportation. Aggregate Production includes the

extraction, processing, and transport of fine and coarse aggregates. Concrete Plant Operations includes batching and

mixing activities, and transport of concrete to the site where it will be placed. On-Site Activities includes placing and

curing concrete. Service Life includes maintenance activities such as repair. End of Life is the in-place rubblizing of

concrete.

To model GGBFS, FA, SF, PLC, RAC and PCAT concrete, modifications were made to this conventional concrete

system boundary. For this comparative study, several simplifying assumptions were made to streamline the study

and reduce the level of effort required at this stage, to facilitate the selection of a software package. Implementation

of these assumptions, however, is not anticipated to compromise the quality of the outcomes of this analysis.

 For GGBFS, a transportation process was added as well as an additional processing stage, as GGBFS

typically requires further grinding before it is incorporated into concrete (Prusinski, Marceau, & VanGeem,

53
 2004). The upstream production of GGBFS is not included in the analysis, as it is a waste product that

would be produced regardless of whether it is used in concrete. Using slag in concrete is a better

environmental choice than landfilling, so excluding these upstream activities is a conservative assumption.

 For FA, a transportation process was added. The upstream production of FA is not included in the analysis;

as with slag this is a conservative assumption.

 For SF, a transportation process was added. The upstream production of SF is not included in the analysis;

as with slag this is a conservative assumption.

 For PLC, Portland limestone cement production was substituted for cement production. Transportation of

raw materials is not included as cement production facilities are typically located at limestone quarries

(Canada Centre for Mineral & Energy Technology and Radian Canada Inc., 1993).

 For RAC, recycled aggregate production was substituted for aggregate production. Recycled aggregate was

assumed to be composed of crushed concrete that contains no foreign or deleterious substances (Butler,

West, & Tighe, 2013).

 For photocatalytic concrete, titanium dioxide production and transportation were added, and the Service

Life stage was modified to include pollutant abatement. Titanium dioxide was assumed to be 5% of cement

in a 5 mm layer (as in Churchill & Panesar, 2013)

 For all concrete types, chemical admixtures were excluded. This is a common assumption due to the low

relative contribution of these materials to the concrete environmental impact (Damineli, Kemeid, Aguiar, &

John, 2010; Nisbet, Marceau, & VanGeem, 2002).

 For all concrete types, maintenance was excluded. This is a common assumption due to the low relative

contribution of this life stage to the concrete environmental impact (Anderson & Silman, 2009).

The system boundary was designed to align with the flow of materials and energy through each system and present a

high level view of what activities are and are not included in the analysis. A critical factor to assess the adequacy of

a software package is the ability to accurately model the proposed system boundary. Data for all the processes in

Figure A1 must be available, or the software package must support the import of data from other sources.

54
4 METHODOLOGY

4.1 General Procedure for Evaluating Software Packages

The general procedure for evaluating software packages is illustrated in Figure A2.

The first task is the Selection of Evaluation Criteria. The most desirable criteria for the software packages are that

they exhibit adequate flexibility, sophisticated and complex analysis, and useful outputs.

 Flexibility: the ability to define custom functional units, system boundaries, and impact assessment

methods and the ability to import and modify LCI data; for this project a software package must have the

flexibility to model Canadian concrete processes

 Sophistication and complexity of analysis: the presence of relevant LCI data and powerful Life Cycle

Impact Assessment (LCIA) methods; the ability to model all life cycle stages; the ability to perform

sensitivity analysis; and the transparency of the processes used

 Output quality: the utility, clarity, and modifiability of the outputs; in particular, the ability to extract raw

quantitative output data

The evaluation criteria were refined in an iterative process as the software packages were used. The result is a

Software Package Evaluation Questionnaire, the completed version of which is shown in Figure A3. The

questionnaire provides a detailed assessment of the flexibility, complexity and output quality of the software

packages. This evaluation questionnaire is generally categorized according to the stages of a LCA as follows: 1)

Goal and Scope Definition, 2) Life Cycle Inventory Analysis, 3) Life Cycle Impact Assessment, 4) Interpretation,

and 5) General Usability. A numerical rating system was developed to compare the software packages. The rating

legend is shown in Table A3; scores range from 0 to 2 and with 24 questions the maximum possible score is 48.

The second task is the Assessment of LCI Data and LCIA Methods. In this phase, the databases and the impact

assessment methods that come packaged with the software packages are identified and assessed. The objective of

this stage is to identify whether software packages (i) have the LCI data required for this analysis, (ii) have a

selection of well-known impact assessment methods, and/or (iii) have the ability to add in LCI data or life cycle

55
impact assessment (LCIA) methods as required. The ability of the software packages to normalize results (for the

comparison of results with different units) and weight results (for the analysis of results in the context of specific

priorities) was also assessed.

The third task is the Modelling of Various Concrete Mixes. As the mix designs are modelled, observations about the

software packages, including their capabilities and the user experience, are recorded. This is shown in Figure A2 as

Updating of LCA Software Package Evaluation Questionnaire, which happens concurrently with the other tasks.

In the fourth task, Comparison of Results, the focus is not on comparing the numerical LCA results from the various

packages. Due to the limitations of some software packages, the same input data may not be used for all the software

packages. This occurs when the databases that are bundled with the software packages do not include the necessary

input data, and the software packages do not allow for the import of custom input data. As a result, the outputs of the

software packages are evaluated based on their clarity and modifiability, as described in Figure A3.

The fifth and final task of this process is the Selection of a Software Package with the highest score after the

application of the Software Package Evaluation Questionnaire.

4.2 Evaluation of Data Quality

Data quality is of critical importance to building a LCI, which is the foundation of any LCA. Data quality is

multidimensional and not necessarily quantitative. Many studies use the semi-quantitative matrix developed by

Weidema and Wesnaes (1996) as shown in Table A4, as it is able to capture this complexity and it is straightforward

for LCA practitioners to quickly and comprehensively assess potential datasets. Table A4 defines data quality based

on five key characteristics: reliability, completeness, temporal correlation, geographical correlation, and further

technological correlation. The first two characteristics speak to the rigour of the data collection, in terms of whether

the data is measured or estimated, the number of sites the data is sourced from, and whether or not the data is

verified. The latter three characteristics speak to the correlation between the scope of the LCA study being

considered and the scope of the dataset, and emphasizes that this correlation should be as high as possible. In this

data quality matrix, possible scores range from a strong score of 1 to a weak score of 5. With five key characteristics

under evaluation, this leads to a best possible total score of 5 and a worst possible total score of 25. Note that

56
‘unknown’ conditions merit the worst scores, corresponding to the right-most column in Table A4. For instance, if

datasets have an unknown age, and are based on unknown production conditions in an unknown area, it is very

difficult to assess their relevance for a specific LCA study. This highlights the difficulty of working with datasets

that are poorly documented, or non-local where production conditions may not be known.

For this research, the data quality matrix was applied to assess potential LCI datasets within each of the software

packages. In addition, research by De Barba Junior et al (2014) built upon the Weidema & Wesnaes (1996) data

quality pedigree matrix to establish data quality indicator (DQI) thresholds for excellent (<10) and very good (<12)

data quality. These thresholds were also used in this evaluation, where the objective was to only use data sets that

met the threshold of very good data quality (total score <12).

5 RESULTS

5.1 Data Quality Results

LCA studies require large quantities of data that are well correlated to the study context. The software packages

were assessed based on the presence of relevant data, the quality of available data, and the ability to import datasets

from other sources in order to compensate for missing or poor data.

Table A5 summarizes the availability of data relevant to the modeling of the six types of green concrete. A single

asterix (*) is used to indicate instances where data is not available with the software package, but the user can import

it. A double asterix (**) is used to indicate instances where the type of green concrete can be modelled, but only as

part of specific products or specific mix designs that cannot be modified by the user. As an example, concrete

containing silica fume can be modelled in SP-B, but only as ‘Lafarge Silica Fume Cement 4KSI’ which is a specific

product containing 10% silica fume and with a compressive strength of approximately 4 KSI. It is not possible to

modify this product or to import alternative mix designs.

The data quality pedigree matrix shown as Table A4 was then applied to the datasets available with each software

package. Table A6 and Table A7 present the data quality results for cement production and slag cement production

respectively. If more than one relevant dataset was assessed for a given software package then the highest scoring

57
dataset is presented. This matrix was also applied to all the other processes identified in the system boundary (Figure

A1). It was found that although all of the software packages scored relatively well for the cement production process

(a very common process), many did not score well for other processes, in particular the processes that are critical to

the modelling of green concrete materials (such as slag).

5.2 Evaluation Questionnaire Results

5.2.1 Flexibility

In the context of LCA software packages, flexibility refers to the extent to which the user can define and modify the

parameters of the study, including the system boundary, functional unit, and weighting of results. Flexibility also

refers to the extent to which the system can be expanded by adding data or impact assessment methods. These

characteristics of flexibility are often dependent on the purpose of the software package; general software packages

are inherently more flexible as they are designed to be appropriate for a range of LCA studies.

The flexibility of SP-A is somewhat limited by the fact that it is not general and it is developed specifically for

roadway applications; the roadway material database does, however, include concrete materials. The functional unit

is a section of road, and it is difficult to model any other functional unit besides volume of concrete. By selecting a

concrete type of a specific strength, and setting the lane length to 0.001 km, the thickness to 1000 mm and the width

to 1 m, a functional unit of 1 m3 of a specific strength can be approximated. The system boundary is similarly rigid,

although the user can select the lifespan, type of concrete, transportation distances for raw materials, and

construction and maintenance activities from pre-programmed options. In this study, it was not possible to create

models that reflected the system boundary shown in Figure A1 for all seven types of concrete using the functional

unit of 35 MPa, due to the limited number of pre-defined concrete products. Custom products can be created, but

only if they are composed of raw materials that are already pre-defined in SP-A’s database. Similarly, the

characteristics of the construction equipment in the library (ex. fuel consumption) can be modified, but no new

construction equipment can be input. It should be noted that the developer of SP-A does indicate on its website that

parties interested in adding a material or system to the databases should contact the developer directly. SP-A results

are not weighted or normalized by the software package.

58
SP-B is one of the least flexible of the five software packages considered in this study. Like SP-A, it is limited by

the fact that it was developed for a specific purpose, in this case building products, and so the database is appropriate

only for projects that align with that purpose. Additionally, it is an online web application that operates based on

drop-down lists of options that the user can select. This format means that users cannot input data or methods that

are not built in to SP-B. This is an issue for this research as SP-B does not already contain all of the processes

identified in Figure A1. Functional units are also pre-defined based on the building elements selected, and it may not

be directly clear to the user what functional unit applies to which building element until the user reviews the online

documentation that accompanies the web application due to a lack of transparency. Users also cannot modify the

system boundary at a process level- either an element is selected and included, or it is not. Default weightings can be

modified based on pre-defined or user-defined schemes. It is easy to run the software on any web browser; however

results cannot be saved by the program and users must print or screenshot their results.

SP-C is one of the most flexible of the five options considered in this study. The software is meant to enable a wide

variety of LCA as it allows for the user to define the system boundary by adding processes to the system, and to

indirectly select the functional unit through the specification of a mix design. Furthermore, the database that is

packaged with the software is extensive, and new processes can be easily added if required. Many impact

assessment methods are available for the user to select from, and users can also customize impact assessment

methods.

SP-D is moderately flexible as it allows users to build life cycle models that include or exclude whatever processes

they deem appropriate to define the system boundary. SP-D builds models that are based on a hierarchy of ‘schema’

(different stages or groups of processes), ‘flows’ (categories of substances, ex. construction materials, chemicals,

infrastructure) and ‘environmental factors’ (the actual materials, ex. gravel, sand, cement). For the modelling of

concrete, SP-D allows for the indirect selection of functional unit through the specification of a mix design. SP-D is

a web application and results are saved to the user’s account.

SP-E is one of the most flexible of the five options considered in this study. Similar to SP-C, the software is

designed for general LCA studies, and is not limited to a specific industry or set of materials. When a user selects a

process, it automatically links to other potentially relevant processes and includes them in the system boundary. This

59
can be an advantage in terms of ease of use in building a model, but it can also be a disadvantage if the user desires

to set the system boundary in a different way. In these cases, the system boundary can be modified manually.

Similar to SP-C and SP-D, the functional unit for concrete can be indirectly defined through the selection of a mix

design.

5.2.2 Complexity

In this study, complexity refers to the level of detail at which the LCA is done, and also the transparency of the

calculations to the user. It is critical that researchers be able to easily understand the software package calculations,

in order to accurately interpret the results.

SP-A contains LCI data that is suitable for road infrastructure (including the types of concrete being studied) and is

specific to Canada (subdivided into 9 Canadian regions); in the context of the data quality indicators discussed

above it generally has a moderate technological correlation and a high geographic correlation to the research project.

SP-A has a fairly detailed user interface that allows for the modelling of activities that occur over the infrastructure’s

lifespan, including maintenance activities, operating energy and pavement-vehicle interaction. There are two options

for the end-of-life stage, including demolition and landfill. There is one impact assessment method that was

developed for this tool, and it is based on the US EPA’s TRACI methodology. The results cannot be weighted

directly. The user interface is somewhat difficult to navigate, and the calculation process is not very transparent.

SP-B has a database built specifically for this tool, and as such it consists of only building construction and

maintenance materials. The data is collected from the industry in the United States, and so it has a moderate level of

geographical correlation and a fairly high level of technological correlation given similarities between the industries

in Canada and the United States (Marceau, Nisbet, & VanGeem, 2006, p. 28). One potential issue with using SP-B

for Canadian applications is that the electricity profiles of different regions of the two countries may be very

different, and as electricity is typically another major source of environmental impact after cement production, this

could significantly affect the results. In the calculation of environmental impact, the weightings can be modified.

The drop-down structure of the tool, while it has the limitations mentioned above, also creates a highly intuitive user

interface. The calculations, however, are not transparent as it is unclear how the results are generated. In addition,

60
during the preparation of the comparative study the web applications experienced errors on several occasions that

required the user to restart work.

SP-C has a large database with multiple additional databases specific to certain industries or regions, which can be

purchased. In addition, the user can also import or add data to their system. The types of concrete being included in

this study are all contained in these databases. SP-C contains most major impact assessment methods, and also

allows users to input their own methods. The software can be used to create parameterized models, perform

sensitivity analysis and compare alternatives. The data entry is intuitive as it consists of building a network diagram

of processes, with logical connections and movements of materials and energy between them. A strength of SP-C is

the transparency of the tool, which is due to the extensive and easily available documentation of all activities, and

the high level of control that users can exercise over the details of their projects. SP-C also provides the most useful

support for users of the software, in the form of a detailed online Learning Centre, and live chat with experts online.

SP-D is packaged with the ecoinvent database version 2.2 (ecoinvent, 2010) which is a large and commonly used

database that is also found in SP-E. It should be noted that the free version of the software has a more limited

database than the full version. With the full version, additional processes can be added by the user and the full

ecoinvent database is included, but no other databases are added. There is more data for the use and disposal life

cycle stages included with the software than many of the other software packages (except SP-E, which includes the

same database). Like SP-A and SP-B, SP-D has only one impact assessment method- in this case, IMPACT 2002+

(Jolliet et al., 2003), a common method that is also included in SP-C and SP-E.

SP-E has a large database that sources data from many different industries and regions, but in general the dataset is

heavily European and there is not much North American data. However, like SP-C, SP-E allows users to import

data. This is a powerful advantage over the other software packages as it allows users to compensate for poor quality

or ‘missing’ data. SP-E has 22 different impact assessment methods available for the user to select from, including

most major and commonly used methods. This is the highest number of any software package studied here. The

software is highly functional, allowing users to build parameterized models and conduct Monte Carlo analyses.

61
5.2.3 Outputs

Comparative study modeling of the concrete types was done for the purpose of testing the software packages and

exploring potential outputs. It should be noted that baseline material inputs have been used with the sole purpose to

compare software packages in context with concrete materials, and not to compare between the material mix designs

themselves. These have not been optimized or evaluated for sensitivity. Due to the limitations of some of the

software packages, it was not possible to ensure consistent LCI data and functional units.

Calculations were conducted for seven types of concrete within the scope of this research in order to ensure that they

could be modelled by each software package; an assortment of output results are shown as samples in Figures 4 to 8.

As noted in Section 5.2.2, not all processes could be modelled in all the software packages due to the limitations of

the software. As a result, although the intent was to model 35 MPa concrete with 20% fly ash replacement in all

software packages, there were two exceptions: (i) in SP-A, the most similar available product was 30 MPa concrete

with 25% fly ash replacement, and (ii) in SP-D, fly ash could not be modelled and so 35 MPa concrete with no

substitution was modelled.

In addition to examining raw graphical outputs, for the purposes of research and the presentation of results, it is also

critical to determine whether the numerical raw outputs are available. Free access to this data is necessary for

processing of the data including normalization, weighting and sensitivity analysis; these are all activities that

commonly occur in LCA.

SP-A produces very simple results that can be aggregated or separated by impact category, shown in Figure A4. A

somewhat confusing aspect of the outputs from this software is that categories that were not included in this

comparative study (ex. maintenance, pavement vehicle interaction) are still shown on the y-axis of the graph, giving

the impression that they might have a ‘0’ rather than ‘N/A’ value. Quantitative data tables are also available and can

be extracted from the software, either aggregated or separated by impact category. Results can be exported to

Microsoft Excel, Microsoft Word, or PDF.

SP-B produces simple and clear graphs that can be displayed in aggregated form or by impact category, as shown in

Figure A5. Results can be shown divided by life stage, by environmental flow, or by embodied energy. A strength of

62
SP-B is that it allows for easy comparison of multiple products and displays these comparative results in a variety of

graphs. Results from SP-B are simultaneously displayed in the web interface as graphs and as quantitative data

tables. The weakness of SP-B is that as it is purely a web application, results cannot be saved easily. Graphs and

tables must be printed or copied for future reference. Graph parameters, such as the axis titles, labels, legend, etc.

cannot be modified. The graphs can be saved as images or copied into another program.

SP-C produces graphs that are separated by impact assessment method and impact category. The available graphs

show the breakdown of each impact category by each process modelled; Figure A6 shows the results for climate

change (global warming potential). The user can modify all of the parameters of the graph, including titles, axes,

colours etc.. Quantitative data tables can also be organized in several different ways. Users can select the exact

parameter that they want to study in detail, and can organize the results based on absolute value or relative

contribution. The software can also perform a weak point analysis, identifying processes and flows with a high

relative contribution to the environmental impact of the system. The graphs and tables can be easily exported or

copied into another program.

SP-D produces a graph displaying the contribution of each life cycle stage to the value of each impact category

studied (ex. climate change). If more than one impact category is selected, values are converted to percentage so that

results with different units can be compared. If only one impact category is selected, as in Figure A7, results are

displayed in their original units. Outputs are organized at the level of life stage and further detail is not available

graphically. Similarly, although a table showing the final value calculated for each category can be copied or

exported to Microsoft Excel, a detailed breakdown of the raw data (ex. quantities for every output from the system)

cannot be extracted and so these outputs are fairly limited. The parameter of the graphs, such as the axis titles,

labels, legend, etc. cannot be modified. Graphs can be saved directly from the program as an image file.

SP-E produces graphs that display the data by impact category, and show results for individual products or a

comparison of several products. Results are displayed as relative percentage impact unless the data is manipulated

further, which is limited but allows for the comparison of multiple impact categories with different units at the same

63
time. Unlike SP-D, if only one impact category is selected graphs will still display results as percentages, as shown

in Figure A8. Quantitative data tables, however, can be easily extracted giving the users the flexibility to interpret

the outputs as they deem appropriate.

6 CONCLUSIONS

This comparative study highlights the importance of selecting a software package that is appropriate for a specific

research project. Note that the software packages reviewed for this study represent a sample of available products,

and other products should be evaluated for other applications. The ability to accurately model the chosen functional

unit and system boundary is an important selection criterion. Three key criteria are defined and explored: adequate

flexibility, sophisticated and complex analysis, and useful outputs.

This study demonstrates a method to select a software package while reducing the level of effort required at the

preliminary stage of a LCA. By prioritizing the most important selection criteria, and making some preliminary

assumptions to simplify models, LCA practitioners can assess the advantages and disadvantages of different tools

without having to fully develop and optimize their models in multiple software packages. This can enable a critical

and rigorous comparison without excessive and redundant duplication of efforts.

The results of this comparative study are shown in Figure A3. The software package with the highest score is SP-C,

with 44 out of a possible 48 points. Its main advantage is that it allows for the user to have a high level of control

over the system being modelled and the calculation methods used. The selected software package will then be used

to complete a full LCA study, including the creation of a detailed LCI, the selection and application of impact

assessment methods, and interpretation of the results within the specific context of the evaluation and comparison of

the environmental impact of various sustainable concrete mix designs.

64
7 Bibliography

Anderson, J. E., & Silman, R. (2009, March). A life cycle inventory of structural engineering design strategies for

greenhouse gas reduction. Structural Engineering International , pp. 283-288.

Butler, L., West, J., & Tighe, S. (2013). Effect of recycled concrete coarse aggregate from multiple sources on the

hardened properties of concrete with equivalent compressive strength. Construction and Building Materials (47),

1292-1301.

Cabral, A. E., Schalch, V., Molin, D. C., & Ribeiro, J. L. (2010). Mechanical properties modeling of recycled

aggregate concrete. Construction and Building Materials , 24, 421-430.

Canada Centre for Mineral & Energy Technology and Radian Canada Inc. (1993). Raw Material Balances, Energy

Profiles, and Environmental Unit Factor Estimates: Cement and Structural Concrete Products. Ottawa: Athena

Sustainable Materials Institute.

Churchill, C., & Panesar, D. (2013). Life-cycle cost analysis of highway noise barriers designed with photocatalytic

cement. Structure and Infrastructure Engineering: Maintenance, Management, Life-Cycle Design and Performance

, 9 (10), 983-998.

Ciroth, A. (2012). Software for Life Cycle Assessment. In M. A. Curran (Ed.), Life Cycle Assessment Handbook: A

Guide for Environmentally Sustainable Products (pp. 143-158). Scrivener Publishing LLC.

Damineli, B. L., Kemeid, F. M., Aguiar, P. S., & John, V. M. (2010). Measuring the eco-efficiency of cement use.

Cement and Concrete Composites , 32, 555-562.

De Barba Junior, D. J., de Oliveira Gomes, J., & Bork, C. A. (2014). Reliability of the Sustainability Assessment.

21st CIRP Conference on Life Cycle Engineering, (pp. 361-366).

Dhir, R., Paine, K., & Dyer, T. (2004). Recycling construction and demolition wastes in concrete. Concrete , 38 (3),

25-28.

ecoinvent. (2010). ecoinvent database, version 2.2. Zurich, Switzerland.

65
Hassan, M. (2010). Evaluation of the durability of titanium dioxide photocatalyst coating for concrete pavement.

Journal of Construction and Building Material , 1456-1461.

Henry, M., Pardo, G., Nishimura, T., & Kato, Y. (2011). Balancing durability and environmental impact in concrete

combining low-grade recycled aggregates and mineral admixtures. Resources, Conservation and Recycling , 55,

1060-1069.

ISO 14040. (1997). Life Cycle Assessment: Principals and Framework. Environmental Management . ISO.

Jolliet, O., Margni, M., Charles, R., Humbert, S., Payet, J. R., & Rosenbaum, R. (2003). IMPACT 2002+: A new life

cycle impact assessment methodology. International Journal of Life Cycle Assessment , 8 (6), 324-330.

Khan, M. I., & Siddique, R. (2011). Utilization of silica fume in concrete: Review of durability properties.

Resources, Conservation and Recycling , 57, 30-35.

Khedr, S. A., & Abou-Zeid, M. N. (1994). Characteristics of Silica-Fume Concrete. Journal of Materials in Civil

Engineering , 6, 357-375.

Kosmatka, S. H., Kerkhoff, B., Hooton, R. D., & McGrath, R. J. (2011). Design and Control of Concrete Mixtures

(8th Edition ed.). Ottawa, ON: Cement Association of Canada.

Marceau, M., Nisbet, M., & VanGeem, M. (2006). Life Cycle Inventory of Portland Cement Manufacture. Skokie,

IL, US: Portland Cement Association.

Marinkovic, S., Radonjanin, V., Malesev, M., & Ignjatovic, I. (2010). Comparative environmental assessment of

natural and recycled aggregate concrete. Waste Management , 30, 2255-2264.

Marsh, B. (2003). High-volume fly ash concrete. Concrete , 37 (4), 54-55.

Morgan, C., & Stevanovic-Briatico, V. (2007). Clean roads to clean air. APWA Reporter , 38-41.

Nisbet, M. A., Marceau, M. L., & VanGeem, M. G. (2002). Environmental Life Cycle Inventory of Portland Cement

Concrete. Portland Cement Association. Portland Cement Association.

66
Ortiz, I. (2003). Life cycle assessment as a tool for green chemistry: application to kraft pulp industrial wastewater

treatment by different advanced oxidation processes. Barcelona: Institut de Cie`ncia i Tecnologia Ambientals, The

Universitat Auto` noma de Barcelona.

Poon, C., & Cheung, E. (2007). NO removal efficiency of photocatalytic paving blocks prepared with recycled

materials. Construction and Building Materials , 21, 1746-1753.

Prusinski, J. R., Marceau, M. L., & VanGeem, M. G. (2004). Life Cycle Inventory of Slag Cement Concrete. Eighth

CANMET/ACI Eighth CANMET/ACI International Conference on Fly Ash, Silica Fume, Slag and Natural

Pozzolans in Concrete.

Scientific Applications International Corporation. (2006). Life Cycle Assessment- Principles and Practice.

Cincinnati, Ohio: US Environmental Protection Agency.

Siddique, R., & Khan, M. I. (2011). Supplementary Cementing Materials. New York: Springer.

Tennis, P., Thomas, M., & Weiss, W. (2011). State-of-the-Art Report on Use of Limestone in Cements at Levels of

up to 15%. Skokie, IL: Portland Cement Association.

Thomas, M. D., Delagrave, A., Blair, B., & Barcelo, L. (2013, December). Equivalent durability performance of

Portland limestone cement. Concrete International , 39-45,65.

Weidema, B., & Wesnaes, M. (1996). Data quality management for life cycle inventories- an example of using data

quality indicators. Journal of Cleaner Production , 4 (3-4), 167-174.

Xu, H., Provis, J., Deventer, J. v., & Krivenko, P. (2008). Characterization of Aged Slag Concretes. ACI Materials

Journal , 105 (2), 131-139.

67
Appendix A-1

Alternative Concrete Materials- Supplementary Information

GGBFS is a by-product of the steel industry, where rapid cooling of blast furnace slag results in fine, glassy

particles that can be used as a direct replacement for Portland cement (Siddique & Khan, 2011). Typical

replacement rates range from 15-50%, and generally GGBFS cement concrete has similar strength

development (Kosmatka, Kerkhoff, Hooton, & McGrath, 2011) and improved durability (Xu, Provis,

Deventer, & Krivenko, 2008) compared to conventional concrete.

FA is a by-product of coal-fired power stations. It is typically used as 10-30% by mass of the total cement

content, but recent development of high-volume FA concrete (HVFAC) has allowed for replacement levels

of more than 50% for use in structural applications (Marsh, 2003). As with GGBFS, the improved

durability of concrete containing FA may also influence the service life of the concrete in certain

applications.

SF is a byproduct that results from the manufacture of silicon or ferrosilicon (Khedr & Abou-Zeid, 1994). It

is typically added in blended cement, replacing 5-8% of Portland cement. As a very fine pozzolanic

material, SF has been shown to improve compressive and bond strength and reduce permeability of

concrete (Khan & Siddique, 2011).

PLC is produced by partially replacing Portland cement (PC) clinker with ground limestone; the

replacement level is greater than the 5% inherent in PC and typically ranges up to 10-15% (Thomas,

Delagrave, Blair, & Barcelo, 2013). Generally, limestone is ground to a higher fineness than typical PC,

which improves the gradation of the cement and can improve workability and increase strength (Tennis,

Thomas, & Weiss, 2011).

