Technical Support Report for Regulatory Action
The Effect of
Dynamometer Inertia Weight Simulation
on Fuel Economy Measurements
February, 1976
Thomas R. Norman
and
Thomas Rarick
Notice
Technical support reports for regulatory action do not
necessarily represent the final EPA decision on regulatory issues.
They are intended to present a technical analysis of an issue and
recommendations resulting from the assumptions and constraints of
that analysis. Agency policy considerations or data received sub-
sequent to the date of release of this report, may alter the recom-
mendations reached. Readers are cautioned to seek the latest analysis
from EPA before using the information contained herein.
Standards Development and Support Branch
Emission Control Technology Division
Office of Mobile Source Air Pollution Control
Office of Air and Waste Management
U.S. Environmental Protection Agency
-------
CONTENTS
1. Introduction 1
1-1 Background 1
1-2 Current Facilities 1
1-3 Problem Statement 3
2. Technical Discussion 4
2-1 Variability 4
2-2 Criteria for Selecting an Acceptable Inertia Weight
Interval Scheme 6
2-3 Effect of Inertia Weight on Fuel Economy 8
2-4 Alternative Solutions 8
2-5 Lead Time 10
3. Dynamometers 11
3-1 Current Dynamometer Configuration 11
3-2 Configuration A 13
3-3 Configuration B 14
3-4 Configuration C 15
3-5 Configuration D 17
3-6 Configuration E 20
3- 7 Summary 21
4. Reference 22
-------
1. Introduction
1.1 Background
Exhaust emissions and fuel economy measurements made by EPA are
done by sampling the vehicle's exhaust while it is operated on a Direct
Drive Variable Inertia Flywheel -(DDVIF) dynamometer. The vehicle is
placed on the rolls of the dynamometer, and the inertia weight and
horsepower specified in the Federal Register (1) for that weight vehicle
are set. The inertia weight is set by engaging the proper set of flywheels
to the dynamometer rolls. Since the simulation of inertia weight is
done by the use of flywheels, the actual test weight of the vehicle
(curb weight + 300 pounds) cannot be simulated. Instead, the nearest
inertia weight setting to the vehicle's test weight is selected.
The Federal Register currently specifies inertia weight intervals
of 250 pounds for vehicles between 1000 and 3000 pounds test weight and
500 pound intervals for vehicles between 3,000 and 5,500 pounds test
weight for light duty vehicles. Proposed for 1978 is the requirement to
test light duty trucks at weight intervals the same as for light duty
vehicles but with the weight extended up to 6500 pounds test weight
(curb weight + 500 pounds). In addition, some trucks above 6500 pounds
test weight may be tested as light duty trucks at the manufacturers option.
Weights above 5500 pounds are in 500 pounds intervals. The inertia
weight-horsepower table for light duty vehicles from the Federal Register
(1) and the proposed table for light duty trucks are given in Table 1-1.
1.2 Current Facilities
The equipment currently used at the EPA certification testing
facility includes six light duty DDVIF dynamometers with a maximum
inertia weight simulation capability of 5500 pounds. These dynamometers
have the capability of simulating inertia weight in 250 pound intervals
from 1750 pounds to 3000 pounds and in 500 pound intervals from 3000
pounds to 5500 pounds. Each of the dynamometers is enclosed in a test
cell. In addition to these dynamometers there is a medium duty DDVIF
dynamometer which is capable of testing in the weight range 4500 to
10,000 pounds in 500 pound intervals. This dynamometer is in the light
duty certification soak area and is currently not enclosed in a test
cell. Similar equipment to that used by EPA is found in the test facil-
ities used by the manufacturers of light duty vehicles and light duty
trucks.
-------
-2-
Table 1-1
Inertia Weight-Road Load Table for
Light Duty Vehicles and Light Duty Trucks
Loaded vehicle
weight, pounds
Equivalent
inertia
weight
pounds
up to 1,225
1,226 to 1,375
1,376 to 1,625
1,626 to 1,875
1,876 to 2,125
2,126 to 2,375
2,376 to 2,625
2,626 to 2,875
2,876 to 3,250
3,251 to 3,750
3,751 to 4,250
4,251 to 4,750
4,751 to 5,250
5,251 to 5,750
5,751 to 6,250
6,251 to 6,750
6,751 to 7,250
7,251 to 7,750
7,751 to 8,250
8,251 to 8,750
8,751 to 9,250
9,251 to 9,750
9,751 to 10,000
1,000
1,250
1,500
1,750
2,000
2,250
2,500
2,750
3,000
3,500
4,000
4,500
5,000
5,500
6,000
6,500
7,000
7,500
8,000
8,500
9,000
9,500
10,000
Road load power at 50 m.p.h.
