Escolar Documentos
Profissional Documentos
Cultura Documentos
(mat)
Here
y
is chosen as
the stress limitation of
stainless steel.
*Ergonomic Const.
Distance between user
head and basin surface:
D
erg
(Stream trajectory u, v,
shape_basin a, )
50mm
;
From common
knowledge, people will
feel uncomfortable if
the basin surface is too
near. This value is set
as 50mm in our design.
User back bending
angle:
(fountain height h, Stream
trajectory max{v}, user stature)
erg
;
erg
is an ergonomic
recommended interval
of back bending angle,
we assume
erg
=[0,
90].
*Hygienic Con.
Drain rate: rate_drain(drain radius R
d
,
shape_basin a, , stream
trajectory u, v) Stream rate;
higher than stream rate
to achieve drain
effectiveness.
*Bottle Filling Con.
Height of the filler h
f
h
bottle
The average bottle
height h
bottle
Radius of the filler
basin
R
f
R
bottle
The average bottle
height R
bottle
Objective Description
S
ergo
+S
splash
: This objective is used in the engineering design
optimization phase where S
ergo
and S
splash
are scores for
ergonomic and splash performance of our design (see
Chapter 6 for more information). This objective is further
introduced in to the logit model where we estimate our
market share in the economic optimization.
Profit P: Profit is our ultimate objective of the economic
optimization based on all the analysis, surveys and
assumptions we made. (see Chapter 7 for more
information)
22
6. The Engineering Design
All pieces of engineering design discussions are presented here: An optimization method
is first used to find the best height for the fountain; then both a mathematical model and a
beta prototype is built to search for the best basin shape in terms of splash reduction and
drain rate improvement; finally, we exam the drain efficiency by determining the drain
radius. During this process, assumptions are made based on both online database and
common sense.
6.1 Design Attributes
Ergonomic Consideration
The performance of a drinking water fountain is highly related to its interaction with
users. From our survey analysis as well as own experience, current designs lack credits to
prove itself as an ergonomically pleasing interface. Unfortunately, as we found out in our
project, designing a fountain in a more ergonomic way somehow conflicts with ADA
requirement, which is an essential attribute in design and market though not mandatory.
Current designs treat this dilemma in two ways. One, adjust the basin to a lower position
to meet ADA requirement of fountain height. The cost of this compromise will be high if
the fountain is installed at some place with rare ADA access like gym. The other way is
to make modulus products, usually two fountain sets with different height to both meet
the needs of ADA requirements and achieve better ergo performance. The team also
discussed the possibility of making the fountain vertically adjustable, and we agreed that
there will be much more trivial things to take care about if we go this way, and moreover,
since adjustable basin involves the new problem of increasing user time consumption, we
dont think it is wise to bring ourselves this problem.
Thus, in our design, we mainly focused on improving the fountain ergo performance by
determine the best basin height for both male and female users. An optimization model
on this issue is later built and calculation carried out. In the model, we defined a scoring
mechanism based on user bending angle, and the target is written as a weighted sum
value of total score.
Hygienic Consideration
This part contains two goals: increasing drain rate and minimizing splash.
Splash
Splash is annoying and it always exits with drinking water fountain. Two ideas correlated
in this issue and will help us to move forward are How to efficiently reduce splash with
a given basin size? and What is the minimum basin boundary to hold the splash?
Intuitively, adding beads at appropriate places in the basin will tackle the first problem
and with a given stream rate and basin depth, it is easy to define a basin by bounding it
23
with maximum splash radius. Since none of the team member is expertise in fluid, we
used experiments to iteratively find the solution for these two problems and the derived
the converged result.
Drain Rate
A major concern about drinking water fountain is its hygiene. Two problems exist:
firstly, lead contamination occurs in public fountain; secondly, the basin surface was
reported to have significant amount of germs. We decided not to include the former one
in our design since it involves piping system. The latter problem is treated as equal to
minimize the droplet remain on the basin surface after use. This design target then helps
us to form the main spine of the basin.
A covered bubbler is also used (like current designs) to prevent lips or flow of water from
mouth to splash on the spout itself.
Strength Consideration
Besides the above, we also build an approximated model to calculate the basin bending
stress by using theoretical analysis and making assumptions to simplify our problem.
This model is further used as a nonlinear constraint in the optimization model.
Please understand that we did not use Catia FEA model in our final optimization model.