RAC replaces virgin aggregate with recycled construction and demolition waste materials (Henry, Pardo,

Nishimura, & Kato, 2011; Dhir, Paine, & Dyer, 2004; Cabral, Schalch, Molin, & Ribeiro, 2010). As mortar

and cement paste may still be attached to recycled aggregate particles, approximately 5% more cement is

68
required in RAC mixes to obtain the same compressive strength and workability, and RAC is not

recommended for aggressive exposures due to uncertain durability performance (Marinkovic, Radonjanin,

Malesev, & Ignjatovic, 2010).

PCAT cement contains titanium dioxide (TiO2) that catalyzes a reaction to convert air pollutants to less

damaging forms. Most research has focused on NOx pollution reduction (Hassan, 2010; Ortiz, 2003; Poon

& Cheung, 2007). It should be noted that in harsh exposure conditions, such as the freeze thaw conditions

and presence of deicing salts in Ontario, mixing photocatalytic cement into a concrete mix and then

applying it to the surface yields adequate durability (Morgan & Stevanovic-Briatico, 2007; Hassan, 2010)

69
List of Tables- Appendix A

Table A1 Main characteristics of LCA software systems (summarized from Ciroth, 2012)

Table A2 Software packages included in the critical and comparative review

Table A3 Rating legend for numerical comparison of software packages

Table A4 Data quality pedigree matrix (reproduced from Weidema & Wesnaes, 1996)

Table A5 Green concrete process data availability in the software packages

Table A6 Data quality pedigree matrix results for cement production

Table A7 Data quality pedigree matrix results for slag cement production
Table A1 Main characteristics of LCA software systems (summarized from Ciroth, 2012)

Platform Web tools vs. desktop tools

Pricing Model Commercial tools vs. freeware
Development Model Open source vs. closed source
Purpose General LCA vs. specialized tools vs. add-ons
 Table A2 Software packages included in the critical and comparative review.

Software Package Developer (Country) Description

(Version)
SP-A Athena Impact Athena Institute (Canada) Free, closed source desktop tool developed
Estimator for specifically for LCA studies of highway and road
Highways (1) elements and projects
SP-B BEES (4) NIST Engineering Free, closed source web tool developed specifically
Laboratory (USA) for LCA studies of building products
SP-C GaBi (6) PE International (Germany) Commercial, closed source desktop tool for general
LCA studies
SP-D Quantis Suite 2.0 Quantis Limited free/full commercial versions, closed source
(Switzerland/Canada) web tool for general LCA studies

SP-E SimaPro7 EarthShift Commercial, closed source desktop tool for general
(The Netherlands) LCA studies
Table A3 Rating legend for numerical comparison of software packages.

Rating Meaning
0 No/ Not at all
1 Somewhat/ Indirectly
2 Yes/ Very
 Table A4 Data quality pedigree matrix (reproduced from Weidema & Wesnaes, 1996)

Score 1 2 3 4 5
Reliability Verified data based Verified data partly Non-verified data Qualified estimate Non-qualified
on measurements based on partly based on (e.g. by industrial estimate
assumptions or non- assumptions expert)
verified measured
data
Completeness Representative Representative data Representative Representative Representative-
data from a from a smaller data from an data from a ness unknown or
sufficient sample number of sites but adequate number smaller # of incomplete data
of sites over an for adequate periods of sites but from sites/shorter from a smaller #
adequate period to shorter periods periods or of sites and/or
even out normal incomplete data from shorter
fluctuations from adequate periods
number of sites
Temporal < 3 years < 6 years difference < 10 years < 15 years Age unknown or
correlation difference to year difference difference > 15 years
of study difference
Geographical Data from area Average data from Data from area Data from area Data from
correlation under study larger area in which with similar with slightly unknown area or
the study area is production similar production area with different
included conditions conditions production
conditions
Further Data from Data from processes Data from Data on related Data on related
technological enterprises, and materials under processes and processes or processes or
correlation processes & study but from materials under materials but materials but
materials under different enterprises study but from same technology different
study different technology
technology
 Table A5 Green concrete process data availability in the software packages

Software Can the software package model concrete containing:

Package
Slag Fly Ash Silica Fume PLC RAC PCAT
SP-A Yes** Yes** Yes** Yes** Yes** No
SP-B Yes** Yes** Yes** Yes** No No
SP-C Yes Yes No* Yes Yes Yes
SP-D Yes Yes No* Yes Yes No*
SP-E Yes No* No* Yes Yes No*
*All data is not available with the software package, but it can be imported by the user.
** It can be modelled but only in specific products or for specific mix designs that cannot be modified by the user.
 Table A6 Data quality pedigree matrix results for cement production.

Software Score Strengths Weaknesses

Package
SP-A 9 High technological and Moderate temporal correlation (2004)
geographical correlation
High completeness
SP-B 14 High technological correlation Moderate geographical correlation (US/Canada data)
(but not limited to cement Low temporal correlation (2000)
production)
Data for cement production is integrated with data from
High completeness other processes and cannot be separated)
SP-C 9 High technological and Moderate temporal correlation (2004)
geographical correlation
High completeness
SP-D 12 High completeness Low geographical correlation (Switzerland)
High temporal correlation (2010)
SP-E 10 High technological correlation Moderate temporal correlation (2004)
High completeness Moderate geographical correlation (US/Canada data)
 Table A7 Data quality pedigree matrix results for slag cement production.

Software Score Strengths Weaknesses

Package
SP-A N/A N/A Not included in existing database so cannot be modelled
SP-B 14 High technological correlation Moderate geographical correlation (US/Canada data)
(but not limited to cement Low temporal correlation (2000)
production)
Data for slag cement production is integrated with data
High completeness from other processes and cannot be separated)
SP-C 10 High technological correlation Moderate geographical correlation (US/Canada data)
High completeness Low temporal correlation (2003)
SP-D 12 High completeness Low geographical correlation (Switzerland)
High temporal correlation (2010)
SP-E 12 High completeness Low geographical correlation (Switzerland)
High temporal correlation (2010)
Figure Captions

Figure A1 System boundary for the LCA of conventional concrete

Figure A2 General procedure for evaluating LCA software packages

Figure A3 Software package evaluation questionnaire and results

Figure A4 Software Package A: LCIA Result for 30 MPa Concrete with 25% Fly Ash Replacement

Figure A5 Software Package B: LCIA Result for 35 MPa Concrete with 20% Fly Ash Replacement

Figure A6 Software Package C: LCIA Result for 35 MPa Concrete with 20% Fly Ash Replacement

Figure A7 Software Package D: LCIA Result for 35 MPa Concrete

Figure A8 Software Package E: LCIA Result for 35 MPa Concrete with 20% Fly Ash Replacement
Figure A1 System boundary for the LCA of conventional concrete
Figure A2 General procedure for evaluating LCA software packages
 Primary Criteria
No. Addressed Questions and Sub-Questions SP-A SP-B SP-C SP-D SP-E
1.0 Goal and Scope Definition
1.1 Flexibility Can system boundaries be defined by the user? 1 0 2 1 2
1.2 Flexibility Can users input any functional unit that they want? 0 0 1 1 1
2.0 Life Cycle Inventory Analysis
Does the software include a database of inventory information for life cycle
2.1 Complexity processes? 2 2 2 2 2
2.1.1 Flexibility Can additional databases be added? 0 0 2 1 2
2.1.2 Complexity Are the databases relevant to Canada? 2 1 1 1 1
2.1.3 Complexity Is the available data relevant to the concrete industry? 2 2 2 2 2
2.1.4 Complexity Is the data updated regularly? 1 1 1 1 1
2.1.5 Complexity Are the concrete processes to be studied included in the database? 1 1 1 1 1
2.2 Complexity Can the use stage of a product be modeled? 2 0 2 2 2
2.3 Complexity Can the disposal phase of a product be modeled? 1 1 2 2 2
3.0 Life Cycle Impact Assessment
3.1 Complexity Does the tool include impact assessment methods? 1 2 2 1 2
3.1.1 Complexity Do the impact assessment methods support weighting? 0 2 2 1 2
3.1.2 Flexibility Can the default weightings be modified? 0 2 2 0 2
3.1.3 Flexibility Can you set a 'cut off' point for what impacts are included? 0 2 2 0 2
3.2 Complexity Can you incorporate other impacts besides environmental ones? 2 2 2 2 2
4.0 Interpretation
4.1 Output Does the software generate graphical representation of results? 2 2 2 2 2
4.2 Output Are the quantitative or physical data outputs readily available? 1 2 2 0 2
4.3 Complexity Can the software be used to perform sensitivity analysis? 1 1 2 1 2
4.4 Output Can the software be used to compare alternatives? 2 2 2 2 2
5.0 General User-Friendliness
5.1 Complexity How intuitive is the data entry? 1 2 2 2 1
5.2 Complexity How transparent is the process? 0 0 2 1 2
5.3 Complexity Does the software have a good user interface? 0 2 2 2 1
5.4 Output/ Flexibility How easy is it to compare alternatives by making small changes? 1 2 2 2 2
5.5 Complexity Is support provided for users of the software? 0 1 2 1 1
TOTAL (MAXIMUM POSSIBLE SCORE = 48 POINTS) 23 32 44 31 41

Figure A3 Software package evaluation questionnaire and results

Figure A4 Software Package A: LCIA Result for 30 MPa Concrete with 25% Fly Ash Replacement
Figure A5 Software Package B: LCIA Result for 35 MPa Concrete with 20% Fly Ash Replacement
Figure A6 Software Package C: LCIA Result for 35 MPa Concrete with 20% Fly Ash Replacement
Figure A7 Software Package D: LCIA Result for 35 MPa Concrete
Figure A8 Software Package E: LCIA Result for 35 MPa Concrete with 20% Fly Ash Replacement
APPENDIX B
MODELLING AND SENSITIVITY OF ONTARIO’S ELECTRICITY GRID
MIX FOR THE LIFE CYCLE ASSESSMENT OF CEMENT-BASED
MATERIALS

Karina E. Setoa, Cameron J. Churchillb, Daman K. Panesara*,

a
Department of Civil Engineering, University of Toronto, 35 St. George St., Toronto, Ontario,
Canada, M5S 1A4
b
Engineering and Society Program, McMaster University, 1280 Main St W, Hamilton , Ontario ,
Canada, L8S 4L7
*
Corresponding author. Email: d.panesar@utoronto.ca

ABSTRACT

Ontario’s industries, including the cement and concrete industry, rely on Ontario’s electricity
supply mix and grid to power many fundamental processes. Ontario’s electricity grid mix is
complex, with nuclear, gas/oil, hydroelectric, wind, biofuel, and solar generation installed
throughout the province, and it is also dynamic. From 2010 to 2015, the Ontario energy
landscape has changed dramatically, partly as a result of i) the government mandate to shut down
all coal burning power plants, ii) an increase in nuclear power capacity, and iii) the growing
importance of renewable energy technologies. Furthermore, the electricity supply mix is
projected to continue to evolve based on policy directives from the Ontario government. Since
industry is so reliant on the electricity grid, an accurate model of the relevant electricity grid is
crucial when evaluating the environmental impact of industrial processes through Life Cycle
Assessment (LCA). In this study, this model is built using a Top-Down Approach where data for
the Canadian electricity grid is adjusted to represent Ontario. LCA results for Portland cement
(PC) concrete, concrete containing fly ash (FA), and concrete containing Portland limestone
cement (PLC) are presented. The sensitivity of the results to changes in the composition of the
electricity supply mix is tested using models for the 2010 and 2025 (predicted) electricity supply
mix. The results show that for all three electricity models, PLC has the best environmental
performance while PC concrete has the worst. The environmental performance of all three

70
concrete materials is highest in 2014 and decreases in 2025, primarily due to the resources
required to increase the proportion of photovoltaic energy in the supply mix.

KEYWORDS

Life cycle assessment, electricity, Ontario, energy policy, Portland limestone cement, fly ash

1 INTRODUCTION

Ontario’s electricity grid mix is complex, with 34 367 MW of nuclear, gas/oil, hydroelectric,
wind, biofuel, and solar generation installed throughout the province (Independent Electricity
System Operator (IESO), 2015). This infrastructure, and the energy it produces, is critically
important to many industries in the province of Ontario, including the concrete industry. The
concrete industry has for the past several decades been working to mitigate the environmental
impacts of concrete production. Methodologies such as life cycle assessment (LCA) can be used
to quantify and analyze these impacts. However, electricity use features very prominently in the
life cycle of concrete, and as such assessing the environmental impact of concrete requires
characterization of the electricity grid.

Electricity grid modelling is of special concern in LCA (Curran, Mann, & Norris, 2005), due to
its ubiquity across industries and its potential for contributing significantly to environmental
impact results. Modelling electricity grids is challenging because the input life cycle inventory
(LCI) data must reflect the high degree of temporal, geographical, and technological specificities
inherent to electricity generation and distribution.

In addition, electricity grids are dynamic, varying with time, geography, and the evolution of
technology. In most jurisdictions, the electricity supply mix is dictated by a multitude of factors
including available resources, economic market forces, and policy directives, all of which have
the potential to change significantly over time. From 2010 to 2015, the Ontario energy landscape
has changed dramatically, partly as a result of i) the government mandate to shut down all coal
burning power plants, ii) an increase in nuclear power capacity, and iii) the growing importance
of renewable energy technologies. Furthermore, the electricity supply mix is projected to
continue to evolve based on policy directives from the Ontario government.
71
The objective of this paper is to propose a method for modelling Ontario’s electricity grid mix, to
identify data gaps and explore critical modelling decisions in the context of Ontario, and to
determine the influence of past and projected electricity supply mixtures on the environmental
performance of concrete.

2 BACKGROUND

2.1 Ontario’s Changing Electricity Grid Mix from 2010 to 2025

Ontario has approximately 34 367 MW of installed generation capacity, and approximately 154
terawatt-hours (TWh) of electricity was generated in 2014 (IESO, 2015). Ontario’s electricity
grid mix includes nuclear, gas/oil, hydroelectric, wind, biofuel, and solar installations that are
located throughout the province. Ontario is also a net exporter of electricity: in 2014, 4.9 TWh
were imported to the province, while 19.1 TWh were exported (IESO, 2015).

With new generation projects commissioned and ageing infrastructure retired every year,
Ontario’s electricity supply mix is constantly evolving. The supply mix has undergone two
substantial policy-directed changes from 2010 to 2014 (Figure B1a and Figure B1b): i) the
implementation of the Feed-In Tariff program to incentivize small renewable energy installations
(Stokes, 2013), and ii) the shutting down of Ontario’s coal power plants (Ministry of Energy,
2014). This has resulted in significant changes in how electricity is generated and distributed to
Ontario residents and industries. As shown in Figure B1a and B1b, between 2010 and 2014
there were increases in the amount of solar (+0.012%), wind (+2.5%), biofuel (+0.19%) and
hydroelectric generation (+4.0%), and decreases in coal (-8.3%) and gas/oil (-4.4%). It should be
noted that the IESO, which records this data, began combining gas and oil generation into one
category in 2013 to capture all dual-fuel plants, and that prior to 2013 there was an ‘other’
category which has since been expanded (IESO, 2015).

Many researchers suggest that changes to the supply mix will continue in the future. Harvey
(2013) argues that wind energy has the potential to largely displace existing Canadian fossil fuel
and nuclear electricity generation. Calvert & Mabee (2015) envision bioenergy crop production
and solar photovoltaic farms designed to co-exist and complement each other at the regional
level, which when combined with existing solar, hydropower and wind generation infrastructure
72
could lead to Ontario achieving a 94% renewable energy grid. Qudrat-Ullah (2014) imagines an
Ontario with “more and affordable green power” (Qudrat-Ullah, 2014, p. 859) based on the
elimination of gas generation and the use of nuclear generation, hydroelectric generation, and
hydroelectric energy storage systems. Further, policy directives at the level of the provincial
government suggest that the supply landscape will continue to change. Figure B1c shows the
projected supply mix in 2025, based on the Ontario Ministry of Energy’s Long-Term Energy
Plan (2013). Figure B1c shows that relative to the 2014 mix, the Ontario government wants, by
2025, i) to increase hydroelectric (+5.0%), gas/oil (+2.4%), wind (+6.6%), biofuel (+2.8%), and
solar (+3.0%) generation, ii) to reduce nuclear generation (-20.0%) and iii) to completely
eliminate coal generation. The rapidly evolving energy mix in Ontario presents challenges when
conducting an LCA, and specifically when choosing an LCI dataset.

2.2 LCI Data for Modelling Electricity Grid Mixes

Several relevant factors that contribute to the complexity of modelling electricity grid mixes in
Ontario are (Curran, Mann, & Norris, 2005):

 Broad geographic scope: Ontario is a large province in terms of land area (917,741 km2
(Ontario Ministry of Finance, 2015)), with energy generators and energy resources
located all over the province. In addition, the province both imports and exports
electricity to its neighbours, including Manitoba, Quebec, New York, Michigan and
Minnesota (IESO, 2015).
 Dynamics of supply dispatch: The generation and distribution of electricity are also
closely tied to consumption patterns- both baseload generators, which provide a
consistent source of electricity, and marginal generators, which are brought on-line to
meet higher than average demand, are relied upon to satisfy dynamic demand patterns.
 Wide variation in emissions and inputs per unit generation across and within fuel types
 Rapid ongoing evolution and regional variety: Ontario’s electricity grid is complex, and
the grid has changed markedly and continues to evolve.
 Long timeframes and importance of electricity consumption: Long-term planning is
essential to meeting demand for electricity, as both ageing and new infrastructure is relied
upon. The electricity grid is of fundamental importance to every industry in Ontario.
73
These factors mean that researchers and LCA practitioners modelling processes that rely on
electricity in Ontario must make several modelling decisions. Some of these decisions are
summarized in Table B1.

To decide whether marginal or average data should be used, it must first be determined whether
an attributional or consequential approach is most appropriate when modelling electricity. This
choice must be aligned with the objective of the study- in this case, to determine the influence of
past and projected electricity supply mixtures on the environmental performance of concrete.
Attributional LCIs describe how entities, such as emissions, raw materials inputs, and
intermediate products, flow within a chosen system and a chosen temporal window. For
electricity grid modelling, the attributional approach assumes that any increase in demand is met
by a proportional increase in electricity production by the entire grid, with each type of
generation contributing an amount corresponding to its portion in the supply mix. Consequential
LCIs focus on the consequences of potential decisions- incremental changes are made to test
incremental responses. For electricity grid modelling, a consequential approach requires an
understanding of how the nuances of supply and demand dynamics affect how generators
operate, and the resulting environmental consequences. As an example, this approach assumes
that marginal generators begin operating when demand exceeds a certain baseline capacity as in
Amor et al. (2014).

The selection of a system boundary is critical in any LCA. A common boundary is geographical-
for example, the province of Ontario. However, given the interconnection of modern electricity
grids, it is also important to consider whether or not imports and exports are included in the
model. Setting a boundary also means determining a threshold for completeness and
representativeness- for example, that the model must include 75% or 95% of all relevant inputs
and outputs. It should be noted that neither ISO 14041 (1998), a standard for goal and scope
definition and inventory analysis in LCA, nor ISO 14040 (2006), a standard for LCA principles
and framework, provide a suggested value for adequate completeness. Rather, this needs to be
determined by LCA practitioners as part of the Goal and Scope Definition phase based on the
aims of the study.

74
Modelling new and non-traditional technologies is often difficult due to a lack of data,
particularly long-term emissions data. In addition, it is difficult to predict how new technologies
might fare in the market in the long-term- for instance, what will their market share be in ten or
twenty years? LCA practitioners need to decide whether new technologies should be included in
the model, and if so, how to compensate for these uncertainties.

Co-product allocation for electricity generators is of particular interest in the context of concrete
products, as a co-product of coal burning, fly ash, is commonly incorporated into concrete
products as a supplementary cementing material. If an LCA practitioner determines that
allocating environmental impacts to various co-products is appropriate, there are several methods
that can be used, including allocation by mass and allocation by economic value as in Babbitt
and Lindner (2008). There is no consensus on allocation methods, however, and applying this
approach does add another layer of complexity to electricity grid modelling.

When modelling the environmental impact of electricity, the focus is often on electricity
generation, including the inputs, efficiencies, and emissions of various generation sources.
However, transmission and distribution infrastructure and processes can also influence
environmental impact. In addition to any capital environmental costs, transmission and
distribution losses influence the amount of electricity that must be produced by generators. Total
generation must always equal the sum of both consumption and losses to meet total demand.

These issues represent a sample of the challenges researchers and LCA practitioners face. The
intent of this work is not to provide definitive answers to these questions, but rather to assess the
influence of certain modelling decisions on the outcomes of LCA of concrete products.

3 SCOPE

The scope of this work is defined based on both consideration of the issues identified in Section
2.2, and the parameters of an LCA of alternative concrete materials, including the system
boundary and functional unit.

The attributional approach is appropriate for this study as it allows the comparison of whole
systems, such as the life cycles for each of the concrete products, and allows for focus on the
75
specific context of Ontario in 2014. As a result, average, rather than marginal, data should be
used, and it is assumed that any variation in demand causes a proportional variation in generation
from all generators.

In selecting a system boundary for modelling Ontario’s electricity grid mix, imports and exports
should be included as the environmental impacts of electricity produced in other jurisdictions
may vary considerably. A temporal system boundary must also be established. For this research,
as the selected approach is attributional, it is appropriate to consider each time period (2010,
2014, and 2025) as a specific temporal system boundary.

The types of generation that are considered in this work include those identified by IESO (2015):
nuclear, hydro, coal, gas/oil, wind, biofuel, and solar. It is assumed that any new technologies
that are not included in these categories comprise only a small portion of overall generation, and
can be neglected for this work, given that the scope of this project is limited to Ontario in 2010-
2025. In the future, if any new technologies represent a large share of the total electricity grid
mix in Ontario then electricity models should be updated accordingly.

For this work, it is assumed that because electricity is the primary useful product that is intended
to be produced by generating technologies, and generating technologies are optimized based on
efficiency of electricity production and not any other product, no co-product allocation is
considered.

In modelling any electricity grid, transmission and distribution losses should be included as they
are a part of the overall consumption. Ontario in particular is a large jurisdiction, with some
energy producers located in Northern Ontario, very far from the large urban centres of Southern
Ontario. Neglecting transmission and distribution losses, therefore, would not be an accurate
representation of the actual electricity grid.

The cradle-to-gate system boundary for LCA of Portland cement concrete products (PC) is
identified in Figure B2. Modifications are made to this conventional concrete system boundary to
model the alternative concrete materials (concrete containing fly ash (FA) and concrete
containing Portland limestone cement (PLC)) that are included in this LCA as follows:

76
  For FA, a transportation process was added. The upstream production of FA is not
included in the analysis, as it is a waste product that would be produced regardless of
whether it is used in concrete. Using FA in concrete is a better environmental choice than
landfilling, so excluding these upstream activities is a conservative assumption.
 For PLC, Portland limestone cement production was substituted for cement production.
Transportation of raw materials is not included as cement production facilities are
typically located at limestone quarries (Canada Centre for Mineral & Energy Technology
and Radian Canada Inc., 1993).

The electricity grid mix plays a key role at several points in the life cycle of concrete products,
including:

 Water Treatment: energy use for the City of Toronto municipal water treatment system,
including chemical manufacturing and water treatment facility operation; approximately
2.58 MJ/m3 of water per year (Racoviceanu, Karney, Kennedy, & Colombo, 2007)
 Cement Production: energy use for cement production including raw materials transport,
crushing, raw grinding, pyroprocessing, and finish grinding; approximately 620 MJ/tonne
of finished cement (Athena, 2005)
 Concrete Plant Operations: energy use for batching and mixing processes; approximately
242 MJ/m3 of 30 MPa concrete (Athena, 2005)
 Aggregate Production: energy use for the extraction and processing of fine and coarse
aggregates; approximately 37.8 MJ/t for the extraction and processing of coarse
aggregate and approximately 59.4 MJ/t for the extraction and processing of fine
aggregate (Canada Centre for Mineral & Energy Technology and Radian Canada Inc.,
1993)

The functional unit of this LCA is a cubic metre (m3) of each type of concrete with a water-to-
cement ratio of 0.4 (+/- 0.05). In this study the objective was to select a consistent functional unit
to isolate the influence of electricity grid modelling decisions.

To assess the influence of changes in the electricity supply mix, the scope of this paper includes
the time periods of 2010, 2014 and 2025 (predicted).

77
4 METHODOLOGY

4.1 Methodology for Building a LCI for Modelling Ontario’s Electricity Grid Mix

Two potential approaches for building a LCI for modelling Ontario’s electricity grid mix were
assessed: i) a bottom-up approach, where LCI data are collected at as fine a resolution as
possible, ideally from each individual generator; and ii) a top-down approach, where LCI data
for Canada (which includes Ontario) is adopted and modified to be appropriate for modelling
Ontario’s electricity grid mix.

To assess which methodology is most appropriate, data quality goals were established for an
Ontario electricity LCI. These data quality goals are categorized based on the data quality
pedigree matrix established by Weidema & Wesnaes (1996) as Reliability, Completeness,
Temporal Correlation, Geographical Correlation, and Further Technological Correlation, and are
shown in Table B2. The aim of evaluating these two methods using data quality metrics is to i)
identify data gaps for the modelling of Ontario’s electricity grid mix, and ii) select a method for
use in an LCA of concrete materials.

The data quality goals presented by Weidema & Wesnaes (1996) are expanded to create more
specific objectives for this research. These are that data i) are measured at each generating
facility and are verified by a third party, ii) cover 95% of Ontario’s electricity supply, and
include 95% of all inputs and outputs, iii) account for imports and transmission and distribution
losses, iv) are post-2014, to reflect major changes including the shutting down of coal power
plants in Ontario, v) are specific to Ontario, or if this is not possible, Canada, and vi) are relevant
to the technologies used in Ontario in 2014. The 95% threshold for completeness was selected to
ensure high completeness, while acknowledging that issues with data availability and collection
which are well-documented in LCA (including in Bjorklund (2002) and May & Brennan (2003))
mean that achieving 100% completeness is likely unrealistic.

4.1.1 Bottom-Up Approach

The Bottom-Up Approach involves compiling data from individual generators to attempt to build
a dataset that is an LCI for the electricity grid for all of Ontario.

78
The National Pollutant Release Inventory (NPRI) was surveyed in order to assess its coverage of
Ontario’s energy producers. The NPRI is a database containing information submitted by
companies that meet published reporting requirements. Companies are required to submit
emissions data for specific substances to NPRI: i) if one or more NPRI substances was
manufactured, processed or otherwise used at the facility during the year, and ii) if the total
number of hours worked at the facility exceeded the 20 000 hour employee threshold
(Environment Canada, 2013). In addition, a report is required if certain activities take place at a
facility including pit or quarry operations, or the use of stationary combustion equipment
(Environment Canada, 2013). There are 363 substances identified as NPRI substances. It should
be noted that carbon dioxide is not a NPRI substance, but that nitrogen oxides, sulphur dioxide,
and carbon monoxide are.

The 2012 NPRI report, which is the latest verified report, was assessed and forty-three Ontario
facilities associated with electric power generation (based on North American Industry
Classification System (NAICS) 6 Sector Name) were identified. Based on the approximate
installed capacity of each facility relative to Ontario’s total installed capacity in 2012 (33,868
MW (IESO, 2015)), it is estimated that the NPRI includes emissions that represent
approximately 76% of the grid capacity. Only 22 substances are accounted for, and carbon
dioxide is not one of those substances despite the fact that it has been shown to have a large
impact on environmental performance in several previous LCAs of concrete materials
(Huntzinger & Eatmon, 2009; Brown, Sadiq, & Hewage, 2014; Chen, Habert, Bouzidi, &
Jullien, 2010). This suggests that there are significant issues with completeness. Furthermore,
raw material inputs are not included at all in the NPRI, and neither are transmission and
distribution losses.

In the context of the data quality goals established in Table B2, an LCI established using the
Bottom-Up Approach has:

 High reliability, as data is recorded at each individual generating facility and is verified
by a third party (NPRI)
 Low completeness, as the data represents only 76% of generators and 22 pollutants, and
raw material inputs, imports, and transmission and distribution losses are not included
79
  Moderate temporal correlation, as although the data is collected annually the verification
process takes time and the latest verified report is for 2012
 High geographical correlation, as the data is specific to the Ontario facilities that are
included in the NPRI
 Moderate technological correlation, as although the data is specific to technologies used
in Ontario in 2012, not all generators or types of generation are included in the NPRI

As a result, an LCI created using this method is insufficient for this research due to a critical lack
of completeness, and moderate temporal and technological correlation, despite the high
reliability and geographical correlation.

