The Economic Effectiveness
of Mandatory Engine Maintenance
for Reducing Vehicle Exhaust Emissions
CRC Extended Phase I Study
\
Interim Report
January 1972
In Support of:
APRAC Project Number CAPE-13-68
for
Coordinating Research Council, Inc.
Thirty Rockefeller Plaza
New York. New York 10020
TRW
SYSTIMS OXOUf
OWE SPACE PARK • RfDONDO B£ACH CALIFORNIA 902«
and
Environmental Protection Agency
Air Pollution Control Office
5600 Fishers Lane
Rockville, Maryland 20852
SCOTT RESEARCH LABORATORIES, INC
P. O. BOX 84I«
SAN BERNARDINO. CALIFORNIA 8S4O6
-------
The Economic Effectiveness
of Mandatory Engine Maintenance
for Reducing Vehicle Exhaust Emissions
CRC Extended Phase I Study
Interim Report
APPROVED BY
RICHARD R. KOPPANG
PROJECT ENGINEER
NEAL A. RICHARDSON
PROJECT MANAGER
January 1972
In Support of:
APRAC Project Number CAPE-13-68
for
Coordinating Research Council. Inc.
Thirty Rockefeller Plaza
New York, New York 10020
TRW
SYSTIMS OXOUf
OWE SPACE PARK • flfDONDO BCACH CALIFORNIA 902IS
and
Environmental Protection Agency
Air Pollution Control Office
5600 Fishers Lane
Rockville, Maryland 20852
SCOTT RESEARCH LABORATORIES, INC
P. O. BOX
SAN BERNARDINO, CALIFORNIA 9IAO9
-------
TABLE OF CONTENTS
1.0 Introduction and Summary. . . . . . . . . . . . . . . . . .. 1
2.0 System Study Results. . . . . . . . . . . . . . . . . . . . . 4
2.1 Analysis of Inspection/Maintenance Procedures Using
Mass Emissions Data. . . . . . . . . . . . . . . . . .. 5
2.1.1 Comparison of Results for Mass/ Concentration
Da ta . . . . . . . . . . . . . . . . . . . . . . . 8
2.1.2 Reoptimization of Procedures. . . . . . . . . . . 10
2.1.3 Relaxation of Optimal Procedures. . . . . . . . . 15
2.1.4 Statistical Analysis of Results. . . . . . . . . 18
2.2 System Sensitivity Studies. . . . . . . . . . . . . . . 24
2.2.1 Inspection/Maintenance Procedure Sensitivity. . . 27
2.2.2 System Configuration. . . . . . . . . . . . . . . 33
2.2.3 Operational Variables. . . . . . . . . . . . . . 37
2.2.4 Weighting Factors for Emission Reduction. . . . . 40
2.3 Regional Evaluation. . . . . . . . . . . . . . . . . . . 41
2.3.1 Regional Characterizations. . . . . . . . . . . . 42
2.3.2 Comparison of Regional Programs. . . . . . . . . 44
3.0 Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . 53
3.1 Mass Emission Analysis. . . . . . . . . . . . . . . . . 53
3.2 Sensitivity Studies. . . . . . . . . . . . . . . . . . . 54
3.3 Regional Impact of Inspection/Maintenance Effectivensss . 55
Append; x A . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
A.l Introduction......... . . . . . . . . . . . . . . . . 56
A.2 Basic System Ground Rules. . . . . . . . . . . . . . . . . . 56
A.3 Basic Input Data. . . . . . . . . . . . . . . . . . . . . . . 57
A.3.1 Emission and Cost Data. . . . . . . . . . . . . . . . 57
A.3.2 Demographic Data. . . . . . . . . . . . . . . . . . . 61
Append; x B . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
B.1 Introduction............. . . . . . . . . . . . . 64
B.2 Estimated Emission Time Histories. . . . . . . . . . . . . . 64
B.2.1 Emission Reduction Model. . . . . . . . . . . . . . . 64
B.2.2 Deterioration of Maintenance. . . . . . . . . . . . . 66
i
-------
TABLE OF CONTENTS (Cont'd)
B.3 Parameter Inspection Models. . . . . . . . . . . . . . . . . . . 71
B. 3. 1 Mi s fi re Model. . . . . . . . . . . . . . . . . . . . . . 71
B.3.2 Positive Crankcase Ventilation Valve (PCV) Model. . . . . 73
B.4 Statistical Inference Model. . . . . . . . . . . . . . . . . . . 75
B.5 Linear Programming Algorithm. . . . . . . . . . . . . . . . . . 78
B.6 Economic-Effectiveness Computer Processor. . . . . . . . . . . . 83
Referen ces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
ii
-------
Figure
2-1
2-2
2-3
2-4
2-5
2-6
2-7
2-8
2-9
2-10
2-11
2-12
2-13
2-14
2-15
2-16
2-17
B-1
B-2
B-3
B-4
B-5
B-6
B-7
B-8
B-9
LIST OF FIGURES
Ti tl e
Mass Emission Time Histories for an Optimum Idle
Inspection and Repair Program. . . . . . . . . . . . . .
Mass Emission Time Histories for an Optimum,
Extensive B Engine Paramet~r Inspection and
Repair Program. . . . . . . . . . . . . . . . . . . . .
Pre and Post Maintenance Emissions Destinations. . . . .
Changes in Parameter Rejection Fractions with Time. . .
Rejection Fractions Resulting from an Emission
Signature Inspection Followed by an Extensive B
Maintenance. . . . . . . . . . . . . . . . . . . . . . .
Effects of Pass/Fail Criteria on Five-Year Average
Emission Reductions. . . . . . . . . . . . . . . . . . .
Sensitivity of Figure of Merit to Engine Parameter
Pass/Fail Criteria. . . . . . . . . . . . . . . . . . .
Influence of Parameter Deterioration Rates on
Procedure Effectiveness. . . . . . . . . . . . . . . . .
Influence of Voluntary Maintenance on Procedure
Effectiveness. . . . . . . . . . . . . . . . . . . . . .
Influence of ICO Adjustment on Procedure Effectiveness
Artist Conception of State Lane Inspection Station
Influence of System Configuration on Procedure
Effectiveness. . . . . . . . . . . . . . . . . . . . .. 36
Influence of Inspection Period on Procedure
Effectiveness. . . . . . . . . . . . . . . . . . . . .. 37
Influence of Labor Costs and User Inconvenience Costs
on Procedure Effectiveness. . . . . . . . . . . . . .. 39
HC Emission Time History for Region IV . . . . . . . . . 49
CO Emission Time History for Region IV . . . . . . . .. 50
NO Emission Time History for Region IV . . . . . . . .. 51
HC Mass Emission Decay. . . . . . . . . . . . . . . . .
CO Mass Emi ss i on Decay. . . . . . . ... . . . . . . . .
NO Mass Emission Decay. . . . . . . . . . . . . . . . .
Percent Misfire at Test Condition Loading. . . . . . . .
Effect of PCV Valve Plugging, Carburetor, Venturi and
Throttle Body Deposits on Exhaust CO and HC Emissions. .
PCV Crankcase Pressure Distribution Pre 1971 Exhaust
Controlled Vehicles. . . . . . . . . . . . . . . . . . .
Statistical Analysis Model Flow Chart. . . . . . . . . .
Li near Programmi ng Interacti ve Scheme. . . . . . . . . .
Economic-Effectiveness Program Design. . . . . . . . . .
iii
Page
13
14
20
25
26
29
31
32
32
34
35
70
70
70
72
76
76
79
82
84
-------
Table
2-1
2-2
2-3
2-4
2-5
2-6
2-7
2-8
2-9
2-10
2-11
2-12
2-13
2-14
2-15
2-16
A-l
A-2
A-3
A-4
A-5
LIST OF TABLES
Title
Page
5
Inspection/Maintenance Procedures Studied. . . . .
Expected Emission Reduction for Typical Parameter
Adj us tments . . . . . .. . . . . . . . . . . . . . .
. . . .
. . . . 7
Comparative Inspection/Maintenance Program Results. . . . 9
Comparison of Absolute Emission Reductions Between
Mass/Concentration Data. . . . . . . . . . . . . . . . . 9
Optimized Inspection/Maintenance Program Performance
Based Upon Mass Emission Data. . . . . . . . . . . . . . . 11
Expected Emission Reductions at End of Five Years
Optimal Inspection Criteria. . . . . . . . . . . . . . . . 16
Optimal Engine Parameter and Mode Emission Pass/Fail
Criteria for Precontrolled Vehicles. . . . . . . . . . . . 16
Relaxation of Optimal Procedures. . . . . . . . . . . . . 17
Influence of Several Confidence Statements on
Estimated Maintenance Effectiveness-Diagnostic,
Extensive B Inspection Emissions. . . . . . . . . . . . . 21
Optimized Inspection/Maintenance Performance Based on
Mass Emission Reductions Stated at the Lower 95%
Confidence Level. . . . . . . . . . . . . . . . . . . . . 23
Typical Input Data Describing a Single Lane State
Inspection Station. . . . . . . . . . . . . . . . . . . . 35
Influence of Emission Weighting Function on Procedure
Effectiveness. . . . . . . . . . . . . . . . . . . . . . . 40
Optimal Parameter Pass/Fail Inspection Criteria for
Selected Regional Programs. . . . . . . . . . . . . . . . 45
Comparison of Regional Programs Employing a Parameter
Inspection Strategy. . . . . . . . . . . . . . . . . . . . 46
Comparison of Regional Programs Employing a Parameter
Inspection Strategy. . . . . . . . . . . . . . . . . . . . 46
Selected Inspection/Maintenance Programs for the
Regions Studied. . . . . . . . . . . . . . . . . . . . . . 52
Emissions Responses to Engine Parameter Changes. . . . . . 58
Equipment and Procedures Required for Diagnosing
Engine Parameter Malfunctions. . . . . . . . . . . . . . . 59
State Inspection Lane Configurations and Costs. . . . . . 60
Franchised Garage Inspection/Maintenance Labor Times. . . 60
Regional Attributes. . . . . . . . . . . . . . . . . . . . 62
iv
-------
Table
B-1
B-2
B-3
B-4
LIST OF TABLES (Cont'd)
Title
The Number of Cars Out of Those Tested for which
Statistically Significant Effects were Determined. . . .
Engine Parameter Deterioration Rates and Voluntary
Maintenance Effectiveness. . . . . . . . . . . . . . . .
Mass Emissions for In-Use Vehicles as Measured by the
1972 Federal Test Procedure. . . . . . . . . . . . . . .
Calculation of Percent Misfire for Seven-Mode Cycle. . .
v
Page
65
69
69
74
-------
1.0 INTRODUCTION AND SUMMARY
An investigation was made of the feasibility of controlling exhaust
emissions through a program of mandatory vehicle inspection and mainten-
ance. This study differed from the one previously reported in that
exhaust emission quantities, i.e., levels and reductions, were estimated
based on a constant volume sampling (CVS) procedure. The developed
economic-effectiveness model provided the vehicle for conducting this
analysis. The model has been designed to consider the following factors
in evaluating the attractiveness of inspection/maintenance:
. extent and frequency of the inspection and maintenance
procedures applied
. quantitative inspection criteria to be applied in identify-
ing and rejecting malfunctioned vehicles
. design and cost of state inspection lanes including their
number, size and location
. user time spent in travelling to and waiting at inspection
and maintenance stations
. time and cost of selective maintenance operations as per-
formed in a certified garage
. the existing and required air quality of the region under
study.
To support this model, empirical data have been developed based upon
statistically designed experiments which defined both the general mainten-
ance states of large vehicle populations and the sensitivity of exhaust
emissions to engine malfunctions. Two basic types of inspection approaches
were evaluated:
. Direct diagnosis of engine parameters (maladjustments and mal-
functions) using conventional or more sophisticated garage-
type equipment.
. Inference of engine parameter deviations from manufacturer's
specification using measurements of exhaust emission levels
(signatures) under differing engine loads.
Both of these procedures were previously evaluated using a figure of
merit based on weighted emission reductions as measured by the 1968 Federal
test procedure (composite, seven-mode). The results of that study
for the Los Angeles region showed that either the direct diagnosis and
-------
maintenance of the idle adjustments (rpm, timing and F/A) or an emissions
signature inspection at idle and load followed by maintenance of the idle
adjustments and ignition system were cost effective.
The results of the study reported herein using emission reductions
based upon constant volume sampling and the seven-mode driving cycle
support the general conclusions of the previous study with regard to Los
Angeles. Conclusions of a more specific riature are:
. The most cost-effective inspection interval is approximately
yearly regardless of the inspection/maintenance procedures
applied.
. Inspection procedures performed in a state lane using the
measurement of emission levels under load are generally more
cost-effective than those performed in franchised garages
using conventional diagnostic instruments, but produce
smaller emission reductions.
. A statistical analysis of the emission reductions for the
baseline and test fleets has revealed that the predicted
differences in mean emission levels is significant at greater
than the 90 percent level of confidence. Furthermore, the
statistically significant emission reductions at a 95 percent
level of confidence are nearly 2/3 of the predicted values.
. The cost-effectiveness of vehicle inspection/maintenance is
highly dependent on the demographic characteristics and air
quality needs of a given region.
. The figure of merit (annularized discounted total program
cost divided by the weighted five-year reductions of emis-
sions) is very sensitive to the relative weighting assigned
to HC, CO and NOx emission reductions and to the extent of
imposed maintenance.
. The figure of merit and emission reductions are highly sensi-
tive to the rate of maintenance deterioration, the effective-
ness of voluntary maintenance of ignition misfire and idle
fuel-to-air ratio adjustment and the reliability of the repair.
. On a percentage basis, emission reductions based upon mass
measurements are smaller than those based upon volumetric
measurements.
. On an absolute basis (tons/year), emission reductions based
upon mass measurements are slightly greater than those based
upon volumetric measurements.
. When using similar inspection/maintenance procedures, larger
percentage reductions for HC and NO and smaller reductions for
CO are estimated based upon mass emissions as compared to
volumetric emissions.
2
-------
. Emission reductions are most sensitive to the inspection
approach used to reject vehicles and to the extent of the
imposed maintenance.
The above conclusions are based upon interim data. The study results
will be upgraded and revised with a consistent set of emissions data based
upon larger populations of vehicles in a final report at the conclusion of
the program. The primary purpose of this interim study is to provide gen-
eral guidance for those states contemplating inspection/maintenance programs
as part of their implementing strategy for meeting the new Federal air
quality standards. Those states actually implementing programs such as
described in this report should consider performing pilot studies on a
small scale to verify that the estimated results presented herein can be
achieved within the framework of their specific requirements and within the
capability of their existing maintenance organizations.
3
-------
2.0 SYSTEM STUDY RESULTS
This section presents a discussion of the results obtained by apply-
ing the revised data bank and economic-effectiveness computer model
described in Appendices A and B. This study consisted of the following
three major parts:
(1) A comparative analysis of economic-effectiveness results
obtained from Phase One using the Federal 1968 open cycle
emission test procedure with results based upon a seven-
mode, closed CVS mass emissions test procedure.
(2) An evaluation of the basic study ground rules and assump-
tions to determine their impact on procedure effectiveness.
(3) An assessment of the feasibility of vehicle inspection/
maintenance in several typical urban centers.
The new economic-effectiveness model provided the means for conduct-
ing this study. Using this model and the developed mass emissions data
base, the performance of each of the previously studied inspection/main-
tenance procedures was simulated over a chosen time period. As in the
Phase One study, two general inspection alternatives were examined with-
in the framework of the economic-effectiveness model: (1) an engine
parameter inspection, followed as necessary, by specific parameter main-
tenance; and (2) a mode emission inspection leading to further diagnosis
and corrective maintenance.
Since each of these two basic approaches contains a large number of
possible variations, candidate strategies were selected in advance of the
actual analysis. The results of the previous study using concentration
based emissions and the input data described in Appendix A were used to
guide this selection. Table 2-1 is a summary of the inspection/maintenance
alternatives examined.