Horsepower
Light Duty
Vehicles (1)
5.9
6.5
7.1
7.7
8.3
8.8
9.4
9.9
10.3
11.2
12.0
12.7
13.4
13.9
14.4
Light Duty
Trucks (2)
9.5
10.3
11.2
12.0
12.8
13.6
14.5
15.3
16.
17,
19.
21.0
22.7
24.3
25.9
27.6
29.2
30.9
32.5
34.2
35.8
37.4
39.1
(1) Light duty vehicles over 5,750 pounds loaded vehicle weight
shall be tested with a 5,500 pound equivalent inertia and a 14.4 horsepower
road load.
(2) Light Duty Truck Regulations are proposed for model year 1978.
-------
1.3 Problem Statement
Thsre are two problems with the current dynamometer equipment used
for testing. First, there is inherent inaccuracy in testing due to the
simulation of the vehicle's weight and road load power in intervals.
Secondly, the Clayton DDVIF dynamometer in use by EPA and the light duty
vehicle and truck manufacturers does not have the capability to test
vehicles above 5500 pounds equivalent inertia weight.
The problem of the inherent inaccuracy due to testing in intervals
can best be visualized by means of an example. Under the current requirements
a vehicle whose test weight is 3250 pounds is tested at 3000 pounds
inertia with a horsepower setting of 10.3 horsepower. Likewise, a
vehicle whose test weight is 3251 is tested at 3500 pounds at a horsepower
setting of 11.2 horsepower. These two vehicles are essentially the same
weight, but the vehicle tested at 3500 pounds is required to move a
weight 500 pounds greater than the vehicle tested at 3000 pounds and
operate against a road load power 0.9 horsepower greater. Thus, the
vehicle tested at 3500 pounds must work harder to operate during the
driving cycle and higher exhaust emissions and lower fuel economy values
would be expected.
The inertia weight and road load power absorbtion simulation are
dependent on the weight of the vehicle and table values based on the
test weight are used. Data presented in this report indicate the effects
on fuel economy of varying both inertia weight and road load power
absorbtion simultaneously according to the requirements of the Federal
Register (1). Thus, for the duration of this report only improvements in
inertia weight simulation will be considered with the assumption that
this would also include similar improvements in road load power simulation.
In the future, road load power simulation could be made more accurate
independent of inertia weight simulation, and thus improvements could be
made beyond those discussed. This report will primarily deal with
improvements in inertia weight simulation and its effect on the accuracy
of fuel economy measurements. It is assumed that any improvement in the
accuracy of measuring fuel economy levels will also improve the accuracy
of exhaust emission measurements. Fuel economy measurements are considered
more critical because these values are used to rank vehicles according
to their fuel economy capabilities. This, in turn, can affect vehicle
marketability. Improving the accuracy of exhaust emission measurements
is also important and this would occur with any improvement in inertia
simulation. However, this report will limit its discussion to fuel
economy accuracy.
The ideal situation would be to test a vehicle at its actual test
weight. However, this level of accuracy in simulation is difficult to
justify due to its relatively high cost and the fact that other sources
-------
-4-
of inaccuracy influencing the test results may bo improved more eout
effectively. The fundamental sources of test variability are discussed
later in this report. There is also variability in the weight of the
same vehicle model from one vehicle to another. Thus, the optimum
situation may be described as simulating the test weight of the vehicle
closely enough such that the inaccuracy in fuel economy measurements due
to weight simulation is small compared to overall test variability, and
of a similar magnitude as other sources of error. Since the cost of
simulating a vehicle's test weight increases as greater accuracy is
achieved, practical considerations need to be kept in mind.
In order to decide what degree of simulation is required, the
magnitude of other errors and the magnitude of overall test variability
must be examined. With this knowledge, a reasonable level of accuracy
may be chosen. Coupling the desired levels of accuracy with information
relating the effect of inertia weight on actual fuel economy values will
provide the means to select an appropriate inertia weight simulation
interval. This is the primary objective of the investigation.
The secondary objective is to examine alternative dynamometer con-
figurations which can be used to achieve improved inertia weight simula-
tion as well as to provide the expanded range of weights necessitated by
the proposed regulations on light duty trucks.
2. Technical Discussion
2.1 Variability
Variability continues to be an important factor when considering
fuel economy and exhaust emission testing. In a general sense, vari-
ability may be interpreted as the inability to duplicate previous results.