4.1.2 Top-Down Approach

The Top-Down Approach involved taking a complete dataset for Canada, and modifying the
dataset to make it appropriate for use in modelling Ontario’s electricity grid mix by manipulating
the proportions of contribution from different types of generation.

The GaBi 6 LCA software package was used in this research (PE International, 2014). The
software platform is packaged with several LCI databases, and access to additional databases can
be purchased. Datasets appropriate for modeling Canada’s electricity grid were found in
Extension database XVII: Full US and in Professional database 2014 are listed in Table B3.

These nine datasets do include import and export activities and transmission and distribution
losses, and they represent at least 95% of all inputs and outputs as recorded in the database
documentation and verified by PE International (2014). However, although the GaBi databases
contain more regionally specific datasets for the United States (ex. ‘East’ or specific states), for
Canada the data is only available at the national level. Since the datasets are separated by mode
of energy production, however, it is possible to create a more regionally specific electricity
model by manipulating the ratio of energy from different sources. This requires the assumption
that the technologies used to produce energy are generally consistent across Canada, with similar
levels of efficiency, inputs and outputs. For this research, this is considered an acceptable
assumption, as the use of average data aligns with the selected attributional approach and the
scope of this work.
80
In the context of the data quality goals established in Table B2, an LCI established using the
Top-Down Approach has:

 Moderate reliability, since data is collected partly from primary industry data and partly
from secondary literature data, and all data is verified by a third party (GaBi)
 High completeness, since the data represents 95% of mass and energy input and output
flows, and 98% of their environmental relevance. Imports and transmission and
distribution losses are also included.
 High temporal correlation, since the dataset is valid for 2014. Furthermore, the data is
available by generation type, so that supply mixes for other years can be modelled by
changing the supply mix proportions.
 High geographical correlation, since the data is for Canada and Ontario is a sub-set of
Canada.
 High technological correlation, since the data is specific to technologies used in Canada
in 2014.

Due to the high completeness and high temporal, geographical and technological correlation of
the GaBi 6 datasets, using the Top-Down Approach creates a higher quality LCI than that created
using the Bottom-Up Approach.

4.2 Life Cycle Impact Assessment (LCIA) Methodology

In order to compile the LCI data and conduct the impact assessment, the GaBi 6 software
package was used (PE International, 2014). GaBi 6 is packaged with several impact assessment
methods, including the International Reference Life Cycle Data System (ILCD) method which
was used for this study (European Commission Joint Research Centre, 2011).

Four impact categories were selected based on their relevance to the assessment of the
environmental impact of concrete and the level of international consensus on their classification
and characterization that has been reached, as described in Stranddorf & Schmidt (2005). Table
B4 presents a description of each impact category as well as the characterization factor used in
the analysis. In LCA, characterization factors are science-based factors that are used to convert

81
LCI data to actual environmental impacts (Scientific Applications International Corporation,
2006).

4.3 Green Indicator Methodology

The life cycle impact assessment results were normalized by dividing the results by the result of
mix PC (concrete with Portland cement) in 2014 (the reference year). These normalized results
were then combined into a single green indicator score that combines the acidification, global
warming, resource depletion, and water depletion potentials. This methodology was selected as a
way to evaluate the comparative environmental performance of different concrete materials
across impact categories that have different units and orders of magnitude. This aligns with the
objective of this study, which is to determine the influence of past and projected electricity
supply mixtures on the environmental performance of concrete. An initial weighting of
0.25:0.25:0.25:0.25 was used to combine these categories, with alternative weightings also tested
to evaluate the sensitivity of the results. A higher green indicator indicates a concrete with a
higher environmental performance (which corresponds to a lower environmental impact).

5 RESULTS AND DISCUSSION

5.1 LCIA Results

Figure B3 shows the green indicator results using the 0.25:0.25:0.25:0.25 weighting of the four
impact categories (Weighting Scheme 1, as shown in Table B5). Each line represents one
concrete type, and the data is plotted in chronological order from left to right (2010, 2014, and
then 2025). The results show a consistent pattern in terms of the relative environmental
performance of the different concrete types, regardless of the electricity dataset used: the 2014
electricity grid gives the most environmentally favourable result (highest green indicator) for all
of the concrete types studied. It is interesting to note that the 2025 supply mix actually shows
worse environmental performance compared to not only the results generated using the 2014
mix, but also those generated using the 2010 mix in all cases. Although from 2010 to 2014 the
environmental performance of concrete containing PC, FA, and PLC increase by approximately
2%, from 2014 to 2025 these values decrease by approximately 10%. This suggests that, in the
context of concrete product life cycles in Ontario, the new proposed supply mix for 2025 may
82
actually be counterproductive to improving environmental performance of concrete. The
sensitivity of this result to the selected weighting scheme is assessed in Section 5.2, and Section
5.3 explores the percentage contribution of each type of generation to the LCIA results in each
impact category.

5.2 Weighting Sensitivity

The life cycle impact assessment (LCIA) results for four impact categories- acidification, global
warming potential, resource depletion, and water depletion- are included in the calculation of the
green indicator. Whenever multiple values are combined to create a single indicator value, some
form of weighting must be applied. There is no consensus on weighting in LCA, and there
should not be as the debate is valuable and critical to interpreting results. There is no way to
definitively determine that, for example, in all cases global warming potential should trump all
other forms of environmental impact. Similarly, there is no way to definitely determine that all
environmental impacts must be treated as equally important- which is the underlying decision
enacted when no weighting is applied. Rather, LCA practitioners need to be aware of the context
that forms the scope of their research, and the context in which their results may be interpreted.
As noted by Bengtsson & Steen, “weighting is not meant to deliver the final verdict about the
environmental performance […] it is meant to give an additional input into the process” (2000, p.
101). In this study, it is acknowledged that the results may be of interest to individuals and
groups with different views on the value of certain environmental systems and processes, and
thus different priorities. Therefore, this paper attempts to address the issue of weighting by
applying a systematic sensitivity analysis to confirm and elaborate on the results presented in
Figure B3 (Weighting Scheme 1).

Four alternative weighting schemes were assessed (Weighting Scheme 2 through 5). These
weighting schemes are shown in Table B5. These weighting schemes were each selected to
emphasize one impact category, which was given a weighting of 0.7 while the other three
categories were all weighted at 0.1. The purpose of selecting these weighting schemes was to
determine whether large changes in weighting of the impact categories affect the influence of the
electricity grid on the environmental impact of alternative concrete materials.

83
Figure B4, Figure B5, and Figure B6 show the green indicator results for Weighting Schemes 1
through 5 for PC, FA, and PLC respectively. These results generally follow the same trends
observed in Figure B3; namely, that the green indicators tend to increase from 2010 to 2014, and
then decrease from 2014 to 2025 to below the 2010 level. One exception is Weighting Scheme
5- Water Depletion-Weighted, where the green indicator values were highest in 2010 and lowest
in 2025 for all types of concrete. Weighting Scheme 4- Resource Depletion-Weighted shows the
most dramatic decline in environmental performance, with the highest negative slope between
2014 and 2025 for all types of concrete. The reasons for these differences are explored by
breaking down the results in Section 5.3.

It should also be noted that the three concrete mixes retain the same relative ranking with each
weighting category, suggesting that the although the differences in the electricity grid mix from
2010 to 2025 are influencing the absolute values of the green indicator, there is no influence on
the relative environmental performance of the concrete mixes.

5.3 Breakdown of LCIA Results

To further explore the trends shown in Figures 4 to 6, the LCIA results for concrete containing
Portland cement are broken down by life cycle process so that the relative influence of each
source of electricity on the environmental impact of alternative concrete materials can be
quantified. Table B6 shows these results, separated by impact category (Acidification, Global
Warming Potential, Resource Depletion, and Water Depletion) and then reference year. Non-
electricity related processes are grouped as one entity (‘Non-Electricity Processes’) while the
electricity related processes are broken down by type of generation. Note that these results
pertain to the entire life cycle of concrete production, and not just the electricity grid mix. As a
result, there are non-zero contributions from coal in 2014 and 2025 due to its presence in other
parts of the life cycle (cement kiln fuel) despite its elimination from the electricity grid mix.

The results show that the relative contributions in 2010 and 2014 are very similar despite the
changes in supply mix and in the calculated green indicator values. In 2025, however, there is a
large increase in the relative contribution of the Photovoltaic process in the Resource Depletion
impact category (from ~0.2% to 32.3%). This correlates to the high negative slope seen in

84
Figures 4 to 6 between 2014 and 2025 for Weighting Scheme 4- Resource Depletion-Weighted.
This occurs with an increase in the proportion of photovoltaic in Ontario’s grid mix from 0.012%
to 3%, as shown in Figure B1, however the relative contribution shown in Table B6 is out of
proportion compared to biomass, which is also projected to increase quite significantly from
0.19% to 3%. This is likely due to the virgin raw material resources that are required to produce
photovoltaic panels and the related infrastructure. As reported by Fthenakis & Kim (2011),
producing various types of photovoltaic technology (including all the types covered by the GaBi
dataset) involves the mining and processing of quartz sand, zinc ores, and copper ores, as well as
minor metals including cadmium, indium, molybdenum, and selenium. In addition, water
resources are also required to process these materials and produce photovoltaic panels, which
also explains the downwards trend shown in Figures 4 to 6 for Weighting Scheme 5- Water
Depletion-Weighted.

6 CONCLUSIONS

Electricity is a critical input into the life cycle of most products and processes. Modelling
electricity grids is challenging due to the complexity of electricity systems, and the high
geographical, temporal, and technological specificity of generation infrastructure. The
conclusions of this paper are relevant to any LCA practitioner or researcher conducting LCA
studies in Ontario.

 Significant data gaps exist in building an LCI for Ontario’s electricity grid using the
‘bottom-up’ approach, as only large generators are required to publicly release their
pollutant emissions to NPRI, and raw material inputs and transmission and distribution
losses are not included.
 Electricity has a large contribution in the life cycle environmental impact of concrete
materials; in this study electricity processes accounted for approximately 37-83% of the
LCIA results in a given impact category and year.
 When environmental impacts are assessed in terms of Acidification Potential, Global
Warming Potential, Resource Depletion Potential, and Water Depletion Potential,
significant changes in the weighting of the impact categories do not have a significant

85
 influence on relative environmental performance of the concrete materials. These results
show that, for all weightings, out of the three concrete materials studied concrete
containing Portland cement has the lowest (worst) green indicator, while concrete
containing PLC cement has the highest (best) green indicator.
 If Ontario’s electricity system shifts to reflect the projected 2025 Ontario supply mix
outlined in the Ministry of Energy’s Long Term Energy Plan (2013), the results suggest
there may be a decrease in the environmental performance of concrete materials. This
trend is especially pronounced for the Resource Depletion impact category, as shown
when the results are weighted to emphasize Resource Depletion.
 The decrease in environmental performance in 2025 appears to be caused by a large
increase in photovoltaic electricity generation, which requires correspondingly larger
amounts of mineral and water resources.
 LCA studies of cement-based materials should account for the evolving electricity grid
mix in the jurisdiction under study. The 2014 electricity grid mix has the highest
correlation to the scope of the study, and so this data will be used for the final LCA.
However, the model should be modified as Ontario’s energy profile continues to evolve.

7 Acknowledgement

The authors acknowledge the MTO HIIFP (2013-2015) program for support of this research.
This research was supported by a contribution from the Ministry of Transportation of Ontario.
Opinions expressed in this report are those of the authors and may not necessarily reflect the
views and policies of the Ministry of Transportation of Ontario.

86
8 Bibliography

Amor, M., Gaudreault, C., Pineau, P.-O., & Samson, R. (2014). Implications of integrating
electricity supply dynamics into life cycle assessment- A case study of renewable
distributed generation. Renewable Energy, 69, 410-419.

Athena. (2005). Cement and Structural Concrete Products: Life Cycle Inventory Update #2.
Ottawa: Athena Sustainable Materials Institute.

Babbitt, C., & Lindner, A. (2008). A life cycle comparison of disposal and beneficial use of coal
combustion products in Florida. International Journal of Life Cycle Assessment, 13(3),
202-211.

Bengtsson, M., & Steen, B. (2000). Weighting in LCA- Approaches and applications.
Environmental Progress, 19(2), 101-109.

Bjorklund, A. (2002). Survey of approaches to improve reliability in LCA. International Journal

of Life Cycle Assessment, 7(2), 64-72.

Brown, D., Sadiq, R., & Hewage, K. (2014). An overview of air emission intensities and
environmental performance of grey cement manufacturing in Canada. Clean
Technologies and Environmental Policy, 16, 1119-1131.

Calvert, K., & Mabee, W. (2015). More solar farms or more bioenergy crops? Mapping and
assessing potential land-use conflicts among renewable energy technologies in eastern
Ontario, Canada. Applied Geography, 56, 209-221.

Canada Centre for Mineral & Energy Technology and Radian Canada Inc. (1993). Raw Material
Balances, Energy Profiles, and Environmental Unit Factor Estimates: Cement and
Structural Concrete Products. Ottawa: Athena Sustainable Materials Institute.

Chen, C., Habert, G., Bouzidi, Y., & Jullien, A. (2010). Environmental impact of cement
production: detail of the different processes and cement plant variability evaluation.
Journal of Cleaner Production, 18, 478-485.

87
Curran, M., Mann, M., & Norris, G. (2005). The international workshop on electricity data for
life cycle inventories. Journal of Cleaner Production, 13, 853-862.

Environment Canada. (2013, July 11). Summary of National Pollutant Release Inventory
Reporting Requirements. Retrieved from Environment Canada- Pollution and Waste:
https://www.ec.gc.ca/inrp-npri/default.asp?lang=en&n=629573FE-1

European Commission Joint Research Centre. (2011). International Reference Life Cycle Data
System (ILCD) Handbook. Instute for Environment and Sustainability. Luxemburg:
Publications Office of the European Union.

Fthenakis, V., & Kim, H. (2011). Photovoltaics- LIfe-cycle analyses. Solar Energy, 85, 1609-
1628.

Harvey, L. (2013). The potential of wind energy to largely displace existing Canadian fossil fuel
and nuclear electricity generation. Energy, 50, 93-102.

Huntzinger, D., & Eatmon, T. (2009). A life-cycle assessment of Portland cement

manufacturing: comparing the traditional process with alternative technologies. Journal
of Cleaner Production, 17, 668-675.

Independent Electricity System Operator (IESO). (2015). Supply Overview. Retrieved February
10, 2015, from IESO: http://www.ieso.ca/Pages/Power-Data/Supply.aspx

International Organization for Standardization. (1998). Environmental management- Life cycle

assessment- Goal and scope definition and inventory analysis. Geneva: ISO.

International Organization for Standardization. (2006). Environmental management- Life cycle

assessment- Principles and framework. Geneva: ISO Copyright Office.

May, J., & Brennan, D. (2003). Application of data quality assessment methods to an LCA of
electricity generation. International Journal of Life Cycle Assessment, 8(4), 215-225.

Ministry of Energy. (2013). Achieving Balance: Ontario's Long-Term Energy Plan. Toronto,
ON: Ministry of Energy.

88
Ministry of Energy. (2014, April 15). Creating Cleaner Air in Ontario: Province has eliminated
coal-fired generation. Retrieved from Ontario Newsroom:
http://news.ontario.ca/mei/en/2014/04/creating-cleaner-air-in-ontario-1.html

Ontario Ministry of Finance. (2015, June 15). Ontairo Fact Sheet June 2015. Retrieved July 5,
2015, from Ontario Ministry of Finance:
fin.gov.on.ca/en/economyecupdates/factsheet.html

PE International. (2014). GaBi 6.0.

PE International. (2014). GaBi life cycle inventory data documentation. Retrieved October 10,
2014, from GaBi: http://www.gabi-software.com/canada/support/gabi/gabi-database-
2014-lci-documentation/

Qudrat-Ullah, H. (2014). Green power in Ontario: A dynamic model-based analysis. Energy, 77,
859-870.

Racoviceanu, A., Karney, B., Kennedy, C., & Colombo, A. (2007, December). Life-Cycle
Energy Use and Greenhouse Gas Emissions Inventory for Water Treatment Systems.
Journal of Infrastructure Systems, 13(4), 261-270.

Stokes, L. (2013). The politics of renewable energy policies: The case of feed-in tariffs in
Ontario, Canada. Energy Policy, 56, 490-500.

Stranddorf, H., & Schmidt, L. (2005). Update on impact categories, normalisation and
weighting in LCA. Danish Ministry of the Environment Environmental Protection
Agency.

Weidema, B., & Wesnaes, M. (1996). Data quality management for life cycle inventories- an
example of using data quality indicators. Journal of Cleaner Production, 4(3-4), 167-174.

89
List of Tables- Appendix B

Table B1Decisions in modelling electricity systems (Curran, Mann, & Norris, 2005)

Table B2 Data quality goals

Table B3 Datasets appropriate for modeling Canada’s electricity grid using the GaBi 6

software package (PE International, 2014)

Table B4 Impact categories

Table B5 Weighting schemes

Table B6 Percentage contribution to LCIA results in each impact category for concrete

containing Portland cement.

 Table B1 Decisions in modelling electricity systems (Curran, Mann, & Norris, 2005)
Marginal vs. average
System boundaries
New and non-traditional
Co-product allocation
Transmission /distribution
Should LCA model the response of the energy supply system to demand?
How wide and broad should the boundaries be to capture environmental flows and data that
are needed for impact?
How should LCA model new technologies, in which the data are highly uncertain, and how
should increased demand for new technologies be accounted for?
How should environmental burdens be allocated across co-products that come from the
same process?
How should transmission and distribution impacts be included in modeling of electricity
generation?
 Table B2 Data Quality Goals
Category Data Quality Goal
(Weidema & Wesnaes, 1996) (Weidema & Wesnaes, 1996)
Reliability Verified data based on measurements
Completeness Representative data from a sufficient
sample of sites over an adequate
period to even out normal fluctuations
Temporal correlation < 3 years difference to year of study
Geographical correlation Data from area under study
Further technological Data from enterprises, processes &
correlation materials under study
Expanded Data Quality Goal
Data that is measured at the source and is verified
by a third party
Data that includes 95% of Ontario’s electricity
supply
Data that includes 95% of all inputs and outputs
Data that reflects changes since 2014
Data from Ontario, or Canada if appropriate
Data for technologies that are used in Ontario in
2014
Table B3 Datasets appropriate for modeling Canada’s electricity grid using the GaBi 6 software package (PE International, 2014)
Dataset
Electricity from nuclear power, Canada
Electricity from wind power, Canada
Electricity from biomass (solid), Canada
Electricity from natural gas, Canada
Electricity from hard coal, Canada
Electricity from hydro power, Canada
Electricity from photovoltaic, Canada
Electricity from heavy fuel oil (HFO), Canada
Electricity grid mix (production mix), Canada
Database
Professional database 2014
Professional database 2014
Professional database 2014
Extension database XVII: Full US
Extension database XVII: Full US
Extension database XVII: Full US
Professional database 2014
Extension database XVII: Full US
Extension database XVII: Full US
 Table B4 Impact categories

Impact Category Description Characterization Factor

Release of protons and leaching out of anions from a Acidification Potential

Acidification
system (mole H+ ion equivalents)
Effect of increasing temperature in the atmosphere due to Global Warming Potential
Global Warming the reflection of radiation by greenhouse gases back to the
surface (kg CO2 equivalents)

Resource Depletion of abiotic non-water resources (ex. minerals or Resource Depletion Potential (kg of
Depletion fossil fuels) antimony (Sb) equivalents)
Water use (reuse, degradation, or incorporation into final Water Depletion Potential (volume
Water Depletion
product) of water used (m3))
 Table B5 Weighting schemes

Weighting Acidification Global Warming Resource Water Depletion

Scheme Potential Depletion
1 0.25 0.25 0.25 0.25
2 0.7 0.1 0.1 0.1
3 0.1 0.7 0.1 0.1
4 0.1 0.1 0.7 0.1
5 0.1 0.1 0.1 0.7
 Table B6 Percentage contribution to LCIA results in each impact category for concrete containing Portland cement.

Percentage Contribution to LCIA Results in each Impact Category (%)

Type of Acidification Global Warming Resource Depletion Water Depletion
Electricity
Generation 2010 2014 2025 2010 2014 2025 2010 2014 2025 2010 2014 2025
Biomass 1.0 1.0 1.3 0.2 0.2 0.2 0.9 1.0 0.8 17.1 17.1 19.6
Coal 54.6 53.0 52.8 31.5 30.3 30.1 14.0 13.3 8.9 6.6 6.1 5.8
Oil 0.2 0.1 0.2 0.2 0.2 0.2 0.3 0.2 0.2 0.0 0.0 0.0
Hydropower 0.0 0.0 0.0 0.1 0.1 0.1 5.2 6.2 5.1 8.6 10.0 11.8
Natural Gas 2.1 2.1 2.2 10.6 10.2 10.6 45.6 43.5 30.4 1.9 1.8 1.8
Nuclear
Power 0.1 0.2 0.1 0.1 0.1 0.1 8.5 9.6 4.5 2.4 2.6 1.7
Photovoltaic 0.0 0.0 0.0 0.0 0.0 0.0 0.0* 0.2* 32.3* 0.0 0.0 0.1
Wind 0.0 0.0 0.0 0.0 0.0 0.0 0.1 0.4 0.6 0.0 0.0 0.0
Non-
Electricity
Processes 42.0 43.5 43.4 57.4 58.9 58.6 25.4 25.8 17.3 63.2 62.3 59.2

*Note that some values are formatted as bold to highlight the significant increase from 2010 to 2025 in a given category.
List of Figures- Appendix B

Figure B1 Ontario’s projected electricity supply mix for a) 2010 (IESO, 2015), b) 2014

(IESO, 2015), and c) 2025 (Ministry of Energy, 2013).

Figure B2 System boundary for LCA of concrete products.

Figure B3 Green indicator results for Weighting Scheme 1 (0.25:0.25:0.25:0.25) for

Portland cement concrete, fly ash concrete, and Portland limestone cement
concrete.

Figure B4 Green indicator results for concrete with Portland cement, all weighting

schemes.

Figure B5 Green indicator results for concrete with FA, all weighting schemes

Figure B6 Green indicator results for concrete with PLC, all weighting schemes
 Wind Other
Gas 1.9% 0.86%
14%

Coal
8.4%

Nuclear
55%

Hydro
20%

a)
Wind Biofuel
0.19% Solar
Gas/Oil 4.4% 0.012%
9.6%
Coal
0.065%

Nuclear
62%
Hydro
24%

b)
Biofuel Solar
3% 3%
Wind
11%
Nuclear
42%
Gas/Oil
12%

Hydro
29%
c)
Figure B1 Ontario's projected electricity supply mix for a) 2010 (IESO, 2015), b) 2014 (IESO, 2015), and c) 2025
(Ministry of Energy, 2013).
Figure B2 System boundary for LCA of concrete products.
 1.4

1.3

1.2

1.1
Green Indicator

0.9

0.8

0.7

0.6
2010 2014 2025
Year

GI-W1-PC GI-W1-FA GI-W1-PLC

Figure B3 Green indicator results for Weighting Scheme 1 (0.25:0.25:0.25:0.25) for Portland cement concrete, fly ash
concrete, and Portland limestone cement concrete.
 1.4

1.3

1.2
Green Indicator

1.1

0.9

0.8

0.7

0.6
2010 2014 2025
Year

GI-PC-W1 GI-PC-W2 GI-PC-W3 GI-PC-W4 GI-PC-W5

Figure B4 Green indicator results for concrete with Portland cement, all weighting schemes.
 1.4

1.3

1.2

1.1
Green Indicator

0.9

0.8

0.7

0.6
2010 2014 2025
Year

GI-FA-W1 GI-FA-W2 GI-FA-W3 GI-FA-W4 GI-FA-W5

Figure B5 Green indicator results for concrete with FA, all weighting schemes
 1.4

1.3

1.2

1.1
Green Indicator

0.9

0.8

0.7

0.6
2010 2014 2025
Year

GI-PLC-W1 GI-PLC-W2 GI-PLC-W3 GI-PLC-W4 GI-PLC-W5

Figure B6 Green indicator results for concrete with PLC, all weighting schemes
APPENDIX C
INFLUENCE OF FLY ASH ALLOCATION METHODOLOGIES IN LIFE
CYCLE ASSESSMENT OF CEMENT-BASED MATERIALS

Karina E. Setoa, Cameron J. Churchillb, Daman K. Panesara*

a
Department of Civil Engineering, University of Toronto, 35 St. George St., Toronto, Ontario, Canada, M5S 1A4
b
Engineering and Society Program, McMaster University, 1280 Main St W, Hamilton , Ontario , Canada, L8S 4L7
*
Corresponding author. Email: d.panesar@utoronto.ca

KEYWORDS

Fly ash (D), cement (D), concrete (E), allocation

90
ABSTRACT

This paper critically compares methods for allocating the environmental impact of fly ash
production in a Life Cycle Assessment (LCA) study of concrete, based on an evaluation of the
variability in LCA outcomes. LCA models of concrete containing 10% fly ash are created based
on four allocation scenarios: i) Baseline- No Allocation, ii) Mass Allocation, iii) Economic
Allocation, and iv) Disposal Avoidance. LCA results are generated and compared to the Baseline
scenario. These results show that depending on the method and assumptions used, the influence
on the results may not be small (eg. ~5% and ~10% for Economic and Mass Allocation
respectively). Based on these results and sensitivity analysis, it appears that in this context no
allocation scenario is more valid than the others as the results are dependent on the allocation
coefficients, which are highly variable. The Baseline scenario is selected for future work to
reduce uncertainties by eliminating the need for additional data sources and assumptions.

1 INTRODUCTION

For the past several decades, strong motivations for the development and implementation of
alternative concrete mix designs other than conventional concrete (comprising 100% general use
(GU) cement) has been an effort to reduce both cost and negative environmental impacts.
Supplementary cementing materials (SCMs) are commonly used to improve the environmental,
economic and in some cases technical performance of concrete products. One such SCM is fly
ash, which is produced as a by-product during the combustion of coal.

Life cycle assessment (LCA), which is defined as “the compilation and evaluation of the inputs,
outputs and potential environmental impacts of a product system throughout its life cycle” [1], is
a useful way to environmentally model the complex processes that are included in the life cycle
of a concrete product. An interesting problem in LCA is the treatment of multifunctional
processes. Specifically, what is the most appropriate way to model processes which produce
more than one product? Coal combustion is an example of a multifunctional process as it is a
single process which produces both electricity and fly ash. That is, for every kilogram of coal
91
that is combusted, a certain amount of both electricity and fly ash are produced, and these
products cannot be produced independently of one another when burning coal. The difficulty for
LCA practitioners is determining which environmental costs (from coal burning) are attributable
to the production of electricity, and which are attributable to the production of fly ash. This
process of distributing environmental impacts is called ‘allocation’. The results of this allocation
can then be used to assess the environmental performance of fly ash further in its life cycle- for
example, when it is incorporated into concrete. To perform an LCA of fly ash concrete,
therefore, it is necessary to determine how to resolve the issue of allocation for fly ash
production.

This paper presents a critical comparison of four allocation scenarios for the allocation of the
environmental impact of fly ash production in an LCA of concrete products; namely i) Baseline-
No Allocation, ii) Mass Allocation, iii) Economic Allocation, and iv) Disposal Avoidance. To
enable these comparisons, LCA outcomes are presented based on an LCIA methodology that
results in the calculation of green indicators for various concrete mix designs, as described in
Section 4.2 and 4.3 of this paper. The objective of this paper is to evaluate the variability in LCA
outcomes as a function of fly ash allocation.