In addition to these main strategies, a program of inspecting and
maintaining only the idle air-to-fuel adjustment was examined for regions
characterized by high CO loadings. The estimated performance characteris-
tics of each of the enumerated strategies was used as a basis for conduct-
ing the comparative analysis required in Parts 1 and 3 above and for the
system sensitivity studies of Part 2.
4
-------
TABLE 2-1
Inspection/Maintenance Procedures Studied
S tra tegy
Engine Parameter Diagnostic
1) Idle (State Lane)
2) Idle (Franchised Garage)
3) Extensive A (Franchised Garage
without Dynamometer)
4) Extensive A (Franchised Garage
Dynamometer)
5) Extensive B (Franchised Garage
Dynamometer)
Inspection/Maintenance Procedure
I/M of:
I/M of:
I/M of:
lCD, rpm
I CO, rpm,
I CO, rpm,
mi sfi re
lCD, rpm, timing,
mi sfi re
lCD, rpm, timing,
misfire, A/P, PCV, A/C
timing
timing,
I/M of:
I/M of:
Emission Signature Analysis
1) Idle (State Lane)
Measure lCD, IHC
Adjust lCD, rpm, timing
Measure lCD, IHC, AHC
Adjust lCD, rpm, timing
Repair misfire
Measure lCD, IHC, AHC, CCO
Adjust lCD, rpm, timing
Repair misfire, PCV, A/C
2) Extensive A (State Lane Emission
Under Load)
3) Extensive B (State Lane Emission
Under Load)
lCD, idle CO emission measurement
IHC, idle HC emission measurement
AHC, loaded mode HC emission measurement
CCO, loaded mode CO emission measurement
2.1 Analysis of Inspection/Maintenance Procedures Using Mass Emissions
Data
To develop results utilizing the mass emission measurements, the
following data sets were converted to a mass emission basis:
. Influence coefficients relating changes in mass emissions to
changes in engine parameter settings obtained from Phase One
Orthogonal Test Program
. Vehicle emission levels based on 1972 Federal Test Procedures
from the Extended Phase One fleet deterioration program
. Emission decay rates based on converted ARB Surveillance Pro-
gram Data.
5
-------
To measure the impact of the mass emissions data on various strategies,
a comparison was first made between the mass emissions data and the con-
centration emissions data obtained from the Phase One orthogonal experi-
ments conducted to develop emission response coefficients. These data
are fundamental to predicting the effectiveness of engine parameter main-
tenance in reducing emissions. In conducting this analysis, an attempt
was made to keep as many of the system factors (e.g., pass/fail criteria)
constant within the economic-effectiveness model as possible. In general,
the mass emissions influence coefficients are consistent and similar to
their corresponding concentration values. For a given change in an engine
adjustment the resultant change in exhaust emissions when measured in mass
units was found to be somewhat smaller than when measured in concentration
units. There are, however, several significant deviations from these
general statements. Two of the most significant influence coefficients,
i.e., aHC/aICO and aCO/aRPM, have undergone a change in sign when re-
computed in terms of mass units and there has been a substantial increase
in absolute magnitude for two of the other influence coefficients, i.e.,
aCO/aICO, aCO/atiming. Table 2-2 shows expected emission reductions in
terms of both mass and concentration units for typical engine adjustments.
These values were computed using the fundamental relationship:
ae.
1
~ei = E ~P. ~ RJ' ...
i J j
(2-1)
where
~ei = expected emission reduction for ith specie
~P. = average adjustment for jth parameter
J
ae.
ap~ = influence coefficient (Table A-l)
J
R.
J
h. 1 . t . f t . f . th ( .
= ve 1C e reJec 10n rac 10n or 1 parameter 1.e., percent
of automobiles that failed the pass/fail criteria for the
jth parameter)
6
-------
TABLE 2-2
Expected Emission Reduction for Typical Parameter Adjustments
(GRAMS/MILE)
lie
6E= l:6Pj C/P. Rj
I
HC CO NO
PARAMETER MASS CONC MASS CONC MASS CONC
ICO + 0.037 -0.252* 8'.09 4.74* 0 -0.036*
RPM 0.131 0.187 -0.374 +0.274* -0.005 -0.006
TIMING 0.082 0.126 -0.482 -0.095* 0.165 0.192
MISFIRE 0.333 0.365 - -
AlP 0.026 0.071 0.282 0.401 0 0.002
PCV 0.039 0.047 1.04 1.56 -0.369 -0.367
AIC 0.021 0.027 1.49 2.12 -0.060 -0.069
.SIGNIFICANT CHANGES IN SIGN OR ABSOLUTE VALUES
To illustrate this computation procedure, consider estimating HC
emission reductions caused by adjusting rco. This example uses a vehicle
rejection rate of 30%, an rco average reduction of 4% and an average
emission response of .0314.
ae.
8e. = 8P. x ap~ x RJ' = (4) x (0.0314) x (.30)
1 J J
8ei = 0.037 grams/mi
A similar and consistent approach was used for computing the remaining
emission reductions shown in Table 2-2. rt should be noted that emission
values with a minus sign signify an increase in emissions. These sign
changes will result in smaller percentage emissions reductions for a given
procedure than were obtained in Phase One.
7
-------
The results of this analysis utilizing mass units show that the
effectiveness of idle adjustments is different from that found in the
previous concentration units study. For example, the results presented
in Table 2-2 show that a timing adjustment reduces the overall effective-
ness of an idle CO adjustment '(i.e., a timing adjustment produces an in-
crease in CO emissions whereas an idle CO adjustment produces a decrease
in CO emissions). The greatest differen~e in predicted reductions
between mass and concentration emissions is attributable to the effect
of an idle CO adjustment. The idle CO adjustment on a mass basis has
negligible effect on HC and NO while on a concentration basis it produced
appreciable emission increases. The net effects are larger mass emis-
sion reductions for HC and NO and somewhat smaller CO emission reductions.
2.1.1
Comparison of Results for Mass/Concentration Data
Table 2-3 presents inspection/maintenance program results based on
comparative mass and concentration data. Shown are the program figures
of merit, average cost per vehicle and resultant emission specie reduc-
tions for the various inspection/maintenance strategies. the figure of
merit is defined as the total annual program cost (discounted capital,
direct and indirect operational and user inconvenience costs) divided by
the weighted four-year average emissions reductions in tons per year
(0.6 ~HC + 0.1 ~CO + 0.3 ~NO). Average vehicle costs include direct
inspection costs for all vehicles (amortized capital and labor) and direct
maintenance costs on rejected vehicles (parts and labor) per year. Emis-
sion reductions are the average values over a four-year program. This
information was derived from computer simulations employing identical
pass/fail criteria for both the mass and concentration emission based
models. This table shows that the figure of merit as computed directly
from mass units is substantially better than the corresponding concentra-
tion result, although the percent emission reductions are generally lower.
This is due primarily to the larger absolute emission reductions achieved.
Table 2-4 shows a comparison of the expected emission reductions for several
candidate strategies. With the exception of the CO reduction for an
Extensive B program, all of the mass based emission reductions on an
absolute basis are larger than their concentration emission counterparts.
8
-------
TABLE 2-3
Comparative Inspection/Maintenance Program Results
fiGURE-Of-MERIT COST PER AVERAGE EMISSION REDUCTIONS
S/fON +++ VEHICLE HC CO NO
Strategy S PERCENT
MASS CONC MASS CONC MASS CONC MASS CONC MASS CONC
ENGINE PARAMETER DIAGNOSTICS
1. IDLE (STATE LANE) 190 320 1.50 1.50 3 0 6 15 0 -7
2. IDLE (FRANCHISED) 260 370 2.50 2.50 3 3 5 13 3 -3
3. EXTENSIVE A + (fRANCHISED) 410/310 460 5.7Q16.50 6.00 9/17 18 5/5 14 :V3 0
4. EXTENSIVE 8++ (fRANCHISED) 490 5040 13.50 12.00 19 22 9 33 0 -5
EMISSION SI GNATURE ANALYSIS
6. IDLE (STATE LANE) 290 430 1.90 2.50 0 2 5 12 2 ~
7. EXTENSIVE A+ (STATE LANE) -/240 360 2.50 4.00 7 ." 5 16 2 -4
8. EXTENSIVE 8++ (STATE LANE) 400 410 6.00 6.00 9 15 7 20 0 -3
(4 YEAR PROGRAM)
+ IDLE PLUS IGNITION TUNEUP, WITHOUT/WITH DYNAMOMETER
++IDLE PLUS IGNITION PLUS INDUCTION TUNE UP
+++ L.A. 8ASIN WEIGHTING fUNCTION
TABLE 2-4
Comparison of Absolute Emission Reductions
Between Mass/Concentration Data
Strategy Mass Concentration
-
HC CO NO HC CO NO
tons/day tons/day
1) Idle 15 282 11 7 277 -6
(Franchised)
2) Extensive B 99 545 0 56 696 -8
(Franchised)
9
-------
These improvements in emission reductions using the mass data are largely
attributable to the differences in influence coefficients previously
discussed.
As anticipated, smaller percentage reductions of hydrocarbons and
CO result for almost all strategies using the mass data. This is caused
in large part by the substantially higher emission levels which are
measured using the 1972 procedure which results in lower percent emissions
reductions. The results for the emission inspection approach, although
more difficult to explain in terms of a change in the model, appear con-
sistent with engine parameter inspection results. The costs for both
basic approaches are similar to those developed from the concentration
data. The double entry for the Extensive A program represents an ignition
tune-up without and with a dynamometer. The use of a dynamometer permits
the diagnosis of ignition system misfire at load and identifies approxi-
mately twice as many ignition failures.
In summary, the following conclusions can be drawn from this compara-
tive analysis:
I No change in the ordinal ranking of inspection/maintenance
procedures occurs upon converting from concentration to closed
seven-mode cycle mass emissions test data.
I Higher absolute HC and NO emission reductions result when
using mass data.
I Slightly lower absolute CO emission reductions result when
mass emissions data are used.
I Similar program costs are obtained for both data sets.
2.1.2 Reoptimization of Procedures
An extension of the previous analysis was to reoptimize the inspection/
maintenance procedures based upon mass emissions data. A five-year time
period was studied to determine if predicted emission levels stabilized at
the end of that period. Projections of the performance of mandatory vehicle
inspection/maintenance programs were limited to no more than five years
because of:
I Uncertainty in estimating the effects of major engine mal-
functions which may occur beyond that period.
10
-------
Lack of experimental emission time history data beyond 50,000
miles (i.e., approximately five years).
Table 2-5 presents the results of the reoptimization for several
strategies. The main information to be obtained from this table is that
the performance criteria (i.e., figure of merit) for the reoptimized pro-
cedures is consistent with the results reported in Section 2.1.1. The
generally lower figure of merit is caused by the larger emission reduc-
tions which result from both the proper evaluation time period and higher
vehicle rejection rates. The higher vehicle rejection rates have resulted
in slightly higher program costs (as shown by the increase in average cost
per vehicle). The ordinal ranking of the procedures has remained unchanged
through the reoptimization. It appears, as discussed later, that higher
rejection rates do not substantially alter the figure of merit; Thus, for
a comparable figure of merit one can achieve higher emission reductions
by specifying a more severe pass/fail criteria. One should also note that
a five-year program does not always result in establishing an emission
equilibrium level. Emissions for the more extensive mainte~ance procedures
.
TABLE 2-5
Optimized Inspection/Maintenance Program
Performance Based Upon Mass Emissions Data
FIGURE-OF-MERIT COST PER EMISSION REDUCT IONS I PERCENT
INSPECTION/MAINTENANCE $/TON +++ VEHICLE HC CO NO
PROCEDURES $
ENGINE PARAMETER' DIAGNOSTICS
1. IDLE (STATE LANE) 155 1.60 3 7 0
2. IDLE (FRANCHISED) 200 3.00 4 6 6
3. EXTENSIVE A + (FRANCHISED) 300/230 6.00/7.00 10/21 6/6 6/6
4. EXTENSIVE B++ (FRANCHISED) 350 13.50 23 12 3
EMISSION SIGNATURE ANALYSIS
5. IDLE (STATE LANE) 225 1.90 0 6 2
6. EXTENSIVE A + (STATE LANE) -/190 2.50 9 6 2
7. EXTENSIVE B++ (STATE LANE) 320 5.75 9 8 0
(FIVE YEAR PROGRAM)
+IDLE PLUS IGNITION TUNEUP, WITHOUT/WITH DYNAMOMETER
++IDLE PLUS IGNITION PLUS INDUCTION TUNEUP WITH DYNAMOMETER
+++ L.A. BASIN WEIGHTING FUNCTION
11
-------
occasionally continue to decrease slowly. It should be noted that the
models' predictive power may be severely taxed as extrapolations further
into the future are attempted. Some of the fundamental assumptions such
as constant owner behavior patterns with respect to maintenance may be
altered by the existence of a program.
The emission time-history profiles shown in Figures 2-1 and 2-2 are
for several levels of imposed maintenance. The so-called baseline fleet
time histories were developed based upon estimating the extent and fre-
quency of voluntary maintenance as well as the rate of maintenance
deterioration. Figure 2-1 shows the results of an idle adjustment pro-
gram which involved direct engine parameter inspection in a franchised
garage and optimal inspection pass/fail criteria. Panels A, Band C show
time history plots generated by the computer model of hydrocarbQns, carbon
monoxide and oxides of nitrogen emissions, respectively. This idle adjust-
ment program has its largest impact on the percentage reduction of CO and
NOx emissions and its smallest impact on HC. The emission levels for the
baseline test have stabilized after the fifth year. The resultant emission
reductions relative to the baseline fleet at the end of the fifth year are
substantially larger than for the first several years of the program.
A similar trend in the time effects on emission reductions can be
seen by examining Figure 2-2. The more extensive inspection/maintenance
program shown in this figure (i.e., inspection and maintenance of idle
plus ignition plus induction variables) yields greater HC and CO emission
reductions than were achieved by the idle parameter inspection/maintenance
program. The CO emission levels appear stable after five years. The
larger HC emission reductions are attributable to correcting misfire be-
cause of maintenance to the ignition system. The model clearly over-
estimates HC emission reductions for the more extensive maintenance pro-
cedures since the predicted absolute emission levels at the end of five
years for vehicles treated in this mandatory inspection/maintenance pro-
gram are lower than for these vehicles when they are new (6.5 gm/mi). This
is a combined data and model deficiency which results from assuming a
linear deterioration rate and selecting what is probably a high value for
this rate. Clearly, the on-going fleet deterioration experimental data will
be required to correct this deficiency. Percentage reduction of hydrocarbon
12
-------
PANE l A
10
~
w
>
W
...J
z ~ 6
0,,-,
V;<-,
~ 1 4
~
w
U
J:
8
~_4 "- --4 - - -4 t- - -4 ~ -
--4
2
o
PANEL B
125
~ 100
w
>
~ 75
5~
V; <-'I 50
~
~
w
o
u
II)
...J
W
>
~
Z~
Q <-' 4
~ 1
~
w
o
z
25
co
-- ~ a- - -4 ~ - --
......
o
8 PANEL C
NO
6
- ~--""'4
-- -~ ... -
-<4 ~--....
BASELINE FLEET
-- - TEST FLEET
2
00
3
4
2
TIME, YEARS
FIGURE 2-1. Mass Emission Time Histories for an
Optimum Idle Inspection and Repair
Program
13
HC
} 0.5 g/m
\
f 10 g,/m
}O.6g,/m
5
-------
~
w
>
W
...J
Z 6
O~
""c}
!Q I 4
~
w
U
:L
PANEL A
HC
3 g/m
PANE L B CO
125
"" 100 t'20 w'm
...J
w
> - -- - - ~ - - - ~..- - - - .~ - - -- -.
W
...J 75
Z
Q~
~? 50
~
w
0 25
u
0
PANEL C NO
~
w
>
W
...J
Z
Q~ 4
~C}
~I
w
o
Z
10
~ ..-- - - -4 "'--
-- ..-
--. ""----4 1"----4
8
2
o
8
- ----
----
---..._- iio___-
BASE LI NE FLEET
--- - - TEST FLEET
6
2
o
o
2 3
TIME, YEARS
4
FIGURE 2-2.