Variability in vehicle testing may be described as the inability to obtain
exactly the same fuel economy or emissions results while striving to re-
create the original testing conditions. There are several fundamental
sources of variability: (1) the vehicle being measured, (2) the measuring
process, (3) the personnel, and (4) the environmental conditions.
1) The Vehicle Being Measured. When considering chassis dynamo-
meter testing, there are many potential sources of variation in the
vehicle itself. There are limited data available to establish the re-
spective contributions to overall test variability.
-------
-5-
a) Vehicle preparation. Vehicle preparation prior to testing
involves physical handling of several components that could affect
subsequent test results. Differences in handling the fuel induction
system, for instance, could result in different evaporative canister
loading. Soak time and soak temperature could affect the condition of
the intake manifold, choke system, and exhaust treatment devices which,
in turn, could affect the test results.
b) Sensitivity inherent in the design of the vehicle. Carburetor
calibration, EGR calibration, evaporative system and fuel system design
are different for various vehicles. Under given test conditions each
system may respond in a different manner. Even in vehicles with identical
components, there exists variability due to slight differences in each
component and their interactive effects on the entire system.
2) The Measuring Process. When reviewing total test variability,
the measuring process has a significant impact. Each piece of equipment
contributes to the test variability depending on its precision and re-
peatability.
Fundamentally, accuracy can be viewed as the level of observation
one chooses to accept. The smallest unit of measurement (also called
"least count") should be selected to adequately reflect the desired
level of observation. A level of observation should be chosen based on
technical needs, but also with an awareness of the equipment available
to meet those needs. A device or system whose accuracy exceeds the
immediate need, however, may prove cumbersome, overly complicated, and
invariably more expensive. Accuracy, as applied to the field of chassis
dynamometer testing, might be explained through the example of inertia
flywheel weight selection. It will be shown later that the existing
limitation in flywheel selection systematically biases the fuel economy
results. The magnitude of this bias is directly related to the accuracy
in simulating the vehicle inertia.
Repeatability is the ability to produce consistent results when sub-
jected to the same conditions. When the repeatability faulters, the in-
tended accuracy becomes less meaningful. In other words, even though a
device or system has the capability of extremely accurate measurements,
this benefit is reduced if the device or system is unable to satis-
factorily repeat the values. One source estimates the maximum vari-
ability of the measuring process to be +5% for replicate testing (2).
Restated, this estimates the maximum repeatability error to be +5%.
Several fundamental sources of equipment-related variability should be
considered:
-------
-6-
a) Dynamometer. The data presented in Table 2-1 estab 11 shea the
estimated error in fuel economy values associated with the current Inertln
increment schemes. This systematic bias may be minimized through better
inertia weight simulation. The repeatability of the current dynamometer
system, however, is yet another, factor contributing to test variability.
It has been estimated that the present direct drive variable inertia fly-
wheel (DDVIF) system has a repeatability error of ±2.5% (2).
b) Constant volume sampler. Factors affecting the variability of
this piece of equipment could be any of the following: temperature and
pressure fluctuations, non-proportionate sampling, varying blower flow
characteristics, leaks or condensation. The repeatability error estimated
for this part of the measuring process is +1.0% (2).
c) Analyzer system. Instrument technology has kept pace reasonably
well with the demand to measure very low concentrations of pollutants.
However, variability still exists and must be considered when performing
fuel economy tests. Common sources of variability include stability
(drift), interference from water vapor or other gases and flow rate
stability. A prime source of variability in this area is the calibration
gas. The variation attributed to this portion of the measuring process
could be expected to be on the order of +0.5% (2).
3) The Personnel. The inability of one person to consistently
duplicate a specific technique is a prime source of variability in test
results. The variability increases with the number of persons utilized
in the testing.
4) The Environmental Conditions. Barometric pressure, humidity,
temperature, and air circulation in the test site may influence the test
results from the vehicle, and thus any variation in these parameters
introduces variability. These factors combined could be expected to
influence the repeatability by +1% (2).
2.2 Criteria for Selecting an Acceptable Inertia Weight Interval
Scheme
The weight of a vehicle affects both the exhaust emissions and fuel
economy capabilities. A decrease in vehicle weight will generally
result in lower exhaust emissions and greater fuel economy. Thus, there
is a real advantage in reducing the weight of a vehicle and the auto
manufacturers should be encouraged to do so. There is currently an
incentive to the auto manufacturers to take enough weight out of their
vehicles such that they will be tested at the next lower weight class.