However, a separated FEA model can be found in the progress report. Commercial
software will be used to link Catia FEA model with our profit model (See Chapter 8) in
our future work. (See Chapter 10)
6.2 Design with Mathematical Methods
Ergonomic Optimization Model
The basic idea is to use back bending angles as an indicator of user satisfaction. A survey
analysis with the help from Dr. Erin Macdonald helped us to quantify user preference on
the bending angle. The result is shown below:
24
Figure 10 Determine users satisfactory by bending angle
It is arguable whether we need more angles on human body besides back bending angle
to better interpret user experience, however, it is reasonable to identify back bending
angle as the major indicator of user satisfaction.
A detailed list of variables, parameters, constants and objectives is shown below:
Table 3 Nomenclature for Ergo Design
Name Description
H = Stature height of the user, 5% -95% for both male and female Constant
h = Basin height, or more precisely, the nozzle height Variable
h
w
= Maximum stream height: max{v} Linking variable
U = The distance from mouth to the top Parameter
r = golden ratio (0.62) Constant
= User bending angle Parameter
S() = Score for a certain Parameter
S
E
= Total ergonomic score Objective
S
M
= Score for male user Parameter
S
F
= Score for female user Parameter
M
= Mean value for male stature height distribution Constant
F
= Mean value for female stature height distribution Constant
M
= Standard deviation for male stature height distribution Constant
F
= Standard deviation for female stature height distribution Constant
P
M
(H) = Probability density of stature height H for male users Parameter
P
F
(H) = Probability density of stature height H for female users Parameter
25
With the assumption that people drink water from the highest point of the stream
(max{v}) and the bending point divides upper and lower body with golden ratio, the
bending angle is given by:
= acos(
hw + h Hr
H(1 r)
)
Based on the regression of the conjoint analysis result, score (utility) of this certain angle
is:
S = 0.072
2
+ 0.541 0.232
And the weighted score is:
S
i
= P
i
(H)S()dH
Where i = F, M.
And thus the total score is given by:
S =
1
2
P
i
(H)S()dH
H
i 95
H
i 5
i=M,F
Constants and parameters are given values based on database and assumptions:
Data Source
U =200mm Assumption
H
F95
= 1736mm, H
F5
= 1521mm Anthropometric Reference Data for
Children and Aults: U.S. Population,
1999-2002
1
H
M95
= 1895mm, H
F5
= 1628mm
F
= 1628mm,
M
= 1767mm
F
= 99.68mm,
M
= 78.03mm
A MATLAB toolbox is then used to find the optimal solution for this program, and the
results are given below:
Table 4 Optimization Results for Nozzle Height
Optimization Results:
h (the nozzle height) = 133.2 cm
S
E
(Total ergonomic score) = 0.52
1
Here we use data for U.S. citizens in the range between 20 29 years old. By keeping
the same methodology, the program can be easily modified for other user groups.
26
Basin Spine Design
In this part we try to minimize the droplet remained on the basin surface. Although there
are other ways to achieve this goal, e.g. changing the material to increase water
repellence, we focus on the basic physical method to maximize the usage of gravity. The
existing Brachistochrone solution helps us to find the shape of the curve down which a
bead sliding from rest and accelerated by gravity will slip from one point to another in the
least time. Although it is arguable since fluid behavior is different from that of solid
bead, however, this method is still applicable because it meets our goal to maximize the
usage of gravity.
The Brachistochrone curve is actually the equations of an inverse cycloid, as shown
below:
x = a( sin())
y = a(1 cos())
This equation set assumes zero friction, which is not the fact in our case. We will resort
to the equation set which enables dynamic friction after we decide the basin surface
material.
In the similar fashion as we did in the ergonomic model, a detailed list of variables,
parameters, constants, constraints and objectives is shown below:
Table 5 Nomenclature for Ergo Water Fountain
h
n
= Nozzle height from drain Variable
l
n
= Nozzle distance from drain Parameter
v = Stream velocity Parameter
= Nozzle shooting angle Variable
= Contact angle between stream and basin surface Objective
a = Character value for basin spine Parameter
= , Character angle for spine at the drain Constant
R = 3.5, Radius of the supporting pipe Constant
D
erg
50mm
, Distance between user head and basin surface Constraint
L
b
= a, Basin length limitation Constraint
ADA = The distance for ADA requirement Constraint
Explanation of the nomenclature is given below:
1. From common knowledge, people will feel uncomfortable is the basin surface is too
near when they are drinking. This value is set as 50mm in our design.
27
2. Basin length limitation: From the point of material cost, the idea length will be
exactly long enough to cover the stream trajectory. With the given design function of
basin spine and the assumption of stream rate and angle, visual experiments are
carried out to find the best spine length.
3. ADA requirement on stream trajectory height: This height should be more than 4
inches from the basin to enable bottle insertion.