Mass Emission Time Histories for an
Optimum, Extensive B Engine Parameter
Inspection and Repair Program
14
. 1 g/m
5
-------
actually exceeds that of carbon monoxide. For these extensive inspection/
maintenance procedures, the differences between baseline and test fleet
NO emissions is insignificant. This results because the positive benefit
of retarding timing is largely off-set by the effect of leaner air-to-fue1
mixtures which result from repair of the induction system.
The percentage emission reductions presented thus far were based on
an average reduction over the five-year time period of the program. One
would obtain larger emission reduction percentages if computations were
made at the end of the program period. Table 2-6 shows a comparison between
the emission reductions expected at the end of the fifth year with those
averaged over the life of the program. In general, the emission reductions
computed at the end of five years are nearly twice as large as those
averaged over the program.
A last step in this reoptimization of the inspection/maintenance
procedures is to compare the optimal pass/fail criteria. Table 2-7 presents
this comparison for both engine parameter and mode emission pass/fail
criteria. Reoptimization of the system has resulted in choosing more
restrictive rejection criteria for both parameters and emissions. For
example, in the case of ICO, a pass/fail criteria of 4% was optimal
using the concentration data, whereas a value of 3% is now estimated.
This lower value will result in larger vehicle rejection fractions which,
in turn, will yield higher emission reductions.
2.1.3 Relaxation of Optimal Procedures
A figure of merit based on the ratio of total program costs to
weighted emission reductions is utilized in the current economic-effective-
ness model for screening procedures. This ratio provides a simple and con-
venient method for ordina11y ranking the various inspection/maintenance strat-
egies. The inspection/maintenance procedure ranked highest based upon the
figure of merit however generally involved a simple inspection procedure and
limited maintenance and does not produce the largest reduction of exhaust
15
-------
TABLE 2-6 Expected Emission Reductions at End of
Five Years Optimal Inspection Criteria
Emission Reductions
Average Over
End of 5 Years. % 5 Years. %
Strategy
HC CO NO HC CO NO
Engine Parameter Diagnostics
1. Idle (state lane) 4 12 0 3 7 0
2. Idle (franchised) 6 11 11 4 6 6
3. Extensive A+ (franchised) 40* 11 11 21* 6 6
4. Extensive B++ (franchised) 43* 21 3 23* 12 3
Emission Signature Analysis
5. Idle (state lane) 0 11 4 0 6 2
6. Extensive A+ (state lane) 12 12 2 9 6 2
7. Extensive B+t (state lane) 13 15 0 9 8 0
+ Idle plus ignition. with dynamometer
++ Idle plus ignition plus induction tuneup (with dynamometer)
* Over predicted due to uncertainties in misfire deterioration
and rejection rate fraction.
rate data
ENGINE SUBSYSTEM/PARAMETERS OPTIMAL PARAMETER VALUE OPTIMAL MODE VALUE
MASS CONe MASS CONC
. IDLE
. IDLE CO 3.0% 4.0% ICO=3'*' ICO = 4%
. RPM 520 RPM 52S RPM IHC:300 IHC = 300
. TIMING 4.4 DEG 10 DEG PPM PPM
.
. IGNITION
. MISFIRE 2.5 % 2.5% AHC=300 AHC = 400
. NOx -- -- PPM -- PPM
--
. INDUCTION
. AIR PUMP PLUGGED PLUGGED ++
+
. PCV 0 P.SI PLUGGED CCO:I% ACO = 1 %
. AIR CLEANER 105 DEG IOS DEG
. CHO K E HEAT RISER -- -- -- --
. CHOKE BLADE SETTING -- -- -- --
L.A. BASIN EMISSION WEIGHTING FUNCTION
+ MODE 6 (15-30 MPH ACCELERATION)
++ CLAYTON CRUISE
16
-------
emissions. The question then is asked whether equivalent exhaust emission
reductions can be achieved with the simpler inspection/maintenance program
by using less optimum rejection criteria while achieving the same cost
effectiveness as the more comprehensive inspection/maintenance procedure.
To make this comparison, the highest ranking inspection/maintenance
strategy is varied from optimum until its figure of merit is equal to the
figure of merit of the next most attractive strategy. This operation was
accomplished by changing the optimized pass/fail criteria developed for
the highest ranking procedure to produce larger emission reductions and
increased cost. In turn, the next most attractive strategy can be relaxed
to the third and so on until all strategies have been intercompared.
To illustrate this technique, consider the information presented in
Table 2-8. Performance measures (e.g., figures of merit) are given for
TABLE 2-8
Relaxation of Optimal Procedures
fiGURE-Of-MERIT COST PER EMISSION REDUCTIONS
VEHICLE HC CO NO
$/TON $ PERCENT
ENGINE PARAMETER DIAGNOSTICS
1. IDLE (fRANCHISED) 200 (230) 3.00(3.75) 4 ( 5) 6 ( 6 ) 6 ( 7 )
2. EXTENSIVE A+ (fRANCHISED) 230 7.00 21 6 6
EMISSION SIGNATURE ANALYSIS
3. EXTENSIVE A* (STATE LANE) 190 (225) 2.50 (3.00) 9 (10) 6 (8) 2 (3)
4. IDLE (STATE LANE) 225 1.90 0 6 2
(fiVE YEAR PROGRAM)
+IDLE PLUS IGNITION TUNEUP WITH DYNAMOMETER
17
-------
the two most attractive procedures for each basic inspection strategy. In
the case of the parameter inspection strategy. the idle program is varied
until its figure of merit is equal that of the Extensive A maintenance
program. The optimal figure of merit for the idle program is $200/ton
while the optimal figure of merit for the Extensive A program is $230/ton.
By adjusting the pass/fail criteria, a new figure of merit of $230/ton was
obtained for the idle program. The associated costs and emission reduc-
tions for both programs can now be directly compared with the figure of
merit held constant. The numbers in parentheses represent system per-
formance data for the modified program. The result of relaxing the idle
program has been to increase both program costs and emission reductions.
The idle program however still does not compare favorably with the Exten-
sive A program in terms of hydrocarbon emission reductions. Selection of
either program for implementation in a given region would depend primarily
on the type of air pollution problem encountered in that region.
A similar comparison was made for the emission inspe~tion strategy-.
In this case, the Extensive A program is more cost effective than the idle
program. Here again, the most attractive program was relaxed until its
figure of merit equalled the next strategy. For the Extensive A program,
its figure of merit was increased from $190/ton to $225/ton. This increase
resulted in a slightly higher emission reductions and program costs. In
this case the Extensive A program is always most attractive.
2.1.4 Statistical Analysis of Results
Predictions of system performance for a program of vehicle inspection/
maintenance are uncertain because of the variability in the data sets used
in the estimate. The ordinal ranking of the procedures using computed
figures of merit may change when the projected emission reductions are re-
duced to reflect the fraction of this reduction which can be achieved at
a stated statistical confidence level. As a consequence, a preliminary
statistical analysis was performed of the results presented in Section
2.1.2.
18
-------
The approach taken in this statistical study required the use of the
emission distributions data developed from the on-going fleet deterioration
experiment. Figure 2-3 shows a comparison of the pre and post tune emis-
sion distributions derived from the deterioration experiment for HC, CO and
NO, respectively. The mean values of these distributions have been super-
imposed on the mean emission levels predicted from the computer model using
the pre and post maintenance data for the oase1ine and test fleets, respect-
ively. Thus, both the mean emission levels and their variances are based on
approximately equivalent states of vehicle maintenance.
A student's statistical test, as described in Appendix A was conducted
using pre and post maintenance emission distri.butions for each specie in
order to estimate the emissions reductions at a specified confidence level.
Table 2- 9 presents a comparison of test and baseline fleet statistical
data, including the expected value of the pre and post tuned emission
means, along with scores and confidence limits on predicted emission re-
ductions at the time of maintenance. As can be seen, the null hypothesis
that there is no difference in mean emission levels between the test and
baseline fleets can be rejected with greater than 90% confidence for both
HC and CO (viz, t (computed) > t (critical at 90% confidence level)). The
null hypothesis cannot be rejected for the case of NO. It should be
pointed out, however, that the candidate inspection/maintenance procedures
were not designed to reduce NO.
It therefore is concluded that the test and baseline emission levels
came from different populations, and consequently that the predicted emis-
sion reductions are not due entirely to chance.
Table 2-9 shows the 90% confidence bands about the estimated mean
emission reductions which were achieved with the Extensive B maintenance.
A comparison of these results with those in column three shows that approxi-
mately 65% of the computer model predicted emission reductions for HC and
CO can be claimed at the lower 90% confidence limit. One therefore would
have a 95% level of confidence that emission reductions will exceed the
value predicted at this lower limit.
19
-------
FIGURE 2-3
PRE AND POST MAINTENANCE EMISSIONS DESTINATIONS
Based on the 1972 Federal Test Procedure
Fleet Deterioration Program
.18 .018 .18
.16 .016 .16
- PRE (BASEt -PRE (BASE)
.14 - - - POST (TEST) - - - POST (TEST)
\ .014 .14
\
.12 .012 .12
.10 .010 .10
N
a
.08 .008 .08
.06 .006 .06
.04 .004
2
4
6
10
8
HC
GRAMS/MILE
12
16
14
20
40
60
80 100
CO
GRAMS/MILE
140
160
120
180
o
o
4
2
6
NO
GRAMS/MILE
8
10
12
-------
TABLE 2- 9
Influence of Several Confidence Statements on
Estimated Maintenance Effectiveness-Diagnostic, Extensive B
Emissions, gms/mi
Inspection
N
......
90%
Maintenance Expected Confidence + "t" Statistic**
Speci e Pre Post Reduction* Band on Reduction Computed Critical @ 90%+
HC 7.1 5.4 1.7 + 0.65 3.3 1.28
-
CO 87.7 73.5 14.2 + 5.1 2.9 1.28
-
NO 6.2 6.2 0 + 0.37 0 1.28
-
*Expected reduction estimated from economic-effectiveness model
+Based on pooled pre and post maintenance variances from the 150 vehicle, fleet
deterioration experiment emissions data and a 90% confidence level
**When t (computed) > t (critical at 90%) there is greater than a 90% confidence that
the expected emission reductions is greater than zero.
-------
The values of inspection/maintenance procedure performance predicted
by the economic effectiveness model and presented in Table 2-5 can now be
adjusted to reflect their minimum anticipated performance. Table 2-10
shows inspection/maintenance procedure performance values which have been
discounted to reflect the smaller emission reductions which can be
achieved with a 95% level of confidence. Ninety-five out of one hundred
times program results would be expected to be at least as good as indicated
in this figure. It is significant to note that the ordinal ranking of
procedures as reflected by the figures of merit has generally changed to
favor those procedures employing the direct diagnosis of engine parameter
malfunctions. The sole exception to this statement is a key mode emission
inspection coupled with adjustment of idle parameters and ignition subsystem
repair. The primary reason for these changes in ranking is the larger emis-
sion reductions achieved with these procedures relative to the pooled esti-
mate of variances in the emissions data. This is dramatically illustrated
by comparing the five-year average emission reductions presented in the
last three columns of Table 2-10 with those of Table 2-5.
Several known and significant deficiencies exist in the economic-
effectiveness model at this point in its development. Specifically, main-
tenance is assumed to be reliably performed and the engine parameter
deterioration rates are inferred from incomplete data sets. Therefore,
those states or cities contemplating the implementation of mandatory
vehicle inspection/maintenance programs are urged to use the results
reported herein as a guide in selecting the procedures which best fit
their needs. A carefully designed pilot program employing large vehicle
(300 to 600) and service organization (30 to 60) samples should be con-
ducted to verify that the results estimated in this report are achievable
within their existing framework of service organization capability and
vehicle owner maintenance habits. The model may then be validated against
these data and used in its predictive mode to determine expected future
emission reductions.
22
-------
TABLE 2-10
Optimized Inspection/Maintenance Performance Based on Mass
Emission Reductions Stated at the Lower 95% Confidence Level
N
W
Inspection/Maintenance Figure of Merit Cost per Emission Reductions, %
Procedures $/ton+++ Vehicle HC CO NO
Engine Parameter Diagnostics
1. Idle (state lane) 393 1. 60 0.2 4.9 -2.0
2. Idle (franchised) 501 3.00 0.2 2.8 4.0
3. Extensive A+ (franchised) 325 7.00 17 3.3 3.2
4. Extensive B++ (franchised) 485 3.50 18 8.4 0.8
Emission Signature Analysis ~
5. Idle (state lane) 1062 1. 90 -3.0 3.5 0.7
6. Extensive A+ (state lane) 293 2.50 6.1 4.0 0.9
7. Extensive B++ (state lane) 517 5.75 5.5 5.4 -2.4
+Id1e plus ignition tuneup, without/with dynamometer
++Idle plus ignition plus induction tuneup
+++L.A. basin weighting function
-------
2.2 System Sensitivity Studies
The purpose of the system sensitivity studies was to establish the
effects of the following variables on procedure effectiveness and selection:
. Engine adjustment deterioration rates
. Engine parameter and mode emission pass/fail
. Frequency of vehicle inspection
. Levels of voluntary maintenance
. Figure of merit emissions weighting factors
. User inconvenience costs
. Labor rates for franchised garage and state lane inspection
. State lane facility configuration
criteria
An important aspect of a sensitivity analysis involves defining a so-
called nominal or standard case. The impact of various changes in system
input on program performance is then measured relative to this reference.
The nominal case selected as a reference for this analysis is an idle para-
meter inspection/maintenance program for the Los Angeles basin. In some
cases assessment was also made of the influence of system variables on
the performance of a program which involved more extensive inspection and
maintenance.
The following sections present a discussion of the results of the
sensitivity analysis. Section 2.2.1 evaluates the impact of inspection
criteria, the nature of voluntary maintenance and parameter deterioration
rates on procedure effectiveness. Section 2.2.2 presents a discussion of
the proposed state lane inspection system and an assessment of the system
configuration tradeoffs. Section 2.2.3 discusses the impact of several
key system operational variables such as the frequency of vehicle inspec-
tion. Finally in Section 2.2.4, an analysis is made of the impact of
variations in the weighting factors assigned to the reduction of emission
species in the system figure of merit upon the selection of optimum
inspection/~aintenance procedures.
24
-------
£:) 60
w
I-
U
~ 40
w
a<:
I-
~ 20
u
a<:
w
Q..
N
U'I
PANEL A
leo =3%
--
.
.
.
.
o
o
1 2 3 4
PROGRAM PERIOD-YEARS
MISFIRE =2.5%
o
o
1 2 3 4
PROGRAM PERIOD-YEARS
FIGURE 2-4
Chanaes in
£:) 60
w
I-
U
W
;J 40
a<:
I-
Z
~ 20
a<:
w
Q..
5
PANEL B
RPM =520
o
o
1 2 3 4 5
PROGRAM PERIOD-YEARS
£:) 6 PANEL E
w
I-
U
w
..,
w
a<:
I- 3
Z
w
U
a<:
w
Q..
PANEl G
£:) 30
w
I-
U
~ 20
w
a<:
I-
~ 10
u
a<:
w
Q..
5
AlP
= res tri cted
o
o
£:) 60
w
I-
U
~ 40
w
a<:
I-
Z
w 20
u
a<:
w
Q..
o
~ 30
I-
U
w
;J 20
a<:
I-
Z
~ 10
a<:
w
Q..
5
o
o
6 PANEL D
£:)
w
I-
U
w
..,
w
a<: 3
I-
Z
w
~
W
Q..
1 2 3 4
PROGRAM PERIOD - YEARS
Ale =105 DEG
".
.
.
.
1 2 3 4
PROGRAM PERIOD - YEARS
parameter Rejection Fractions With Time
o
o
5
PANEL C
TIMING
=4.4 DEG
1 2 3 .. 5
PROGRAM PERIOD-YEARS
PANEL F
PCV =0 in H20
.-
.
.
.
.