Because of the size of the current weight intervals, It is very difficult
to take enough weight (as much as 500 pounds) out of some vehicles. The
manufacturers are expected to concentrate primarily on reducing the
weight of vehicles which are just above the cut-off weight between
-------
-7-
inertia weight classes so that the goal of having the vehicles tested at
the lower weight can be achieved for a minimum of effort, therefore,
one criterion for establishing the proper weight intervals is to have
them small enough so that there is an incentive to reduce weight in all
vehicles. Any reduction in the current interval size will help to in-
crease the incentive to reduce weight.
*
If the manufacturer does take a small amount of weight out of a
vehicle and thereby qualifies to be tested at a lower weight, the resulting
test will show a much larger apparent increase in the fuel economy of
that vehicle than was actually achieved by the small weight reduction.
For example, consider the case of a 2,376 pound vehicle which is required
to be tested at an inertia weight setting of 2500 pounds. Data indicate
that a test on the same vehicle at an inertia weight setting of 2250
pounds would result in a 1.0 mpg higher fuel economy value. Thus, as
the vehicle is currently tested, the result is 0.5 mpg lower than the
actual fuel economy of that vehicle. By reducing the vehicle's weight
one pound to 2375 pounds, the vehicle is then required to be tested at
2250 pounds. The small reduction in weight of one pound would have a
very small effect on the actual fuel economy of the vehicle, but the
apparent.fuel economy benefit is 1.0 mpg better than the result when it
was tested at 2500 pounds, and it is roughly 0.5 mpg better than the
vehicle's actual fuel econony. Thus, there currently exists the problem
that when a vehicle's weight is near the cut-off weight between two
inertia weight classes, the measured or apparent fuel economy can be
lower or higher than the vehicle's actual fuel economy. This problem is
inherent when inertia weight intervals are used. It is desirable to
reduce these inherent errors as much as practicable and this can be
accomplished by reducing the weight interval size. Therefore, reducing
the interval size can provide weight reduction incentives for more
vehicles, and more accurate test results. Achieving smaller and smaller
inertia weight increments is progressively more expensive and progres-
sively complicates the test. Therefore, the largest acceptable interval
must be determined.
The selection of the most desirable interval is somewhat subjective
and depends on which criterion seems to be the most important (i.e., incen-
tive for weight reduction, test accuracy, lowest equipment cost, test sim-
plicity) . The accuracy of the fuel economy test is such that the final
values are currently rounded to the nearest whole mile per gallon. The
fuel economy value that is reported is, therefore, within +0.5 mpg of the
value measured. One could select an inertia weight interval scheme such
that the error associated with testing the vehicle at the weight class
immediately above or below the proper weight class would result in an
error no larger than 1.0 mpg. An inertia weight interval scheme which
allowed anything greater than a 1.0 mpg error would have an inherent
-------
Vehicle
Weight
2000
#50
2500
2750
3000
3500
4000
4500
5000
5500
% change
in weight
to get to
next class
12.5
U^t
10.0
9.1
16.7
14.3
12.5
11.1
10.0
9.1
% change in
Fuel Economy
Ref. 1
5.0
4.1
2.7
2.4
5.9
6.9
7.0
5.0
4.9
5.1
Ref. 2
4.1
4.1
4.2
4.0
7.8
6.4
5.4
3.9
2.8
2.2
Change in Fuel
Economy (mpg)
Ref. 3
It5
1.1
.6
.6
1.1
1.1
1.1
.8
.6
.6
Ref. 2
V
1.2
1.1
.9
1.0
1.4
1.0
.8
.6
.4
.4
Table 2-1 C hanges in Fuel Economy Due to Testing Vehicle at Next
Highest Inertia Weight Class.
-------
-8-
possible error of 2.0 mpg due to the fact the fuel economy values are
rounded to the nearest whole mile per gallon. Unless the vehicle could
be tested at Its actual weight there would always exist the possibility
that a 1.0 mpg error could occur when the measured fuel economy value is
rounded. Therefore, limiting the fuel economy error to 1.0 mpg due to
errors in inertia weight simulation is desirable and further limiting
this error would merely serve to reduce the probability that a test at
the next lowest weight class would result in a 1.0 mpg higher reported
fuel economy value. The incentive to reduce weight could, in fact, be
reduced if an interval scheme were used that would result in low
probabilities that the reported value would increase by 1.0 mpg if the
vehicle could be tested at the next lowest weight class.
2.3 Effect of Inertia Weight on Fuel Economy
The data in Table 2-1 indicate the actual and percent errors which
result when a vehicle of the weight shown is tested at the next highest
inertia weight. The data shown from Reference (3) is based on 127 data
entries. Data from this source will be the one primarily used in any
further evaluation. Similar data from Reference (4) indicates that,
while the values are not identical with those from Reference (3), they
are very similar. Comparing these data also serves to illustrate another
important fact concerning fuel economy errors as they relate to inertia
weight changes. There is a wide scatter of such data between vehicles.