Figure 11 ADA requirements for bottle filling
4. The lower end of the basin spine should have a descent direction to ensure drain is
effective. For our equation set, this indicates [0, ]. Here we fix
d
= for at
the drain.
Figure 12 The contact angle
5. The nozzle shooting angle should not be zero, i.e. shooting directly upwards is not
allowed, according to the regulation. Here we set
6
,
2
.
6. From our design concept, the nozzle is place on a pipe of 3.5 inches. Refer to
Engineering Analysis for details about this pipe.
7. The angle of the stream hitting the basin (the contact angle) should be minimized to
reduce the splash. We found certain research report to prove our intuitive guess.
A mathematical model is then built in MATLAB to play with and optimize the contact
angle. The results are shown below.
28
Figure 13 Design with different nozzle position
Figure 14 Design with different shooting angle
Conclusions:
1. From our calculations, h
n
should be at least 60mm high from the bottom to meet the
ADA requirement;
2. With given stream velocity and angle, the minimum is reached when the nozzle is
placed at the highest possible position.
3. The minimum is reached at the lower boundary of .
0 0.05 0.1 0.15 0.2
-0.12
-0.1
-0.08
-0.06
-0.04
-0.02
0
0.02
0.04
Nozzle at 1.5r
Nozzle at 2r
0 0.05 0.1 0.15 0.2
-0.12
-0.1
-0.08
-0.06
-0.04
-0.02
0
0.02
0.04
Shooting Angle = 30deg
Shooting Angle = 45deg
29
Figure 15 Final design characteristics
Drain Radius Design
Sufficient drain radius will ensure the whole system performance. Here we calculate the
radius based on the model weve discussed above. Notice that this is only a first order
model where we treat droplets as solid rigid bodies and basin as ideally smooth surface;
while the stream pressure at the nozzle and drain are the same (we cut a hole in the drain
pipe to meet this goal). Thus the drain radius is given by the XXX:
2
1 =
2
2
, where R
n
is the radius of the nozzle, v1 and v2 are stream velocity at the nozzle and the
drain accordingly.
Notice that the stream lost part of its velocity when contacting the basin surface, and
there is a height gap between the nozzle and the drain. Based on the optimized design, the
relationship between v1 and v2 is then given as:
1
2
1
2
=
1
2
3
2
+(2
)
1
= 3cos()
0 0.05 0.1 0.15 0.2 0.25
-0.14
-0.12
-0.1
-0.08
-0.06
-0.04
-0.02
0
0.02
0.04
0.06
x/m
y
/
m
Nozzle
Shooting angle = 30deg
Stream velocity = 1.6m/s
Nozzle position from drain:
hn = 3.5 inches
ln = 0
Contact angle = 37.0deg
Basin spine design:
Total length: 200.9mm
Total Height: 127.9mm
30
1
2
1
2
+ 2 =
1
2
2
2
Given values of v1(1.6m/s), a (64mm), h
n
(89mm) and (37deg) according to the design
results, the minimum value for R
d
is calculated as 2.7mm.
However, we find out later in the test run that this radius is far from enough the produce
sufficient drain rate. The problem could be that the pressure at the drain is practically
higher than that of the nozzle, thus lead to the decrease of v2 and higher R
d
.
Strength Consideration
The simplified strength analysis model is given below:
The dimensions of the basin cross-section at the bottom, the inertial moment, and the
main axis are determined by the results from Experimental Methods section (See
section 6.3); the length of the basin is determined by Basin Spine Design; the applied
load is set as 80kg and the safety coefficient as 1.5. All values are listed below:
Table 6 Strength Analysis Var./Para.
Var./Para. Description Value Units
Parameters
M Applied load 80 kg
mu Safety coefficient 1.5 /
g Gravity constant 9.8 m/s
2
L Basin length Determined by Basin Spine
Design
mm
c Maximum distance from the
principal axis
Determined by Experiments mm
I Moment of inertia of the cross-
section
Determined by Experiments mm
4
Variable
t Basin thickness / mm
31
The stress constraint is then given by:
=
()
6.3 Design with Experimental Methods
We studied several reports on the topic of droplets behaviors. However, formulas given
in those papers are not coherent with our experimental results. The problem might be the
difference between the Reynolds number/Weber number of fountain stream and that of
single droplets used in those papers. Thus, we conducted our design with experiments
based on our beta prototype.
Beta prototype
Due to the nature of the project, it was determined that a system that could help emulate
water flow into the basin shape would be of great value. This prototype aims at
demonstrating the design concept can be realized into an actual product. It is intended as
a first step to demonstrate that the architecture proposed can be built with real materials
at an appropriate cost, and that it is likely to meet most of the design requirements,
including functionality and user appeal.