1 23.. 5
PROGRAM PERIOD - YEARS
-------
FIGURE 2-5
Rejection Fractions Resulting from an Emission
Signature Inspection Followed by an Extensive B
Maintenance
PANEL A PANEL B
60 Q 60
Q
w ICO =3% w IHC =300 ppm
I- I-
u (CUT POI NT) u (CUTPOINT)
w 40 w 40
..., ...,
w . . w
~ . ~
I- J.. I-
Z 20 .,... Z 20
w w
U ~
~
w W
Q.. Q..
00 1 2 3 4 5 00 1 2 3 4 5
PROGRAM PERIOD"" YEARS PROGRAM PERIOD"" YEARS
N
m
PANEL C
Q 60
w
I-
U
~ 40
w
~
I-
~ 20
u
~
W
Q..
Q
w
I-
U
w
...,
w
~
I-
Z
w
u
~
W
Q..
60
AHC =300 ppm
(CUTPOJNT)
40
20
o
1 2 3 4 5
PROGRAM PERIOD"" YEARS
PANEL 0
.
CCO =1%
(CUTPOJNT)
. .
.
.
00
1 2 3 4 5
PROGRAM PERIOD"" YEARS
-------
2.2.1
Inspection/Maintenance Procedure Sensitivity
The sensitivity of system performance to the basic input data and
assumptions which characterize the inspection/maintenance procedures is
presented herein. The extent of voluntary maintenance and criteria for
inspection appear to be the most significant variables in terms of their
impact on procedure effectiveness.
Fraction of Vehicles Rejected
Figure 2-4 shows the percentage of vehicles rejected based upon
various engine parameter maladjustments and malfunctions for an Extensive
B maintenance program over a five year interval. This figure is for an
optimal pass/fail criteria using a direct engine parameter inspection pro-
cedure. At the first inspection interval, the state of maintenance of
vehicles varies widely resulting in high initial vehicle r~jection frac-
tions with values ranging between 20-50% of the vehicle population. At
subsequent inspection intervals engine adjustments in the population are
under better control and the vehicle rejection fraction decreases. Very
little decrease in rejection fraction with time is found for idle fue1-
to-air ratio while a substantial decrease is found for timing. This sensi-
tivity of the rejection fraction to individual engine adjustments relates
directly to their deterioration rates and to the degree to which they are
out-of-specification at program initiation. For example, a large percent-
age of vehicles have timing maladjustments most of which are corrected
during the first inspection/maintenance cycle. Since timing adjustments
deteriorate slowly. subsequent inspections with a fixed pass/fail level
find fewer and fewer malfunctions of this type. Over five years, the
rejection fraction for timing was found to drop from 50 to 10%. For the
more highly leveraged, but infrequent malfunctions such as misfire, air
pump failure and air cleaner or pev component restrictions, a more nearly
constant rejection fraction occurs over the five-year program period.
Figure 2-5 presents the fraction of vehicles rejected by an emission
inspection using both idle and loaded mode measurements of eo and He. As
can be seen, the rejected fractions based upon eo measurements are
27
-------
relatively insensitive to the program duration. This is because the
emission inspection approach is not a reliable procedure for diagnosing
the induction system malfunctions under study and a large number of errors
of omission occur. Usually, CO emissions measured under load identify
only 30 to 40% of the actual malfunctions of interest in the induction
system. Since a relatively small percent of these component malfunctions
are identified and corrected with these inspection procedures, the re-
jection fraction remains high. As anticipated, rejection fractions for an
inspection of CO emissions at idle are similar to those for a direct engine
parameter inspection.
Inspections using both idle and loaded HC emission levels result in
a vehicle rejection level which decreases sharply with inspection program
duration. This is because the first inspection finds most of the misfires
within the general population with few errors of omission being committed.
The mean level of HC emissions measured during vehicle inspections decreases
substantially because of the large HC response to misfire, thus low levels
of rejection result during subsequent inspection intervals.
Pass/Fail Criteria
The influence of pass/fail criteria for the three idle adjustments
of timing, rpm and idle CO upon the average emission reductions achieved
over five years by an idle engine parameter inspection is shown in Figure
2-6. The effect of varying the pass/fail criteria is determined by vary-
ing one of the three idle inspection pass/fail criteria while holding the
other two at optimal levels. Optimum HC reductions apparently do not
occur until the rpm and idle CO pass/fail criteria are set close to manu-
facturer's specification. The effect of the timing pass/fail criteria
upon HC and CO emissions is very small. The sensitivity of NO emissions
to timing however tends to drive the optimal timing rejection point toward
lower values. The optimal pass/fail criteria for the three idle para-
meters, therefore, are close to the specification value for each.
28
-------
Idle Parameter Inspection and Maintenance Program
PANEL PANEL PANEL
Z A Z B Z C
Q 10 RPM = 525 Q 10 RPM = 525 Q 10 RPM = 525
I- TIMING = 4.4 0BTOC U TIMING = 4.4 0BTOC I- TIMING = 4.4 °BTOC
u u
:::::> spec :::::> :::::>
0 0 o . . . .
w ~
w w ~ 0<:
0<: 5 0<: 5 5
::R. ::R. ::R.
0
0 0
l l l
u a a
J: 0 U 0 Z 0
0 3 0 2 3 4 5 0 2 3 4 5
ICO % ICO% ICO%
z PANEL ) PANEL Z PANEL
D Z E
Q 10 ICO = 3.0 % Q 10 ICO=3.0 % QlO F ICO==3.0 %
I- TIMING = 4.4 °BTOC I- TIMING == 4.4 °BTOC I- TIMING == 4.4 0BTOC
u u u
:::::> :::::>
0 :::::> 0 ---
0 .
w spec w . . . w ....
0<: 5 0<: 5 0<: 5
eft. ...-/ ::R.
~ 0
\0 l l
u a
J: b z 0
450 500 550 600 450 500 550 600 450 500 550 600
RPM RPM RPM
PANEL Z PANEL PANEL
G H Z I
Z a a
Q 10 ICO = 3.0 % U 10 ICO == 3.0 % ;:: 10 ICO=3.0%
I- RPM = 520 :::::> RPM == 520 u RPM == 520
u 0 :::::>
:::::> 0
0 spec w w ~
0<: 0<:
W . . .
0<: 5 ::R. 5 ::R. 5
0 0
eft. l l
l a a
u u Z
J: 0 0 0
2 4 6 8 2 4 6 8 2 4 6 8
TIMING degrees BTOC TIMING degrees BTOC TIMING degrees BTOC
FIGURE 2-6 Effects of Pass/Fail Criteria on Five-Year
Average Emission Reductions
-------
,
,
/
The sensitivity of emission reductions to variation in the pass/fail
criteria near their optimal values is not substantial because a failure
criteria placed close to specification leads to rejecting a large vehicle
population in which the average maladjustment is small. The product of
the average malfunction deviation from specification and failure frequency
therefore remains relatively constant with small variations of pass/fail
criteria.
As shown in Figure 2-7, the figure of merit for idle adjustments is
less sensitive to variations of the pass/fail criteria than were the emis-
sions reductions. The pass/failure criteria which yield optimum values
for the figure of merit however also tend to produce maximum emissions
reductions. Similarly the value of the figure of merit is not sensitive
to pass/fail criteria placed on PCV and air cleaner flow restriction.
Parameter Deterioration Rates and the Extent of Voluntary Maintenance
The influence of uncertainties in the basic input data was also studied.
Figure 2-8 shows the sensitivity of the figure of merit (Panel A) and emis-
sions reductions (Panel B, C and D) to the deterioration rates of idle fuel-
to-air ratio (ICO) and misfire when these rates are varied by ~30% of the
nominal rates shown in Table A-2. The figure of merit is based on Region I
weighti~g factors and an idle adjustment program. The rate of deteriora-
tion of idle fuel-to-air ratio tends to effect procedure performance most
significantly. It should be noted that the ~30% variation of deterioration
rates from nominal was selected arbitrarily for this study. A similar
analysis for the program involving an engine parameter inspection and the
maintenance of idle adjustments, ignition system components causing misfire
and the induction system components showed that the figure of merit and HC
emission reductions were extremely sensitive to the ignition system deteri-
oration rate. Generally, those engine parameters most strongly influencing
emissions are those which were found-to be most sensitive to uncertainties
in deterioration rates.
The effects of various levels of voluntary maintenance on the effect-
iveness of a mandatory inspection/maintenance program were also studied.
Figure 2-9 shows the results of changing the effectiveness of voluntary
maintenance. Varying the effectiveness of voluntary maintenance by
30
-------
Idle Parameter Inspection and Maintenance Program
200 .. ~ . 200 . (9) . 200 .... ~ ...
I- I- I-
~ c.: c.:
w
w w ~
~ ~ I
I I IL.
IL. IL. 9100
9 100 9100
w
w w c.:
c.: c.: ~
~ ~
l') l') l')
i:i: RPM '"'520 i:i: ICO =3% i:i: ICO '"'3%
TIMING = 4.4 0STOC TIMING = 4.4 0BTOC RPM =- 520
0 II I I 0 l I I I 0 II I I I I,
2 3 4 450 500 550 600 3 4 5 6 7
ICO RPM TIMING
Extensive S Inspection and Maintenance Program
w
.....
400
400
-.
(i)
.
I-
c.:
w
~
I
IL.
9 200
w
c.:
~
l')
u:
.
~.. -
t:
c.:
w
~
IL.
0200
w
c.:
~
Q
IL.
o
II
-0.4 -0.2
Alc = 105 0
I I
o 0.2 0.4
PCY inches of water
o
o
PCV=0"H20
8 Optimal parameter settings
based upon the figure of
merit using Region I
weighting factors
40 80 120 160
AIR CLEANER degrees
FIGURE 2-7 Sensitivity of Figure of Merit to Engine Parameter
Pass/Fail Criteria
-------
/
Idle Parameter Inspection and Maintenance Program
PANel A
400
....
"'
~
~ 200
.1.
'"
=>
o
u::
PANEL B
10
z
o
;::
u
=>
Q
...
'" 5
~
I
U
:z:
z
o
;::
u
=>
~
~ 5
I
o
U
It
,
,
,
,
,
,
,II
,
,-'
,
.
..
.
o
lOW
HIGH
NOMINAL
FIGURE 2-8
KEY
lOW: -30%OF NOMINAL RATES
HIGH: 30%OF NOMINAL RATES
-ICO
---- MISFIRE
o
lOW
HIGH
NOMINAL
PANEL C
10
PANEL D
10
r
z
o
;::
u
=>
Q
...
..
.,. 5
I
o
Z
-A
---~.
o
o
lOW
HIGH
lOW
HIGH
NOMINAL
NOMINAL
Influence of Parameter Deterioration
Rates on Procedure Effectiveness
PANel A
400
Idle Parameter Inspection and Maintenance Program
KEY
LOW: -3a'/oOF NOMINAL EXTENT
HIGH: 3a'/oOF NOMINAL EXTENT
....
..
~
~ 200
=>
£!
...
/
o
PANel B PANEL C PANEL D
10 10 10
~ Z ~ Z
0 Q
;:: ;:: ~
U U ....
=> => u
=>
Q ~ 5 a
...
'" 5 ::1 5
"I- .'1- if'.
I . . . I I
U 0 0
:z: U Z
0 0 0
lOW lOW HIGH
NOMINAL NOMINAL
FIGURE 2-9
Influence of Voluntary Maintenance on Procedure Effectiveness
32
-------
~30% of the baseline values resulted in a 150% change in the figure
of merit for the idle adjustment program. Most of this change can be
attributed to the large impact of voluntary maintenance on CO emissions.
Further data to characterize the frequency and extent of voluntary main-
tenance are required in view of the sensitivity of the effectiveness of
a program of mandatory vehicle inspection and maintenance to these factors.
Maintenance Effectiveness
Once specific malfunctions have been identified by vehicle inspection
the next issues are the reliability with which a repair or adjustment can
be effected by a service organization and the effect of imperfect mainten-
ance performance on a mandatory program. Program effectiveness is detri-
mentally affected when repairing agencies either fail to make a required repair
or inadequately make it because of poor mechanic skill or diagnostic
equipment limitations. One of the more poorly maintained adjustments
identified in the engine parameter survey is idle fuel-to-air ratio (idle
CO). This parameter is likely to be set rich by service organizations to
minimize subsequent customer complaints. The effect of failure to set
idle CO to the average specification value of 2.5% is shown in Figure 2-10.
As can be seen, setting idle CO one percent richer than average specifica-
tion can almost completely negate the effectiveness of an idle adjust pro-
gram for controlling CO emissions. It therefore appears important that
franchised service organizations undergo a certification procedure to
assure adequate standards of performance.
2.2.2 System Configuration
The purpose of this section is to demonstrate the capacity of the
Economic-Effectiveness model to determine the optimal design of a state
lane inspection system.
Figure 2-11 is an artist's conception of a state lane inspection
station. This particular one-lane inspection station configuration is
equipped with a dynamometer for making loaded mode emission measurements.
It has been designed to process approximately 50,000 vehicles annually.
The total capital cost for this one-lane station with the attendant
33
-------
400
PANEL A I
spec
I DLE PROGRAM
PARAMETER INSPECTION
...
02
~
I
9200
UJ
'"
::J
~
LL
lean
rich
ICO setti ng %
10 10 10
Z Z z
Q Q Q
... ... ...
u u u
:> :> ::J . . .
o 0 0
UJ 5 UJ 5 UJ 5
'" -. '" '"
rfI. . .- 1fI. :i1
I 0
I I
U 0 0
:I: U Z
3 3
ICO % ICO % ICO %
FIGURE 2-10
Influence of ICO Adjustment on Procedure Effectiveness
computer and measuring equipment is $45,000. A computer readout showing
inspection times, costs and other salient information for the state lane
system is given in Table 2-11. The inspection times used in the analysis
were 1.5 minutes for an idle emission inspection and 2.2 minutes for a
loaded-mode emission inspection.
A key tradeoff in designing a state lane inspection system is the
number of lanes at each inspection station and the number of stations
required to provide the total number of inspection lanes for a given
region. Obviously the more lanes each station has, the fewer the required
number of stations. This decision involves trading off user inconvenience
costs with inspection station capital and labor costs. Labor and capital
34
-------
FIGURE 2-11
Artist Conception of State Lane Inspection Station
TOTAL NUMBER Of STATE OPERATED' 'INSPECTION l#iN-E-S[-~--4'O
NUMBER OF LANES PER SITE
1
TOTAL NUMBER OF SITES
40
VEHICLE INSPECTIO~ TIMES
IDLE
LOADED
1.50 "'IN
2.20 "'IN
~ i
~EQUIPMENT REQUIREMENTS
NDIR
COMPUTER
M I SC .
DYNO
AND COSTS
:2000.00
Isoao.ao
3000.00
7000.00
JOlLARS/LANE
JOLLARS/LA\lE
DOLLARS/LANE
:)OLLARS/LA\JE
INfORMATION PROCESSING COSTS
j
'~
.50 )OLLARS/CA~.
'USER TIME COSTS
2.00 )OLLARS/H~.
TOTAL USER TIME
NUMBER OF STATE EMPLOYEES
13.18 "tIN
2
'I'"
.. SI'A11orf SIZE
600.30 SQ. FT.
TABLE 2-11 Typical Input Data Describing a Single
Lane State Inspection Station
35
-------
costs are directly proportional to the number of inspection lanes while
user inconvenience costs vary inversely with the number of lanes.
Figure 2-12 shows the results of trading off the total number of
lanes and the number of lanes per station for a prescribed state lane
system. Panel A shows the influence of the number of lanes per station
IDLE PROGRAM
!:: 400 rNEL A EMISSION INSPECTION
~. ~
~ 200G
.L
~ 0 TOTAL LANES . 40
o 1 2 3 4
LANES/STATION
PANEL B
400
I-
..
~
~ MINIMUM
.L 200 NUMBER REQUIRE D
..
:>
C>
ii:
.
t .