Summarizing and extrapolating these data show that, with the current
inertia weight increments, actual errors as large as 1.6 mpg can occur
when testing is done at an inertia weight above or below the proper one
for the vehicle's test weight and these errors occur at the lowest
inertia weight class that is currently used for testing (1750 pounds).
In addition, a maximum percent error as large as 7% can occur at the
4000 pound inertia weight class. An error as large as 1.6 mpg is unac-
ceptable with respect to the previous discussion in that it could allow
a change in the measured fuel economy of 2 mpg when a vehicle is tested
at the next lowest weight class. Such errors could create situations in
which the relative ranking of vehicles is incorrect. Therefore, an
inertia weight class scheme which reduces these errors should be proposed,
2.4 Alternative Solutions
1) Take No Action. As was stated earlier, the current situation
is one in which there exists a possibility of errors as large as 2.0 mpg
resulting from small reductions in vehicle weight. The marketability
of vehicles is becoming increasingly dependent on the fuel economy of
the vehicle relative to other vehicles, and thus taking no action to im-
prove the current inertia weight simulation is undesirable. Also, the
fuel economy values generated by the test should be reasonably accurate
so that the consumer can predict his expected operating costs prior to
-------
-9-
purchasing a vehicle. The percent errors associated with the current
weight interval scheme are such that substantial errors as large as 3.5%
in this estimate can be made. For these reasons, some improvement in
simulation should be pursued.
2) Establish Correction Factors. If accurate correction factors
could be established, the errors due to systematic inertia weight errors
could be greatly reduced. This alternative would not require modification
of the current dynamometers used by EPA or by the industry, and thus pro-
jected equipment costs would be zero. The cost to establish good correction
factors could, however, be substantial. Studies done by Murrell (5) indi-
cate that the error varies with weight and thus the correction factor
would have to vary depending on the weight of the vehicle. In addition
the correction factor could also be a function of other vehicle parameters
such as engine size and possibly axle ratio. It would be very costly to
establish tables of correction factors based on all of these factors, and
the values would have to be verified in the future should substantial
vehicle modifications be made. If great pains were not taken to assure
that the correction factors were accurate, manufacturers could contend
that the correction factors applied to their vehicles were unfair.
Further, manufacturers could be expected to optimize vehicle design to
take advantage of the parameters employed in a correction factor which
could negate the validity of the basis upon which they were generated.
Such contentions could grow more serious due to new legislation which re-
quires certain fuel economy levels to be met. For these reasons it is
not recommended that improving the current inertia weight simulation be
accomplished by the use of correction factors.
3) Use Smaller Inertia Weight Increments. In order to simulate
inertia weight closer to the actual test weight of a vehicle, additional
inertia weight flywheels will be required on existing dynamometers.
Electric dynamometers would also provide this feature. The cost of add-
ing flywheels to the dynamometers currently used both at EPA and through-
out the vehicle industry would be much lower than purchasing new electric
dynamometers and thus the addition of flywheels is the only alternative
which will be considered further. The addition of a 125 pound flywheel
to the current dynamometers would approximately halve the current,errors.
The addition of both a 62.5 pound flywheel plus a 125 pound flywheel
would cut errors roughly by one-fourth. The addition of yet smaller
flywheels would progressively reduce errors. Similarly, however, the
addition of one flywheel would roughly double the number of inertia
weight classes and the number of dynamometer settings. Likewise, two
flywheels would roughly quadruple the number of classes. Also, the
equipment cost would increase with an increasing number of additional
flywheels. Thus, adding flywheels can reduce the errors associated with
inertia weight simulation, but it is desirable to only add enough fly-
wheels to obtain an acceptable level of accuracy.
-------
-10-
Figure 2-1 shows the required intervals needed to meet various
accuracy criteria. The errors discussed are errors resulting from
testing at the next highest or next lowest weight interval. This
figure was constructed based on the data in Table 2-1. This figure
shows that a 5% maximum error is achievable by merely extending the
250 pound intervals to the 4500 pound weight class. However, this
still allows for up to a 1.6 mpg error for the lightest weight vehicles.
The addition of a 125 pound flywheel will limit errors to a maximum
of 1.0 mpg and 5%. With the addition of a 125 pound flywheel, the
maximum error occurs at roughly the 3000 pound weight class. If the
use of 125 pound intervals is extended to 4000 pounds, the maximum
error occurs at both the lowest weight class and the 4000 pound weight
class. This maximum error is 3% or 0.8 mpg when the vehicle is tested
at the next lower or higher weight class. Thus, a 125 pound flywheel
can reduce errors substantially and can meet the requirement of an
error no greater than 1.0 mpg even at the lowest weights.