2 BACKGROUND

2.1 Fly Ash as a Cement Replacement in Concrete Products

Fly ash is a by-product of coal-fired power stations, and the majority of it has been traditionally
treated as a waste product. Since the 1930s, fly ash has been used in concrete due to its durability
and structural benefits, including low permeability, adequate sulfate and chloride ingress
resistance, low long-term shrinkage and creep, and effective mitigation of alkali-silica reaction
(ASR) [2]. It is typically used as 15-30% by mass of the total cement content, but recent
development of high-volume fly ash concrete (HVFAC) has allowed for replacement levels
greater than 50% and use in common structural applications [2]. Its environmental benefits stem
from both the repurposing of a waste product, and the replacement of cement content in a given
concrete mix design. The replacement of cement content can significantly improve the overall
92
environmental performance of a concrete mix design. This is because cement is the component
of concrete that has the highest environmental impact, with 6-7% of global CO2 emissions
attributed to its production [3]. O’Brien et al [4] have shown that replacing cement with fly ash is
environmentally beneficial from a greenhouse gas emissions perspective even when the
transportation distances between the coal plant and the cement plant are very large (ex. “more
than a quarter of the way around the world by road, and much further by rail and sea”
[4]). The potentially improved durability of concrete containing fly ash at certain replacement
rates may also extend the service life of the concrete.

Table C1 summarizes LCA studies from the literature that include concrete containing fly ash.
The table describes the source, functional unit, geographical context, and allocation procedures
of eight different studies. It can be seen that in general, allocation procedures are not commonly
applied. Three general reasons are typically cited: i) there is often a lack of data about the inputs
and outputs of fly ash production and other processes associated with coal burning (as in [5]), ii)
fly ash is produced as a waste product and collected to reduce particulate matter emissions and
improve air quality regardless of how it is used (as in [6]), iii) it is assumed that the
environmental impact of fly ash production is relatively very small compared to the
environmental impact of the overall concrete life cycle (as in [7]).

2.2 Allocation Methods

A multifunctional process is “an activity that fulfils more than one function” [8, p. 197], as
shown in Figure C1. Many industrial processes, including the combustion of coal, can be
described as multifunctional. It should be noted that the multiple products may not have equal
value or usefulness- for example, a process may produce both a primary desired product and also
multiple waste products. It is necessary to decide how to account for this multifunctionality to
ensure appropriate assessment of environmental impacts. If a selected method of accounting for
multifunctionality is not appropriate for the specific context of the LCA, then environmental
impacts may be skewed, either higher or lower. For example, if co-products or waste products
are not accounted for, the environmental impact of a single product may appear higher, as the

93
environmental burden of the entire process is attributed to only that product. The question,
therefore, is not only whether or not to account for these impacts, but how this should be done.

Several methods can be used for allocation in multifunctional processes as described in ISO
14041 [9] and as illustrated in Figure C2.

1. Wherever possible, allocation should be avoided by applying:

a. The Subdivision Method: dividing the process into subprocesses and assessing
subprocesses independently; or
b. The System Expansion Method: expanding the product system to include the
additional functions related to the coproducts.
2. Where allocation cannot be avoided, the Mass Allocation Method should be applied:
inputs and outputs of the system should be partitioned between different functions in a
way which reflects the underlying physical relationships between them.
3. Where physical relationship alone cannot be established or used as the basis for
allocation, the inputs should be allocated between the products and functions in a way
that reflects other relationships between them. One example of this is the Economic
Allocation Method, which is based on relative economic value of products.

It should be noted that although ISO 14041 presents this procedure and suggests a preference of
methods, ISO 14041 does not evaluate these methods in terms of feasibility, or explain the effect
of selecting one method or the other on the results of an LCA. Some comments on each of these
methods are presented.

2.2.1 Subdivision Method

The Subdivision Method is the first approach suggested in ISO 14041, and is the most
theoretically straightforward procedure. This method involves separating the multifunctional
process described in Figure C1 into separate processes that each correspond to one of the
products as shown in Figure C2. The actual implementation of this method, however, does
present many challenges. This method assumes that the multifunctional process is simply an
amalgamation of separate single function processes, and that data is available at a high enough
94
resolution to enable the quantification of the inputs and outputs that comprise both sub-
processes. Both of these assumptions are problematic for two primary reasons.

Firstly, many multifunctional processes are simply impossible to separate into single function
processes due to a lack of independence. Consider the example of coal combustion, where
multiple products including energy and fly ash are produced. There is no unique coal combustion
process that only produces energy, while another only produces fly ash. Rather, the two products
are both dependent on a single coal combustion process.

Secondly, data collection and quality validation is always a challenging aspect of LCA and this is
compounded when data is required for more processes and at a higher level of detail when
compared to single function processes, as in the subdivision method. If this data is not collected
by the relevant industry partners, or has not been studied at this resolution in the literature, then it
will be unavailable and the modelling of independent sub-processes will not be possible.

Despite the theoretical simplicity of this method, owing to the challenges, the Subdivision
Method is rarely able to completely resolve the issue of multifunctional processes [8]. The
method of subdivision is not included in the sensitivity analysis presented in this paper.

2.2.2 System Expansion Method

The System Expansion Method avoids allocation by extending the system boundary in Figure C1
to include alternative products and means of production. The LCI data for these alternatives
(shown in Figure C2 as Sub-Process C and Product C) are then subtracted from that of the
multifunctional process to calculate the remaining portion of the multifunctional process data-
the portion that can be attributed to Product A.

To use this method for fly ash production, data would be required for a process equivalent to
electricity production. However, because of the variability in environmental and economic
efficiency of various electricity production processes, and the dynamic nature of electricity
supply grids across time and geography, using the system expansion method could lead to
drastically different results depending on the selected alternative product. For example, if it is

95
assumed that the alternative to the coal combustion process was electricity production from solar
energy, the LCI data will be very different than if the alternative was electricity production from
nuclear energy.

Similar to the Subdivision Method, to successfully implement the System Expansion Method
significant amounts of alternative data are required. Furthermore, many assumptions must be
made in order to decide what alternative is an adequate replacement for the required function. As
a result, this method is difficult to implement and subject to debate and challenge during the
LCA review process. The method of system expansion is included in this review, but is not
included in the sensitivity analysis.

2.2.3 Mass Allocation Method

The Mass Allocation Method is an example of partitioning based on underlying physical
relationships as described in ISO 14014 [9]. In this method, environmental impacts for the whole
multifunctional process are calculated. These impacts are then allocated according to the relative
mass of the co-products, as shown in Figure C2. For products that are not typically measured
using units of mass, such as electricity, a mass equivalent can be calculated if input masses are
known by applying the principle of the conservation of mass, as in Chen et al. [10].

The advantage of this method over the Subdivision and System Expansion Methods is that it
does not demand the collection and analysis of large amounts of extra data. Only the masses
must be known, and then the allocated impacts can be quickly calculated. The disadvantage of
this method is that there is no consensus that this method is appropriate for use in LCA, and
some researchers [11, 12] believe that this method is an over-simplification of the realities of
multifunctional processes. In addition, the outcomes of this method are solely dependent on the
relative masses of the products. For example, if the mass of a waste product is much larger than
all the other products, this could lead to a manufacturer claiming that the environmental impact
of their primary product is very low.

96
2.2.4 Economic Allocation Method
The Economic Allocation Method is an example of partitioning based on other relationships
between products and functions as described in ISO 14014 [9]. In this method, as with the Mass
Allocation Method, environmental impacts for the whole multifunctional process are calculated.
These impacts are then allocated according to the relative economic value of the co-products as
shown in Figure C2.

The Economic Allocation Method has a similar advantage to the Mass Allocation Method, in
that it avoids the need for the collection and analysis of large amounts of extra data. However, it
can be argued that the Economic Allocation Method is in some ways more problematic than the
Mass Allocation Method due to the temporal and geographical variability of economic value. For
example, a kilogram of coal in 2014 is not worth the same amount in Canada as in China, and
neither is it worth the same amount as it as in 1964. LCA studies are already highly context-
specific, but introducing economic allocation methods further constrains the relevance of the
results. Considering economic value, however, may remedy a key issue with the Mass Allocation
Method, as the primary product will likely always have the highest economic value. As a result,
even if a large mass of waste product is produced, for example, the environmental impacts will
still be primarily assigned to the main product.

3 LCA SCOPE DEFINITION

The cradle-to-grave system boundary for LCA of %100 GU concrete (referred to hereafter as
100GU) is identified in Figure C3. The processes for 100GU represent the baseline system
boundary. Each process is comprised of input and output data such as energy use, raw material
use, and emissions to air, land, and water, that correspond to the activities described as follows.
The Electricity Grid Mix process, which is shown in Figure C3 as encompassing the other life
stages due to its important role in many processes, is modelled based on Ontario’s 2014
electricity grid mix and includes nuclear, hydroelectric, natural gas, and alternative means of
energy production. The Water Treatment process is the extraction and processing of water.
Cement Production is the extraction and transportation of raw materials, the manufacturing of
97
cement including blending, grinding, and pyroprocessing, and cement transportation. Aggregate
Production includes the extraction, processing, and transport of fine and coarse aggregates.
Concrete Plant Operations includes batching and mixing activities, and transport of concrete to
the site where it will be placed. Service Life includes maintenance activities, which are included
in this paper as additional quantities of materials and energy (estimated as 20%) that are required
over the entire life of the concrete. End of Life is the in-place rubblizing of concrete.

This system boundary is modified to model concrete with 10% cement replacement by fly ash
(GU-10FA in this paper) by adding a transportation process. Fly ash transportation is modeled
based on fly ash sources and modes of transportation that are typical for the MTO, where the
worst-case transportation scenario (937 km of diesel-powered truck transportation from Pleasant
Prairie, Wisconsin to Toronto, Ontario) was determined based on discussion with the MTO (D.
Rhead & H. Schell, personal communication, August 14, 2014). Based on a previous LCI, it is
assumed that 1.18 MJ/t-km are consumed for road transportation by diesel-powered truck [5].
The relevant factors for calculating emissions are shown in Table C2.

The system boundary is also modified based on the type of allocation selected for each of the
following four scenarios explored:

i. Baseline scenario with no allocation: no additional modifications are required

ii. Allocation of environmental impact by mass: a multifunctional coal combustion process
is added
iii. Allocation of environmental impact by economic value: a multifunctional coal
combustion process is added
iv. Allocation of benefits from disposal avoidance: a landfill avoidance process is added

The functional unit of this LCA is a cubic metre (m3) of each type of concrete (conventional
concrete containing 100% GU cement, and concrete containing 10% fly ash) with a water-to-
cement ratio of 0.4 (± 0.05). This functional unit is kept consistent in order to isolate the
influence of allocation modelling decisions. Mix designs used as inputs for this work, which

98
represent typical MTO mix designs (D. Rhead & H. Schell, personal communication, August 14,
2014) are shown in Table C3.

Four impact categories were selected- Acidification, Global Warming Potential, Resource
Depletion and Water Depletion. These categories were selected for the analysis based on i) their
relevance to the assessment of the environmental impact of concrete, as determined in
discussions with the MTO (D. Rhead & H. Schell, personal communication, December 4, 2014),
and ii) the level of international consensus on their classification and characterization that has
been reached, as described in Stranddorf and Schmidt [13]. As shown in Figure C3, the overall
LCA project is for concrete materials, and not solely fly ash production, and so the impact
categories were selected accordingly. Further information about these impact categories can be
found in Section 4.2.

4 METHODOLOGY

4.1 Allocation Scenarios

Four allocation scenarios are explored: i) baseline scenario- no allocation, ii) allocation of
environmental impact by mass, iii) allocation of environmental impact by economic value, iv)
allocation of benefits from landfill avoidance.

Where allocation is applied, LCI data for coal mining, preparation and combustion is sourced
from Babbitt & Lindner [14]. Although Babbitt & Lindner [14] present an LCI for electricity
production in Florida, it is deemed to be acceptable for use in this research due to the similarities
between production technologies and raw material sources in Canada and the U.S. Furthermore,
there is no known published LCI for coal combustion in Canada.

4.1.1 Baseline Scenario

In the baseline scenario, no allocation scenario is applied. This scenario is based on the
assumptions described in Table C1- namely, that i) adequate data is difficult to obtain and
validate, ii) fly ash is produced regardless of whether or not it is used in concrete, iii) the

99
resulting impact is likely to be small- and therefore, the environmental impact of its production
should not be included.

4.1.2 Mass Allocation Scenario

In this scenario, environmental impacts for coal mining, preparation and combustion are
allocated between fly ash and electricity based on Eq. 1 which gives the formula for the mass
allocation coefficient (Cm) [10]. In Eq. 1, fly ash is the by-product and electricity is the main
product, and mass quantities are represented using the symbol ‘m’. Note that ‘m total’ corresponds
to the sum of the mass of all products of the system, including the main product and all by-
products.

Eq.1

Cm = m by-product/m total

Mass quantities were calculated based on LCI data. Babbitt & Lindner [14] stated that for every
1000 kg of coal combusted, 216 kg of coal combustion products are formed, including 80 kg of
fly ash. Based on these values and the principle of conservation of mass, it was determined that
the electricity produced has a mass equivalent of 784 kg (similar to Chen et al. [10]). Using the
mass allocation coefficient equation, Cm is calculated as 0.093. This means approximately 9.3%
of the environmental impact of coal mining, preparation and combustion is attributable to the
production of fly ash, using this model of allocation.

4.1.3 Economic Allocation Scenario

In this scenario, environmental impacts for coal mining, preparation and combustion are
allocated between fly ash and electricity based on Eq. 2 which gives the formula for the
economic allocation coefficient (Ce) (Chen et al., 2010). In Eq. 2, fly ash is the by-product and
electricity is the main product and economic value in 2015 $CAN is represented by the symbol
‘$Value’. Note that ‘$Value total’ corresponds to the sum of the economic value of all products of
the system, including the main product and all by-products.

100
Eq. 2

Ce = ($Value) by-product/$Value total

The amount of electricity generated from coal burning is calculated based on LCI values [14] as
2689 kWh per tonne. Where required to calculate economic value, mass quantities were
calculated using the same methodology as described for the calculation of Cm. Electricity costs
were taken from Toronto Hydro [15]. It should be noted that in addition to any variation in
electricity price between jurisdictions or from to year to year, Toronto Hydro also has different
rates depending on consumption and time of use. To be conservative (with a higher C e), the off-
peak rate of $0.08/kWh (which is lower than the on-peak rate) was used in the calculation. The
value of fly ash in Ontario was determined to be approximately $120/tonne, with potential
variation due to transportation distances (S. Zupko (Lafarge), personal communication, July 3,
2015). The sensitivity of the results to changes in the allocation coefficients is discussed in
Section 5.3. Using the economic allocation coefficient equation, Ce is calculated as 0.043. This
means approximately 4.3% of the environmental impact of coal mining, preparation and
combustion is attributable to the production of fly ash, using this model of allocation.

4.1.4 Disposal Avoidance Scenario

In this scenario, environmental impacts for coal mining, preparation and combustion are not
included. This scenario represents the opposite end of the spectrum, where the environmental
benefits of avoiding disposal of fly ash are accounted for. LCI data for the disposal of fly ash
used for the calculations in this scenario is from Babbitt and Lindner (2005), who present an LCI
for the disposal of multiple coal combustion products. To separate the LCI data attributable to fly
ash, relative masses of the various coal combustion products are used. This method is chosen
because waste disposal is typically managed based on the mass of waste, and furthermore the
economics of disposal processes (ex. landfill fees within a specific jurisdiction) are often directly
linked to mass of waste products.

101
4.2 Life Cycle Impact Assessment Methodology

In order to compile the LCI data and conduct the impact assessment, the GaBi 6 software
package was used [16]. GaBi 6 is packaged with several impact assessment methods, including
the International Reference Life Cycle Data System (ILCD) method which was used for this
study [17]. As described in Section 3, four impact categories were selected. Table C4 presents a
description of each impact category as well as the characterization factor used in the analysis. In
LCA, characterization factors are science-based factors that are used to convert LCI data to
actual environmental impacts (Scientific Applications International Corporation, 2006).

4.3 Green Indicator Methodology

LCIA results may be normalized to allow for the calculation of relative impacts, and to allow for
the interpretation of results from impact categories with different units. In this work, results are
normalized to the results for 100GU, meaning that the numerical results for GU-10FA in each
impact category are divided by the numerical results for 100GU in each impact category. The
result is a unit-less value that is equal to 1 for 100GU. In any impact category, i) if the
normalized result is equal to 1, the material has equal environmental performance compared to
100GU, ii) if the normalized result is <1, the material has worse environmental performance
compared to 100GU, and iii) if the normalized result is >1, the material has better environmental
performance compared to 100GU.

The normalized results were then combined into a single green indicator score that combines the
acidification, global warming, resource depletion, and water depletion potentials. An initial
weighting of 0.25:0.25:0.25:0.25 was used to combine these categories, with alternative
weightings also tested to evaluate the sensitivity of the results. A higher green indicator indicates
a concrete with a higher environmental performance (and a lower environmental impact).

102
5 RESULTS AND DISCUSSION

5.1 LCIA Results

Table C5 shows the green indicator results for GU-10FA under the four allocation scenarios
described: i) Baseline Scenario, ii) Mass Allocation Scenario, iii) Economic Allocation Scenario,
and iv) Disposal Avoidance Scenario. As described in Section 4.3, the results are normalized to
100GU, which therefore has a score of 1. Since one of the objectives of using fly ash is to lower
the environmental impact of concrete compared to when 100% GU cement is used, a green
indicator value above one is desirable as it indicates a superior environmental performance. As
shown in Table C5, for all allocation scenarios the green indicator values are higher than 1,
indicating a higher (better) environmental performance in the combined four impact categories
than 100GU.

These results show that the Baseline Scenario and the Disposal Avoidance Scenario generate
near identical results. This is likely due to both the relatively small contribution of the disposal
stage to the fly ash life cycle impacts compared to coal combustion [14], and the relatively small
contribution of the fly ash processes to concrete life cycle impacts.

Allocation of environmental impacts from coal mining, preparation and combustion by both the
Mass Allocation and Economic Allocation Scenarios lowers the green indicator compared to the
Baseline Scenario. However, the drop in the Mass Allocation scenario compared to the Baseline
value is more pronounced (a difference of ~10%, whereas for Economic Allocation the
difference is ~5% and for Disposal Avoidance ~0%). These results seem to indicate that one of
the reasons commonly given for not including allocation in LCAs of concrete containing fly ash-
that the impact of fly ash production is relatively very small compared to the overall concrete life
cycle- is potentially not valid.

These results highlight the importance of rigorous assessment of different allocation methods, as
LCA results may be used to inform decision makers and the allocation methods used can result
in entirely different conclusions. In general, however, the results suggest that regardless of which

103
allocation scenario is applied, the environmental performance of fly ash concrete is still
improved compared to conventional concrete. These results are expanded on through sensitivity
analysis.

5.2 Sensitivity Analysis of Weighting

The life cycle impact assessment (LCIA) results for four impact categories- acidification, global
warming potential, resource depletion, and water depletion- are included in the calculation of the
green indicator. The results presented in Table C5 are for an equal weighting of these four
categories (Weighting Scheme 1 (0.25:0.25:0.25:0.25)). A sensitivity analysis of the weighting
scheme can confirm and elaborate on these results. Evaluating different weighting schemes can
also provide decision makers, who might prioritize the impact categories differently, with
context for interpreting the green indicator results.

Four alternative weighting schemes were assessed (Weighting Scheme 2 through 5). These
weighting schemes are shown in Table C6. These weighting schemes were each selected to
emphasize one impact category, which was given a weighting of 0.7 while the other three
categories were all weighted at 0.1. The purpose of selecting these weighting schemes was to
determine whether large changes in weighting of the impact categories affect the influence of the
choice of allocation method on the green indicator results for fly ash concrete.

Figure C4 shows the results using the five different weighting schemes. These results show that
generally, there is little change in the green indicator values for the Baseline and Disposal
Avoidance Allocation. For the Mass Allocation and Economic Allocation scenarios, which used
higher allocation coefficients (9.3% and 4.3% respectively), different weightings change the
results more significantly. This indicates that where allocation coefficients are higher, variations
due to changes in the weighting schemes are amplified. This suggests that in these scenarios
where allocation coefficients are higher, LCA practitioners should assess the sensitivity of their
results to alternative weighting schemes when single indicators are used.

104
Furthermore, Figure C4 shows that when the variability of the green indicator results with each
weighting scheme is taken into account, the results appear very similar for the Baseline and
Disposal Avoidance scenarios. This suggests that incorporating Disposal Avoidance has little
effect on LCIA results for LCA of fly ash concrete.

5.3 Sensitivity Analysis of Allocation Coefficients

One of the main criticisms of allocation is that the allocation coefficients can potentially be
highly variable. For mass allocation, for example, differences in raw materials and production
conditions and processes can lead to different amounts of materials being produced from facility
to facility, and even within the same facility from year to year [14]. For economic allocation,
there are many factors that influence the price of different products; two examples include supply
and demand dynamics and the influence of the changing value of the Canadian dollar on
transportation costs.

To examine the sensitivity of the results, additional allocation coefficients were modelled. Chen
et al [10] studied allocation of fly ash five years ago (in 2010) and in the context of France. The
mass allocation coefficient was calculated as 12.4%, while the economic allocation coefficient
was calculated as 1%. These numbers represent an approximately 3% increase and 3% decrease
compared to the mass and economic allocation coefficients presented in Section 4.1. The
resulting green indicators are plotted with the previously calculated green indicators and shown
in Figure C5. Figure C5 shows a linear relationship between allocation coefficient and green
indicator, which is expected since the fly ash content remains constant. A decrease of 3.1% in the
mass allocation coefficient (12.4 to 9.3%) resulted in a 3.0% increase in the green indicator
(from 0.93 to 0.97). An increase of 3.3% in the economic allocation coefficient (1 to 4.3%)
resulted in a 3.8% decrease in the green indicator (from 1.12 to 1.05). These results indicate that
changing both mass and economic allocation coefficients can have significant influence on the
results, even when fly ash production only represents a small input in the overall concrete life
cycle (10% of cement content). However, as results for both scenarios are equally variable these
results do not suggest whether one allocation scenario is more appropriate than the other. Rather,

105
this sensitivity analysis reinforces the common criticism that when allocation methods are
applied, results are sensitive to allocation coefficients that are highly variable.

6 CONCLUSIONS

In context with the use of fly ash in concrete, this paper demonstrates the importance of
comparing different allocation methods, including a ‘no allocation’ method. It is beneficial to
assess the influence of allocation approaches on the LCA results prior to finalizing an LCA
methodology. This paper compared four allocation methods in the specific context of an LCA of
fly ash concrete in Ontario. The conclusions of this paper are relevant to any LCA practitioner or
researcher modelling multifunctional processes.

 A review of the literature shows that allocation methods are not frequently applied when
fly ash is used in concrete because: i) there is lack of data, ii) fly ash is generated and
collected regardless of how it is used, and iii) the relative environmental impact of fly ash
processes are expected to be small compared to the overall concrete life cycle.
 Compared to the Baseline scenario, when the Disposal Avoidance Allocation scenario is
applied the change in environmental performance is very small (~0%). This aligns with
the common assumption that the relative environmental impact of fly ash processes is
expected to be small.
 Compared to the Baseline scenario, when the Economic Allocation and Mass Allocation
scenarios are applied the changes in environmental performance of the overall concrete
life cycle are more significant (~5% and ~10%, respectively). This is contrary to the
common assumption that the relative environmental impact of fly ash processes in
concrete LCA is expected to be small.
 When different weightings of impact categories are considered, the Mass and Economic
Allocation scenarios are more sensitive than the other scenarios considered. Conversely,
for the Baseline and Disposal Avoidance scenarios the results are generally stable even
when different weightings are applied. This indicates that the Mass and Economic
Allocation scenario leads to results that are sensitive to modelling decisions. This
106
 suggests that LCA practitioners should assess the sensitivity of their results to alternative
weighting schemes when single indicators are used.
 Based on these results, and an assessment of the sensitivity of the results to changes in
allocation coefficients, it appears that no allocation scenario is more valid than the others
in application to an LCA of fly ash concrete. This is because all of the allocation
coefficients vary with changes in parameters that are highly variable, such as market
value of goods, exchange rates, raw material input quality, and production processes.
Therefore, the Baseline scenario will be selected for future work as this will reduce the
uncertainty of the results by eliminating the need for additional data sources and
modelling assumptions.
 Future work should include studying other supplementary cementitious materials such as
ground granulated blast furnace slag and silica fume to determine if the Baseline scenario
is most appropriate for these materials as well.

7 ACKNOWLEDGEMENTS

The authors acknowledge the MTO HIIFP (2013-2015) program for support of this research.
This research was supported by a contribution from the Ministry of Transportation of Ontario.
Opinions expressed in this report are those of the authors and may not necessarily reflect the
views and policies of the Ministry of Transportation of Ontario.

107
8 BIBLIOGRAPHY

[1] ISO 14040, Life Cycle Assessment: Principals and Framework, ISO, 1997.

[2] B. Marsh, "High-volume fly ash concrete," Concrete, vol. 37, no. 4, pp. 54-55, April
2003.

[3] C. Shi, A. F. Jimenez and A. Palomo, "New cements for the 21st century: The pursuit of
an alternative to Portland cement," Cement and Concrete Research, vol. 41, pp. 750-763,
2011.

[4] K. R. O'Brien, J. Menache and L. M. O'Moore, "Impact of fly ash content and fly ash
transportation distance on embodied greenhouse gas emissions and water consumption in
concrete," International Journal of Life Cycle Assessment, vol. 14, pp. 621-629, 2009.

[5] Athena, "Cement and Structural Concrete Products: Life Cycle Inventory Update #2,"
Athena Sustainable Materials Institute, Ottawa, 2005.

[6] T. Garcia-Segura, V. Yepes and J. Alcala, "Life cycle greenhouse gas emissions of
blended cement concrete including carbonation and durability," International Journal of
Life Cycle Assessment, no. 19, pp. 3-12, 2014.

[7] D. Huntzinger and T. Eatmon, "A life-cycle assessment of Portland cement

manufacturing: comparing the traditional process with alternative technologies," Journal
of Cleaner Production, vol. 17, pp. 668-675, 2009.

[8] T. Ekvall and G. Finnveden, "Allocation in ISO 14041- a critical review," Journal of
Cleaner Production, vol. 9, pp. 197-208, 2001.

108
[9] International Organization for Standardization (ISO), Environmental management- Life
cycle assessment- Goal and scope definition and inventory analysis, Geneva: ISO, 1998.

[10] C. Chen, G. Habert, Y. Bouzidi, A. Jullien and A. Ventura, LCA allocation procedure
used as an incitative method for waste recycling: An application to mineral additions in
concrete, vol. 54, 2010, pp. 1231-1240.

[11] A.-M. Tillman, "Significance of decision-making for LCA methdology," Environmental

Impact Assessment Review, vol. 20, pp. 113-123, 2000.

[12] B. Weidema and J. Schmidt, "Avoiding co-product allocation in life-cycle assessment

revisited," Journal of Industrial Ecology, vol. 14, no. 2, pp. 192-195, 2010.

[13] H. Stranddorf and L. Schmidt, "Update on impact categories, normalisation and

weighting in LCA," Danish Ministry of the Environment Environmental Protection
Agency, 2005.

[14] C. Babbitt and A. Lindner, "A life cycle inventory of coal used for electricity production
in Florida," Journla of Cleaner Production, vol. 13, pp. 903-912, 2005.

[15] Toronto Hydro, "Electricity Rates and Charges," 1 May 2015. [Online]. Available:
https://www.torontohydro.com/sites/electricsystem/residential/yourbilloverview/Pages/El
ectricityRates.aspx. [Accessed 9 July 2015].

[16] PE International, GaBi 6.0, 2014.

[17] European Commission Joint Research Centre, "International Reference Life Cycle Data
System (ILCD) Handbook," Publications Office of the European Union, Luxemburg,
2011.

[18] M. A. Nisbet, M. L. Marceau and M. G. VanGeem, "Environmental Life Cycle Inventory

109
 of Portland Cement Concrete," Portland Cement Association, 2002.

[19] F. Collins, Inclusion of carbonation during the life cycle of built and recycled concrete:
influence on their carbon footprint, vol. 15, 2010, pp. 549-556.

[20] E. Crossin, "Comparative life cycle assessment of concrete blends," Centre for Design
RMIT University, Melbourne, 2012.

[21] K. Celik, C. Meral, A. Gursel, P. Mehta, A. Horvath and P. Monteiro, "Mechanical

properties, durability, and life-cycle assessment of self-consolidating concrete mixtures
made with blended portland cements containing fly ash and limestone powder," Cement
and Concrete Composites, vol. 56, pp. 59-72, 2015.