- 1 LANE/STATION
@ 2 LANES/STATION
o
o
10 20 30 40 50 60
TOTAL NUMBER OF LANES
FIGURE 2-12
Influence of System Configuration on Procedure Effectiveness
on the figure of merit for a system with 40 total lanes. As can be
seen, the lowest figure of merit for the stated procedure occurs at
one lane per station. This is because user inconvenience costs,
when valued at $2jhr., tend to dominate system capital costs.
Labor costs, moreover, do not directly impact this tradeoff
since they are primarily related to the total number of vehicles
inspected. Panel B shows the results of a similar tradeoff exet'cise,
36
-------
except that the total number of lanes has been varied, while holding the
number of lanes per station constant. Here the total inspection costs
nearly balance user inconvenience costs and result in an invariant figure
of merit. A minimum of 35 complete lanes are required to process the con-
trol vehicles (1966-1970) for Region I. When the number of lanes per
station is increased to two, for a 50-lane system, the associated figure
of merit is comparable with the results shown in Panel A.
2.2.3 Operational Variables
Frequency of vehicle inspection is one of the most significant opera-
tional variables in a program of vehicle inspection/maintenance. Figure
2-13 shows the results of a parametric analysis of various inspection
periods. Panel A reveals that a yearly inspection is most cost effective
800 PANEL A
!:: 600
'"
...
~
I
~ 400
.L
'"
::;)
~ 200
PANEL B
10
z
o
>=
u
::;)
S 5
'"
'/I.
I
U
%:
z
o
6
::;)
s s
'"
00
'/I.
I
o
U
INSPECTION PERIOD- MONTHS
,,0
,
~ ,P"
o I
\;:d .'
V/
00 6 12 18
INSPECTION PERIOD- MONTHS
PANEL C
10
o
o
KEY
IDLE PROGRAM
. - - - - - EXTENSIVE B
--- DISCONTINUITY
!O
PANEL D
6
INSPECTION PERIOD - MONTHS
z
o
;::
u
::;)
S 5
'"
'/I.
I
o
Z
~
o
o
INSPECTION PERIOD - MONTHS
FIGURE 2-13
Influence of Inspection Period on Procedure Effectiveness
37
-------
for an idle parameter maintenance program. For a more extensive main-
tenance program, the optimal inspection period appears to be ten months.
Panels B through D of Figure 2-13 show predicted emission reductions
for the idle parameter maintenance program. These emission reductions
represent the average values achieved over a five-year period of program
operation. As expected, the more frequent inspection periods, e.g., six
months, yield larger reductions for all three emission species. In the
cases of HC and CO, Panel Band C, the emission reductions tend towards
zero as the frequency of inspection approaches 18 months. This is because
the voluntary maintenance program undergone by the baseline fleet becomes
equivalent in terms of achieving emission reductions to the 18-month idle
parameter inspection/maintenance procedure. The relatively high rate of
idle CO and rpm deterioration with time result in a substantial HC and CO
sensitivity to inspection frequency. The trend towards zero emission re-
ductions at 18 months is not observed for NO. The voluntary maintenance
program does not strongly affect NO emissions (Table A-2) and consequently
reductions produced by the inspection/maintenance procedures will appear
as an improvement. The low rate of change in NO emission reductions with
respect to frequency of inspection can be attributable to the relative low
timing deterioration rate.
Because labor costs dominate capital costs for a state lane or fran-
chised garage inspection system, it becomes important to assess the impact
of various labor rate structures on the program figure of merit. Panels
A, Band C of Figure 2-14 show the influences of franchised garage labor
rates, user time costs and state lane wage rates on program effectiveness.
The nominal values used in this analysis are denoted by a circled dot.
For the idle parameter maintenance program, the figure of merit appears
most sensitive to user time costs and least sensitive to state lane wage
rates. As a measure of this sensitivity, the derivatives for the figure
of merit, with respect to each of these variables, (i.e., ~F/~$), has been
computed for the range studied. Results of these calculations show that
the franchised garage labor rates (Panel A) have twice as much impact per
unit change as state lane labor costs (Panel C). Costs associated with
user inconvenience have even a larger impact. This is because user costs
are computed based upon the total vehicle population while, for example
38
-------
FIGURE 2-14
Influence of Labor Costs and User Inconvenience
Costs on Procedure Effectiveness
IDLE PROGRAM
EMISSION INSPECTION
PANE L A
400
PANEL B
400 EXTENS\YE
t: i).
01: .
W
~ ~~ = 21 \DLE
I .
LL. @
o
I 200 .
W
01:
::>
~
LL.
0
0 1 2 3
USER TIME COSTS"" $/HOUR
I-
-
01:
w
~
I
LL.
o 200
I
w
01:
::>
~
LL.
w
~
"* == 16
@ NOMINAL VALUE
o
~ I I
B 10 12
LABOR TIME CHARGE", $/HOUR
PANEL C
400
200
.
~ .-
'=-" £:. F
"Z$= 8
o
o 2 4 6
INSPECTION LABOR RATES"'" $/HOUR
-------
garage repair costs are incurred on rejected vehicles only. To evaluate
the influence of user inconvenience costs more fully, a sensitivity study
was also performed for an extensive maintenance program and is shown along
with the results of an idle parameter maintenance program on Panel B. Al-
though the figure of merit has increased, its rate of change with user in-
convenience cost is about the same as that of the idle program. Again this
is attributed to the dominance of operating costs in a vehicle inspection/
maintenance program.
2.2.4 Weighting Factors for Emission Reduction
One of the more uncertain assumptions used in the analysis is involved
in specifying values for the emission weighting factors which are used for
combining the computed reductions of each emission specie into a program
figure of merit. Weighting factors which have often been used for the Los
Angeles basin involve weighting HC by six-tenths, CO by one-tenth, and NO
by three-tenths. This weighting of the relative importance of emission species
reflects the particular air pollution problem in this region. Regions with a
different type of air pollution problem would be expected to assign a different
weighting to specie emission reductions.
To improve the understanding of the influence of these weighting
factors on procedure effectiveness, a parametric study was performed.
Presented in Table 2-12 are the results of the study for three different
weighting functions. Case 1 employs the values for the Los Angeles basin
TABLE 2-12
Influence of Emission Weighting Function on
Procedure Effectiveness (WHC' WCO' WNO)
Engine Parameter
Inspection
Figure of Merit
l.
2.
3.
Idle (franchised)
Extensive A+ (franchised)
Extensive B++ (franchised)
Case 1
(0.6,0.1,0.3)
200
230
350
Case 2 Case 3
(0.9,0,0.1) (0.1,0.9,0)
570 35
250 75
470 80
+ Idle plus ignition with dynamometer
++ Idle plus ignition plus induction tune-up (with dynamometer)
40
-------
which .have been studied. The emission factors for Case 2 place a heavy
emphasis on reducing HC, while Case 3 stresses CO reduction. Results for
Case 3 show a substantial reduction in the figure of merit but no change
in the ordinal ranking of the procedures. This lower figure of merit is
the result of the large weighting factor and emission reduction, in tons,
for CO. For Case 2 a change in the ordinal ranking of the procedures has
resulted. The Extensive A program is now substantially more attractive
than the idle program. This can be attributed to the high HC weighting
factor and the fact that the Extensive A program was designed primarily
for reducing hydrocarbons. The figures of merit for Case 2 are poorer
than for the Case 1. This again is caused by the much lower emission re-
ductions, in tons, obtained for HC for an equivalent dollar expenditure.
It is concluded that placing primary emphasis on HC reductions results in
selecting a procedure which diagnoses and corrects engine misfire while
placing emphasis on CO reductions primarily leads to the selection of an
idle CO adjustment program. It should be noted that the most cost-
effective procedure may not be selected if regional authorities require
emission reductions in excess of those estimated for optimal strategies
in order to achieve their air quality goals.
2.3
Regional Evaluation
While the air quality problem is national in scope, its specific
characteristics vary considerably between basic geographic regions. The
effectiveness of mandatory vehicle inspection/maintenance for controlling
emissions therefore must be analyzed on a regional basis. It is conceiv-
able that a general program design could be used regardless of the air
quality problem of a particular region. However, it is much more likely
that particular regional characteristics will necessitate a specifically
designed program to achieve maximum cost-effectiveness. From the other
viewpoint, it is essential to investigate the sensitivity of an optimal
inspection/maintenance program design to regional variables in order to
determine the advantage of programs tailored to the needs of specific
regions.
41
-------
Regional analyses were therefore performed using the cost-effective-
ness computer model. Their purpose was to obtain preliminary data on the
regional sensitivity of optimal inspection/maintenance program design. In
addition to the Los Angeles basin (Region I), three other urban areas were
selected for analysis. This sample was specially chosen to provide a wide
range of regional air quality and demographic conditions. Los Angeles has
a photochemical smog problem that is related to high HC and NOX concentra-
tions. Region II was selected because CO is the major pollutant with HC
being of secondary concern. Region III involves a rather balanced situa-
tion. While there is an incipient HC problem, air quality degradation is
due rather equally to HC, CO and NO. Region IV's altitude of approximately
5000 feet above sea level makes its air quality problem different still.
Because of the reduced air pressure at such altitudes, the fuel metering
of the IC engine carburetor is considerably altered. The resultant high
fuel-to-air ratio leads to considerable unburned HC, which is by far the
dominant pollutant. It is evident then, that this small sample of four
regions provides a wide range of air quality problems.
2.3.1
Regional Characterizations
Presented below is a list of important data required to characterizing
the various regions:
. Emission Surveillance Data
. Demographic Data
- Regional Area (effective)
- Average Travel Time
- Vehicle Population Density
- Vehicle Powertrain Distributions
- Economic Factors
. Vehicle General Maintenance State
. Vehicular Emission Reduction Goals
. Emission Specie Weighting Function
42
-------
Regional differences were characterized in the model in several basic
ways. First, the different abatement goals for the three pollutant
species were reflected in different regional weighting functions for
total emission calculations. These emission weighting factors are shown
in Table A-5. For example, in Region I where high HC concentrations
occur, HC was assigned a relative weight of 0.6, CO of 0.1 and NO of 0.3.
On the other hand, in the case of Region II where CO is of prime concern,
CO was assigned a relative weight of 0.6, HC of 0.1 and NO of 0.3. Weight-
ing factors were assigned in a similar manner to Regions III and IV to
reflect specffic air quality objectives. The effect of these weighting
factor variations is to force the optimal program design towards one which
minimizes emissions of the most heavily weighted emission specie.
A second source of regional differences was the set of motor vehicle
population characteristics. Surveillance data taken in Region I were used
to estimate present emission levels for each region (Reference 5). Result-
ant vehicle emission rates in grams/mile for the emission controlled por-
tion of the automobile fleet appear in Table A-5. These data vividly show
the unique problem of Region IV caused by the altered air-to-fue1 metering
at high altitudes. Emissions of CO and HC are dramatically higher with
commensurate low values of NO when compared with the other regions. The
HC and CO emission levels for Regions II and III are slightly lower than
that of Region I because the emission controlled vehicles in these regions
have accumulated fewer miles than the vehicles in Region I. These lower
mileage levels are attributable to the relatively newer population of
emission controlled cars (i.e., 1968-1970) in regions outside the state of
California.
Regional variations in the population of uncontrolled, controlled and
post-1970 vehicles were accounted for in the model. Roughly, two-thirds
of the vehicle population in Regions II, III and IV were uncontrolled in
1971, as compared with less than one-half of the Region I population. The
high proportion of controlled vehicles in the Region I population is a
direct reflection of the more restrictive California emission control
statutes adopted during the 1960's.
43
-------
The geographical size of the region and average driving speed were
al~o included in regional characterization. These data were particularly
crucial in assessing the user inconvenience time and therefore in developing
inspection system configurations associated with the inspection/maintenance
program. Another important regional factor was the absolute size of the
vehicle population. In the regions considered, the emission controlled
vehicle population varied from four million cars for Region I to 600,000
cars for Region IV.
Finally, regional differences in costs should be incorporated in the
model. While the preliminary regional analysis results presented below
does not reflect these differences, provision has been made in the model
to include regional variations of construction and labor costs. Varia-
tions in labor and construction costs between regions would lead to
different labor intensities (or capital intensities) in the optimal
regional program designs.
These regional data, particularly regional vehicle and air quality
characteristics, provided the basis for the regional evaluation of
optional inspection/maintenance programs which is presented in the
following section.
2.3.2 Comparison of Regional Programs
The approach taken in assessing the impact of regional differences
on optimum vehicle inspection/maintenance programs consisted of:
. Designing an optimal program for each region using the given
data set
. Comparing the performance characteristics of optimal programs
between regions.
To identify accurately the most cost-effective program within a
region requires an evaluation of both those inspection/maintenance alter-
natives previously used (i.e., idle and extensive) as well as several
"customized" strategies. These "customized" strategies generally take
the form of an idle CO inspection/maintenance program for regions character-
ized by high CO weighting factors. An alternative for areas having a
critical hydrocarbon problem is to incorporate an idle misfire analysis
into the basic idle program. It will be seen that both of these procedures
44
-------
yield results which are somewhat more cost-effective for Regions II and
IV than using the procedures previously identified.
The main problem in designing cost-effective program(s) for each
region involved determining optimum specific pass/fail inspection
criteria. Numerical values for these criteria were obtained using the
developed linear programming algorithm. Table 2-13 shows the derived
inspection pass/fail criteria for several inspection/maintenance alterna-
tives. As could be expected, those regions with a high CO problem tend
to reject a larger percentage of CO related engine parameters (e.g., idle
CO), whereas regions experiencing an HC problem tend to fail more vehicles
based on rpm, timing and misfire.
Parameter I (Idle) II (Idle CO) ~ IV (Idle +
Ignition
ICO (%) 3 3 3 3
rpm (rpm) 525 N. I. N. I. 560
Timing (Deg-BTDC) 4.4 N. I. N. I. 3
Misfire (%) N. I. N. I. N. I. 2.5
N.I. - Not Inspected
TABLE 2-13
Optimal Parameter Pass/Fail Inspection
Criteria for Selected Regional Programs
Utilizing these pass/fail criteria together with the regional data,
a series of computer simulation runs were made for the four regions.
Tables 2-14 and 2-15 present the results of the computer evaluation for
both idle and extensive programs of inspection/maintenance. The basic
case consisted of a direct engine parameter inspection undertaken in a
franchised garage. A similar set of results would have been achieved if
an emission signature inspection program had been analyzed.
45
-------
TABLE 2-14
Comparison of Regional Programs Emp10ying a
Parameter Inspection Strategy
IDLE MAINTENANCE
REGION FIGURE-OF-MERIT COST PER EMISSION REDUCTIONS
VEHICLE PERCENT
$/TON $
HC CO NO
I 6
IDLE 200 3.00 4 6
IDLE (MISFIRE) 300 6.00 10 6 6
II
IDLE CO ONLY 35 2.50 1 7 0
IDLE 50 3.00 3 6 0
III
IDLE CO ONLY 55 2.50 1 7 0
IDLE 70 3.00 3 6 0
IV
IDLE 940 3.00 1 1 13
IDLE (MISFIRE) 550 6.00 6 1 13
PARAMETER INSPECTION
TABLE 2-15
Comparison of Regional Programs Employing a
Parameter Inspection Strategy
REGION FIGURE-OF-MERIT COST PER EMISSION REDUCTIONS
VEHICLE PERCENT
$/TON $
HC CO NO
I
A 230 7.00 21 6 6
B 350 13.50 23 12 3
II
A 110 6.80 21 6 0
B 120 13.50 23 11 -4
III
A 140 6.80 21 6 0
B 165 13.50 23 11 -4
IV
A 275 7.00 16 1 13
B 380 13.25 17 4 6
EXTENSIVE A AND B MAINTENANCE
PARAMETER INSPECTION
A
B
IDLE PLUS IGNITION WITH DYNAMOMETER
IDLE PLUS IGNIT ION PLUS INDUCTION
46
-------
Tables 2-14 and 2-15 list three system performance variables used
for regional contrasts. They are:
. Figure of merit in terms of total cost per weighted ton
reduction
. Average cost per vehicle per inspection
. Emission specie average five-year reductions in percent.
