Escolar Documentos
Profissional Documentos
Cultura Documentos
Concrete Materials
by
Karina E. Seto
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.
ProQuest LLC.
789 East Eisenhower Parkway
P.O. Box 1346
Ann Arbor, MI 48106 - 1346
ii
Karina E. Seto
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
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
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 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
1 Introduction ................................................................................................................................ 1
3 Background ................................................................................................................................ 4
4.2.2 Processes Included in the Life Cycles of Alternative Concrete Constituents ....... 25
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
6 Conclusions .............................................................................................................................. 35
8 References ................................................................................................................................ 41
vi
List of Tables
Table 1 Environmental and performance properties of alternative concrete constituents. ............. 4
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 8 Relevant factors for the calculation of transportation energy consumption and emissions
(Athena, 2005) .............................................................................................................................. 24
List of Figures
Figure 1 LCI building methodology ............................................................................................. 12
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
Impact of the selection of functional unit on the life cycle assessment of green
concrete
Key conclusions are presented in Section 6, and potential future research is discussed in
Section 7.
2
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.
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.
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)
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.
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
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
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.
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
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
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
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
Scope
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
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
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.
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.
Equation 1 describes how the functional unit for this study is calculated.
Eq. 1
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
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
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.
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.
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
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.
Eq. 2
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
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.
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.
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 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
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
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
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
Athena. (2005). Cement and Structural Concrete Products: Life Cycle Inventory Update #2. Ottawa: Athena
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,
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
Collins, F. (2010). Inclusion of carbonation during the life cycle of built and recycled concrete: influence on their
Coulon, R., Camobreco, V., Teulon, H., & Besnainou, J. (1997). Data quality and uncertainty in LCI. International
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.
De Barba Junior, D. J., de Oliveira Gomes, J., & Bork, C. A. (2014). Reliability of the Sustainability Assessment.
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
De Schepper, M., Van den Heede, P., Van Driessche, I., & De Beile, N. (2014). Life cycle assessment of completely
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.
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,
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
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.
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 -
ISO. (2006b). ISO 14044 International Standard. In: Environmental Management - Life Cycle Assessment -
Jonsson, A., Bjorklund, T., & Tillman, A.-M. (1998). LCA of concrete and steel building frames. International
Kawai, K., Sugiyama, T., Kobayashi, K., & Sano, S. (2005, October). Inventory data and case studies for
Khan, M. I., & Siddique, R. (2011). Utilization of silica fume in concrete: Review of durability properties.
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
Kosmatka, S. H., Kerkhoff, B., Hooton, R. D., & McGrath, R. J. (2011). Design and Control of Concrete Mixtures
Lee, K.-M., & Park, P.-J. (2005). Estimation of the environmental credit for the recycling of granulated blast furnace
Li, C., Cui, S., Nie, Z., Gong, X., Wang, Z., & Itsubo, N. (2015). The LCA of portland cement production in China.
Marceau, M., Nisbet, M., & VanGeem, M. (2006). Life Cycle Inventory of Portland Cement Manufacture. Skokie,
Marinkovic, S., Radonjanin, V., Malesev, M., & Ignjatovic, I. (2010). Comparative environmental assessment of
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
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
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
Ortiz, I. (2003). Life cycle assessment as a tool for green chemistry: application to kraft pulp industrial wastewater
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.
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
Scientific Applications International Corporation (SAIC). (2006, May). Life Cycle Assessment: Principles and
W10/EPAonLCA2006.pdf
Shi, C., Jimenez, A. F., & Palomo, A. (2011). New cements for the 21st century: The pursuit of an alternative to
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
Tennis, P., Thomas, M., & Weiss, W. (2011). State-of-the-Art Report on Use of Limestone in Cements at Levels of
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
Weckenmann, A., & Schwan, A. (2001). Environmental Life Cycle Assessment with Support of Fuzzy-Sets.
Weidema, B., & Wesnaes, M. (1996). Data quality management for life cycle inventories- an example of using data
Wilburn, D., & Goonan, T. (1998). Aggregates from natural and recycled sources. US Geological Survey, US
Xing, S., Xu, Z., & Jun, G. (2008). Inventory analysis of LCA on steel- and concrete-construction office buildings.
Xu, H., Provis, J. L., Deventer, J. S., & Krivenko, P. V. (2008, Mar/Apr). Characterization of aged slag concretes.
Yang, K.-H., Jung, Y.-B., Cho, M.-S., & Tae, S.-U. (2014). Effect of supplementary cementitious materials on
doi:http://dx.doi.org/10.1016/j.jclepro.2014.03.018
APPENDIX A
CRITERIA FOR THE EVALUATION OF LIFE CYCLE ASSESSMENT
APPLICATION TO CONCRETE
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
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
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
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,
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
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.
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;
For SF, a transportation process was added. The upstream production of SF is not included in the analysis;
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,
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
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, &
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
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
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
Output quality: the utility, clarity, and modifiability of the outputs; in particular, the ability to extract raw
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
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
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
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
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
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
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
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
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,
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
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
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
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
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
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
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
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
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
64
7 Bibliography
Anderson, J. E., & Silman, R. (2009, March). A life cycle inventory of structural engineering design strategies for
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
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
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.