[22] M. Dolatabadi, Properties and Performance of Photocatalytic Concrete, Toronto, ON:

Library and Archives Canada, 2013.

110
List of Tables- Appendix C

Table C1 Literature review of fly ash concrete LCI and LCA studies

Table C2 Relevant factors for the calculation of fly ash transportation energy consumption
and emissions [5]

Table C3 Mix designs

Table C4 Impact categories

Table C5 Green indicator results for GU-FA10 using four allocation scenarios.

Table C6 Weighting schemes

 Table C1 Literature review of fly ash concrete LCI and LCA studies
Source Functional Unit
Nisbet et al. [18] 3 FUs: 1 m3 of ready mixed concrete,
100 concrete masonry units, 1 m3 of
precast concrete
Athena [5] 1 m3 of 15, 20 or 30 MPa ready
mixed concrete; precast double ‘T’
beam, precast hollow deck, standard
concrete blocks, cement mortar
Huntzinger & 1 ton of cement x 4 types: traditional, United States No allocation
Eatmon [7] blended (natural pozzolans), cement +
sequestration in waste materials,
cement + recycled CKD
Chen et al. [10] Binding equivalent value of fly ash,
slag, and CEM 1
Collins [19] Concrete bridge (life of 100 years,
then crushed and recycled in bridge)
Crossin [20] 1 m3 of AS 1379:2007 40 MPa
concrete for 50 years
Garcia-Segura et al. Reinforced concrete building column
[6] (3x0.3x0.3m) during its lifetime
Celik et al. [21] 1m3 of ready mixed SCC concrete
Geographical Are the Upstream Production Impacts Allocated
Context to Fly Ash?
(Reason Given)
United States No allocation
(Upstream profiles of SCMs not included)
Canada Environmental benefit of cement replacement
included
No allocation of upstream production impacts
(No data, minor distortion)
(Assumed no associated environmental impact,
produced regardless of use, most likely
insignificant)
France Three allocation scenarios explored: no
allocation, mass allocation, and economic
allocation
Australia No allocation
Australia System expansion used for fly ash, credit for
avoided waste treatment
Spain No allocation
(Collected for air quality regardless of use and
needs little processing)
United States Major material production and preparation
processes for fly ash included
Not clear what allocation method was used if any
Table C2 Relevant factors for the calculation of fly ash transportation energy consumption and emissions [5]
Energy Emissions Factors [kg/GJ]
CO2 SO2 NOx VOC CH4 CO
70.7 0.102 0.807 0.0869 0.0217 0.443
 Table C3 Mix designs
Mix Design ID
Quantity (kg/m3) 100GU* GU-10FA
Water 155 128
GU Cement 300 316
SL 53 -
FA - 35
Coarse agg. 1064 1037
Fine agg. 779 777
* Base case scenario [22]
 Table C4 Impact categories
Impact Category Description Characterization Factor
Release of protons and leaching out of anions from a Acidification Potential
Acidification
system (mole H+ ion equivalents)
Effect of increasing temperature in the atmosphere due to
Global Warming Potential
Global Warming the reflection of radiation by greenhouse gases back to the
(kg CO2 equivalents)
surface
Resource Depletion of abiotic non-water resources (ex. minerals or Resource Depletion Potential (kg of
Depletion fossil fuels) antimony (Sb) equivalents)
Water use (reuse, degradation, or incorporation into final Water Depletion Potential (volume
Water Depletion
product) of water used (m3))
Table C5 Green indicator results for GU-FA10 using four allocation scenarios.
Scenarios Normalized* Green Indicator
Baseline 1.28
Mass Allocation 1.15
Economic Allocation 1.22
Disposal Avoidance 1.28
*Normalized to 100GU
 Table C6 Weighting schemes
Weighting Scheme Acidification Global Warming Potential Resource Depletion Water Depletion
1 0.25 0.25 0.25 0.25
2 0.7 0.1 0.1 0.1
3 0.1 0.7 0.1 0.1
4 0.1 0.1 0.7 0.1
5 0.1 0.1 0.1 0.7
List of Figures- Appendix C

Figure C1 Multifunctional process diagram

Figure C2 Graphical representations of a) Subdivision Method, b) System Expansion

Method, c) Mass Allocation Method, and d) Economic Allocation Method.

Figure C3 System boundary for LCA of concrete products.

Figure C4 Green indicator results for GU-10FA for five Weighting Schemes and four
allocation scenarios.

Figure C5 Sensitivity of green indicator results for GU-10FA with equal weighting (W1) to
allocation coefficient.
Figure C1 Multifunctional process diagram
 a)

b)

c)

d)

Figure C2 Graphical representations of a) Subdivision Method, b) System Expansion Method, c) Mass Allocation
Method, and d) Economic Allocation Method.
Figure C3 System boundary for LCA of concrete products.
 1.4

1.2

1
Green Indicator

0.8

0.6

0.4

0.2

0
Baseline Mass Allocation Economic Disposal Avoidance
Allocation Allocation
Allocation Scenarios

W1 W2 W3 W4 W5

Figure C4 Green indicator results for GU-10FA for five Weighting Schemes and four allocation scenarios.
 1.4
Green Indicator, Normalized to 100GU

1.2

R² = 0.998
1

0.8

0.6

0.4

0.2

0
0 2 4 6 8 10 12 14
Allocation Coefficient (%)

Figure C5 Sensitivity of green indicator results for GU-10FA with equal weighting (W1) to allocation coefficient.
APPENDIX D
IMPACT OF THE SELECTION OF FUNCTIONAL UNIT ON THE LIFE
CYCLE ASSESSMENT OF GREEN CONCRETE

Karina E. Setoa, Daman K. Panesara*, Cameron J. Churchillb

a
Department of Civil Engineering, University of Toronto, 35 St. George St., Toronto, Ontario, Canada, M5S 1A4
b
Engineering and Society Program, McMaster University, 1280 Main St W, Hamilton , Ontario , Canada, L8S 4L7
*
Corresponding author. Email: d.panesar@utoronto.ca

Acknowledgement

The authors acknowledge the MTO HIIFP (2013-2015) program for support of this research.
This research was supported by a contribution from the Ministry of Transportation of Ontario.
Opinions expressed in this report are those of the authors and may not necessarily reflect the
views and policies of the Ministry of Transportation of Ontario.

111
ABSTRACT

Purpose: The importance of the selection of an appropriate functional unit in the context of LCA
of concrete materials is explored by examining the sensitivity of LCA results to the choice of
functional unit. Methods: Six functional units with varying degrees of complexity are included
in the analysis: i) volume of concrete, ii) 28-day compressive strength, iii) 28-day rapid chloride
permeability (RCP), iv) binder intensity; defined as amount of cementing materials (kg) per unit
of 28-day compressive strength (MPa), v) combination of compressive strength and RCP, and vi)
combination of binder intensity and RCP. The 28-day compressive strength and 28-day RCP are
used as measures of concrete functional performance in terms of strength and durability
respectively. LCA models are created based on three concrete mix designs containing slag, silica
fume, and limestone cement as cement replacement, and one concrete mix design for
conventional concrete. Results: LCIA results are generated and then interpreted as
Environmental Efficiency Indicators (EEIs) calculated for each concrete type, using each of the
six functional units. For the same concrete mix designs, EEI results vary significantly (up to
845%) depending on the choice of functional unit. Conclusions: Choice of functional unit can
significantly influence the value of EEIs. Despite the variation, mix designs containing slag,
limestone cement and silica fume show a consistently improved environmental efficiency over
the base case (100% general use cement concrete) for all functional units. Concrete containing
silica fume has the highest EEI results and shows the most significant variability for different
functional units due to relatively better functional performance as measured by lower RCPT
values and higher compressive strength values.

KEYWORDS

Functional unit, life cycle assessment, concrete, compressive strength, rapid chloride
permeability, supplementary cementitious materials, limestone cement

112
1 INTRODUCTION

The importance of the selection of an appropriate functional unit in the context of life cycle
assessment (LCA) of concrete materials is explored by examining the sensitivity of LCA results
to the choice of functional unit. An LCA model is established based on four mix designs
containing conventional concrete constituents and varying amounts of silica fume, ground
granulated blast furnace slag, and limestone cement; these mix designs are commonly used by
the Ministry of Transportation of Ontario for transportation infrastructure. Six different
functional units are explored, which incorporate functional performance requirements of
concrete, including 28-day compressive strength and rapid chloride permeability test (RCPT)
value. A methodology to combine LCA results with each of the proposed functional units is
proposed, and the resulting indicator is the Environmental Efficiency Indicator (EEI). The
sensitivity of the EEIs for each mix design to the choice of functional unit is assessed.

2 BACKGROUND

For the past several decades, one strong motivation for the development and implementation of
alternative concrete mix designs other than conventional concrete has been an effort to reduce
both cost and negative environmental impacts. Supplementary cementing materials (SCMs) are
commonly used to improve the environmental, economic and in some cases technical
performance of concrete products. Examples of SCMs include silica fume (SF), which is
produced as a by-product of silicon production, and ground granulated blast furnace slag (SL),
which is produced as a by-product of steel production. Another approach is to replace general
use cement (GU) with cement containing additional limestone (GUL). Both the use of SCMs and
the use of GUL propose to reduce the environmental impact of concrete materials by reducing
the amount of cement in a given concrete mix design. Cement is the most environmentally
intensive component of concrete, with its production accounting for approximately 6-7% of
global CO2 emissions (Shi, Jimenez, & Palomo, 2011) and requiring approximately 3 GJ of
energy per ton of clinker (Van den Heede & De Belie, 2012).

Life cycle assessment (LCA), which is defined as “the compilation and evaluation of the inputs,
outputs and potential environmental impacts of a product system throughout its life cycle” (ISO
113
14040, 2006), is a useful way to environmentally model the complex processes that are included
in the life cycle of a concrete product. An important step in defining the scope of an LCA is to
select the functional unit for the analysis. A functional unit “defines the quantification of the
identified functions (performance characteristics) of the product” (ISO 14040, 2006).

The purpose of the functional unit is to provide a basis for the quantification of all inputs and
outputs (i.e. a reference point for which data is collected), and to allow for comparison of LCA
results based on equivalent functional performance of different processes or products. The
challenge for practitioners is how to select a functional unit when a product or process has
multiple identified functions, and ultimately how to determine the influence of such a choice on
the results. Despite the fundamental importance of the functional unit in LCA studies, the ISO
standards that relate to LCA (ISO 14040 and ISO 14041) provide little guidance on how to
resolve these challenges.

ISO 14041 (1998) states that functional units shall be i) consistent with the goal and scope of the
study and ii) clearly defined and measurable. ISO 14041 (1998) also briefly refers to scenarios
where more than one function is performed by a system, stating that when two systems are
compared, it must be ensured that the functional units for each system account for the same
number of functions. The ISO standards, however, are meant to serve as a framework and not as
explicit guidelines for LCA studies. As a result, they provide little information about how to
identify the most appropriate functional unit, or include more than one functional unit in the
analysis, when a system performs more than one function.

Published concrete LCA studies shown in Table D1 provide a summary of a variety of types of
functional units that have been used in these studies. This review organizes the LCA studies into
three categories based on the level of complexity of the functional unit, where i) simple
functional units are based on one variable, ii) moderately complex functional units are based on
two variables, and iii) complex functional units are based on three or more variables.

Based on this review, the following observations were made with respect to functional units in
concrete LCAs:

114
  Despite reference to a ‘functional’ unit, many LCAs use units of analysis that don’t
correspond to a performance characteristic of concrete. Often, physical characteristics
such as mass or volume are used (as in Brown et al., 2014 and Lee & Park, 2005). This
occurs most often in studies that use simple functional units (based on one variable).
 Compressive strength (typically 28-day) is the most consistently applied functional unit
that reflects a performance characteristic (as in Athena, 2005 and Prusinski et al., 2004).
 Many functional units are tied to a specific type of structural element and/or a specific
environmental exposure (as in Kawai et al., 2005 and Knoeri et al., 2013).

Functional units selected for use in LCA must be aligned with the useful functions of the product
in question; therefore, in the context of transportation infrastructure in Ontario, two common
performance characteristics of concrete materials are described. Concrete is a versatile material
that can serve many functions. Most commonly, concrete is specified based on compressive
strength and durability requirements that vary depending on the application and on the type of
materials included in the mix designs.

Plain (unreinforced) concrete has a relatively high compressive strength and relatively low
tensile strength. In structural applications, it is typically reinforced with steel bars. The steel
provides tensile strength capacity, while the concrete surrounding the steel protects it from
substances such as chlorides that might cause corrosion.

Ontario Provincial Standards (2014) has developed a Material Specification for Concrete –
Materials and Production (OPSS 1350 (2014)) which provides a methodology for evaluating the
compressive strength of concrete and is used for Ministry of Transportation of Ontario (MTO)
projects. This includes casting concrete cylinders for compressive strength testing for
acceptance. When an individual compressive strength test falls below the specified minimum 28-
day compressive strength by more than 4.0 MPa, as specified in OPSS 1350, additional testing of
cores extracted from the concrete may be required. For both cylinders and cores, compressive
strength is measured through destructive tests according to CSA-A23.2-14C-14 (2014) at
laboratories pre-qualified by the MTO.

115
Concrete durability is a key aspect of its functional performance, particularly for reinforced
concrete applications where the protection of reinforcing steel from corrosion is critical to
maintaining the integrity of the structural system. Durability is multi-dimensional, and for
concrete applications can include resistance to damage from abrasion, freeze-thaw damage,
sulphate attack, and chloride ingress. Often, durability of concrete is correlated to the
permeability of the materials, as the transport of deleterious substances through concrete is a
critical issue that may negatively influence the service life of concrete (Boddy, Bentz, Thomas,
& Hooton, 1999).

The rapid chloride permeability test (RCPT) is a common test that primarily measures
conductivity as an indicator of permeability through concrete. In Canada the ASTM C 1202
(2012) test is the standard. MTO requires that cores be removed for the evaluation of rapid
chloride permeability in high performance concrete and SF overlay applications, and RCPT may
also be required for other concrete as specified.

3 GOAL AND SCOPE DEFINITION

3.1 LCA Goal

The importance of the selection of an appropriate functional unit in the context of LCA of
concrete materials is explored by examining the sensitivity of LCA results generated based on
six functional units:

i. Volume of concrete;
ii. 28-day compressive strength;
iii. 28-day RCP;
iv. Binder intensity, defined as amount of cementing materials per unit of 28-day
compressive strength;
v. Combination of compressive strength and RCP; and
vi. Combination of binder intensity and RCP.

116
The choice of functional unit is the main variable explored in this work and the objective of this
paper is to assess the sensitivity of LCA results, interpreted as Environmental Efficiency
Indicators (EEIs), to the selection of different functional units.

3.2 LCA Scope Definition

3.2.1 Mix Designs

Four concrete mix designs are considered, three of which contain cement replacements (SL, SF
and GUL) that have historically been used by the MTO in transportation infrastructure. The mix
designs and their corresponding experimental results for 28-day compressive strength and RCPT
are shown in Table D2. The mix design in the leftmost column, designated 100GU, is taken as
the base case as it represents a conventional concrete mix design with general use cement and no
SCMs. There are two types of cement used in these mixtures: general use cement (GU), and
GUL with a limestone content of approximately 12%.

The cradle-to-grave system boundary for LCA of conventional concrete, which includes GU
cement, fine and coarse aggregate, and water, is identified in Figure D1. The processes for
conventional concrete represent the baseline system boundary. Each process is comprised of
input and output data such as energy use, raw material use, and emissions to air, land, and water,
that correspond to the activities described as follows. The Electricity Grid Mix process, which is
shown in Figure D1 as encompassing the other life stages due to its important role in many
processes, is modelled based on Ontario’s 2014 electricity grid mix and includes nuclear,
hydroelectric, natural gas, and alternative means of energy production. The Water Treatment
process is the extraction and processing of water. Cement Production is the extraction and
transportation of raw materials, the manufacturing of cement including blending, grinding, and
pyroprocessing, and cement transportation. Aggregate Production includes the extraction,
processing, and transport of fine and coarse aggregates. Concrete Plant Operations includes
batching and mixing activities, and transport of concrete to the site where it will be placed.
Service Life includes maintenance activities, which are included in this paper as additional
quantities of materials and energy (estimated as 20%) that are required over the entire life of the
concrete. End of Life is the in-place rubblizing of concrete.

117
This system boundary is modified to model concrete containing SF, SL and GUL as follows:

 To model SF, a SF transportation process is added. The mode of transportation is

assumed to be road transportation by diesel-powered truck. Based on a previous LCI
(Athena, 2005), it is assumed that 1.18 MJ/t-km are consumed for road transportation by
diesel-powered truck. The relevant factors for calculating emissions are shown in Table
D3. The upstream production of SF is not included in the analysis, as it is a waste product
that would be produced regardless of whether it is used in concrete. Using SF in concrete
is a better environmental choice than landfilling, so excluding these upstream activities is
a conservative assumption.
 To model SL, a SL transportation process is added using the same methodology as for SF
transportation. In addition, a processing stage is added as raw slag typically requires
further grinding before it is incorporated into concrete (Prusinski et al., 2004). The
upstream production of SL is not included in the analysis; as with SF this is a
conservative assumption.
 To model GUL, GUL production was substituted for GU production. Transportation of
raw materials is not included as cement production facilities are typically located at
limestone quarries (Canada Centre for Mineral & Energy Technology and Radian Canada
Inc., 1993).

MTO utilizes materials from a variety of pre-qualified sources that are compiled into a database
titled Designated Sources for Materials (Ministry of Transportation of Ontario, 2014). In this
study, all transportation distances are taken as worst-case — that is, the transportation scenario
with the highest environmental impact was selected. Where the modes of transportation are the
same from scenario to scenario (i.e. all by truck), the worst-case scenario corresponds to the
longest transportation distance. Selecting the worst-case transportation scenario for the analysis
is a conservative assumption. The transportation scenarios are presented in Table D4.

118
4 METHODOLOGY

4.1 Life Cycle Impact Assessment (LCIA) Methodology

In order to compile the LCI data and conduct the impact assessment, the GaBi 6 software
package is used (PE International, 2014). GaBi 6 is packaged with several impact assessment
methods, including the International Reference Life Cycle Data System (ILCD) method which
was used for this study (European Commission Joint Research Centre, 2011). Four impact
categories- Acidification, Global Warming Potential, Resource Depletion and Water Depletion-
were selected for their analysis based on MTO’s assessment of priority (H. Schell and D. Rhead,
personal communication, December 5 2015) and the level of international consensus on the
classification and characterization that has been reached for each category, as described in
Stranddorf and Schmidt [14]. Table D5 presents a description of each impact category as well
as the characterization factor used in the analysis. In LCA, characterization factors are science-
based factors that are used to convert LCI data to actual environmental impacts (Scientific
Applications International Corporation, 2006).

4.2 Green Indicator Methodology

The life cycle impact assessment results are factored by dividing the results for 100GU, the base
case mix design, by those for the alternative materials as shown in Eq. 1. The results for 100GU
are placed on the numerator, because a higher LCIA result (ex. higher acidification, higher
global warming potential) is less desirable. Therefore, a factored result that is greater than one
corresponds to a material that is less environmentally impactful than 100GU.

In Eq. 1 and 2 that follow, the index ‘i’ refers to the four impact categories included in this study,
as indicated in Table D1. The index ‘j’ references each of the four mix designs included in this
study, as indicated in Table D2. The index ‘x’ references each of the six functional units included
in this study, as indicated in Table D6. These equations therefore present a methodology for
calculating GIs (and in Section 4.4, EEIs) that incorporate LCA results from all impact
categories and are specific to each mix design and functional unit.

Eq. 1

119
 𝐿𝐶𝐼𝐴 𝑟𝑒𝑠𝑢𝑙𝑡𝑖𝑚𝑝𝑎𝑐𝑡 𝑐𝑎𝑡𝑒𝑔𝑜𝑟𝑦 𝑖, 𝑏𝑎𝑠𝑒 𝑚𝑖𝑥 𝑑𝑒𝑠𝑖𝑔𝑛
𝐹𝑎𝑐𝑡𝑜𝑟𝑒𝑑 𝑟𝑒𝑠𝑢𝑙𝑡𝑖,𝑗 =
𝐿𝐶𝐼𝐴 𝑟𝑒𝑠𝑢𝑙𝑡𝑖𝑚𝑝𝑎𝑐𝑡 𝑐𝑎𝑡𝑒𝑔𝑜𝑟𝑦 𝑖, 𝑎𝑙𝑡𝑒𝑟𝑛𝑎𝑡𝑖𝑣𝑒 𝑚𝑎𝑡𝑒𝑟𝑖𝑎𝑙 𝑗
where i= 1 to 4 and corresponds to impact category
and j = 1 to 3 and corresponds to alternative materials

These normalized results are then combined into a single green indicator score that combines the
acidification, global warming, resource depletion, and water depletion potentials, as shown in
Eq. 2. An initial equal weighting of 0.25:0.25:0.25:0.25 is used to combine these categories.
Alternative weighting schemes are also tested to evaluate the sensitivity of the results, as
discussed in Section 5.2. A higher green indicator indicates a concrete with a higher
environmental performance (and a lower environmental impact).

Eq.2

𝑊𝑒𝑖𝑔ℎ𝑡𝑒𝑑 𝑟𝑒𝑠𝑢𝑙𝑡𝑗 = ∑ 𝐹𝑎𝑐𝑡𝑜𝑟𝑒𝑑 𝑟𝑒𝑠𝑢𝑙𝑡𝑖,𝑗 × 𝑤𝑒𝑖𝑔ℎ𝑡𝑖𝑛𝑔

𝑖=1

where i= 1 to 4 and corresponds to impact category

and j = 1 to 3 and corresponds to alternative mix designs

4.3 Functional Unit Calculation

Although the green indicators calculated using the methodology of Section 4.2 provide some
indication of the relative environmental performance of the materials, they do not give insight
into the functional performance of the materials. In this section, the methodology for calculating
functional units is described. Equations are given for the calculation of the six functional units.
These equations are then used to calculate EEIs, as shown in Section 4.4, by multiplying the
weighted LCIA result by the functional unit. Six functional units are considered which vary in
complexity as a result of the number of variables that are simultaneously considered. These are
summarized in Table D6.

4.3.1 FU1: Volume (1 m3 of Concrete)

Volume is not a truly functional unit, as it provides information only about the physical
properties of a particular product, rather than the performance of that product. However, it is
included here in recognition that despite these limitations it is simple to employ and is one of the
most common functional units employed in cement and concrete LCAs (as shown in Table D1).

120
Eq. 3 shows how FU1 is calculated based on the relative volumes of the base material and each
alternative concrete material that is modelled. Note that the base material is on the numerator;
this is because a higher volume of concrete is worse for the environment due to the additional
materials required. Therefore, if the alternative concrete requires more material it should lower
the EEI.

Eq. 3

𝑣𝑜𝑙𝑢𝑚𝑒 (𝑚3 )𝑏𝑎𝑠𝑒 𝑚𝑎𝑡𝑒𝑟𝑖𝑎𝑙

𝐹𝑈 1,𝑗 = =1
𝑣𝑜𝑙𝑢𝑚𝑒 (𝑚3 ) 𝑎𝑙𝑡. 𝑚𝑎𝑡𝑒𝑟𝑖𝑎𝑙 𝑗

where j corresponds to each alternative mix design, as described in Table D2

4.3.2 FU2: Volume and Compressive Strength

Compressive strength is one of the most important performance characteristics of concrete, and
28-day compressive strength is a value that is commonly specified and evaluated to ensure the
quality of concrete products. As a result, it is an appropriate functional unit for this work. Eq. 4
shows how FU2 is calculated based on the relative 28-day compressive strength values of the
base material and each alternative concrete material that is modelled. Note that the value for the
alternative concrete material is on the numerator; this is because a higher compressive strength is
a more desirable performance characteristic for concrete. Concrete typically provides
compressive strength capacity in structural systems and using higher strength concrete has the
potential to both improve functional performance and reduce the amount of concrete needed to
obtain the same performance (De Schepper et al., 2014). Therefore, if the alternative concrete
has a higher compressive strength, this should raise the EEI.

Eq. 4

𝑠𝑡𝑟𝑒𝑛𝑔𝑡ℎ (𝑀𝑃𝑎)𝑎𝑙𝑡. 𝑚𝑎𝑡𝑒𝑟𝑖𝑎𝑙 𝑗

𝐹𝑈 2,𝑗 =
𝑠𝑡𝑟𝑒𝑛𝑔𝑡ℎ (𝑀𝑃𝑎) 𝑏𝑎𝑠𝑒 𝑚𝑎𝑡𝑒𝑟𝑖𝑎𝑙

where j corresponds to each alternative mix design, as described in Table D2

121
4.3.3 FU3: Volume and Durability
As with compressive strength, durability is a critical performance characteristic of concrete,
particularly when the concrete is steel reinforced, or when it is exposed to harsh conditions such
as impact, freeze-thaw cycling, or the presence of ions such as chlorides or sulphates. 28-day
RCPT is a value that is commonly specified and evaluated to ensure the quality of concrete
products. As a result, it is an appropriate functional unit for this work. Eq. 5 shows how FU3 is
calculated based on the relative 28-day RCPT values of the base material and each alternative
concrete material that is modelled. Note that the value for the base material is on the numerator;
this is because a higher RCPT value represents a worse performance for concrete. Therefore, if
the alternative concrete has a lower RCPT value, this should raise the EEI.

Eq. 5

𝑅𝐶𝑃𝑇(𝐶)𝑏𝑎𝑠𝑒 𝑚𝑎𝑡𝑒𝑟𝑖𝑎𝑙
𝐹𝑈 3,𝑗 =
𝑅𝐶𝑃𝑇 (𝐶)𝑎𝑙𝑡. 𝑚𝑎𝑡𝑒𝑟𝑖𝑎𝑙 𝑗

where j corresponds to each alternative mix design, as described in Table D2

4.3.4 FU4: Volume and Binder Intensity

Binder intensity is the ratio of the mass of binder material in a concrete mix design (kg) to the
compressive strength of the concrete (MPa), for 1 m3 of concrete. This combines physical
characteristics (volume and mass) with a performance characteristic (compressive strength). Eq.
4 shows how FU5 is calculated based on the relative binder intensities of the base material and
the alternative concrete material. Note that the value for the base material is on the numerator;
this is because a higher volume of concrete is worse for the environment due to the additional
materials required to achieve a unit quantity of 28-day compressive strength. Therefore, if the
alternative concrete requires more material it should lower the EEI.

Eq. 6

𝑏𝑖𝑛𝑑𝑒𝑟 𝑖𝑛𝑡𝑒𝑛𝑠𝑖𝑡𝑦 (𝑘𝑔/𝑀𝑃𝑎)𝑏𝑎𝑠𝑒 𝑚𝑎𝑡𝑒𝑟𝑖𝑎𝑙

𝐹𝑈 4,𝑗 =
𝑏𝑖𝑛𝑑𝑒𝑟 𝑖𝑛𝑡𝑒𝑛𝑠𝑖𝑡𝑦(𝑘𝑔/𝑀𝑃𝑎) 𝑎𝑙𝑡. 𝑚𝑎𝑡𝑒𝑟𝑖𝑎𝑙 𝑗

where j corresponds to each alternative mix design, as described in Table D2

122
4.3.5 FU5: Volume, Compressive Strength and Durability
FU5 represents one of the most complex types of functional units, as it combines three variables,
two of which are important performance characteristics for concrete materials. This allows the
functional unit to capture multiple dimensions of these materials, which are representative of
acceptance criteria. Eq. 7 illustrates how FU5 is calculated by multiplying together the values for
FU2 and FU3. As a result, the functional unit, and the resulting EEI, has its highest value when
the strength value is higher and the RCPT value is lower.