An analysis of Table 2-14 indicates that an idle CO only inspection/
maintenance policy yields the most cost-effective strategy for both
Regions II and III. The corresponding figure of merit for Region II is
nearly an order of magnitude less than for the Region I idle maintenance
programs. This is caused by the high weighting factor given CO for
Region II. In terms of CO emission reductions, a program in which only
idle CO is maintained is preferred over a total idle maintenance program
because adjustments in both timing and rpm tend to increase CO. Examina-
tion of the predicted emission reductions shows a seven percent decrease
for a program of idle CO maintenance only whereas a total idle maintenance
program yields a six percent reduction over the five years. Because of
driveability problems resulting from adjusting idle CO only, that frac-
tion of vehicles with low idle speed may also require adjustment.
This requirement should not significantly influence the previously stated
conclusions.
A comparison of the performance characteristics of an idle plus un-
loaded misfire inspection program with the "standard" idle program yields
somewhat ambivalent results. For Region I, the addition of an unloaded
misfire diagnosis cannot be justified in terms of its cost effectiveness
(a figure of merit of $200/ton for idle maintenance versus $300/ton for
idle maintenance with misfire). In the case of Region IV, however, a sub-
stantially improved figure of merit has been achieved with the addition of
an idle misfire inspection. The improved figure of merit can be directly
attributed to the larger hydrocarbon emission reductions achieved and to
the corresponding large hydrocarbon emission weighting factor assigned to
Region IV. The generally lower percent reduction of He and CO emissions
for Region IV is primarily caused by the much higher vehicle emission
levels. The converse is the case for NOX'
47
-------
The results for the more extensive maintenance programs tend to show
less performance variation across regions. The resultant figures of merit
for each region are more nearly the same than for the idle programs. In
the case of Regions II and III the figure of merit has increased almost
twofold whereas Region IVls figure of merit has decreased by at least two-
thirds. Again for Region IV, this can be explained by the heavy emphasis
placed on hydrocarbons in the figure of merit. The corresponding hydro-
carbon reductions produced by the extensive maintenance programs appear to
be about the same across all regions. The substantial differences in
regional CO and NO emission reductions can be attributed to different pass/
fail criteria and vehicle emission levels.
Figures 2-15 through 2-17 show Region IV emission time histories for
HC, CO and NO, respectively for an Extensive B program. These graphs were
generated directly by the Economic Effectiveness Computer Program. Where-
as the test fleet emission profiles appear similar to those developed for
other regions, the baseline fleet emission history is somewhat different in
shape (Figure 2-2). This difference can be attributed directly to the use
of Region lis (Los Angeles) voluntary maintenance schedule. in Region IV.
A logical future development would be the construction of unique mainten-
ance models for each region. It should also be noted that the significantly
richer carburetion at altitude will probably effect the value of the emis-
~
sion response coefficients which were measured at a near sea level eleva-
tion. The CO and NO emission reductions projected for timing, rpm, air
cleaner and PCV maintenance therefore are probably overestimated.
In summary. the viability of mandatory vehicle inspection/maintenance
programs depends upon the demographic characteristics of a given region.
For regions characterized by a high CO problem (e.g., Region II) either a
program in which an idle ~ir-to-fuel adjustment is made or the standard
idle maintenance program appears very cost effective. Use of these same
programs for a region having a high HC problem (e.g., Region IV) produces
results which have poor cost-effectiveness. Controlling HC emissions
requires a more extensive maintenance program, For Region IV the exten-
sive maintenance treatment generates the most cost-effective results both
in terms of hydrocarbon reduction and in terms of the figure of merit. A
primary conclusion from these results is that abs01ute CO emission
48
-------
~
~
-I
.~
m
~ l
~
o
-~
:r:
:VI
-
0.00
0.00 ...
1 . 71 +
3.43 +
5.14 ...
6.86 +
8.57 ...
10.29 ..
12.00 ...
13.71 +
15.43 +
11.14 +
18.86 +
20.51 +
22.29 +
24.00 ...
25.71 +
21.43 +
29.14 +
30.86 +
32.57+
3.4.29 ...
36.00 +
31.71 ...
39.43 +
41.14 ...
42.86 ...
44.51 +
46.29 +
48.00 +
49.71 +
51.43 ...
53.14 +
5Ct . 86 +
56.57 +
58.29 +
60.00 +
- - --- ------
-_.---
1.67
+
+.
+
+
+
+
+
+
.+
+
+
+
+.
...
...
+
...
+
+
...
+
+
+
+
+
+
+
+
+
...
...
...
...
+
...
...
--,------
FIGURE 2-15 HC Emission Time History for Region IV
HC DATA HISTORY
EMISSION(~PM) VS. TIMEiMO.)
3.33
+
+
...
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
...
+
+
+
...
+
...
...
+
5.00
6.67
S.33
: : : ~
+ + ...,
+ + .. I
+ + +,
+ + ... I
+ + ... f
+ + ... I'
... + ... /
+ + ... /
+ + .. /
... ... .~
+ ... ;/t
... ... ~4 ..
+ + /..
... ..../>..
... ... j +
+ + I +
... + i ...
... ...~. +
+ + l ..
+ ... I ...
... + ¥ +
+ + ~ ...
+ ... I I-
+ I +
+ .4- ...
... J... ..
+ I + I-
+ I + ...
+ / + +
+ i- + +
... j-~ ... I-
+ I ~ ... +
"'---~__J.~__n~~_.__---_._-"- - . ...- __4__-
...
+
+
...
+
+
+
+
.-
10.0D
... .-.
... :J
+ r
+ "
+ 3
... -
... .CJ
+ CJ
I +10
I :z
\ +- :I:
+!-
+ :'"
.....
+'0
+
... ~
+ 1-
+.m
... ><
... ;.....
m
... Z
... '"
... -
... <
... m
... ."
+
+ ."
... ~
+ 0
... (i)
... ~
+ »
+ ~
...
tXJ
»
VI
m
u-
- --- ----_.~
-------
FIGURE 2-16
CO Emission Time History for Region IV
CO DATA HISTORY EMISSION(GPM) VS. 1IMf(Mn.}
0.00 25.00 50.00 75.00 100.00 125. 00 150.GD
0.00 + ... ... .. + .. .. In
1.71 ... + ... ... .. + +
3.43 ... + ... ... + +- ... 0
5.14 ... + .. + ... .. + m
f.86 + + ... + ... .. + ~
8.57 + + + + + .. + -
+ ... .. ... + .. + '"
12.00 + + ... + + ... + '"
-
13.11 + + .. + + .. + 0
+ + + ... .. .. + Z
17.14 + + ... + + .. +
18.86 + + ... .. + .. + %
+ + + ... + .. ..' -
, 22.29 '"
' .. + + ... + ... + ....
-If
.~i 24.00 + + ... + + ... + 0
c.n", 25. 1 + .. ... ... .. .. + ::a
o l 27.43 + + + + + t ... 1-
~ 29.14 '" + ... + + t ..
0, + + ... + + .. ... m
Z' 32.51 + + + + + .. + ><
-I: ....
:J:! 34.29 + + ... + + .. .. m
VI'
6.00 + + + + + .. + Z
37.11 + + + +- ... .. .. '"
39.43 + + .. ... .. + ..
41.14 + + + + + .. + <
42.86 ... ... .. + + ..' + m
44.51 .. + + + + ... UI
46.29 ... +- + ... .. ...
48.00 + + + .. + ... l"'a
\::a
49.11 .. + +- ... + + 10:
.. + .. + + + +
53. 14 + + + + + .. g:II' (i)
» +
54. 86 .. + + ... ... .. VI' ., + ::a
",
+ ... + + ... I -i .. + I:J:-'
58.29 + + ... + + ' m to ... '~
60.00 ... + + + ... A\~ .. ...
-------
FIGURE 2-17
NO Emission Time History for Reg'ion IV
---- -----"__m-
---~. ~ - --_c ---.
~j? ~,r:j" t; 1/' :;~ T£!. r;.l:>~~
NO DATA HISTORY
EMISSION(GPM) 1S. TIMF(MO.)
0.00
1.67
3.33
5.00
6.67
d.33
1U.08
0'1
....
'0.00.' ':'" . '~:j(R': - \~~~~~:: ~.;~': 5: -
.. t'" + ... ~ + Z
1.71 .. ... ... ... ... ~ ...
3.43 .. ., + ~ .. + ... ... ... '0
'
5.14 .. + +- + +- " .. ... m
6.86 .. ... I + + + ... ... ~
8.57 .. + I + ... ... ... ...
-
10..29 " ; ... , ... .. ... + ... ''''
12.0.0. + .. r + + + ... + '"
-
13.71 .. .. ... .. .. .. + :0
15.43 + + I '" + + .. + Z
17.14 ... + I + + + .. '+
18.86 ... .. . .. ... ... ... + :E:
,- 20.57 .. '" ... .~ ... ... ... .. + -
. .-1; 22.29 + ... 1 ... ... ... .. ... ~ '"
:~: " '~
~m 24.0.0. + ... # ... ... ,+ .. + '0
l 25.71 .. ... I '" + ... .. +
~ 27.43 + ... I '" ... ... .. + ,;g
o 29.14 ... ... I '" ... ... ... + 1-
"!Z: 30..86 .. ... !' ... ... ... .. ... 1m
,-I I
,:I: 32.57 ... ...' ,, '" ... ... ~ + ,)(
V'I' 34.29 .. L '" ... ... ... .. ... '~
J
36.00 '" ... " ... +- ... .. + :m
37.71 ... ... I '" ... ... .. + Z
39.43 .. ... , + ... + .. ... '"
-
41.14 + ... I '" ... + .. + <
42.86 .. + I '" ... ... .. ... m
44.57 .. " ~ . .. .. ... ... .. '01
46.29 ... + , " + + .. +
48.0.0. ... + I ... ... + +- + "'a
49.71 ... ... , + + + .. ... ;g
51.43 .. ... , + ... .. ... .. 0
53.14 ... ... I ... + + t + (i)
54. ,86 " +-A ~ ... ... + + + ;g
,I
56.57 ... ... m» + ... ... +- ... ~
VI, ,
58.29 ... ... -I V'I + + ... .. + ~
.m
60.0.0 + 'N ' - .. ...
... .. ~ +
.
.
-------
reductions (tons) from automobiles can be achieved for considerably less
cost than comparable hydrocarbon reductions. Although no specific pro-
cedures were examined for reducing NO, it appears that the cost of a
comparable NO reduction falls somewhere in between the cost of reducing
CO and HC. The most attractive inspection/maintenance strategies from
Tables 2-14 and 2-15 for the four regions are summarized below in Table
2- 16 .
TABLE 2-16
Selected Inspection/Maintenance Proqrams
for the Regl0ns Studled
Region
I
Strategy Principal Specie Reduced
Idle parameter inspection/mainten- HC, CO, NO
ance or load mode emission inspec-
tion and Extensive A maintenance
II
III
Idle CO only inspection/maintenance
Idle CO only inspection/maintenance
CO
CO
IV
Direct inspection of idle para-
meters and ignition misfire and
Extensive A maintenance
HC
52
-------
3.0 CONCLUSIONS
Data from previous sections generally support the conclusions of the
original study with respect to the cost and effectiveness of inspection/
maintenance procedures. The detailed conclusiQns of this three-part
study are summarized in this section.
3.1
Mass Emission Analysis
The primary objective of the mass emission analysis was to compare
the effectiveness of inspection procedures as determined using the
newer constant volume sampling (CVS) mass emissions measurements
with those determined using concentration emission measurements.
Although differences in figures of merit and emission reductions
result from using the different emissions test methods, the
economic-effectiveness ranking of the several inspection/mainten-
ance crocedures investigated was not changed.
. Absolute emissions reductions (tons) calculated on a mass
basis are usually slightly larger than when calculated on
a concentration basis, thus resulting in a better figure
of merit.
. Percentage reduction of emissions from a test fleet when
referenced to their respective baseline fleet emission
levels are larger for NO and smaller for HC and CO when
measured on a mass basis.
Reoptimization of the inspection/maintenance procedures utilizing
mass emissions data resulted in slightly greater emissions reductions than
were achieved in the earlier analyses. For the idle adjustment inspection/
maintenance programs, the increased emission reductions ranged from one to
several percent of the baseline fleet level. The most significant impact
of reoptimization occurred for the case of a direct engine parameter
inspection followed by extensive maintenance. For this case, increased
emissions reductions of three and four percent were achieved for NO and
HC, respectively. CO emission reductions remained about the same. Higher
vehicle rejection fractions and therefore costs resulted, but the net
effect was to improve the figure of merit. Additional conclusions derived
from the model using Los Angeles basin emission weighting factors are as
follows:
53
-------
. Inspection of emissions in selected operating modes finds
less than half of the timing~ PCV and air cleaner malfunc-
tions within the vehicle population.
. The most cost effective inspection/maintenance procedure
(figure of merit of 155) is a direct inspection of idle
rpm and idle CO in a state inspection lane followed by
adjustment of these parameters. Five-year average HC, CO
and NO emission reductions are 3, 7 and 8%, respectively.
. The second most cost effective inspection/maintenance pro-
cedure (figure of merit of 190) employs idle mode HC and
CO and loaded mode HC emission inspection in a state inspec-
tion lane followed by the adjustment of the idle parameters
and the repair of ignition system components causing misfire.
Five-year average HC, CO and NO emission reductions of 9, 6
and 2% are achieved, respectively.
. The largest emission reductions are obtained with a direct
engine parameter inspection followed by an Extensive B
maintenance program, all performed by franchised garages.
This program is 5 to 8 times more costly than the most cost
effective inspection/maintenance procedure, however, the HC
and CO emission reductions achieved are nearly double those
obtained using the Extensive A state lane em~ssion inspection
procedure.
. Emission reductions at the end of five years are approximately
twice those obtained on the average over a five-year period.
. The emission means observed in the pre and post tune popula-
tion as measured by the fleet deterioration experiment are
statistically different at the 90% confidence level.
. The Extensive A procedure becomes most cost effective when
the statistically significant emission reductions are used
in place of the predicted mean value emission reductions
in the figure of merit.
3.2 Sensitivity Studies
Sensitivity studies were performed to determine the
input data and assumptions on the results of this study.
figure of merit was found to be sensitive to:
. The weighting factors assigned to emission species,
when CO reductions are weighted heavily. i.e., 0.9.
. Direct program costs such as maintenance, labor and
convenience.
influence of the
In general, the
particularly
user in-
Both the emission reductions and the figures of merit achieved were found
to be sensitive to:
54
-------
. The effectiveness with which repair is made by a service
organization. For example, failure to set average fue1-
to-air ratio 1% richer than average specification re-
sulted in nearly complete ineffectiveness of an idle
maintenance program to control CO.
. The effectiveness of voluntary maintenance as it is currently
occurring. Obviously, as vehicles are better maintained
voluntarily. a mandatory program becomes less necessary.
. The estimated rate of deterioration of the engine parameters,
particularly idle fue1-to-air-ratio and misfire.
. The frequency of imposed inspection and maintenance. Once
yearly is nearly optimum.
3.3 Regional Impact of Inspection/Maintenance Effectiveness
The economic-effectiveness of mandatory programs of inspection/main-
tenance is most influenced by the emission reduction weighting factors
assigned for various regions. Weighting factors which stress CO reduction
generally result in a program that does little to control HC and NO emis-
sions. Large absolute reductions of CO occur, thereby resulting in
extremely good figures of merit (35 to 55 $/ton). Those procedures which
adjust idle fue1-to-air ratio to specification are most effective. CO
emission reductions can be approximately doubled by maintaining components
of the induction and air reactor systems, but at four times the cost and
with figures of merit which are twice as large as that of the simpler idle
adjustment program.
In regions where photochemical smog is a problem and HC and NO emis-
sion reductions are heavily weighted, timing adjustments and ignition
system repair become important. Little reduction of CO emissions is
desired since NO emissions would be adversely affected. For Region IV
where HC is most heavily weighted, maintenance of idle timing and ignition
system misfire is over three times more cost effective than performing idle
adjustments (idle CO and rpm).