Ideally it is portable, so it can be used for demonstration purposes as well.
Figure 16 Block diagram of beta prototype
As can be seen in Figure 16, there is a submersible water pump contained in a reservoir
that feeds through a wall into the basin. The water is collected and returned into the
container creating a closed loop.
32
Figure 17 Beta prototype top view
The basin shape of this prototype is built with copper rods as its skeleton and a
waterproof curtain as its surface. The side view of the skeleton is formed by previous
discussion on spine. The curtain is clipped to the skeleton at five fixed points. This
fixture helps us to easily reshape the basin by changing its length, width and depth.
Figure 18 Beta prototype side view
Basin Boundary and Depth Design
The idea here is to determine the basin boundary with the maximum splash radius along
the spine, and thus minimize the basin surface area. Besides, it is intuitively true that we
33
can reduce splash by using deeper basin, however, this method also increase the basin
surface area. Thus, our goal is to make the best compromise between splash and surface
area via experiments.
In detail, we separate the clips into two sets: the upper three (A, B and C as shown in the
figure below) and the lower two. By loosening or tightening these two sets of clips, we
can examine the different performance of splash and drain rate. The results are shown
below:
Figure 19 Beta prototype fixture
Table 7 Design Methods for Experiments
Method Observation
Case Tighten Loosen Splash/Drain Surface area
1 A B C D E Upper side splash 0
2 D E A B C Upper side splash
Water remains in the basin
+
3 All / Returning water goes out of
the basin
-
4 / All Good performance ++
Conclusions:
1. Both in case 1 and 2, upper side splash occurs because the stream almost hits the
upper boundary of the basin, which is true in the previous math model. So the basin
spine should be extended to hold this part of splash.
A
B C
D E
34
2. In case 2, water remains in the basin because the basin surface is almost flat thus the
drain is not efficient. So clip D E should be loosened appropriately to enhance drain
rate.
3. In case 3, water goes out of the basin because depth is not enough and thus enables
water to flow in other paths. From the point, we decide to add beads to create a
definite drain path. Also, when placed at proper positions, beads also help to reduce
splash.
4. In case 4, we obtain a fairly good performance by loosening all the clips, however,
this also increases the total surface and in turn the material cost.
Although there is no quantitative method available, iterative tries can help us to find a
qualified performance for both splash and drain rate. Both basin boundaries and depth are
decided by the experiment and used in the CAD model.
6.4 Conclusion
As a conclusion, the design flow chart is shown below. This is further used in the
microeconomic model where the whole picture is completed.
Figure 20 Design flow
35
7. Choice Based Conjoint Analysis
7.1 Logit model
Briefly, a logit model with CBC (choice-based conjoint) analysis serves to estimate the
probability of people choosing one product among the others, and further estimate the
market share of this product. This probability is given as:
, where i represents the i
th
consumer, j represents the j
th
product and v
ij
is the utility value
of product j according to consumer i (here we assume that consumer preferences are
identical, so i can be omitted). In our model, the detailed form of %market is given
below:
%
, where
is the utility value of choosing not to buy any product given.
7.2 Composition of the utility and CBC results
The utility values can be further decomposed into beta values which indicate peoples
preferences on different design attributes. The beta values used in this model is listed
below:
Table 8 Beta values
Beta Description
Beta_color User preference on the fountain color
Beta_filler User preference on bottle filler efficiency
Beta_ergo The same as S
E
in Chapter 6
Beta_splash An alternative expression of contacting
angle , it is defined as:
(-200tan()/pi+50)/50
Beta_price User preference on the price
The values of these betas are obtained by conducting a CBC survey. The survey results
with regression curves are shown below:
36
Figure 21 Beta values for bending angles
Figure 22 Beta values for colors
Figure 23 Beta values for filling times
Figure 24 Beta values for prices
37
8. The Microeconomic Model
Economic analysis for fE consists of estimating the annual US market size of drinking
water fountain, estimating fixed and variable costs for producing water fountain,
estimating market share via logit model and then maximizing profits using optimization.
The complete model is show below while methods used in brackets. Details are discussed
in the following.
Figure 25 Microeconomic model
8.1 Market Size
The market size is given by an authenticate business report
1
. Here we use the number
from year 2007 as the total market size of drinking water fountain. And we assume that
the premium market our product will aim at is roughly 15% of the total market.