Eq. 7

𝐹𝑈 5,𝑗 = 𝐹𝑈2,𝑗 × 𝐹𝑈3,𝑗

where j corresponds to each alternative mix design, as described in Table D2

4.3.6 FU6: Volume, Binder Intensity and Durability

FU6 represents one of the most complex types of functional units, as it combines four variables,
two of which are important performance characteristics for concrete materials. This allows the
functional unit to capture multiple dimensions of these materials, which are representative of
acceptance criteria. Eq. 8 illustrates how FU6 is calculated by multiplying together the values for
FU3 and FU4. As a result, the functional unit, and the resulting EEI, has its highest value when
both the binder intensity (kg of cementing material per MPa of 28-day compressive strength) and
the RCPT value are lower.

Eq. 8
𝐹𝑈 6,𝑗 = 𝐹𝑈3,𝑗 × 𝐹𝑈4,𝑗

where j corresponds to each alternative mix design, as described in Table D2

4.4 Environmental Efficiency Calculation

Following the calculations outlined in Sections 4.2 and 4.3, EEIs can be calculated as shown in
Equation 8. For every alternative concrete material (j), and for each functional unit (x), a specific
EEI can be calculated by multiplying together the weighted result, calculated using Eq. 2, and the
functional unit, calculated using one of Eq. 3 to 8.

Eq. 8

123
 𝐸𝐸𝐼𝑗,𝑥 = 𝑊𝑒𝑖𝑔ℎ𝑡𝑒𝑑 𝑟𝑒𝑠𝑢𝑙𝑡𝑗 × 𝐹𝑈𝑥,𝑗

where j corresponds to each alternative mix design, as described in Table D2

and where x corresponds to each functional unit, as described in Table D6

5 RESULTS AND DISCUSSION

5.1 Environmental Efficiency Indicator Results

EEI results for all four mix designs, using the six functional units and an equal weighting of
impact categories, are presented in Figure D2. As discussed in Section 4, 100GU represents the
base case scenario and all EEIs are calculated normalized to the results of that mix design. This
is why for all functional units, 100GU has a value of 1.

The results show that for all functional units, the incorporation of SL, of GUL and of SF
improves environmental efficiency compared to the base case, as the EEI values are all above 1.
It is also observed that generally, the mix containing SF and SL (GU-8SF-25SL) has the highest
environmental efficiency (highest EEI) of all the mix designs included in this study. One
exception is the EEI calculated using FU1, where the environmental performance is slightly
lower for GU-8SF-25SL compared to the mix containing GUL and SL (GUL-25SL). This is
likely due to the transportation of materials for SF. It should also be noted, however, that in this
outlying case, FU1 is a physical characteristic functional unit (volume). When performance
characteristics are taken into account, as for FU2-FU6, it is clear that the concrete mix containing
SF has a superior environmental performance, as measured with the EEI.

It is useful to examine the functional unit results in more detail to explain the large increase in
environmental efficiency of GU-8SF-25SL. As shown in Table D6, for FU3, FU5 and FU6 in
particular the functional units are approximately four times higher for GU-8SF-25SL than for the
other materials included in this study. As FU3 is calculated based on RCPT results, FU5 is
calculated based on compressive strength results combined with RCPT results, and FU6 is
calculated based on binder intensity combined with RCPT results, the superior durability
performance of GU-8SF-25SL (as demonstrated in Table D2) is the reason for the large increase
in environmental efficiency for these two functional units. These results show the importance of

124
using functional units that incorporate performance characteristics, and where relevant, multiple
performance characteristics, and are not limited to physical characteristics.

Table D7 shows the percentage difference for each mix design between the EEI results for each
functional unit, as compared to the results for FU1. It can be seen that the EEIs vary significantly
(at least ~10% for every mix design). For GU-8SF-25SL the EEIs vary by up to 845% (the
results for FU6 compared to the results for FU1). This highlights the importance of the selection
of the functional unit, in particular for concrete with unusually high performance characteristics
such as concrete containing SF.

5.2 Weighting

As described in Section 4.2, the life cycle impact assessment (LCIA) results for four impact
categories- acidification, global warming potential, resource depletion, and water depletion- are
included in the calculation of the green indicator. The results presented in Figure D2 are for an
equal weighting of the different impact categories (Weighting Scheme 1 (0.25:0.25:0.25:0.25)).
A sensitivity analysis of the weighting scheme can confirm and elaborate on these results.
Evaluating different weighting schemes can also provide decision makers, who might prioritize
the impact categories differently, with context for interpreting the green indicator results. As
noted by Bengtsson and Steen, “weighting is not meant to deliver the final verdict about the
environmental performance […] it is meant to give an additional input into the process” (2000,
p. 101). In this study, it is acknowledged that the results may be of interest to individuals and
groups with different views on the value of certain environmental systems and processes.
Therefore, this paper attempts to address the issue of weighting by applying a systematic
sensitivity analysis to confirm and elaborate on the results presented in Figure 3 (Weighting
Scheme 1).

Four alternative weighting schemes were assessed (Weighting Scheme 2 through 5). These
weighting schemes are shown in Table D8. These weighting schemes were each selected to
emphasize one impact category, which was given a weighting of 0.7 while the other three
categories were all weighted at 0.1. The purpose of selecting these weighting schemes was to

125
determine whether large changes in weighting of the impact categories affect the relative
environmental performance, as measured by EEIs, of the concrete mix designs.

The result of this sensitivity analysis was that the general trends discussed in Section 5.1, in
terms of the relative performance of the materials and the variation with different functional
units, did not change. That is, GU-25SL, GUL-25SL, and GU-8SF-25SL all had improved
environmental efficiency compared to the base case 100GU, and GU-8SF-25SL had the highest
environmental efficiency compared to 100GU, with the exception of when FU1 (volume) is used
as the functional unit. Table D9 shows the percent difference between the Weighting Scheme 1
result and the results for each alternative weighting scheme, for each concrete mix design. It can
be seen that in general, the results changed very little (<4%). Based on this analysis, these results
are not sensitive to the weighting scheme. Changing the weighting scheme to reflect the
priorities of different stakeholders, as described by Bengtsson & Steen (2000), is in this case
unnecessary as it does not provide any additional insight into these results.

6 CONCLUSIONS

This paper demonstrates the importance of comparing different functional units, including
assessing their influence on LCA results, before selecting a final LCA methodology. This paper
compared six functional units in the specific context of an LCA of concrete containing GU,
GUL, SL, and SF in Ontario. The conclusions of this paper are relevant to any LCA practitioner
or researcher modelling cement and concrete materials.

 Varying the functional unit and using functional units that incorporate performance
characteristics can significantly influence the value of EEIs for each concrete mix design
relative to the value of EEIs with FU1 of 1 m3 (ex. the results for FU6 were 845% higher
than the results for FU1). This is because of the significant variability in functional
performance for various mix designs depending on their constituents.
 Despite the variation with different functional units, mix designs containing SL, GUL and
SF have been shown to have a consistently improved environmental efficiency over the
base case (100GU) for all functional units. This is because of the reduction in cement

126
 content due to cement replacement, improvements in durability (all other mix designs
compared to 100GU), and improvement in strength (GU-8SF-25SL).
 In addition, the concrete mix design containing SL and SF (GU-8SF-25SL) has been
shown to generally have the highest environmental efficiency of the four mix designs
included in this analysis. This is likely because of its high functional performance
(highest compressive strength and lowest RCP). One exception is for FU1 (1 m3 of
concrete). This suggests that to accurately assess the environmental efficiency of concrete
materials it is important to select more sophisticated functional units, which incorporate
performance characteristics specific to the product under study (concrete), such as
compressive strength and durability as measured by RCPT.
 The EEI results were assessed for their sensitivity to the choice of weighting scheme. It
was determined that the relative ranking of the materials in terms of environmental
performance did not change, even when weighting schemes were changed significantly.

127
7 Bibliography

[1] C. Shi, A. F. Jimenez and A. Palomo, "New cements for the 21st century: The pursuit of an
alternative to Portland cement," Cement and Concrete Research, vol. 41, pp. 750-763, 2011.

[2] P. Van den Heede and N. De Belie, "Environmental impact and life cycle assessment (LCA)
of traditional and 'green' concretes: Literature review and theoretical calculations," Cement
& Concrete Composites, vol. 34, pp. 431-442, 2012.

[3] International Organization for Standardization, " Environmental management- Life cycle
assessment- Principles and framework," ISO, Geneva, 2006.

[4] International Organization for Standardization (ISO), Environmental management- Life

cycle assessment- Goal and scope definition and inventory analysis, Geneva: ISO, 1998.

[5] Ontario Provincial Standards, Material Specification for Concrete - Materials and
Production, Ontario: Ontario Provincial Standards, 2014.

[6] Canadian Standards Association, "Obtaining and testing drilled cores for compressive
strength testing," 2014.

[7] A. Boddy, E. Bentz, M. Thomas and R. Hooton, "An overview and sensitivity study of a
multimechanistic chloride transport model," Cement and Concrete Research, no. 29, pp.
827-837, 1999.

[8] ASTM International, Standard Test Method for Electrical Indication of Concrete's Ability to
Resist Chloride Ion Penetration, West Conshohocken, PA: ASTM International, 2012.

[9] Athena, "Cement and Structural Concrete Products: Life Cycle Inventory Update #2,"
Athena Sustainable Materials Institute, Ottawa, 2005.

[10] Canada Centre for Mineral & Energy Technology and Radian Canada Inc., "Raw Material

128
 Balances, Energy Profiles, and Environmental Unit Factor Estimates: Cement and Structural
Concrete Products," Athena Sustainable Materials Institute, Ottawa, 1993.

[11] Ministry of Transportation of Ontario, "Hydraulic Cements and Supplementary Cementing

Materials," 1 December 2014. [Online]. Available:
http://applications.roadauthority.com/mpl/mpl.asp?MPIShortName=MTO+DSM. [Accessed
7 August 2015].

[12] PE International, GaBi 6.0, 2014.

[13] European Commission Joint Research Centre, "International Reference Life Cycle Data
System (ILCD) Handbook," Publications Office of the European Union, Luxemburg, 2011.

[14] H. Stranddorf and L. Schmidt, "Update on impact categories, normalisation and weighting
in LCA," Danish Ministry of the Environment Environmental Protection Agency, 2005.

[15] M. Bengtsson and B. Steen, "Weighting in LCA- Approaches and applications,"

Environmental Progress, vol. 19, no. 2, pp. 101-109, 2000.

[16] J. E. Anderson and R. Silman, "A life cycle inventory of structural engineering design
strategies for greenhouse gas reduction," Structural Engineering International, pp. 283-288,
March 2009.

[17] D. Brown, R. Sadiq and K. Hewage, An overview of air emission intensities and
environmental performance of grey cement manufacturing in Canada, vol. 16, 2014, pp.
1119-1131.

[18] K. Celik, C. Meral, A. Gursel, P. Mehta, A. Horvath and P. Monteiro, "Mechanical

properties, durability, and life-cycle assessment of self-consolidating concrete mixtures
made with blended portland cements containing fly ash and limestone powder," Cement and
Concrete Composites, vol. 56, pp. 59-72, 2015.

[19] C. Chen, G. Habert, Y. Bouzidi and A. Jullien, Environmental impact of cement production:

129
 detail of the different processes and cement plant variability evaluation, vol. 18, 2010a, pp.
478-485.

[20] C. Churchill and D. Panesar, "Life-cycle cost analysis of highway noise barriers designed
with photocatalytic cement," Structure and Infrastructure Engineering: Maintenance,
Management, Life-Cycle Design and Performance, vol. 9, no. 10, pp. 983-998, 2013.

[21] D. Huntzinger and T. Eatmon, "A life-cycle assessment of Portland cement manufacturing:
comparing the traditional process with alternative technologies," Journal of Cleaner
Production, vol. 17, pp. 668-675, 2009.

[22] K.-M. Lee and P.-J. Park, "Estimation of the environmental credit for the recycling of
granulated blast furnace slag based on LCA," Resources, Conservation and Recycling, no.
44, pp. 139-151, 2005.

[23] S. Marinkovic, V. Radonjanin, M. Malesev and I. Ignjatovic, "Comparative environmental

assessment of natural and recycled aggregate concrete," Waste Management, vol. 30, pp.
2255-2264, 2010.

[24] M. A. Nisbet, M. L. Marceau and M. G. VanGeem, "Environmental Life Cycle Inventory of

Portland Cement Concrete," Portland Cement Association, 2002.

[25] N. Tosic, S. Marinkovic, T. Dasic and M. Stanic, Multicriteria optimization of natural and
recycled aggregate concrete for structural use, vol. 87, 2015, pp. 766-776.

[26] S. Xing, Z. Xu and G. Jun, "Inventory analysis of LCA on steel- and concrete-construction
office buildings," Energy and Buildings, no. 40, pp. 1188-1193, 2008.

[27] C. Chen, G. Habert, Y. Bouzidi, A. Jullien and A. Ventura, LCA allocation procedure used
as an incitative method for waste recycling: An application to mineral additions in concrete,
vol. 54, 2010, pp. 1231-1240.

[28] M. Hassan, "Life-cycle assessment of titanium dioxide coatings," in Building a Sustainable

130
 Future – Proceedings of the 2009 Construction Research Congress, Seattle, 2009.

[29] A. Jonsson, T. Bjorklund and A.-M. Tillman, "LCA of concrete and steel building frames,"
International Journal of Life Cycle Assessment, vol. 3, no. 4, pp. 216-224, 1998.

[30] C. Knoeri, E. Sanye-Mengual and H.-J. Althaus, "Comparative LCA of recycled and
conventional concrete for structural applications," The International Journal of Life Cycle
Assessment, vol. 13, no. 5, pp. 909-918, June 2013.

[31] C. Li, S. Cui, Z. Nie, X. Gong, Z. Wang and N. Itsubo, "The LCA of portland cement
production in China," International Journal of Life Cycle Assessment, vol. 20, pp. 117-127,
2015.

[32] J. R. Prusinski, M. L. Marceau and M. G. VanGeem, "Life Cycle Inventory of Slag Cement
Concrete," in Eighth CANMET/ACI Eighth CANMET/ACI International Conference on Fly
Ash, Silica Fume, Slag and Natural Pozzolans in Concrete, 2004.

[33] F. Collins, Inclusion of carbonation during the life cycle of built and recycled concrete:
influence on their carbon footprint, vol. 15, 2010, pp. 549-556.

[34] E. Crossin, "Comparative life cycle assessment of concrete blends," Centre for Design
RMIT University, Melbourne, 2012.

[35] M. De Schepper, P. Van den Heede, I. Van Driessche and N. De Beile, "Life cycle
assessment of completely recyclable concrete," Materials, no. 7, pp. 6010-6027, 2014.

[36] T. Garcia-Segura, V. Yepes and J. Alcala, "Life cycle greenhouse gas emissions of blended
cement concrete including carbonation and durability," International Journal of Life Cycle
Assessment, no. 19, pp. 3-12, 2014.

[37] K. Kawai, T. Sugiyama, K. Kobayashi and S. Sano, "Inventory data and case studies for
environmental performance evaluation of concrete structure construction," Journal of
Advanced Concrete Technology, vol. 3, no. 3, pp. 435-456, October 2005.

131
[38] K.-H. Yang, Y.-B. Jung, M.-S. Cho and S.-U. Tae, "Effect of supplementary cementitious
materials on reduction of CO2 emissions from concrete," Journal of Cleaner Production,
pp. 1-10, 2014.

[39] B. L. Damineli, F. M. Kemeid, P. S. Aguiar and V. M. John, "Measuring the eco-efficiency

of cement use," Cement and Concrete Composites, vol. 32, pp. 555-562, 2010.

[40] R. Wang, M. Eckelman and J. Zimmerman, "Consequential environmental and economic

life cycle assessment of green and gray stormwater infrastructures for combined sewer
systems," Environmental Science & Technology, vol. 47, pp. 11189-11198, 2013.

[41] M. Dolatabadi, Properties and Performance of Photocatalytic Concrete, Toronto, ON:

Library and Archives Canada, 2013.

132
List of Tables- Appendix D

Table D1 Review of functional units in known cement and concrete LCAs

Table D2 Mix designs and hardened properties used as input for the LCA and functional unit
calculations

Table D3 Relevant factors for the calculation of transportation energy consumption and
emissions

Table D4 Worst-case MTO transportation scenarios for

Table D5 Impact categories

Table D6 Calculated functional units

Table D7 Percentage difference between EEIs for FUs incorporating performance characteristics
vs. FU1 (volume)

Table D8 Weighting schemes

Table D9 Variation in EEI results with various impact category weighting schemes
 Table D1 Review of functional units in known cement and concrete LCAs

Source Functional Unit

I. SIMPLE FUNCTIONAL UNIT (SINGLE VARIABLE)
Anderson & Silman (2009) Six floor steel or concrete building
Brown et al. (2014) 1 tonne of cement
Celik et al. (2015) 1 m3 of ready mixed self-consolidating concrete
Chen et al. (2010a) 1 kg of cement
Churchill & Panesar (2013) 1 noise barrier system- 9 panels (3500 mm x 500 mm x 132 mm each)
Huntzinger & Eatmon (2009) 907 kg (2000 lbs) of cement x 4 types: traditional, blended (natural pozzolans), cement
+ sequestration in waste materials, cement + recycled CKD
Lee & Park (2005) 1 kg of cement clinker, cement, slag powder, silicate fertilizer
Marinkovic et al. (2010) 1 m3 of ready-mixed natural aggregate & recycled aggregate concrete
Nisbet et al. (2002) 1 m3 of ready mixed concrete, 100 concrete masonry units, 1 m3 of precast concrete
Tosic et al. (2015) 1 m3 of ready mixed natural and recycled aggregate concrete
Xing et al. (2008) 1 m2 building area, steel and concrete office structures
II. MODERATELY COMPLEX FUNCTIONAL UNIT (2 VARIABLES)
Athena (2005) 1 m3 of 15, 20 or 30 MPa ready mixed concrete; precast double ‘T’ beam, precast
hollow deck, standard concrete blocks, cement mortar
Chen et al. (2010) 1 kg of binding equivalent fly ash, slag, and CEM 1
Hassan (2009) 5 mm thick surface mix of ultrafine TiO2 anatase, cement, filler, and water
(0.1:1:3:0.4) applied on 1 lane-kilometre of pavement surface
Jonsson et al. (1998) 1 m2 floor area for seven frame cases over the lifetime of a building
Knoeri et al. (2013) 1 m3 of a specific strength class of concrete (lean, indoor and outdoor)
Li et al. (2015) 1 tonne of Portland cement, 1 tonne of Portland cement with 42.5 MPa of strength
grade
Prusinski et al. (2004) 1 m3 and 1 yd3 of 20 and 35 MPa ready mixed concrete, 50 and 70 MPa precast
concrete, concrete block mix
III. COMPLEX FUNCTIONAL UNIT (MORE THAN THREE VARIABLES)
Collins (2010) Concrete bridge (40 MPa, 100 yr life + recycling, CO2 carbonation included)
Crossin (2012) 1 m3 of AS 1379:2007 40 MPa concrete for 50 years
De Schepper et al. (2014) Mass of concrete for 1 MPa of strength and one year of service life
Garcia-Segura et al. (2014) 25 MPa reinforced concrete building column (3x0.3x0.3m) during its lifetime
Kawai et al. (2005) Pre-stressed concrete bridge, overflow dike of dam, retaining wall, secondary lining in
tunnel with specific strength requirements and service lives
Yang et al. (2014) Binder intensity and CO2 intensity concepts (Damineli, Kemeid, Aguiar, & John, 2010)
in terms of the unit strength (1 MPa) of concrete
Wang et al. (2013) 4047 m2 of permeable pavement & other types of runoff systems sized for runoff
associated with 2.5 cm of rainfall generated over a 0.4 hectare (1 acre) watershed
Table D2 Mix designs and hardened properties used as input for the LCA and functional unit calculations

Mix Design ID
*j= - 1 2 3
Quantity (kg/m3) 100GU** GU-25SL GUL-25SL GU-8SF-25SL
Water 180 155 150 140
GU 429 300 - 276
GUL - - 300 -
SL - 53 54 100
SF - - - 24
Coarse aggregate 900 1064 1071 1000***
Fine aggregate 875 779 781 800***
Air entraining 52 - - -
admixture (mL/kg of
cementing material)
28 day compressive 48.5 36 34.7 56.5
strength (MPa)
28 day RCPT (C) 3186 2135 1969 421
* Index values corresponding to Eq. 1, 2 and 8.
** Base case scenario (Dolatabadi, 2013)
*** Aggregate quantities are estimated based on typical MTO proportions.
Table D3 Relevant factors for the calculation of transportation energy consumption and emissions (Athena, 2005)

Energy Emissions Factors [kg/GJ]

CO2 SO2 NOx VOC CH4 CO
70.7 0.102 0.807 0.0869 0.0217 0.443
 Table D4 Worst-case MTO transportation scenarios for SL and SF

Material Route Total Distance (km) Mode of Transportation

SL Hamilton, ON to Picton, ON 498 Truck
Picton, ON to Toronto, ON
SF Niagara Falls, NY to St Basile, QC 1623 Truck
St Basile, QC to Toronto, ON
 Table D5 Impact categories

*i= Impact Category Description Characterization Factor

1 Release of protons and leaching out of anions from a Acidification Potential
Acidification
system (mole H+ ion equivalents)
2 Effect of increasing temperature in the atmosphere Global Warming Potential
Global Warming due to the reflection of radiation by greenhouse
gases back to the surface (kg CO2 equivalents)

3 Resource Depletion of abiotic non-water resources (ex. Resource Depletion Potential (kg
Depletion minerals or fossil fuels) of antimony (Sb) equivalents)
4 Water use (reuse, degradation, or incorporation into Water Depletion Potential
Water Depletion
final product) (volume of water used (m3))
* Index values corresponding to Eq. 1 and 2.
 Table D6 Calculated functional units

Variables Included Type of Functional Unit Functional Units

*x= GU-25SL GUL-25SL GU-8SF-25SL
1 FU1 Volume Simple 1 1 1
2 FU2 Volume & Compressive Moderately complex 0.74 0.72 1.16
Strength
3 FU3 Volume & RCPT Moderately complex 1.49 1.62 7.57
4 FU4 Volume, binder content, Complex 0.90 0.87 1.25
and strength
5 FU5 Volume, compressive Complex 1.11 1.16 8.82
strength & RCPT
6 FU6 Volume, binder content, Complex 1.35 1.40 9.45
strength & RCPT
* Index values corresponding to Eq. 8.
Table D7 Percentage difference between EEIs for FUs incorporating performance characteristics vs. FU 1 (volume)

Percentage Difference (%)

GU-25SL GUL-25SL GU-8SF-25SL
FU2 -25.8 -28.5 16.5
FU3 49.2 61.8 657
FU4 -9.88 -13.4 24.8
FU5 10.8 15.8 782
FU6 34.5 40.2 845
 Table D8 Weighting schemes

Weighting Scheme Acidification Global Warming Potential Resource Depletion Water Depletion
1 0.25 0.25 0.25 0.25
2 0.7 0.1 0.1 0.1
3 0.1 0.7 0.1 0.1
4 0.1 0.1 0.7 0.1
5 0.1 0.1 0.1 0.7
 Table D9 Variation in EEI results with various impact category weighting schemes

% Difference Compared to Weighting Scheme 1

Weighting Scheme GU-25SL GUL-25SL GU-8SF-25SL
2 -0.581 -0.653 -1.51
3 2.56 3.56 3.45
4 0.321 0.246 0.417
5 -2.30 -3.16 -2.35
List of Figures- Appendix D

Figure D1 System boundary for LCA of general use cement concrete products

Figure D2 EEI results for four mix designs using six functional units
Figure D1 System boundary for LCA of general use cement concrete products
Figure D2 EEI results for four mix designs using six functional units
APPENDIX E
LIFE CYCLE ASSESSMENT AND DEVELOPMENT OF
ENVIRONMENTAL EFFICIENCY INDICATORS FOR CONCRETE
CONSTITUENTS

Karina E. Setoa, Daman K. Panesara*, Cameron J. Churchillb,

a
Department of Civil Engineering, University of Toronto, 35 St. George St., Toronto, Ontario, Canada, M5S 1A4
b
Engineering and Society Program, McMaster University, 1280 Main St W, Hamilton , Ontario , Canada, L8S 4L7
*
Corresponding author. Email: d.panesar@utoronto.ca

KEYWORDS

Compressive strength (C), durability (C), concrete (D), silica fume (D), life cycle assessment

133
ABSTRACT

Life Cycle Assessment (LCA) is applied to develop a methodology for calculating

Environmental Efficiency Indicators (EEI) for concrete products. Eight mix designs containing
conventional concrete constituents as well as varying amounts of i) ground granulated blast
furnace slag, ii) fly ash, iii) silica fume, iv) general use limestone cement, v) photocatalytic
cement, and vi) recycled aggregate are included in this study. An LCA model is established
based on inputs and outputs over the life cycle of a concrete product. Hardened properties (28-
day compressive strength and rapid chloride permeability testing) are incorporated to calculate
EEIs for each mix design, relative to the base case of conventional concrete. The results show
that SCMs and limestone cement improve EEI compared to the base case. Photocatalytic cement
improves EEI only if SCMs are incorporated. Concrete containing recycled aggregate has a
lower (worse) EEI compared to the base case.

1 INTRODUCTION

In this paper, the concepts of LCA are applied to develop a methodology for calculating
Environmental Efficiency Indicators (EEIs) for concrete products. An LCA model is established
based on the constituents of concrete and eight mix designs. A methodology for combining the
LCA outputs with the functional performance of the concrete products, as measured by the
hardened properties corresponding to each mix design, is proposed. The result is the
development of EEIs, which allow for the comparison of diverse concrete materials on the basis
of both life cycle environmental impact and functional performance.

2 BACKGROUND

The production of cement-based products is responsible for a significant portion of global CO2
emissions. In particular 6-7% of global CO2 emissions [1] can be attributed to the production of
cement, which requires 3 GJ of energy per ton of clinker [2] and is the component of concrete
that has the highest environmental impact. For the past several decades, one strong motivation
for the development and implementation of alternative concrete mix designs other than
conventional concrete has been an effort to reduce both cost and negative environmental impacts.
134
These efforts have resulted in the proliferation of concrete mix designs in the literature and in the
industry that are claimed to be ‘greener’ or more sustainable than traditional concrete mix
designs.

2.1 Alternative Concrete Constituents

Table E1 provides examples of several approaches used to achieve ‘green’ concrete and
describes the six alternative concrete constituents included in this LCA: i) ground granulated
blast furnace slag (SL), fly ash (FA), silica fume (SF), limestone cement (GUL), recycled
aggregate concrete (RAC), and photocatalytic concrete (PCAT). Table E1 includes a brief
summary of some of the literature on compressive strength, durability, and environmental
performance, as compared to conventional concrete, which contains GU cement and virgin
aggregate. Compressive strength and durability performance are included in this review as they
are key indicators of concrete functional performance. Plain (unreinforced) concrete has a
relatively high compressive strength and relatively low tensile strength. In structural
applications, concrete is typically reinforced with steel, which provides tensile strength capacity,
while the surrounding concrete protects the steel from substances such as chlorides that might
cause corrosion. Concrete durability, including resistance to damage from abrasion, freeze-thaw
damage, sulphate attack, and chloride ingress, is therefore often critical to maintaining the
integrity of a structural system. Often, concrete durability is correlated to the permeability of the
materials, and the potential for the transport of deleterious substances through concrete. The
rapid chloride permeability test (RCPT) is a common test that measures conductivity as an
indicator of permeability through concrete.

2.2 LCA of Cement and Concrete

Life cycle assessment (LCA), which is defined as “the compilation and evaluation of the inputs,
outputs and potential environmental impacts of a product system throughout its life cycle” [3], is
a useful way to environmentally model the complex processes that are included in the life cycle
of a concrete product, and rigorously assess different mix designs. There are four stages in an
LCA, which guide the structure of this paper: i) Goal and Scope Definition, ii) Life Cycle
Inventory, iii) Life Cycle Impact Assessment, and iv) Interpretation.