55
-------
APPENDIX A
STUDY GROUND RULES AND DATA BASE
A. 1
Introduction
Presented and discussed herein are the study ground rules and assump-
tions used in conducting the current Economic Effectiveness study. Also
presented are the emissions weighting factors and regional demographic
characteristics utilized in the analysis.
A.2 Basic System Ground Rules
The general study ground rules are:
. The economic-effectiveness of the procedures investigated is
determined by calculating a figure of merit, which is defined
as follows:
Figure of -
merit -
annularized, discounted total program cost
~ Wi[baseline emissions(eb.)-test emissions(eti)]
. ,
i=l
(A-l)
. Only vehicles with emission control equipment (air reactor
and engine modification) are considered (i.e., 1966-1970).
. The effects of vehicles entering and leaving the population
because of attrition and new production is not considered.
. Mean values of emissions predicted for vehicle populations
treated by both the engine parameter and emission inspection
procedures are allowed to vary with time and with the extent
of maintenance, however, the variances about these means are
assumed to be time invariant.
. Basic maintenance of those engine adjustments not covered in
the enforced maintenance procedure is assumed to be performed
voluntarily by the vehicle owner.
. All maintenance is performed in franchised garages. Manda-
tory maintenance is limited to restoring the following para-
meters to manufacturer's specification: idle adjustments of
fuel-to-air ratio, timing and rpm; components of the ignition
system when causing misfire; the induction system components
of air cleaner and positive crankcase ventilation (PCV) system
when severely restricted; and components of the air injection
system when failed.
56
-------
. Inspection procedures requiring large capital expenditures
(i.e., the purchase of equipment for either remote sensing
or exhaust emission measurements) are always performed with-
in a state operated system.
. All vehicles failing a state inspection are reinspected by
the maintenance organization.
. All maintenance by service organizations is assumed to be
reliably performed and only those idle adjustments adversely
affecting emissions are restored to specification.
. Pass/fail inspection criteria are optimized for the first
inspection interval and remain invariant with time.
. The cost of maintenance labor and parts are based upon
Chiltons flat rate manual.
. State inspections require 1.5 minutes for the idle mode test
and 2.2 minutes for the idle plus loaded mode tests. State
labor rates average $4/hour.
. Emissions reductions are based on measurements made with the
constant volume sampling (CVS) closed seven-mode test pro-
cedure.
. The initial maintenance state of the vehicle population, the
rate of engine parameter deterioration and the effectiveness
of voluntary maintenance were assumed to be identical for
the four regions studied.
A.3 Basic Input Data
The basic cost and emissions data described in the previous study,
References 1 and 2, were used in this study. These data are summarized
in Tables A-l, A-2, A-3 and A-4.
A.3.1 Emission and Cost Data
The seven-mode CVS mass emissions response coefficients from the 11
power trains tested in the definitive orthogonal experiment were weighted
by their approximate fraction in a California population to obtain the
population average reported in Table A-l. Although these coefficients are
similar to those based on concentration measurements, they are sufficiently
different to effect predictions of emissions reductions for inspection/
maintenance procedures.
57
-------
TABLE A-l Emissions Responses to Engine Parameter Changes
(Definitive Orthogonal Experiment)
+
Emissi on, Ae Emission Response to Enqine Parameter Changes, ae
aP
m nQ e RPM P r C eaner a e
Ae/deg L\e /rpm /1n H20* Ae/deg** Ae/%
HC, g/m 0.055 -0.006 0.48 0.0010 0.031 1.11
CO, g/m -0.32 0.020 12.6 0.072 6.75
NO, g/m 0.11 0.00029 -0.52 -0.0024 0.034
01 49 mph CO, % -0.0030 0 0.12 0.0028 0.017
00
33 mph CO, % 0.00137 0 0.74 0.0039 0.031
Idle
HC, ppm 7.97 -0.55 -15.1 0.18 8.27 80.4
CO, % 0.0012 0.00034 0 O. 00018 0.87
0-15 moh HC, ppm 6.41 -0.067 29.8 0.094 1.25 115.0
+ units of numer.ator qiven in first column
* pev restriction measured by crank case pressure at idle
**air cleaner restriction measured by AC tester in degrees
-------
TABLE A-2
Equipment and Procedures Required for Diagnosing
Engine Parameter Malfunctions
U'1
1.0
En~ir.e Parameter Emission Signature
Subsystem Inspection
Equipment Procedure Equipment Procedure
Idle
... rpm Tachometer Idle rpm NDIR HC analyzer Idle HC
- Timing Timi ng 1 i ght Basi c Timing NDIR HC analyzer Idle HC
- 'Fue l-to-Ai r Fuel-to-air or Idle CO NDIR CO analyzer Idle CO
NDIR CO a.nalyzer
Ignitfon-Misfire Engine electronic Misfire at idle NDIR HC analyze~! 45-mph HC
analyzer/dynamo- and 50-mph road- dynamometer
meter (optional) load
Induction
- PCV Pressure gage Idle crankcase NDIR CO analyzer! 45-mph CO
pressure dynamometer
- Air Cleaner AC ai r cleaner Pressure drop NDIR CO analyzer! 45-mph CO
te s te r across element dynamometer
Ai r Reactor Fuel-to-ai r or Di rect vi'sua 1 NDIR CO analyzer! 45-mph CO
- inspection &
NDIR CO analyzer change in CO dyn-amometer
upon discon-
necting the.
air pump
-------
TABLE A-3 State Inspection Lane Configurations and Costs
No. Inspec- No. Informa- No. SupDort Data ACQuisition
tion lanes tion Officers Personnel Item Costs, $/Site*
(T) (2) I
1 1 1 20 K 22 K
2 2 2 33 K 37K
3 2 3 45 K 51 K
* Add $ 7,000/lane for dynamometer for loaded mode inspection.
(1) One emission specie
(2) Two emission species
TABLE A-4 Franchised Garage Inspection/Maintenance Labor Times
Subsystem Engine Parameter Time (hours)*
Inspection 'Maintenance
Idle Adjustments ICO and rpm 0.10 0.20
Timi ng 0.05 0.10
Secondary Ignition Plug, distributor 0.25 1.30
and wire harness
Induction Air cleaner uO.05 0.2
PCV 0.05 0.3
Air injection 0.15 0.6
*A11 labor charged at $10.00/hour
60
-------
Franchised garages are assumed to use commercially available instru-
mentation and equipment to perform vehicle inspections, although a new
procedure for diagnosing air reactor system malfunction is required. The
equipment, instrumentation and procedures required for diagnosing engine
parameter malfunctions are presented in Table A-2. The influence of the
number of inspection lanes in a state inspection facility upon manpower
requirements and equipment costs is shown in Table A-3. The productivity
of data acquisition equipment and personnel increases with the number of
test lanes because of time sharing. For example, two information officers
with automated data acquisition equipment can handle three inspection lanes.
Inspection and maintenance labor times are shown in Table A-4. These times
were developed using flat rate manuals and rough time studies.
A.3.2 Demographic Data
The effectiveness of mandatory inspection/maintenance programs was
studied in four urban centers with contrasting air quality problems. A
summary of the demographic data for the four urban centers is presented
in Table A-5. Data fundamental to characterizing a region include air
quality goals, motor vehicle exhaust emission levels as measured by a
vehicle surveillance program and the number and number density of the
vehicle population to be inspected and maintained.
The estimated vehicle mass emission levels for Region I for a 1966-
1970 exhaust-controlled population with an average odometer reading of
32,000 miles (Reference 3) were used as the basic data for all regions
because of a lack of information on automobile emission levels for Regions
II, III and IV. Further, it was assumed that the engine parameter deteri-
oration rates for these three regions were identical to Region I. Since
exhaust controls were not utilized in Regions II, III and IV until 1968,
their mass emission levels at the start of a mandatory inspection/main-
tenance program were assumed to be equivalent to Region I at an average
odometer reading of 22,000 miles. To include the effect of altitude on
the estimated mass emission levels of Region IV, the emission levels for
an odometer reading of 22,000 miles were adjusted (Reference 4) by the
fOllowing amounts:
61
-------
He = 1.3 x Region
CO = 1.6 x Region
NO = 0.5 x Region
I mass
I mass
I mass
emission
emission
1 evel
1 evel
1 evel
emission
The emission weighting factors of Table A-5 were selected based on a
literature survey as well as upon discussions with regional air pollution
control personnel. They are considered as only rough measures of the
desired air quality. In Region II, for example, traffic -is extremely
congested and approximately one-third of the total driving time is spent
idling. This causes relatively high CO concentrations at ground level.
Thus, a high weighting was placed upon CO emissions reductions for this
region.
TABLE A-5
Regional Attributes
REGION
I II III IV
ATTRIBUTES
0 VEHICLE POPI.J~ATlON <4.0X 1r1' 2.8 x 106 0.87 X 106 0.6 X 106
o REGIONAL AREA (SQ. MI.) 1250 1212 62 700
o POPULATION DISTRIBUTION (1971) ~~
0 PRE CONnOl 44 64 71 63
o CONTROL 51 32 27 33
o POST 1970 5 <4 2 <4
o EMISSION WEIGHTING FACTORS
0 HC 0.6 0.1 0.3 0.9
o CO 0.1 0.6 0.<4 0.1
o NOx 0.3 0.3 0.3 O.
o VEHICLE EMISSION LEVELS (G/M)
0 HC 7.1 7.0 7.0 9.1
o CO 87.7 86.0 86.0 137.4
o NOx 6.2 6.4 6.4 3.2
62
-------
The total automobile population and metropolitan regional area were
taken from Reference 5 for Region I and Reference 6 for Region II. The
percentage breakdown of the automobile population for the four regions by
the categories of uncontrolled, controlled and 1971 automobiles were de-
rived from Reference 7. The number of maintenance garages in Region I was
provided by the local Highway Patrol.
63
-------
APPENDIX B
ECONOMIC EFFECTIVENESS COMPUTER MODEL
B. 1
Introduction
This appendix presents a discussion of the major economic effective-
ness model revisions and program developments undertaken during this study.
These revisions and developments include:
. The emissions prediction model has been revised to calculate
emission reductions based on closed seven-mode cycle (CVS)
mass emission data.
. A statistical analysis model was developed to estimate
confidence limits on predicted emission reductions.
. A linear program optimizer was developed for determining
optimum parameter and mode emission pass/fail criteria.
. Regional demographic models were formulated for characteriz-
ing various urban areas.
B.2 Estimated Emission Time Histories
The estimation of emission time histories is central to the calcula-
tion of emission reductions. As a consequence of converting the model to
a mass basis, both maintenance and deterioration models required revision.
B.2.1
Emission Reduction Model
Emission reductions achieved by parameter inspection/maintenance is
calculated with a procedure essentially identical to the method used pre-
viously to estimate composite emission reductions (Reference 1). Calcula-
tions were performed to assess whether the interaction of two engine
adjustments upon emissions should be considered in the model. The fre-
quency of interactions found to be statistically significant at or above
the 90% level of confidence is summarized in Table B-1. As can be seen,
the most frequently occurring interactions (i.e., four out of eleven
vehicles) are idle rpm with idle CO (BE) for HC; and a confounded inter-
action of timing with restricted air cleaner or idle rpm with PCV for CO
and NO (AD=BC). The above interactions were not found to have a signif-
icant impact on predicted mass emission reductions. The effect of the
64
-------
TABLE B-1
The Number of Cars Out of those
Tested for which Statistically
Significant* Effects were Determined
Definitive Orthoqonal Experiment
Main Effects
A B C D E F
Composite Timing Idle RPM PCV (8)** Ai r Cleaner I dl e CO Air Pump (3)**
HC 10 11 4 3 3 3
CO 6 6 5 8 9 3
NO 11 3 6 6 3 1
r~ass
HC 8 9 2 3 2 3
CO 7 8 3 7 8 3
NO 11 1 3 3 2 1
Interactions
Composite AB=CD + ~.C=BD + AD=BC + AE BE CE DE
--
HC 5 0 1 0 ? 0 1
..
CO 3 3 4 1 5 2 1
NO 2 4 5 2 1 1 2
Mass
HC 3 1 3 0 4 2 1
CO 2 2 4 2 2 3 2
NO 2 2 4 3 0 1 1
*Statistically significant at or above the 90% confidence level
**Maximum of 11 data points per parameter with exception of PCV
(8 data points) and air pump (3 data points)
+Confounded interaction effect--there is an equal probability of
either interaction
65
-------
timing-air cleaner interaction on estimates of NO mass emission reductions
was found to be the most significant of the frequently occurring inter-
actions. Even here, neglecting the interaction effect resulted in only
a 9% error in the prediction of the emission change caused by maintaining
the air cleaner and timing, simultaneously. Therefore, as was done on the
previous study, interaction effects were not considered when calculating
emissions changes resulting from maintenance.
B.2.2 Deterioration of Maintenance
After maintenance, engine adjustments will deteriorate resulting in
both mass and inspection mode emission changes. The previous model re-
flected this effect as a direct change in composite emissions, the change
being prorated to each of the three major subsystems (idle, ignition and
induction). Although this approach is consistent with the existing ARB
(Reference 3) and AAA (Reference 8) composite emissions (seven-mode) data
banks, it is inconsistent with the mass emissions data (CVS, seven-mode)
used in this study. Since parameter deterioration data will be available
at the conclusion of the Extended Phase One study, a modeling approach
which is consistent with this data, while bridging deficiencies in the
current data base, was selected.
The approach taken was to use the mass emission influence coefficients,
ae/ap, derived from the definitive orthogonal experiment in conjunction
with the best estimate of the rate of deterioration of the individual para-
meters to calculate the emission deterioration rate. A residual term is
retained to reflect the emission deterioration caused by engine parameters
not considered in this study such as compression loss as well as carburetor
and cylinder deposit effects. An emission deterioration equation of the
following form was developed:
6e. 6e I 3 10
1 - -; + ~ ~
6M - 6M 0 ;=1 j=l
[ae; 6Pj]
ap. x 6M
J
(B- 1 )
66
-------
where:
Ae.
1
AM
Ae; I
6M 0
a e.
1
ap.
J
Ap.
::.:.J.
6M
=
population average mass emission deterioration as a
function of mileage, emissi~n level/mileage, {grams/
mil e)/mil e
=
component of emissions change because of engine
deterioration effects not considered in this study,
emission level/mileage, {grams/mile)/mile
=
mass emission change caused by a change in parameter
level, ~Pj {grams/mile)/parameter level change
=
parameter level average deterioration as a function of
mileage, parameter level change/mile
The above formulation necessitated converting the deterioration rates
of the previous study which were based on subsystem values to direct para-
meter deterioration rates. This was performed by solving for the engine
parameter deterioration term in Equation B-1:
Ap.
J
6M
K..
lJ
=
[Aei - 6ei ] / oei
K;j AM AM op.
o J
(B-2)
=
factor for apportioning emission changes at the sub-
system level to the several engine parameters signi-
ficantly influencing a specific emission species~ ei.
This approach generally results in a set of linear equations which are
insufficient in number to solve explicity for ~p./~M and K'J'. Therefore,
J 1
it was necessary to begin by selecting values of K.. in order to compute
lJ
parameter deterioration rates. These values were then used with the
voluntary maintenance model to predict emission time histories on a CVS
mass basis. The predicted emission time histories were next compared against
the ARB emission surveillance data (Reference 3). The selected values of
67
-------
K.. were then adjusted to bring these two sets of data into correlation.
T~~le B-2 shows the engine parameter deterioration rates calculated using
the described procedure.
Figures B-1 through B-3 show the ARB surveillance data transformed to
a 1972 Federal mass emissions basis for HC, CO and NO, respectively.