The addition of a 62.5 pound flywheel will insure errors less than
0.5 mpg and 2.5%. But, since errors no greater than 1.0 mpg can be
achieved by using only a 125 pound flywheel, the use of a 62.5 pound
flywheel would primarily add cost and complexity. Thus, the recommended
course of action is to add a 125 pound flywheel to the dynamometer and
use the 125 pound intervals from 1500 to 4000 pounds, the 250 pound
intervals from 4000 to 5500 pounds, and the 500 pound intervals above
5500 pounds. This would require the addition of a 125 pound inertia
weight wheel to current light duty DDVIF dynamometers. This inertia
weight scheme insures that errors in the actual fuel economy values
due to the inertia weight simulation are no greater than 0.4 mpg at
any test weight.
Currently, there are few vehicles which are required to be tested
below the 1750 pound weight class. However, in the future, vehicles
are expected to become increasingly lighter and, therefore, it is
recommended that the capability to test vehicles at weights down to
1500 pounds be obtained.
2.5 Lead Time
If smaller inertia weight increments are required for testing
vehicles in the future, it will take time to produce, deliver, and
install the needed equipment for dynamometers at EPA and throughout
the industry. Some manufacturers may require slight modifications
while others may require modifications as extensive as those needed
at the EPA test facility.
Currently, there are approximately 300 dynamometers in the field
of the type used at,EPA. It would take at least 120 days to produce
and deliver the equipment to modify 20-30 dynamometers. In 180 days
-------
Inertia Weight, Ib.
20f° , 300° , 40|00 | 500° , ^000
1. Current LDV Requirements ^ 250 Ib. *>U - 500 Ib.
2. Maximum 1.0 mpg Error -« 125 Ib. »|* 250 Ib. *4-« 500 Ib:
3. Maximum 0.5 mpg Error -*—62.5 Ib. *H 125 1B*P 250 Ib.— »-| ^ 5QQ lb-
4.. Maximum 5% Error
250 Ib.
—-J-S 500 lb.-
5. Maximum 2.5% Error ^62.5+« 125 Ib. H- 250,11..,* | • 500 Ib.
Ib.
t 500 Ib.
6. Maximum 1.0 mpg and 5% Error — 125 Ib. »"-* 250 Ib.
7. Maximum 0.5 nipg and 2.5% Error -*— 62.5 -*j-* 125 Ib. *•!•* 250 lbrr-»
«—500 Ib.
Ib.
8. Recommended — 125 lb- ««-U 25° lb* » i* 500 Ib.
I - 1 - 1 - 1 - 1 - 1 - 1
2000 3000 4000 5000 6000
Inertia Weight, Ib.
Figure 2-1 Inertia Weight Increment Ranges Required to meet various criteria.
-------
-11-
equipment could be supplied for 75 dynamometers, and it would take
roughly 360 days to equip 175 dynamometers. Thus, based on these
estimates by Clayton Manufacturing Co., it would take approximately one
year to equip half of the dynamometers in the field.
In addition to these production and delivery times, time will be
required for equipment installation. This time will depend on the
extent of the modification. If no modification to the dynamometer pit
is required, it would take about one week to install the needed equip-
ment. Depending on the extent of pit modifications required, this time
changes.
3. Dynamometers
The following dynamometer layouts are intended to familiarize the
reader with the existing light duty vehicle (LDV) certification dynamo-
meter configuration and the dynamometer configurations which might be
employed to satisfy the following requirements:
1) Inertia weight simulation up to 6,500 pounds.
2) Finer increments in the inertia weight simulation (to improve
fuel economy error).
Each configuration has associated with it certain advantages and
disadvantages. These considerations have been addressed where possible.
It is understood, however, that each dynamometer user has a specific set
of requirements. What may be an advantage to one user, may be a dis-
advantage to another user. These dynamometer configurations have been
evaluated with respect to the requirements previously mentioned and
apply specifically to the facilities at the Motor Vehicle Emission
Laboratory, Ann Arbor, Michigan.
3.1 Current Dynamometer Configuration.
Light duty vehicle certification work is performed exclusively on
the Clayton Direct Drive Variable Inertia Flywheel (DDVIF) dynamometer
equipment. Figure 2-2 displays the general equipment design.