135
Known published concrete LCA studies have been reviewed, and are summarized in Table A1
(appended). Some key outcomes of this literature review with regards to scope, functional unit,
and LCA results in previous published concrete LCA studies are presented:

Scope

 Often, the scope of concrete material LCAs is limited to ‘cradle-to-gate’ analysis (i.e. as
in [4, 5, 6]). This means that the scope is limited to the production of either cement or a
specific concrete product. Less commonly, LCAs may be ‘cradle-to-grave’, which
includes all life stages, or ‘cradle-to-cradle’, which includes the recycling of all materials
at the end of life of a product.
 There has not been a full LCA that has focused on the Canadian context, or more
specifically on the province of Ontario.
 Typically the upstream profiles of admixtures and supplementary cementing materials
(SCMs) are not included within the scope of concrete LCAs [7, 6]

Functional unit

 Many studies use physical characteristics of concrete such as volume or mass (ex. 1 m3 or
1 kg) as the functional unit [8, 9]. This neglects important functional requirements for
concrete, such as strength and durability.

LCA results

 The energy required in these systems (both electricity grid and other fuels) is often a key
contributor to environmental impact [5].
 The production of cement (and in particular pyroprocessing) is typically the process that
contributes the most to the overall environmental impact of cement and concrete products
[1]

2.3 Environmental Metrics

There is substantial literature on the development of metrics that evaluate environmental

performance. A correspondingly large number of indices have been proposed; some are more
general and some are developed for highly specific applications.

Rametsteiner et al. [10, p. 62] argues that the role of sustainability indicators “is to structure and
communicate information about key issues and their trends considered relevant for sustainable
development”; the authors acknowledge that the development of sustainability indicators extends

136
beyond technical calculations and necessarily includes political or philosophical value
judgements about how sustainability is defined and measured. In light of this complexity,
Moldan et al. [11] argue that it is not the calculation of indicators that is the challenge, but rather,
the selection of appropriate indicators and the proper application and interpretation of these
metrics. Fernandez-Sanchez & Rodriguez-Lopez [12] also argue that the many ISO standards
that have been developed with respect to sustainability indicators serve as a framework, rather
than a methodology for creating and selecting indicators. This work shows that there is a need
for straightforward methodologies for developing sustainability or green indicators, and that
these must be clearly linked to the context of a specific project and set of stakeholder priorities.

Several indicators which are specific to cement-based materials have been proposed. Hanson UK
[13], a cement manufacturer, has developed environmental targets for the year 2020 which take a
multi-dimensional approach to sustainability that includes four main aspects: i) ‘people’, ii)
‘carbon’, iii) ‘waste and raw materials’, and iv) ‘water and biodiversity’. Fernandez-Sanchez &
Rodriguez-Lopez [12] propose engaging stakeholders (including engineers, managers, and
contractors) through an ‘analytic hierarchy process’ that allows them to rank sustainability
priorities in a ‘sustainable breakdown structure’. Zhong & Wu [14] define environmental
sustainability of concrete based on four primary factors: i) material reduction such as by using
recycled materials and/or reuse of structural elements, ii) CO2 emission/energy consumption
during construction, iii) water consumption during construction, and iv) noise pollution during
construction. It is clear that in the concrete industry there is an understanding that environmental
performance is multi-dimensional, and that metrics must capture this aspect.

There has been less development of indicators that incorporate functional performance. This is
perhaps because the functional performance of concrete products can vary widely based on the
materials used, the type of application, and the environmental exposure, amongst other factors.
Pons & de la Fuente [15] address this issue by using a Multi-Criteria Decision Making model to
develop a Sustainability Index which incorporates the cross-sectional shape, compressive
strength, and compacting system (mechanical vibration or use of self-compacting concrete) of
structural concrete columns. Their work highlights the intersections between these variables and
environmental performance, but does not account for durability performance- which is critical

137
for many concrete elements, in particular those exposed to the outdoor environment. Evidently,
there are many diverse approaches to the measurement and evaluation of concrete environmental
performance.

3 LCA GOAL AND SCOPE DEFINITION

3.1 Goal

In this paper EEIs are proposed as metrics for the assessment of both environmental and
functional performance. This paper presents a methodology for LCA and the development of
EEIs for concrete materials. The objective of this paper is to critically compare six alternative
concrete constituents based on their environmental and functional performance.

3.2 Scope

3.2.1 Mix Designs

This paper includes concrete mix designs that are used by the Ministry of Transportation of
Ontario (MTO). MTO utilizes concrete that contains SCMs (SL, FA, and SF) and GUL, which
replaces general use cement, in transportation infrastructure. Data from the literature was also
used to model two more novel concrete constituents that are of interest to MTO and are the
subject of trials: PCAT concrete [16] and RAC [17]. Eight mix designs are shown in Table E2.
The first mix design, designated 100GU, is taken as the base case. There are two types of cement
used in these mixtures: general use cement (GU), and general use limestone cement (GUL) with
a limestone content of approximately 12%.

3.2.2 System Boundary and Assumptions

Figure E1 presents the cradle-to-grave system boundary for conventional concrete (indicated in
this paper as 100GU), which is the base case mix design for this work, and includes Portland
cement, fine and coarse aggregate, and water. The processes for conventional concrete represent
the baseline system boundary. Each process is comprised of input and output data such as energy
use, raw material use, and emissions to air, land, and water, that correspond to the activities
described as follows. The Electricity Grid Mix process, which is shown in Figure E1 as

138
encompassing the other life stages due to its important role in many processes, is modelled based
on Ontario’s 2014 electricity grid mix and includes nuclear, hydroelectric, natural gas, and
alternative means of energy production. The Water Treatment process is the extraction and
processing of water. Cement Production is the extraction and transportation of raw materials, the
manufacturing of cement including blending, grinding, and pyroprocessing, and cement
transportation. Aggregate Production includes the extraction, processing, and transport of fine
and coarse aggregates. Concrete Plant Operations includes batching and mixing activities, and
transport of concrete to the site where it will be placed. End of Life is the in-place rubblizing of
concrete. This system boundary is modified to model the alternative concrete constituents as
described in Section 4: LCA Life Cycle Inventory.

3.2.3 Functional Unit

The purpose of the functional unit is to provide a basis for the quantification of all inputs and
outputs (i.e. a reference point for which data is collected), and to allow for comparison of LCA
results based on equivalent functional performance of different processes or products. Most
commonly concrete is specified based on strength and durability requirements that vary
depending on the application and on the type of materials included in the mix designs and the
geographic location (as it directly relates to exposure conditions). As MTO typically specifies
concrete used for transportation infrastructure according to compressive strength (MPa) and
durability (typically measured in Coulombs (C) using the rapid chloride permeability test
(RCPT)) requirements, these two properties are included in the functional unit. Table E3 presents
the experimental results for compressive strength and RCPT corresponding to the mix designs
shown in Table E2. Equation 1 describes how the functional unit for this study is calculated.

Eq. 1

𝑠𝑡𝑟𝑒𝑛𝑔𝑡ℎ (𝑀𝑃𝑎)𝑎𝑙𝑡. 𝑚𝑎𝑡𝑒𝑟𝑖𝑎𝑙 𝑗 𝑅𝐶𝑃𝑇(𝐶)𝑏𝑎𝑠𝑒 𝑚𝑎𝑡𝑒𝑟𝑖𝑎𝑙

𝐹𝑈𝑗 = ∗
𝑠𝑡𝑟𝑒𝑛𝑔𝑡ℎ (𝑀𝑃𝑎) 𝑏𝑎𝑠𝑒 𝑚𝑎𝑡𝑒𝑟𝑖𝑎𝑙 𝑅𝐶𝑃𝑇 (𝐶)𝑎𝑙𝑡. 𝑚𝑎𝑡𝑒𝑟𝑖𝑎𝑙 𝑗

where j corresponds to each alternative mix design, as described in Table E2

The strength value for the base material is on the denominator because, in general, a higher
compressive strength is a more desirable performance characteristic for structural concrete.
139
Concrete typically provides compressive strength capacity in structural systems and using higher
strength concrete has the potential to both improve functional performance and reduce the
amount of concrete needed to obtain the same performance [18]. Therefore, if a mix design
containing alternative concrete constituents has a higher compressive strength, this should raise
the EEI. Under the same logic, for durability, the value for the base material is on the numerator;
this is because a higher RCPT value indicates a concrete that is more permeable and therefore
more vulnerable to the ingress of harmful chemicals such as sulphates or chlorides. A higher
RCPT value, therefore, represents a lower durability performance for concrete. Therefore, if a
mix design containing alternative concrete constituents has a lower RCPT value compared to the
base case, this should raise the EEI. As a result, the functional unit, and the resulting EEI, has its
highest value when the strength value is higher and the RCPT value is lower. This methodology
for calculating the functional unit corresponds with the scenario that stronger and more durable
concrete is better for the environment, also as described in Van den Heede & De Belie [2].

4 LCA Life Cycle Inventory

The Life Cycle Inventory (LCI) for each type of concrete, representing the inputs and outputs
throughout the product’s life cycle, was compiled based on a baseline LCI for conventional
concrete containing GU cement and virgin aggregate (represented in this paper by the mix
100GU), which was then modified to model each constituent.

4.1 Processes Included in the Life Cycle of Conventional Concrete

4.1.1 Cement Production Process

A report released by the Athena Sustainable Materials Institute [4] contains detailed LCI data for
the production of cement. The raw material, energy use, atmospheric emissions, liquid effluent,
and water usage data presented in the Athena report was used to create the Cement Production
process in all of the LCA models. This dataset provides an overview of the Canadian Portland
cement industry, and is based on data from specific regions of the industry. The data for the
Central Region of Canada (Manitoba and Ontario) was used for this research, as it is most
relevant to the jurisdiction of the MTO.

140
4.1.2 Electricity Grid Mix Process
The electricity grid mix process is based on the Canadian electricity grid mix process from the
GaBi Extension database XVII: Full US, which includes biogas, biomass, hard coal, heavy fuel
oil, hydropower, natural gas, nuclear, photovoltaics, wind and waste-to-energy. To increase the
geographical and temporal correlation to the scope of this work, the ratio of the individual types
of generation was modified to reflect the supply mix of Ontario in 2014 (shown in Figure E2).

4.1.3 Water Treatment Process

Racoviceanu et al. [19] used the economic input-output life-cycle assessment model and the
GHGenius model to quantify the energy use and greenhouse gas emissions for the City of
Toronto municipal water treatment system. Several processes were considered including
chemical manufacturing, chemical transportation, and water treatment facility operation.
Although the scope of the study was somewhat limited, given that only energy use and
greenhouse gas emissions were considered, a review of the National Pollutant Release Inventory
(NPRI) has shown that there are no emissions recorded for the four Toronto water treatment
plants (which are: R.C. Harris, R.L. Clark, F.J. Horgan, and Island). This suggests that none of
the 363 NPRI substances (including NOx, SO2 or particulate matter) are released [20].

4.1.4 Aggregate Processes

There are four aggregate processes developed for the models used in this research: i) fine
aggregate production, ii) coarse aggregate production, iii) fine aggregate transportation, and iv)
coarse aggregate transportation. To build each of these processes, data for energy requirements,
raw material requirements, atmospheric emissions and liquid effluents were collected for the
extraction, processing, and transportation of the fine and coarse aggregates [4]. Based on a
previous LCI [4], it is assumed that 1.18 MJ/t-km are consumed for road transportation by
diesel-powered truck. Emissions factors for road transportation by truck are taken from Athena
[4] and are shown in Table E4. For all materials the worst-case transportation scenario,
corresponding to the route that has the highest environmental impact, are shown in Table E5. For
this work, where all the transportation processes use the same mode (truck), the route with the
longest distance indicates the worst-case transportation scenario.

141
4.1.5 Concrete Production Process
The concrete production process represents the stage in the life cycle where the various concrete
‘ingredients’, including water, cement, fine aggregate and coarse aggregate, are combined in
order to create concrete. The mix proportions are entered into the parameterized model according
to the mix designs described in Table E2. In addition to the raw material requirements, the
batching and mixing processes require energy; Prusinski, Marceau & VanGeem [21] calculated a
value of 247 MJ for the production of a cubic metre of concrete (this value is consistent even
with the addition of SCMs).

4.1.6 Concrete Maintenance Process

Concrete maintenance activities are highly dependent on the specific concrete application and
environmental conditions that the concrete is exposed to over its life. In terms of life cycle
assessment, however, they can be represented as additional quantities of materials and energy
that are required over the entire life of the concrete. As operational and cost considerations are
not within the scope of this work, it is appropriate to combine all maintenance activities into a
single activity. In order to estimate these impacts, the concept of a Maintenance Factor was used,
where a Maintenance Factor value of 20% means that over the entire life of the concrete, 20%
more of the materials and energy required for initial placement must be input in order to maintain
the concrete. This value was selected in consultation with MTO (D. Rhead & H. Schell, personal
communication, Dec. 4 2014) and is applied to all of the mix designs included in this study.
Although detailed information about maintenance processes could inform Maintenance Factors
that are specific to each given mix design, these data were unavailable for this work and
therefore the 20% approximation was applied.

4.1.7 End-of-Life Process

‘End of life’ is the final stage in a cradle-to-grave LCA. Potential disposal options for concrete
products include recycling, landfill, and reuse. Discussions with the MTO (D. Rhead & H.
Schell, personal communication, Dec. 4 2014) revealed that MTO infrastructure is typically
crushed in place for reuse as fill. This activity, as with all on-site activities, is assumed to be
powered by diesel. The energy required to crush one tonne of concrete is estimated as 34 MJ

142
[22]. The emissions resulting from the use of diesel were estimated using the emission factors
shown in Table E4.

4.2 Processes Included in the Life Cycles of Alternative Concrete Constituents

4.2.1 Slag Processing and Transportation Processes

SL, a by-product of the steel industry, requires processing before it can be incorporated into
concrete. Energy and emissions for this processing are quantified by Marceau, Nisbet &
VanGeem [23]. Upstream energy, raw materials, and emissions are not included in this process
as SL is generated regardless of whether it is incorporated into concrete, disposed of in a landfill,
or any other usage. Using SL in concrete is environmentally beneficial, and so not including any
upstream processes is conservative. This is an assumption commonly used in concrete LCAs [24,
25]. SL transportation is modeled based on SL sources and modes of transportation that are
typical for the MTO, where emissions factors are shown in Table E4 and the worst-case
transportation scenario is shown in Table E5.

4.2.2 Fly Ash Transportation Process

FA is a by-product of coal burning. It is a fine substance that does not require further processing
before it is incorporated into concrete. As with SL, upstream energy, raw materials, and
emissions are not included in this process, which is a common conservative assumption. FA
transportation is modeled based on FA sources and modes of transportation that are typical for
the MTO, where emissions factors are shown in Table E4 and the worst-case transportation
scenario is shown in Table E5.

4.2.3 Silica Fume Transportation Process

SF is a by-product of silicon production. It is a very fine substance that does not require further
processing before it is incorporated into concrete. As with SL and FA, upstream energy, raw
materials, and emissions are not included in this process, which is a common conservative
assumption. SF transportation is modeled based on SF sources and modes of transportation that
are typical for the MTO, where emissions factors are shown in Table E4 and the worst-case
transportation scenario is shown in Table E5.

143
4.2.4 Limestone Process
For limestone cement production, transportation of raw materials is not included. This is because
cement production facilities are typically located at limestone quarries, and so there is no
transportation requirement for these materials [24]. As a result, the only added process for
limestone cement production is for the grinding of raw materials [4].

4.2.5 Titanium Dioxide and Photocatalytic Processes

Three processes are required to model photocatalytic concrete: titanium dioxide production,
titanium dioxide transportation, and photocatalytic pollutant abatement. Titanium dioxide
production was modeled using data from Ortiz [26]. The production process modeled was the
chlorine process for TiO2 production, which is more commonly used in North America and is
therefore most appropriate for this context. The assumptions made by Ortiz (2003) to build the
portion of the LCI that pertains to transportation, such as the transportation modes and distances,
are based on Spanish production, and are not appropriate for this research. As a result, a new
transportation process was created based on the emissions factors shown in Table E4 and the
distance shown in Table E5. A photocatalytic NOx abatement rate of 6 mg/h/m2 was selected
[27], and this is included in the model during as part of the Service Life phase, where a service
life of fifty years is assumed.

4.2.6 Recycled Aggregate Process

To model recycled aggregate production, a crushing process is required. This process was
created using the same data as that used for the end-of-life rubblization stage described
previously, the energy required to crush 1 tonne of concrete is estimated as 34 MJ [22] and it is
assumed that diesel is used to power the equipment. Emissions factors for diesel burning are
from Athena [4].

4.2.7 Admixture Production Process

Data for air entraining admixtures has been gathered from an environmental product declaration
prepared by the European Federation of Concrete Admixtures Associations [28]. The
environmental product declaration includes the raw materials, emissions to air, emissions to soil,
and emissions to water.
144
5 LCA Life Cycle Impact Assessment (LCIA) Methodology

In order to compile the LCI data and conduct the impact assessment, GaBi 6 software package
was used [29]. GaBi 6 is packaged with several impact assessment methods, including the
International Reference Life Cycle Data System (ILCD) method which was used for this study
[30]. Four impact categories- i) acidification, ii) global warming potential, iii) resource depletion
and iv) water depletion- were selected for their analysis based on MTO’s assessment of priority
(H. Schell and D. Rhead, personal communication, December 5 2014) and the level of
international consensus on the classification and characterization that has been reached for each
category, as described in Stranddorf and Anders Schmidt [31]. Table E6 presents a description
of each impact category as well as the characterization factor used in the analysis. In LCA,
characterization factors are science-based factors that are used to convert LCI data to actual
environmental impacts [32].

6 LCA Interpretation

This section describes the methodology used to interpret the LCA results by calculating two
indicators: 1) a Green Indicator, which uses weighting to combine results from multiple impact
categories into a single value; and 2) an EEI, which combines the Green Indicator (GI) with the
functional unit to provide a measure of both environmental and technical performance. For both
of these indicators, the results for the mix designs containing the alternative concrete constituents
are calculated relative to the results for the 100GU mix design, which are normalized to 1. This
allows for a quick determination of whether the mix designs containing alternative concrete
constituents offer an improvement over conventional concrete (EEI and GI >1) or whether they
have a higher environmental impact for a given functional performance (EEI and GI <1).

In the equations that follow, the index ‘i’ refers to the four impact categories included in this
study (acidification, global warming potential, resource depletion, water depletion) as described
in Table E6. The index ‘j’ references each of the eight mix designs included in this study as
described in Table E2. These equations therefore present a methodology for calculating GIs and
EEIs that incorporate LCA results from all impact categories and are specific to each mix design.

145
6.1 Green Indicator Methodology

The life cycle impact assessment results were factored by dividing the results for 100GU, the
base case mix design, by those for the alternative materials as shown in Eq. 2. The results for the
base material are placed on the numerator, because a higher LCIA result (ex. higher
acidification, higher global warming potential) is less desirable, and a higher value for the green
indicator corresponds to a reduced environmental impact.

Eq. 2

𝐿𝐶𝐼𝐴 𝑟𝑒𝑠𝑢𝑙𝑡𝑖𝑚𝑝𝑎𝑐𝑡 𝑐𝑎𝑡𝑒𝑔𝑜𝑟𝑦 𝑖, 𝑏𝑎𝑠𝑒 𝑚𝑎𝑡𝑒𝑟𝑖𝑎𝑙

𝐹𝑎𝑐𝑡𝑜𝑟𝑒𝑑 𝑟𝑒𝑠𝑢𝑙𝑡𝑖,𝑗 =
𝐿𝐶𝐼𝐴 𝑟𝑒𝑠𝑢𝑙𝑡𝑖𝑚𝑝𝑎𝑐𝑡 𝑐𝑎𝑡𝑒𝑔𝑜𝑟𝑦 𝑖, 𝑎𝑙𝑡. 𝑚𝑎𝑡𝑒𝑟𝑖𝑎𝑙 𝑗

where i= 1 to 4 and corresponds to impact category

and j = 1 to 8 and corresponds to alternative materials

These normalized results were then combined into a single green indicator score that combines
the acidification, global warming, resource depletion, and water depletion potentials, as shown in
Eq. 3. An initial weighting of 0.25:0.25:0.25:0.25 was used to combine these categories.
Alternative weightings also tested to evaluate the sensitivity of the results, as discussed in
Section 7.2.1. A higher green indicator indicates a concrete with a higher environmental
performance (and a lower environmental impact).

Eq. 3

𝐺𝑟𝑒𝑒𝑛 𝑖𝑛𝑑𝑖𝑐𝑎𝑡𝑜𝑟𝑗 = ∑ 𝐹𝑎𝑐𝑡𝑜𝑟𝑒𝑑 𝑟𝑒𝑠𝑢𝑙𝑡𝑖,𝑗 × 𝑤𝑒𝑖𝑔ℎ𝑡𝑖𝑛𝑔

𝑖=1

where i= 1 to 4 and corresponds to impact category

and j = 1 to 8 and corresponds to alternative materials

6.2 Environmental Efficiency Calculation

Following the calculations outlined in Section 6.1, the EEI can be calculated as shown in Eq. 4.
For every alternative concrete material (j), a specific EEI can be calculated by multiplying
together the green indicator result from Eq. 3 and the functional unit from Eq. 1.

146
Eq. 4

𝐸𝐸𝐼𝑗 = 𝐺𝑟𝑒𝑒𝑛 𝑖𝑛𝑑𝑖𝑐𝑎𝑡𝑜𝑟𝑗 × 𝐹𝑈

where j = 1 to 8 and corresponds to alternative materials

7 RESULTS AND DISCUSSION

7.1 EEI Results

EEI results for eight mix designs are presented in Figure E3. As discussed in Section 3, 100GU
represents the base case scenario and all EEIs are calculated normalized to the results of that mix
design; this is why 100GU has a value of 1. It can be seen that generally, most of the concrete
mix designs containing alternative concrete constituents have EEIs that are greater than one,
indicating an improved environmental efficiency compared to the base case (100GU).

There are two exceptions where mix designs have EEIs that are less than one: i) GU-PCAT 5, the
mixture containing photocatalytic cement and with no SL used as cement replacement, and ii)
100RAC, which contains recycled aggregate. These exceptions are likely due to the relatively
higher RCPT (and lower durability) of these two mix designs, as shown in Table E3. This is an
interesting result as all of these materials have been proposed to improve environmental
performance. Furthermore, Figure E3 shows that the only mix containing SF (GU-8SF-25SL)
has an environmental efficiency that is far superior compared to the other materials. This is due
to its relatively higher compressive strength and lower RCPT value compared to the other
mixtures, as shown in Table E3. The influence of the hardened properties (as incorporated into
the functional unit) on the EEI results is illustrated in Figure E4, which includes both the green
indicator results (which do not account for the functional unit) and the EEI results. Figure E4
shows that incorporating the strength and durability performance of the materials to measure
environmental efficiency, rather than measuring only environmental performance, significantly
influences the relative ranking of the materials.

147
7.2 Sensitivity Analysis

7.2.1 Weighting
As described in Section 6, the life cycle impact assessment (LCIA) results for four impact
categories- acidification, global warming potential, resource depletion, and water depletion- are
included in the calculation of the green indicator. The results presented in Figure E3 are for an
equal weighting of the different impact categories (Weighting Scheme 1 (1:1:1:1)). A sensitivity
analysis of the weighting scheme can confirm and elaborate on these results. Evaluating different
weighting schemes can also provide decision makers, who might prioritize the impact categories
differently, with context for interpreting the green indicator results. As noted by Bengtsson and
Steen, “weighting is not meant to deliver the final verdict about the environmental performance
[…] it is meant to give an additional input into the process” [33, p. 101]. This quotation
highlights two important points about weighting in LCA: 1) weighting should not be used to
selectively determine the ‘final verdict’, as this makes the results vulnerable to bias, and 2)
testing the model using a variety of weighting schemes that favour one impact category or
another can provide valuable information about the sensitivity of the model, and the robustness
of the results. For example, if it is found that results do not change significantly when many
different weighting schemes are considered, then the results can be said to be robust and
independent of the priorities of different stakeholders. If results do change significantly, then
assessing various weighting schemes can provide insight for various stakeholders about how the
results may be interpreted from different perspectives. In this study, it is acknowledged that the
results may be of interest to individuals and groups with different views on the value of certain
environmental systems and processes. Therefore, this paper attempts to address the issue of
weighting by applying a systematic sensitivity analysis to confirm and elaborate on the results
presented in Figure 3 (Weighting Scheme 1).

Four alternative weighting schemes, as shown in Table E6, were assessed (Weighting Scheme 2
through 5). These weighting schemes were each selected to emphasize one impact category,
which was given a weighting of 0.7 while the other three categories were all weighted at 0.1. The
purpose of selecting these weighting schemes was to determine whether large changes in the

148
relative importance assigned to each of the four impact categories affect the relative
environmental performance, as measured by EEIs, of the concrete mix designs.

The result of this sensitivity analysis was that the general trends discussed in Section 7.1, in
terms of the relative performance of the mixtures, did not change. That is, i) GU-25SL, GUL-
25SL, GU-8SF-25SL, GU-10FA, GUL-10FA, and GU-25SL-5PCAT all had improved
performance compared to 100GU, ii) the other mixtures (GU-5PCAT and GU-100RAC) had
lower environmental efficiency compared to 100GU, and iii) GU-8SF-25SL had a much higher
environmental efficiency compared to all the other materials. Table E8 shows the percent
difference between the Weighting Scheme 1 result and the results for each alternative weighting
scheme, for each concrete mix design. It can be seen that in general, the results presented in
Section 7.1 changed very little (<4%), indicating that for a given mix design, there is little
variability when different weighting schemes are applied.

However, there are two notable exceptions. The mix designs containing PCAT (GU-5PCAT and
GU-25SL-5PCAT) show much higher variability with the changes in weighting scheme (up to
~21% and 28% respectively). These two mix designs show especially high variability for
Weighting Schemes 2 and 4. For Weighting Scheme 2, which heavily weights the Acidification
Potential impact category, this is likely due to the important property of PCAT concrete in
catalyzing a reaction that reduces NOx (and therefore reduces acidification potential). For
Weighting Scheme 4, which heavily weights the Resource Depletion impact category, this high
variability is likely due to the fact that additional resources need to be extracted to produce
photocatalytic cement (containing TiO2) as compared to conventional cement or SCMs. This
sensitivity analysis, therefore, has identified that these two materials are more sensitive to LCA
modeling decisions such as the choice of a weighting scheme.

7.2.2 Transportation
Transportation distances of materials can potentially be highly variable. The availability of
material supply may change over time, especially when the materials are waste products like
SCMs. For example, the shutting down of coal plants in the Province of Ontario [34] could limit
the availability of FA in the long-term, and potentially necessitate further transportation
distances. Logistical factors, such as the commercial relationships between certain material
149
suppliers and processers, may also influence transportation distances. That is, the closest
processing facility to a plant may not be the facility that receives the material, if there is no
agreement in place between the two parties. In addition, minor events, such as weather events
and construction restrictions or closures, add additional variability that can affect any
transportation activity.

Due to the high potential for variability, the sensitivity of the results to changes in transportation
distances was studied. As discussed in Section 4, worst-case transportation scenarios were input
into the model as a base case for this work. To assess the sensitivity of the results, best-case
transportation scenarios were also modelled. Since the same transportation mode was used in all
of these scenarios (truck), the best-case transportation scenarios correspond to the routes with the
shortest total distances. These transportation scenarios are outlined in Table E9. Table E10
shows the percentage difference in EEI results between the best-case and worst-case
transportation scenarios for the purposes of comparison (the worst-case transportation scenarios
are also displayed in Figure E3). The results show that even with differences in transportation
distance ranging up to approximately 1470 km as shown in Table E5 and Table E9, for a given
mix design the difference between the LCA results calculated with the best case transportation
distance and the LCA results calculated with the worst case transportation distance is less than 6
% in all cases.