De Barba Junior, D. J., de Oliveira Gomes, J., & Bork, C. A. (2014). Reliability of the Sustainability Assessment.
Dhir, R., Paine, K., & Dyer, T. (2004). Recycling construction and demolition wastes in concrete. Concrete , 38 (3),
25-28.
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.
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
Marceau, M., Nisbet, M., & VanGeem, M. (2006). Life Cycle Inventory of Portland Cement Manufacture. Skokie,
Marinkovic, S., Radonjanin, V., Malesev, M., & Ignjatovic, I. (2010). Comparative environmental assessment of
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
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
Poon, C., & Cheung, E. (2007). NO removal efficiency of photocatalytic paving blocks prepared with recycled
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.
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
Thomas, M. D., Delagrave, A., Blair, B., & Barcelo, L. (2013, December). Equivalent durability performance of
Weidema, B., & Wesnaes, M. (1996). Data quality management for life cycle inventories- an example of using data
Xu, H., Provis, J., Deventer, J. v., & Krivenko, P. (2008). Characterization of Aged Slag Concretes. ACI Materials
67
Appendix A-1
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,
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
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,
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,
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 A4 Data quality pedigree matrix (reproduced from Weidema & Wesnaes, 1996)
Table A7 Data quality pedigree matrix results for slag cement production
Table A1 Main characteristics of LCA software systems (summarized from Ciroth, 2012)
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
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 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
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
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.
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.
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.
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.
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).
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).
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.
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.
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.
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.
Independent Electricity System Operator (IESO). (2015). Supply Overview. Retrieved February
10, 2015, from IESO: http://www.ieso.ca/Pages/Power-Data/Supply.aspx
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 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 B3 Datasets appropriate for modeling Canada’s electricity grid using the GaBi 6
Table B6 Percentage contribution to LCIA results in each impact category for concrete
Marginal vs. average Should LCA model the response of the energy supply system to demand?
System boundaries How wide and broad should the boundaries be to capture environmental flows and data that
are needed for impact?
New and non-traditional How should LCA model new technologies, in which the data are highly uncertain, and how
should increased demand for new technologies be accounted for?
Co-product allocation How should environmental burdens be allocated across co-products that come from the
same process?
Transmission /distribution How should transmission and distribution impacts be included in modeling of electricity
generation?
Table B2 Data Quality Goals
Dataset Database
Electricity from nuclear power, Canada Professional database 2014
Electricity from wind power, Canada Professional database 2014
Electricity from biomass (solid), Canada Professional database 2014
Electricity from natural gas, Canada Extension database XVII: Full US
Electricity from hard coal, Canada Extension database XVII: Full US
Electricity from hydro power, Canada Extension database XVII: Full US
Electricity from photovoltaic, Canada Professional database 2014
Electricity from heavy fuel oil (HFO), Canada Extension database XVII: Full US
Electricity grid mix (production mix), Canada Extension database XVII: Full US
Table B4 Impact categories
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
*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
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
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
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
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
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
KEYWORDS
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
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]).
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.
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.
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.
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.
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.
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:
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
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.
99
resulting impact is likely to be small- and therefore, the environmental impact of its production
should not be included.
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.
100
Eq. 2
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.
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).
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
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.
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.
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.
[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.
[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].
[17] European Commission Joint Research Centre, "International Reference Life Cycle Data
System (ILCD) Handbook," Publications Office of the European Union, Luxemburg,
2011.
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.
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 C5 Green indicator results for GU-FA10 using four allocation scenarios.
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
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.
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.
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:
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
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).
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
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.
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
Eq. 4
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
Eq. 6
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
Eq. 8
𝐹𝑈 6,𝑗 = 𝐹𝑈3,𝑗 × 𝐹𝑈4,𝑗
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
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.
[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.
[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.
[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.
[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.
[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.
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.
132
List of Tables- Appendix D
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 D7 Percentage difference between EEIs for FUs incorporating performance characteristics
vs. FU1 (volume)
Table D9 Variation in EEI results with various impact category weighting schemes
Table D1 Review of functional units in known cement and concrete LCAs
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)
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
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
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
KEYWORDS
Compressive strength (C), durability (C), concrete (D), silica fume (D), life cycle assessment
133
ABSTRACT
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.
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.
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]
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.
3.1 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
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.
Eq. 1
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].
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.
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).
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).
142
[22]. The emissions resulting from the use of diesel were estimated using the emission factors
shown in Table E4.
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].
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
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
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
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.
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.
[13] Hanson UK, "Summary of KPI performance against 2020 targets," 2015. [Online].
Available: http://www.hanson-sustainability.co.uk/en. [Accessed 24 August 2015].
[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.
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.
[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.
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.
[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.
[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.
[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.
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.
[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.
161
(cont’d) Table A Literature review of concrete LCI and LCA studies.
[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 E4 Relevant factors for the calculation of transportation energy consumption and
emissions
Table E8 Variation in EEI results for each mix design with various impact category weighting
schemes.
Table E10 Sensitivity analysis results for best-case and worst-case transportation scenarios (%
difference)
Table E1 Alternative concrete constituents: description and performance
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.
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
12
11
10
9
(normalized to 100GU)
8
7
6
5
4
3
2
1
0
GI EEI
Figure E4 EEI and green indicator (GI) results for eight concrete mix designs.