-------
-12-
.(>•'
i
k
ti
10
utitT
Figure 2-2 Current Equipment Design
This design consists of several fundamental systems:
1) Rolls. Two rolls are used to provide the interface between the
vehicle drive wheels and dynamometer.
2) Power Absorption Unit. This component provides the capability
of simulating actual road load at the vehicle drive wheels. Basically,
it is comprised of a water brake which is capable of absorbing a maximum
of 50 horsepower.
3) Inertia Weight Flywheels. The inclusion of these weights
provides for the simulation of actual vehicle inertia. The current
system of weights is capable of simulating from 1,750 pounds to 5,500
pounds. These weights can be easily switched in or out to provide the
test weights shown in Table 2-2.
-------
-.13-
TEST WEIGHT
1750
2UUU
2250
2500
2750
3UUU
3500
4000
4500
5UUU
5500
FLYWHEEL WEIGHTS
1750
1C
X.
X
X
K
X.
t
X
X
X.
X
250
X
X
X
X
X
X
X
X
500
X
X
X
X
X
1000
X
X
X
X
X
2000
X
X
X
X
Table 2-2 Engagement Chart for Current Equipment
The dynamometer configurations to follow are similar to the current
equipment in basic design. All configurations shown are of the Clayton
type. Let it be restated that the relative merits of these configurations
are based specifically on the comparison with the equipment currently
used in the Motor Vehicle Emission Laboratory.
3.2 Configuration A.
Shown in the diagram below (Figure 2-3), is a configuration which
would satisfy both requirements set forth earlier.
Figure 2-3 Equipment Design for Configuration A
-------
-14-
The system could be placed in the existing dynamometer enclosure
without alterations. The system would be capable of testing up to 6,500
pounds inertia and would provide a 250 pound increment up to 5,000
pounds, and a 500 pound increment up to 6,500 pounds. However, one
serious drawback to this system is that it is incapable of simulating
inertia weights less than 3,000 pounds. Table 2-3 displays the inertia
weight combination available with this system.
Test Weight
3000 -.
^?sn ;-
3500 "
l25<
400<
3
3
4250
4500
4750
5000
5SOO
6000
fisno
FLYWHEEL WEIGHTS
2750
X
V
X
y
X
V
X
•v
X
5t
X
250
X
X
X
X
L x
X
^
X
500
X
X
X
X
X
X
1000
X
X
X
X
X
X
2000
X
X
X
X
X
Table 2-3 Engagement Chart for Configuration A
The equipment cost to modify an existing dynamometer in this manner
would be approximately $1,650. Installation would be estimated-to be
$1,200, bringing the total package cost to $2,850.
3.3 Configuration B.
Shown in Figure 2-4 is a dynamometer system which would complement
configuration^
Figure 2-4 Equipment Design for Configuration B
-------
-15-
The modification could be done without altering the dynamo-
meter enclosure. The system would be capable of simulating
inertia weights from 1,750 to 3,500 pounds. Inertia weight
increments would be 125 pounds up to 3,000 pounds inertia, and
250 pounds up to 3,500 pounds inertia. The corresponding
drawback to this configuration is that inertia weight simula-
tion is limited to 3,500 pounds..
Test Weight
1750
187S
?000
2125
2250
2375
2500
2625
2750
2875
3000
3250
T500
FLYWHEEL WEIGHTS
1750
X
•^
X
X
X
X
X
X
X
X
X
TT
- X
125
X
X
X
X
X
250
X
X
X
X
X
X
500
X
X
X
X
X
Y
1000
X
X
X
X
X
Table 2-4 Engagement Chart for Configuration B
The equipment cost would be approximately $600 to modify
an existing dynamometer. The installation cost would be approxi-
mately $1200, bringing the estimated cost of the package
to $1,800.
3.4 Configuration C.
Figure 2-5 displays a dynamometer system which could be
obtained with slight modification to the dynamometer enclosure.
-------
-16-
Figure 2-5 Equipment Design for Configuration C
This system would be capable of simulating inertia weights from
1,500 to 5,250 pounds. Inertia weight %inulation up to 4,000 pounds
would be in increments of 125 pounds. Alsove 4,000 pounds inertia, the
increment would be 250 pounds. Table 2-5 displays the combinations
available.
-------
-17-
T^st Weight
1500
lfi?S
17 SO
1R75
2000
2125
2250
2175
2500
2625
2750
2875
jnnn
1125
39^0
1175
1500
3625
3/bU
3875
^nnn
4250
Asnn
4750
snnn
s^sn
FLYWHEEL WEIGHTS
1500
X
V
X
•x
X
X
X
X
5f
X
X
X
•Jf
X
X
X
X
X
X
X
V
X
•x
X
V
X
125
X
X
X
X
X
£
X
X
X
X
250
X
X
X
X
X
X
X
X
X
X
X
X
X
500
X
X,
X
X.