8 CONCLUSIONS

This paper presents an LCA of eight concrete mix designs that include various alternative
concrete constituents: SL, FA, SF, GUL, PCAT cement, and recycled aggregate. This paper also
presents a methodology for calculating EEIs, which incorporate LCIA results and functional
performance. The conclusions of this paper are relevant to any LCA practitioner or researcher
modelling cement and concrete materials. The key findings include:

 The mix designs containing SCMs, limestone cement, or PCAT with cement replacement
(GU-25SL, GUL-25SL, GU-8SF-25SL, GU-10FA, GUL-10FA and GU-25SL-5PCAT)
have EEIs that are greater than 1 when normalized to the base case (100GU), indicating
an improved environmental efficiency compared to the base case. This is because in these
150
 mix designs, there is significant (at least 10%) cement replacement. As cement is the
most environmentally intensive component of concrete, cement replacement is an
effective way to improve the life cycle environmental efficiency of concrete products.
 The mix design containing PCAT with no SCMs (GU-5PCAT) has an EEI that is less
than one when normalized to the base case (100GU), indicating a reduced environmental
efficiency compared to the base case. This is likely due to the relatively low functional
performance of these materials, and in particular the high RCPT values which indicate
poorer durability performance. In addition, there is little cement replacement in GU-
5PCAT, and PCAT cement is not a waste product so it does have upstream production
environmental impacts which are included in this analysis.
 The mix design containing recycled aggregate (GU-100RAC) has an EEI that is less than
one when normalized to the base case (100GU), indicating a reduced environmental
efficiency compared to the base case. This is likely due to the relatively low functional
performance of this material, and in particular the high RCPT value which indicate
poorer durability performance. Furthermore, a high replacement rate was used (100% of
coarse aggregate replaced by recycled aggregate) which reduced the functional
performance of concrete. This highlights the importance of incorporating functional
performance into measures of environmental efficiency. In addition, there is no cement
replacement in GU-100RAC, and in fact recycled aggregate concrete typically requires
additional cement compared to an equivalent mix with virgin aggregate to achieve similar
strength performance [35].
 GU-25SL has an improved environmental efficiency (EEI>1) compared to 100GU, and
GU-25SL-5PCAT has an improved environmental efficiency compared to GU-5PCAT.
This suggests that SL improves environmental efficiency. Again, this is because cement
is environmentally intensive and its replacement can improve the environmental
efficiency of concrete mix designs.
 The only mix containing SF (GU-8SF-25SL) has an environmental efficiency far superior
to the other materials: EEI = 12.16 compared to the next highest value of EEI = 2.29
(GU-25SL-5PCAT) and the lowest value of EEI = 0.25 (GU-RA100). Although the
replacement rate for SF is not as high (8%) as for other SCMs, this material can

151
 significantly influence strength and permeability and so this result is likely due to its high
strength and low permeability. In addition, the combined cement replacement rate (8%
SF and 25% SL) is the highest out of all the mix designs.
 Large changes in weighting generally have little influence on the EEI results (<2%),
except for mixtures containing PCAT, which unlike many of the other alternative
concrete constituents modelled, requires the additional extraction and processing of
virgin materials. These EEI results suggests that for mixtures containing PCAT, LCIA
results are more sensitive to modeling decisions and sensitivity analysis should be used to
explore the rigorousness of results and the implications for stakeholders and decision-
makers.
 The sensitivity of the results to the transportation distances was also explored by
generating results using best-case and worst-case transportation scenarios. Changes in
transportation distance up to approximately 1470 km have an overall influence on the
results for each given mix design that is <6%. This suggests that the transportation
processes do not have a large contribution to the overall LCIA results.

9 Acknowledgement

The authors acknowledge the MTO HIIFP (2013-2015) program for support of this research.
This research was supported by a contribution from the Ministry of Transportation of Ontario.
Opinions expressed in this report are those of the authors and may not necessarily reflect the
views and policies of the Ministry of Transportation of Ontario.

152
10 Bibliography

[1] C. Shi, A. F. Jimenez and A. Palomo, "New cements for the 21st century: The pursuit of an
alternative to Portland cement," Cement and Concrete Research, vol. 41, pp. 750-763, 2011.

[2] P. Van den Heede and N. De Belie, "Environmental impact and life cycle assessment (LCA)
of traditional and 'green' concretes: Literature review and theoretical calculations," Cement
& Concrete Composites, vol. 34, pp. 431-442, 2012.

[3] International Organization for Standardization, " Environmental management- Life cycle
assessment- Principles and framework," ISO, Geneva, 2006.

[4] Athena, "Cement and Structural Concrete Products: Life Cycle Inventory Update #2,"
Athena Sustainable Materials Institute, Ottawa, 2005.

[5] D. Huntzinger and T. Eatmon, "A life-cycle assessment of Portland cement manufacturing:
comparing the traditional process with alternative technologies," Journal of Cleaner
Production, vol. 17, pp. 668-675, 2009.

[6] M. A. Nisbet, M. L. Marceau and M. G. VanGeem, "Environmental Life Cycle Inventory of

Portland Cement Concrete," Portland Cement Association, 2002.

[7] B. L. Damineli, F. M. Kemeid, P. S. Aguiar and V. M. John, "Measuring the eco-efficiency

of cement use," Cement and Concrete Composites, vol. 32, pp. 555-562, 2010.

[8] H. K. Stranddorf, L. Hoffmann and A. Schmidt, "Update on Impact Categories,

Normalisation and Weighting in LCA- Selected EDIP97-data," Danish EPA, 2005.

[9] S. Marinkovic, V. Radonjanin, M. Malesev and I. Ignjatovic, "Comparative environmental

assessment of natural and recycled aggregate concrete," Waste Management, vol. 30, pp.
2255-2264, 2010.

153
[10] E. Rametsteiner, H. Pulzl, J. Alkan-Olsson and P. Frederiksen, "Sustainability indicator
development- Science or political negotiation?," Ecological Indicators, pp. 61-70, 2011.

[11] B. Moldan, S. Janouskova and T. Hak, "How to understand and measure environmental
sustainability- Indicators and targets," Ecological Indicators, pp. 4-13, 2012.

[12] G. Fernandez-Sanchez and F. Rodriguez-Lopez, "A methodology to identify sustainability

indicators in construction project management- Application to infrastructure projects in
Spain," Ecological Indicators, pp. 1193-1201, 2010.

[13] Hanson UK, "Summary of KPI performance against 2020 targets," 2015. [Online].
Available: http://www.hanson-sustainability.co.uk/en. [Accessed 24 August 2015].

[14] Y. Zhong and P. Wu, "Economic sustainability, environmental sustainability and

constructability indicators related to concrete- and steel- projects," Journal of Cleaner
Production, pp. 1-9, 2015.

[15] O. Pons and A. de la Fuente, "Integrated sustainability assessment method applied to

structural concrete columns," Construction and Building Materials, pp. 882-893, 2013.

[16] M. Dolatabadi, Properties and Performance of Photocatalytic Concrete, Toronto, ON:

Library and Archives Canada, 2013.

[17] D. Prasada Rao and P. Sindhu Desai, "Experimental investigations of coarse aggregate
recycled concrete," International Journal of Advances in Engineering & Technology, pp.
1522-1530, 2014.

[18] M. De Schepper, P. Van den Heede, I. Van Driessche and N. De Beile, "Life cycle
assessment of completely recyclable concrete," Materials, no. 7, pp. 6010-6027, 2014.

[19] A. Racoviceanu, B. Karney, C. Kennedy and A. Colombo, "Life-Cycle Energy Use and
Greenhouse Gas Emissions Inventory for Water Treatment Systems," Journal of
Infrastructure Systems, vol. 13, no. 4, pp. 261-270, December 2007.

154
[20] Environment Canada, "Summary of National Pollutant Release Inventory Reporting
Requirements," 11 July 2013. [Online]. Available: https://www.ec.gc.ca/inrp-
npri/default.asp?lang=en&n=629573FE-1 .

[21] J. R. Prusinski, M. L. Marceau and M. G. VanGeem, "Life Cycle Inventory of Slag Cement
Concrete," in Eighth CANMET/ACI Eighth CANMET/ACI International Conference on Fly
Ash, Silica Fume, Slag and Natural Pozzolans in Concrete, 2004.

[22] D. Wilburn and T. Goonan, "Aggregates from natural and recycled sources," US Geological
Survey, Denver, 1998.

[23] M. Marceau, M. Nisbet and M. VanGeem, Life Cycle Inventory of Portland Cement
Manufacture, Skokie, IL: Portland Cement Association, 2006.

[24] CANMET & Radian Canada, Raw material balances, energy profiles and environmental
unit factor estimates: Cement and structural concrete products, Ottawa, ON, 1993.

[25] K.-H. Yang, Y.-B. Jung, M.-S. Cho and S.-U. Tae, "Effect of supplementary cementitious
materials on reduction of CO2 emissions from concrete," Journal of Cleaner Production,
pp. 1-10, 2014.

[26] I. Ortiz, Life cycle assessment as a tool for green chemistry: application to kraft pulp
industrial wastewater treatment by different advanced oxidation processes, Barcelona:
Institut de Cie`ncia i Tecnologia Ambientals, The Universitat Auto` noma de Barcelona,
2003.

[27] C. Churchill and D. Panesar, "Life-cycle cost analysis of highway noise barriers designed
with photocatalytic cement," Structure and Infrastructure Engineering: Maintenance,
Management, Life-Cycle Design and Performance, vol. 9, no. 10, pp. 983-998, 2013.

[28] European Federation of Concrete Admixtures Associations, "EFCA Publications," 2010.

[Online]. Available: http://www.efca.info/publications.html. [Accessed 10 January 2015].

155
[29] PE International, GaBi 6.0, 2014.

[30] European Commission Joint Research Centre, "International Reference Life Cycle Data
System (ILCD) Handbook," Publications Office of the European Union, Luxemburg, 2011.

[31] H. Stranddorf and L. Anders Schmidt, "Update on impact categories, normalisation and
weighting in LCA," Danish Ministry of the Environment Environmental Protection Agency,
2005.

[32] Scientific Applications International Corporation, "Life Cycle Assessment- Principles and
Practice," US Environmental Protection Agency, Cincinnati, Ohio, 2006.

[33] M. Bengtsson and B. Steen, "Weighting in LCA- Approaches and applications,"

Environmental Progress, vol. 19, no. 2, pp. 101-109, 2000.

[34] Ministry of Energy, "Creating Cleaner Air in Ontario: Province has eliminated coal-fired
generation," 15 April 2014. [Online]. Available:
http://news.ontario.ca/mei/en/2014/04/creating-cleaner-air-in-ontario-1.html.

[35] L. Butler, J. West and S. Tighe, "Effect of recycled concrete coarse aggregate from multiple
sources on the hardened properties of concrete with equivalent compressive strength,"
Construction and Building Materials, no. 47, pp. 1292-1301, 2013.

[36] J. E. Anderson and R. Silman, "A life cycle inventory of structural engineering design
strategies for greenhouse gas reduction," Structural Engineering International, pp. 283-288,
March 2009.

[37] D. Brown, R. Sadiq and K. Hewage, An overview of air emission intensities and
environmental performance of grey cement manufacturing in Canada, vol. 16, 2014, pp.
1119-1131.

[38] K. Celik, C. Meral, A. Gursel, P. Mehta, A. Horvath and P. Monteiro, "Mechanical

properties, durability, and life-cycle assessment of self-consolidating concrete mixtures

156
 made with blended portland cements containing fly ash and limestone powder," Cement and
Concrete Composites, vol. 56, pp. 59-72, 2015.

[39] C. Chen, G. Habert, Y. Bouzidi, A. Jullien and A. Ventura, LCA allocation procedure used
as an incitative method for waste recycling: An application to mineral additions in concrete,
vol. 54, 2010, pp. 1231-1240.

[40] F. Collins, Inclusion of carbonation during the life cycle of built and recycled concrete:
influence on their carbon footprint, vol. 15, 2010, pp. 549-556.

[41] E. Crossin, "Comparative life cycle assessment of concrete blends," Centre for Design
RMIT University, Melbourne, 2012.

[42] T. Garcia-Segura, V. Yepes and J. Alcala, "Life cycle greenhouse gas emissions of blended
cement concrete including carbonation and durability," International Journal of Life Cycle
Assessment, no. 19, pp. 3-12, 2014.

[43] M. Hassan, "Life-cycle assessment of titanium dioxide coatings," in Building a Sustainable

Future – Proceedings of the 2009 Construction Research Congress, Seattle, 2009.

[44] A. Jonsson, T. Bjorklund and A.-M. Tillman, "LCA of concrete and steel building frames,"
International Journal of Life Cycle Assessment, vol. 3, no. 4, pp. 216-224, 1998.

[45] K. Kawai, T. Sugiyama, K. Kobayashi and S. Sano, "Inventory data and case studies for
environmental performance evaluation of concrete structure construction," Journal of
Advanced Concrete Technology, vol. 3, no. 3, pp. 435-456, October 2005.

[46] C. Knoeri, E. Sanye-Mengual and H.-J. Althaus, "Comparative LCA of recycled and
conventional concrete for structural applications," The International Journal of Life Cycle
Assessment, vol. 13, no. 5, pp. 909-918, June 2013.

[47] K.-M. Lee and P.-J. Park, "Estimation of the environmental credit for the recycling of
granulated blast furnace slag based on LCA," Resources, Conservation and Recycling, no.

157
 44, pp. 139-151, 2005.

[48] C. Li, S. Cui, Z. Nie, X. Gong, Z. Wang and N. Itsubo, "The LCA of portland cement
production in China," International Journal of Life Cycle Assessment, vol. 20, pp. 117-127,
2015.

[49] N. Tosic, S. Marinkovic, T. Dasic and M. Stanic, Multicriteria optimization of natural and
recycled aggregate concrete for structural use, vol. 87, 2015, pp. 766-776.

[50] R. Wang, M. Eckelman and J. Zimmerman, "Consequential environmental and economic

life cycle assessment of green and gray stormwater infrastructures for combined sewer
systems," Environmental Science & Technology, vol. 47, pp. 11189-11198, 2013.

[51] S. Xing, Z. Xu and G. Jun, "Inventory analysis of LCA on steel- and concrete-construction
office buildings," Energy and Buildings, no. 40, pp. 1188-1193, 2008.

[52] S. H. Kosmatka, B. Kerkhoff, R. D. Hooton and R. J. McGrath, Design and Control of

Concrete Mixtures, 8th Edition ed., Ottawa, ON: Cement Association of Canada, 2011.

[53] H. Xu, J. L. Provis, J. S. J. v. Deventer and P. V. Krivenko, "Characterization of aged slag

concretes," ACI Materials Journal, vol. 105, no. 2, pp. 131-139, Mar/Apr 2008.

[54] B. Marsh, "High-volume fly ash concrete," Concrete, vol. 37, no. 4, pp. 54-55, April 2003.

[55] K. R. O'Brien, J. Menache and L. M. O'Moore, "Impact of fly ash content and fly ash
transportation distance on embodied greenhouse gas emissions and water consumption in
concrete," International Journal of Life Cycle Assessment, vol. 14, pp. 621-629, 2009.

[56] M. I. Khan and R. Siddique, "Utilization of silica fume in concrete: Review of durability
properties," Resources, Conservation and Recycling, vol. 57, pp. 30-35, 2011.

[57] P. Tennis, M. Thomas and W. Weiss, "State-of-the-Art Report on Use of Limestone in

Cements at Levels of up to 15%," Portland Cement Association, Skokie, IL, 2011.

158
[58] A. M. Ramezanianpour and R. D. Hooton, "Sulfate resistance of Portland-limestone
cements in combination with supplementary cementitious materials," Materials and
Structures, vol. 46, pp. 1061-1073, 2013.

[59] M. Hassan, "Evaluation of the durability of titanium dioxide photocatalyst coating for
concrete pavement," Journal of Construction and Building Material, pp. 1456-1461, 2010.

[60] Independent Electricity System Operator (IESO), "Supply Overview," 2015. [Online].
Available: http://www.ieso.ca/Pages/Power-Data/Supply.aspx. [Accessed 10 February
2015].

159
 Table A Literature review of concrete LCI and LCA studies.

Source LCA System Functional Unit Geographical Alternative

Scope Boundary Context Concrete
Constituent
[36] LCI only Cradle to Steel or concrete building United States Recycled
EIO- grave; no concrete
LCA maintenance

[4] LCI only Cradle to gate 1 m3 of 15, 20 or 30 MPa ready mixed Canada Fly ash, slag,
concrete; precast double ‘T’ beam, silica fume
precast hollow deck, standard
concrete blocks, cement mortar
[37] LCI only Cradle to gate 1 tonne of cement Canada -
for cement
manufacture
[38] LCA Cradle to gate 1m3 of ready mixed SCC concrete United States SCC with fly ash
and limestone
[8] LCA Cradle to gate 1 kg of cement Europe -
for cement
manufacture
[39] LCA Cradle to gate Binding equivalent value of fly ash, Europe Slag and fly ash
for cement slag, and CEM 1
manufacture
[27] LCI and Cradle to Noise barrier system Ontario Slag and
costs grave photocatalytic
only
[40] LCA Cradle to Concrete bridge (life of 100 years, Recycled
cradle then crushed and recycled in bridge) concrete, slag,
fly ash
[41] LCA Cradle to gate 1 m3 of AS 1379:2007 40 MPa Australia Proprietary
+ end of life concrete for 50 years blends; slag, fly
(no use or ash, recycled
maintenance) aggregate, silica
fume
[18] LCA Cradle to Total amount of concrete necessary to Belgium Recycled
cradle deliver 1 MPa of strength and one aggregate,
year of service life concrete reused
for cement
production

160
 (cont’d) Table A Literature review of concrete LCI and LCA studies.

Source LCA System Functional Unit Geographical Alternative

Scope Boundary Context Concrete
Constituent
[42] LCI for Cradle to Reinforced concrete building column Spain Fly ash, slag
GHG grave (+ post- (3x0.3x0.3m) during its lifetime
only demolition)
[43] LCA Cradle to gate 5 mm thick surface mix of ultrafine United States Photocatalytic
for TiO2 TiO2 anatase, cement, filler, and water
(concrete not (0.1:1:3:0.4) applied on pavement
included) surface
[5] LCA Cradle to gate 1 ton of cement x 4 types: traditional, United States Natural
for cement blended (natural pozzolans), cement + pozzolans
manufacture sequestration in waste materials, (volcanic ash,
cement + recycled CKD rice husk ash, fly
ash)
[44] LCA Cradle to 1 m2 of floor area for seven frame Sweden/ -
grave cases (including in-situ cast and Nordic
precast concrete frames) nations
[45] LCA Cradle to gate Prestressed concrete bridge, overflow Japan Slag
dike of dam, retaining wall, secondary
lining in tunnel
[46] LCA Cradle to gate 1 m3 of a specific strength class of Switzerland Recycled
concrete concrete
Lean, indoor and outdoor structural
concrete
[47] LCA Cradle to gate 4 FUs: 1 kg of cement clinker, South Korea Slag
for cement Portland cement, slag powder, silicate
and slag fertilizer
powder
[48] LCA Cradle to gate 2 FUs: 1 t of portland cement, 1 t of China and -
for cement Portland cement with 42.5 MPa of Japan
strength grade
[9] LCA Cradle to gate 1 m3 of ready-mixed natural aggregate Serbia Recycled
for concrete concrete and recycled aggregate aggregate
including concrete concrete
transport
[6] LCI only Cradle to gate 3 FUs: 1 m3 of ready mixed concrete, United States Fly ash, slag,
100 concrete masonry units, 1 m3 of silica fume
precast concrete

161
 (cont’d) Table A Literature review of concrete LCI and LCA studies.

Source LCA System Functional Unit Geographical Alternative

Scope Boundary Context Concrete
Constituent
[21] LCI only Cradle to gate 5 FUs: 20 and 35 MPa ready mixed United States Slag
concrete, 50 and 70 MPa precast
concrete, concrete block mix
[49] LCA Cradle to gate 1 m3 of ready mixed natural and Serbia Recycled
recycled aggregate concrete aggregate
[50] LCA Cradle to gate 4047 m2 of permeable pavement Northeastern Permeable
+ maintenance (sized to store runoff associated with United States pavement
and 2.5 cm of rainfall generated over 1
stormwater acre watershed)
runoff and Other FUs include other types of
treatment runoff systems designed for the same
capacity
[51] LCI only Cradle to 1 m2 building area Shanghai, -
grave Steel and concrete office structures China

[25] LCA Cradle to gate Binder intensity and CO2 intensity South Korea Slag, fly ash,
for concrete concepts [7] in terms of the unit silica fume
strength (1MPa) of concrete

162
List of Tables- Appendix E

Table E1 Alternative concrete constituents: description and performance

Table E2 Mix designs

Table E3 Hardened properties

Table E4 Relevant factors for the calculation of transportation energy consumption and
emissions

Table E5 Worst-case MTO transportation scenarios for concrete constituents

Table E6 Impact Categories

Table E7 Weighting Schemes

Table E8 Variation in EEI results for each mix design with various impact category weighting
schemes.

Table E9 Best-case MTO transportation scenarios for concrete constituents

Table E10 Sensitivity analysis results for best-case and worst-case transportation scenarios (%
difference)
 Table E1 Alternative concrete constituents: description and performance

Description Performance in Concrete

SL By-product of steel Strength: similar strength development to GU concrete (Kosmatka et al. [52])
industry Durability: generally improved compared to GU concrete (Xu et al. [53]); however,
Direct replacement scaling potential in deicing exposures at high replacement levels
for cement Environmental: reduction in cement use and use of a waste material
FA By-product of coal- Strength: may improve strength at recommended replacement rates [54]
fired power stations Durability: good durability and effective mitigation of alkali silica reaction (ASR)
Direct replacement [54]; however, scaling potential in deicing exposures at high replacement
for cement Environmental: replaces cement content, repurposes waste; beneficial even with
large transportation distances [55]
SF By-product from the Strength: improved as it is a very fine pozzolanic material [56]
manufacture of Durability: improved resistance to abrasion, freeze thaw, chloride penetration,
silicon/ferrosilicon sulfate attack, and ASR due to low permeability [56]
Typically added in Environmental: replaces cement content, repurposes waste, improves durability
blended cement
GUL Produced by i) Strength: if ground finer than cement, generally has similar strength development
partially replacing [57]
clinker with ground Durability: sufficient durability with adequate SCMs (CSA >40%) [58]
limestone, or ii)
blending cement Environmental: replaces cement content
and limestone
during batching
RAC Replacement of Strength/durability properties: highly dependent on quality of aggregate, which is
virgin aggregate highly variable depending on the source of the waste material; 5-10% more cement
with construction recommended for strength development [35]
and demolition Environmental: replaces virgin aggregate & repurposes waste; processing is energy
waste intensive; transportation is key variable [46].
PCAT Titanium dioxide Strength: little effect on strength in typical application [16]
photocatalyst Durability: adequate durability in typical application where cement is used in
Typical application: surface layer of concrete [59].
TiO2 blended into Environmental: titanium dioxide catalyzes a reaction that converts air pollutants (ex.
cement (~5%) NOx, SOx) to less damaging forms
 Table E2 Mix designs
Mix Design ID
*j = - 1 2 3 4 5 6 7 8
Quantity 100GU* GU- GUL- GU-8SF- GU-10FA GUL- GU- GU-25SL- GU-
(kg/m3) * 25SL 25SL 25SL 10FA 5PCAT 5PCAT 100RAC
Water 180 155 150 140 128 147 180 180 283
GU Cement 429 300 - 276 316 - 407 305 565
GUL Cement - - 300 - - 319 - - -
SL - 53 54 100 - - - 107 -
FA - - - - 35 35 - - -
SF - - - 24 - - - - -
TiO2 - - - - - - 22 16 -
Coarse agg. 900 1064 1071 1000*** 1037 1036 900 900
Recycled - - - - - - - - 1955
coarse agg.
Fine agg. 875 779 781 800*** 777 771 869 876 966
Air entraining 52 - - - - - 52 52 -
admixture
(mL/kg
cement matl)
* Index values corresponding to Eq. 1 to Eq. 4
** Base case scenario [16]
*** Aggregate quantities are estimated based on typical MTO proportions.
 Table E3 Hardened properties
Mix Design ID
Hardened 100GU* GU- GUL- GU-8SF- GU-10FA GUL- GU- GU-25SL- GU-
Property 25SL 25SL 25SL 10FA 5PCAT 5PCAT 100RAC
28 day f’c 48.5 36.0 34.7 56.5 51.2 50.6 42.3 52.2 27.7
(MPa)
28 day 3186 2135 1969 421 3614 4024 5436 2002 5759
RCPT (C)
* Base case scenario
Table E4 Relevant factors for the calculation of transportation energy consumption and emissions [4]

Energy Emissions Factors [kg/GJ]

CO2 SO2 NOx VOC CH4 CO
70.7 0.102 0.807 0.0869 0.0217 0.443
 Table E5 Worst-case MTO transportation scenarios for concrete constituents

Material Route Total Distance (km) Transportation Mode

SL Hamilton to Picton, Picton to Toronto 498 Truck
FA Pleasant Prairie, WI to Toronto 937 Truck
SF Niagara Falls, NY to St Basile, QC, St Basile, QC to Toronto 1623 Truck
TiO2 Essexville, MI to Toronto 487 Truck
Coarse agg. Typical route 200 Truck
Fine agg. Typical route 200 Truck
 Table E6 Impact Categories

*i= Impact Category Description Characterization Factor

1 Release of protons and leaching out of anions from a Acidification Potential
Acidification
system (mole H+ ion equivalents)
2 Effect of increasing temperature in the atmosphere Global Warming Potential
Global Warming due to the reflection of radiation by greenhouse
gases back to the surface (kg CO2 equivalents)

3 Resource Depletion of abiotic non-water resources (ex. Resource Depletion Potential (kg
Depletion minerals or fossil fuels) of antimony (Sb) equivalents)
4 Water use (reuse, degradation, or incorporation into Water Depletion Potential
Water Depletion
final product) (volume of water used (m3))
* Index values corresponding to Eq. 2 and 3.
 Table E7 Weighting Schemes

Weighting Scheme Acidification Global Warming Potential Resource Depletion Water Depletion
1 0.25 0.25 0.25 0.25
2 0.7 0.1 0.1 0.1
3 0.1 0.7 0.1 0.1
4 0.1 0.1 0.7 0.1
5 0.1 0.1 0.1 0.7
 Table E8 Variation in EEI results for each mix design with various impact category weighting schemes.

% Difference Compared to Weighting Scheme 1 (0.25:0.25:0.25:0.25) for Each Mix Design

Weighting GUL- GU-8SF- GUL- GU- GU-25SL- GU-
Scheme GU-25SL 25SL 25SL GU-10FA 10FA 5PCAT 5PCAT 100RAC
2
(0.7:0.1:0.1:0.1) -0.58 -0.66 -1.54 -0.78 -0.73 21.24 27.47 0.12
3
(0.1:0.7:0.1:0.1) 2.50 3.44 3.33 1.54 2.41 -5.16 -7.81 -1.94
4
(0.1:0.1:0.7:0.1) 0.32 0.25 0.42 0.27 0.25 -19.50 -23.56 -0.95
5
(0.1:0.1:0.1:0.7) -2.35 -3.26 -2.41 -1.06 -2.04 -6.09 -13.08 2.65
 Table E9 Best-case MTO transportation scenarios for concrete constituents

Material Route Total Distance (km) Transportation Mode

SL Hamilton to Mississauga, Mississauga to Toronto 76 Truck
FA Belle River, MI to Toronto 332 Truck
SF Niagara Falls, NY to Mississauga, Mississauga to Toronto 145 Truck
TiO2 Essexville, MI to Toronto 487 Truck
Coarse agg. Typical route 80 Truck
Fine agg. Typical route 80 Truck
 Table E10 Sensitivity analysis results for best-case and worst-case transportation scenarios (% difference)

GU-25SL GUL-25SL GU-8SF-25SL GU-10FA GUL-10FA GU-5PCAT GU-25SL-5PCAT GU-100RAC

1.65 2.21 2.92 1.38 1.83 1.14 5.98 1.74
List of Figures- Appendix E

Figure E1 System boundary for LCA of 100GU concrete

Figure E2 Ontario's electricity supply mix for 2014 .

Figure E3 EEI results for eight concrete mix designs.

Figure E4 EEI and green indicator (GI) results for eight concrete mix designs.
Figure E1 System boundary for LCA of 100GU concrete
Figure E2 Ontario's electricity supply mix for 2014 [60].
 13
Environmental Efficiency Indicator (normalized to 100GU)

12

11

10

Concrete Mix Design

Figure E3 EEI results for eight concrete mix designs.

 13
Environmental Efficiency Indicator or Green Indicator

12
11
10
9
(normalized to 100GU)

8
7
6
5
4
3
2
1
0

Concrete Mix Design

GI EEI

Figure E4 EEI and green indicator (GI) results for eight concrete mix designs.