Figure 26 Market size estimation
1
SBI Market Profile: Plumbing Fixtures and Fittings a market intelligence report,
Find/SVP published products, NY, 1998 (Note: 2007 estimate only)
4230041600
45700
42300
44400
50400
52000
57000
64000
1992 1993 1994 1995 1996 1997 1998 2003 2007
US Drinking Fountain
Shipments
38
8.2 Cost Estimates
Since the product is similar to current products in the market, cost estimates were
obtained by comparing the prices of competitive products. Also, prices of various
replacement parts were obtained that helped to calculate the variable cost of the product.
One thing needs to mention is that the basin price is based on its volume, thus the unit
cost is not a constant. The following table shows the variable costs of components.
Costs independent from design variables
Table 9 Components cost
Variable costs
Component Price
Bubbler 20
Trigger 18
Bottle filler 120
Chiller 440
Supporting 225
LCD screen 80
Basin **
Total 903+
Manufacturing costs
Labor 100
Packaging 25
Shipping 50
Total 175
Total Cost/Unit 1078+
Fixed Costs
Casting Mold 300000*
Assembly Eq 100000
Rent 25000
Salary 250000
Overhead 50000
Total 725000
* We use cast here because the basin shape we created is hard to stamp.
* *The basin price is based on its volume.
Costs dependent on design variables
The basin price is based on its volume and hence linked to basin spine, width and
thickness. An analytical model is built as follows:
39
=
+
2
, where w1 and w2 are determined by experiments, t is the basin thickness and
is
the total length of the spine curve, which is given by the following calculation:
=
2
+
2
0
=
1 + (
)
2
0
= (1 )
2
+
2
0
= 4
The unit cost for the basin is approximately $5/kg, and the density of the material is set at
7.8e-6kg/mm
3
. Thus the price for one basin is:
=
+
2
4 (7.8 10
6
kg
mm3
) $5/kg
8.3 Market Share Estimates
8.3.1 Assumption
Predicting market share is always challenging but still essential. In this part we introduce
the Logit model and CBC analysis results (from Chapter 7) to address this problem.
However, some critical assumptions should be highlighted at the first place:
1. The preference of drinking water fountains is coherent throughout consumers and
users. This might be the most arguable assumption we made. However, this
assumption helps us to simply the problem.
2. There are only a few competitors in the market and the products families of these
competitors are both limited in number and stable in design attribute.
3. Design attributes will not change, i.e. people only consider about certain attributes
we analyzed and will not turn to other ones.
8.3.2 Utilities from competitors and fE
Based on the conjoint analysis results in Chapter 7, we calculated utilities from our
product (
)
*
:
To address to the ADA requirement on fountain height, we resort to two installation
fixtures: one fountain installed at ADA height (80cm), or bi-leveled fountains with both
ADA height and optimized height. The utilities of both fixtures are listed below:
40
Table 10 Utility for fE
fE fixture 1 Exterior
color
Bottle
filling time
Bending
angle
(Ergo)
Contacting
angle
(Splash)
Price Utility
Cooler N/A Chapter 6 Chapter 6 Variable
Beta_color Beta_filler Beta_ergo Beta_splash Beta_price
0.34 0.53 0.52 0.04 / /
fE fixture 2 Exterior
color
Bottle
filling time
Bending
angle
(Ergo)
Contacting
angle
(Splash)
Price Utility
Cooler N/A Chapter 6 Chapter 6 Variable
Beta_color Beta_filler Beta_ergo Beta_splash Beta_price
0.34 0.53 0.52 0.04 / /
*
:
Two premium market competitors utilities are estimated by determining their beta
values. Both of these are from ELKAY Manufacturing Co., Competitor 1 has bi-leveled
fountains which is similar to our selling strategy; Competitor 2 bears some critical
characteristics that are identical to our design (recessed cooler, redesigned basin). Also,
we assume the ADA requirement is complied with in all cases. The resulting utilities are
shown below:
Table 11 Utilities for competitors
Competitor 1 Exterior
color
Bottle
filling time
Bending
angle
(Ergo)
Contacting
angle
(Splash)
Price Utility
Natural N/A According
to ADA
requirement
Beta value
estimated
directly
$1962
Beta_color Beta_filler Beta_ergo Beta_splash Beta_price
-0.08 -0.57 -0.57 (h =
100cm)
-0.3 0.65 -0.87
Competitor 2 Exterior
color
Bottle
filling time
Bending
angle
(Ergo)
Contacting
angle
(Splash)
Price Utility
Natural N/A According
to ADA
requirement
Beta value
estimated
directly
$1800
Beta_color Beta_filler Beta_ergo Beta_splash Beta_price
-0.08 -0.57 -1.87 (h =
80cm)
-0.3 0.97 -1.85
41
*