TIJ
X
X
X
V
X
X
X
1000
^
X
X
X
•ft
X
X
X
X '
X
X
X
2000
.
X
X
X
X
X
X
X
X
Jf • '.
K
Table 2-5 Engagement Chart for Configuration C
The estimated equipment cost to modify an existing dynamometer
would be $2,000. Installation would be expected to be on the order of
$2,000, bringing the total package cost to $4,000. However, since this
dynamometer system does not provide testing capability up to 6,500
pounds inertia, it would have to be accompanied by one or more
dynamometers capable of doing so in order to fulfill the stated
objectives.
3.5 Configuration D.
This dynamometer package would also necessitate modifying the
dynamometer enclosure. The inertia weight simulation capability would
be from 1,500 pounds to 6,500 pounds. This capability may be achieved
using either of the two following configurations shown in Figures 2-6
and 2-7.
-------
-18-
Figure 2-6 Equipment Design for Configuration D
—*f—\
Figure 2-7 Alternative Equipment Design for Configuration D
-------
-19-
Inertia weight could be simulated in Increments of 125 pounds up to
4,000 pounds, 250 pounds up to 5,000 pounds; and 500 pounds up to 7,500
pounds. Table 2-6 displays the total inertia weight simulation
capability.
TEST WEIGHT
1500
1625
1750
187 5
2000
2125
2250
2375
2500
<7V*
2875
3rmS{M>
ftX>o
6500
7OOO
7500
iFLyWHEEL WEIGHTS
1500
X
TT
X
X
X
X
X
X
x
X
•ir
X
X
X
X
X
X
x
X
x
X
x'
Jf
X
X
V
X
X
x
X
125
X
X
X
X
X
250
x
X
. X
X
*
X
X
T' -
}£
X
.....
X
X
. » .
x
x
x
X
500
V
V
Tf
V
X
X
- .Jt~. . .
X
x
x
V
X
X
1000
X
X
x
X
V
X
x
X .
X
X
X
V
X
X
2000
X
V
X
x
X
v
X
X
x
X
X
X
X
2JLQO
x
X
X
X
X
Table 2-6 Engagement Chart for Configuration D
-------
-20-
The estimated equipment cost associated with this configuration is
$10,000 to modify an existing dynamometer. Installation would cost
approximately $3,000. Thus, the total package would cost approximately
$13,000 per dynamometer.
3.6 Configuration E.
This configuration would encompass the features specified in
configuration D. However, the blank hub shown in Figures 2-6 and 2-7
would be utilized with an additional 62.5 pound inertia weight flywheel.
This would provide the inertia weight simulation capability shown in
Table 2-7.
Test Weight
1500
isfi?..s
1625
16R7.-S
1750
1819. S
1875
1937^-5
2000
2062-5
2125
79R7.S
7250
2312.5
2375
2437.5
?soo
2625
2750
2875
1000
3125
1250
3375
3500
3750
4000
4250
4500
£7Sfl
5000
5500
6000
6500
7000
7500 .
FLYWHEEL WEIGHTS
1500
X
V
X
X
X
X
X
X
X
X
X
X
X
X
X
X
y
X
V
X
X
X
X
Y
X
V
V
X
V
X
X
X
X
X
X
X
62.5
\
X
L_ X
X
X
X
X
X
X
125
1 X
V
"
X
V
X
X
X
X
X
X
X
X
250
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
500
X
•Jf
X
X
X
X
X
X
1000
X
X
X
~ X
X X
X
X
X
X
X
X
X
X
! x
•"
X
X
X
X
X
2000
2500
i
i
X
- 1
x !
X |
X
X
X X i
X !
X
X
X
X
X
X:::
X
x
X
X
k
Table 2-7 Engagement Chart for Configuration E
-------
-22-
References
1. Federal Register, Vol. 39, No. 133, Section 85, July 10, 1974.
2. Paulsell, C. D. and R. Kruse, "Test Variability of Emission and Fuel
Economy Measurements Using the 1975 Federal Test Procedure," SAE
Paper No. 741035.
3. Murrell, J. D., "Factors Affecting Automotive Fuel Economy," SAE
Paper 750958.
4. Marks, C., and G. Niepoth, "Car Design for Economy and Emissions,"
SAE Paper No. 750954.
5. U.S. Environmental Protection Agency, "Factors Affecting Automotive
Fuel Economy," Office of Air and Waste Management, MSAPC, September
1975.
------- |