These data were developed by first transforming to CVS mass units using
the following equations:
-4 2
HCmass = -.036 + 0.0381 HC7M - 0.64 x 10 HC 7M + 0.339
(B-3)
-7 3
x 10 HC 7M
(R2 = 0.71)
2 3
COmass = 28.6 + 38.1 C07M - 13.4 CO 7M + 1.89 CO 7M
2
(R = 0.75) (B-4)
NOmass
= 4.68 - 0.0072 N07M +
-8 3
- 0.148 x 10 NO 7M
-5 2
0.743 x 10 NO 7M
(R2 = 0.68) (B-5)
These equations were developed from a least squares curve fit of the
seven-mode cold start CVS mass data to the hot seven-mode composite
data acquired during the previous program (Reference 2).
Finally, the mass emission data base was adjusted to reflect the
higher emissions levels measured with the 1972 Federal Test Procedure.
The ARB data were curve fitted and transformed to CVS using equations
B-3 through B-5. The transformed data were then transposed by adjusting
the vertical scale to match the emission levels actually measured with
the 1972 Federal Test Procedure at an average actual accumulated mileage
of 32,000 miles. These premaintenance 1972 mass emissions measurements
were obtained from the ongoing fleet deterioration experiment, Table B-3.
As a consequence of both voluntary and enforced maintenance and the
subsequent deterioration of this maintenance, the population means for
the engine parameter settings and the mass and mode emission levels must
be adjusted. These adjustments are made immediately prior to the next
inspection interval using the following equations:
68
-------
Parameter
ICO
rpm
Timing
Misfire
Air pump
PCV
Air cleaner
TABLE B-2
Engine Parameter Deterioration Rates and
Voluntary Maintenance Effectiveness
Deterioration Rate Emission Decrease for +
Voluntary Maintenanc~
~p .
--l.. HC CO NO
~M
1 x 10-4 %/mile 0 9 0
-7.5 x 10-4 rpm/mile 2 0 0
6.7 x 10-4 deg/mile 2 -1 0
5.8 x 10-5 %/mile 8 0 0
2.5 x 10-6 psig/mile 0 0 0
5.4 x 10-6 in H20/mile 0 1 0
2.7 x 10-3 deg/mile 0 1 0
+ % reduction relative to premaintenance level
HC
CO
NO
TABLE B-3
Mass Emissions for in-use Vehicles as
Measured by the 1972 Federal Test
Procedure (gm/mi)
(CRC APRAC CAPE-13 Fleet Deterioration
Experiment)
Uncontro 11 ed
1971
4.3
51.0
6.3
Pre 1971 Controlled
11.3
122 . 0
4.4
7.1
88.0
6.2
69
-------
8.0
IU
...
~
"-
VI
~ 7.0
~
.1
~
...
Z 6.0
o
Vi
VI
~
IU
5.0
o
100
IU
...
~
"-
VI
~ 90
~
!
...
IU
>
IU
...
Z 80
o
Vi
VI
~
IU
8.0
~
~
VI
~ 7.0
~
!
...
IU
~
...
Z 6.0
o
Vi
VI
~
II.!
5.0
o
-1966-1970 EXHAUST CONTROLLED VEHICLES - ARB DATA
8
FIGURE B-1
HC Mass Emission Decay
co MASrM".ON DECAY
UPPER LIMIT
MEAN VALUE
I
- 1966-1970 EXHAUST CONTROLLED VEHICLES - ARB DATA
70
o
8
FIGURE B-2
CO Mass Emission Decay
NO MASS EMISSION DECAY
LOWER LIMIT
I
-1966-1970 EXHAUST CONTROLLED VEHICLES" ARB DATA
FIGURE B-3
NO Mass Emission Decay
70
48
48
48
-------
Mean Parameter Setting
Pj,M+LlM
= Mean Parameter Setting
(8-6)
Fi na 1
P. M
J ,
Initial
+ (Mean Parameter Setting Decay - Parameter Maintenance)
ap.
- RjLlP m+t,j + aMJ LlM
where: ap.
Mean parameter setting decay = ~ (from equation 8-2)
Parameter maintenance = change in mean parameter level
from voluntary and enforced maintenance.
resulting
Mean Emission Level
en,M+LlM
= Mean Emission Level
(8-7)
Fi na 1
en,M
Initial
+ (Mean Emission Level Decay - Emission Reduction)
+ 10 aen
E ap.
j=l J
(ap. )
art LlM - LlPj
Lle.
1
8.3 Parameter Inspection Models
The misfire and positive crankcase ventilation valve inspection
models were modified to improve the precision of the results and to
reflect more accurately the engine parameter field survey (Reference 1).
8.3.1
Misfire Model
The previously used misfire model did not allow both the frequency
and extent of misfire to vary non-linearly with engine load and speed,
Figure 8-4. When misfire effects were calculated for the seven-mode
cycle, misfire was assumed to be constant and equivalent to misfire at
the average 50 mph road load for the test cycle. The misfire influence
coefficient, LlHC/(% misfire), was developed from the statistically
designed experiment using an electronic firing line interrupter which
simulated a constant level of misfire independent of load. To reflect
the sensitivity of misfire to engine load, the Federal seven-mode cycle
71
-------
1.2
1.0
w
""
...
~ 0.8
#.
a
w
~ 0.6
~
0.4
0.2
IDLE
ENGINE PARAMETER SURVEY
(225 VEHICLE TEST POPULATION)
10
20
30
MPH
40
50
60
I
o
8 25
EQUIVALENT HORSEPOWER
(estimated)
I
35
FIGURE B-4
Percent Misfire at Test Condition
Loading
72
-------
weighting factors were used to develop an approximate population weighted
average misfire rate. All closed throttle modes (idle and deceleration)
were assumed to have a misfire rate equivalent to that at no load. The
resulting calculation is shown in Table 8-4. A mode weighted value for
misfire of 0.68% misfire is estimated. This value, when multiplied by
the influence coefficient, predicts the emission reductions achieved when
all of the misfiring vehicles within the population are maintained. This
weighting procedure is appropriate for concentration based emission
measurements, but is only an approximation for mass based HC emissiQns.
Inspection at heavier loads will tend to find incipient misfires at
low load, thus decreasing the probability of a substantial frequency of
high load misfires on the following inspection. To reflect this situa-
tion, the population weighted percent misfire was assumed to decrease
exponentially to half its original value according to the following
equation:
M = M e-kN
o
k = time constant yielding 50% of M in five years
Mo = initial population weighted misfire rate, 0.68%
N = number of inspection intervals
8.3.2 Positive Crankcase Ventilation Valve (PCV) Model
The earlier PCV model assumed for simplification that this device
either failed or operated nominally. This assumption is consistent with
the method of malfunction simulation used in the orthogonal test, but
does not reflect actual PCV valve degradation. Data from Reference 10
(Figure 8-5) indicates that HC and CO emissions vary approximately
linearly with PCV volumetric flow rate and that flow rate reductions
occur gradually because of deposit buildup. To reflect this condition
using the data in-hand (i.e., crankcase pressure measurements from the
parameter field survey) the PCV inspection model was revised to describe
the time, varying frequency and extent of malfunction of this device.
73
-------
TABLE B-4
Calculation of % Misfire for Seven Mode Cycle
(1966-1970 Exhaust Controlled Vehicles)
""
~
(A) (B)
Seven Mode Horsepower/Mode % Misfire/Mode Federal Register % Misfire/Mode
Federal Cycle Mode Weightinq Factor
(Fi~ure 3-4) (A x B)
Id1e 0 0.27 0.042 0.01134
0-30 mph Acceleration 25 0.75 0.244 0.18300
30 mph Cruise 8 0.32 0.118 0.03776
30-15 mph Deceleration 0 0.27 0.062 0.01674
15 mph Cruise 3 0.27 0.050 0.01350
15-50 mph Acceleration 30 0.90 0.455 0.40950
50-0 mph Dece 1 era ti on 0 0.27 0.029 0.00783
TOTAL 1.000 0.67967
-------
Crankcase pressure was the only indicator of PCV maintenance state
obtained in the engine parameter survey. The distribution function for
crankcase pressure (inches of water) is shown in Figure 8-6 and was used
to establish pass/fail criteria for the PCV system. PCV system flow rate
and, therefore, emissions were assumed to vary linearly with crankcase
pressure since data relating PCV flow to crankcase pressure under load
were not available.
B.4 Statistical Inference Model
The results derived from the economic effectiveness model are basic-
ally deterministic. That is, the predicted emission reductions are
computed utilizing fixed relationships for each operational step. Obvious-
ly, a real program of vehicle inspection/maintenance involves aGtivities
which cannot be characterized deterministically. For example, consider
the impact of maintenance effectiveness on overall program performance.
Presently the model assumes that all maintenance undertaken by the service
organization is performed with complete reliability. In actuality, garages
will not be able to set all engine adjustments exactly to manufacturers'
specification. The resulting uncertainty in maintenance effectiveness is
directly translated into uncertainties in predicted emission reductions.
To evaluate the implications of the program uncertainties, a
statistical inference model was developed. The function of this model
was to:
. Test whether model predicted emission reductions are
statistically significant (i.e., greater than zero)
. Estimate confidence limits on statistically significant
emission reductions
. Estimate confidence limits on the emission time history
profiles.
The model utilizes the emission predictions for both the baseline and
test fleets to perform these various statistical tests.
The main sources of uncertainty in predicting the performance of
candidate inspection/maintenance programs can be divided into two sets.
The first set consists of those uncertainties associated with the
inspection/maintenance process, namely:
75
-------
CARBURETOR DEPOSITS
- - - - WITH
WITHOUT
60,000
I I I
IDlE CONDITION -'
,~
, ...
,
,..
",.
,
~
~
.--
CRUISE C~NDITION -
--~' ~_I-
40,000
z
o
<0
...:
~ 20,000
~
...
...
I
In
Z
o
<0
...:
«
u
~
o
>-
:I:
10,000
8,000
6,000
4,000
5,000
4,000
3 000
, 0 20 40 60 80 100
PCV VALVE PLUGGING - %
FIGURE B-5
12 IDLE CONDITION
eft
~ 10
o
~
I
~ 8
X
o
Z 6
o
~
Z 4
o
<0
...:
5 2 -.
o
o
20 40 60 80
PCV VALVE PLUGGING - %
Effect of PCV Valve Plugging,
Carburetor, Venturi and Throttle
Body Deposits on Exhaust CO and
HC Emissions (Reference 10)
ENGINE PARAMUER SURVEY
227 VEHICLE TEST POPULATION
70
60
so
~ 40
I
>-
u
Z
~
~ 30
LOG NORMAL CURVE DISTRIBUTION
-------
. Inspection measurement variance (i.e., instrumentation)
. Maintenance reliability
The other set entails those uncertainties in the actual emissions
data base. These uncertainties are due to the fact that attributes of
the total population have been estimated from relatively small samples.
the case of influence coefficients derived from the orthogonal experi-
ments, the selected 11 cars were assumed to be characteristic of the
total population. The chief sources of uncertainty in the emissions data
base are:
In
. correlation of emissions-to-engine parameter variation
(i.e., residual error from orthogonal experiments)
. vehicle manufacturer-to-manufacturer emissions variability
. emissions variability caused by out of specification engine
parameters (a function of maintenance state).
The emission distributions for each specie are estimated during each
inspection/maintenance period using the distributions for the previous
period convoluted with the maintenance effects and associated error terms.
These distributions are then statistically compared with those for the
baseline fleet.
The standard "tl! score statistical test is used in comparing these
distributions. The null hypothesis used to test whether the two samples
came from different populations, i.e., that the means are significantly
di fferent, is:
HO: uT= Us
where: uT = test mean
Us = base mean
The alternative hypothesis is:
H1: uT < Us
If the analysis produces a positive test, the null hypothesis can be
rejected and that a difference in means does exist (accept H1)'
Utilizing the computed statistic, we can also establish confidence limits
around the predicted emission reduction at each time point.
77
-------
a < x < b
where a = lower confidence
b = upper confidence
x = XB - XT
limit
limit
If two-sided 90% confidence limits are placed on the distribution then
it can be stated that there is a 90% chance that the predicted emission
reductions will fall within these limits. If the upper and lower tails
of the distribution curves are similar it can also be stated that
there 1S a 95% change that the emission reductions will exceed the value
a.
The figure of merit can now be recomputed using the statistically
significant emission reductions. Once this has been accomplished, the
various inspection/maintenance procedures can be reordered based on the
revised figure of merit. The reordered set can then be compared to the
init1al set to determine whether the relative attractiveness of the candi-
date procedures has been altered.
Presently, the statistical inference model (Figure B-7) has been in-
corporated into the economic effectiveness processor. Because of the
lack of necessary input data, e.g., deterioration and maintenance un-
certainties, the model will not reproduce the actual emission specie
distributions. In the interim, the experimental data developed from the
fleet deterioration test has been used to estimate confidence limits on
the predicted emission reductions (see Section 2.1.4).
B.5 Linear Programming Algorithm
The determination or an optimum set of engine parameter or mode
emission pass/fail criteria for use in analyzing various inspection/
maintenance procedures represents a complex problem. Presently, the
updated economic effectiveness processor contains a total of 30 unique
engine parameter settings (i.e., 10 parameters x 3 control types) and
12 unique mode emission distributions (i.e., 4 modes x 3 control types).
78
-------
rr:EM1~51~~ S:;;:TURE) ---- "DiScRiMINANT ;ODE' l
I INS EC QNONlY I ~MI~StON SIGNATURE.' I
I DISTRIBUTION OF 1
I '.....~-!lc.l[£...S!!!lfl.EU_/ I
I re;;S$;Q'N ~~4;;ETERl I
L-~~~~::_J
I ("'coRRfiAr;o~\ r:----:i I
EIIROR -__--IGENE~TE J-.--- CONVOLUTION
I \ I TABLE J I
..._~~o_-_/ L___- ---.J
L...---------- 1_-
ERROR SOURCES:
. DECAY PARAMETER
. MEASUREMENT
. MAINTENANCE
. CORRELATION !PA.RAMETEII, SPECIE!
. INSPECTION
SYMBOL KEY
9.5% CONFIDENCE THAT 11M PROGRAM
Ri:OUCES EMISSIONS 8Y AT LEAST X
I FOR EACH SPECIE
Ho. "'RASE - IIYEST < X AND WEIGHTED
TOTAL SUM
I OPERATION ~
-DIRECT PARAMETER
INSPECTION
----EMISSION SIGNA.TURE
INSPECTION
~"A..05
PIlOBABILITY
DENSITY BASE - TEST
DIFFERENCE EMISSION SPECIE
.. IGRAMS/MllE)
o X'
"8-"'1
If X" z X, HYPOTH!:SIS 15 REJECTEO AND X
IS THE EMISSION HDUCTION
0NPUT OUTPUT)
(CON\OLUTlON)
FIGURE B-7
Statistical
Analysis Model
Flow Chart
79
-------
It is the pass/fail criterion which ultimately determines the rate of vehicle
rejection and thus, to a considerable degree, the cost and improvement in
emissions achieved by each candidate inspection/maintenance strategy.
Determination of the optimum parameter or emission mode pass/fail criteria
is therefore critical to both the basic system design and the relative
merit of a given procedure. Presented in the following paragraphs is a
brief overview of the approach used in determining optimal criteria.
Linear programming offers a useful, although approximate solution to
selecting optimum pass/fail criteria. A computational subroutine, there-
fore, was devised to estimate locally optimal cutpoints by means of a
linear programming algorithm. The relationship between the objective
function of the linear program and its independent variables, Xj' used
in this application, i.e., engine parameter adjustments, is given below.
I: I:
Z = i j
ae.
ap~ Wi'(R/, Pj)
J
ae.
ap' (influence coefficient), change in emission of ith specis per unit
j change in jth parameter setting
Wi = relative weight assigned to ith emission specie
R.llP. = optimum average value of parameter "j" relative to the total
J J vehicle population.
The objective function, Z, is the weighted reduction in
achieved by mandatory maintenance and the problem is to
values of RjllPj which maximize Z.
The maximization may be subject, however, to two basic classes of
inequality constraints:
. The emission reduction for each specie may be constrained
to exceed a minimum level, otherwise, the procedure is of
no interest. This requirement may be stated by the in-
equality constraint
I: ae
. aP (R.llP.) > b.
J j J J - ,
where bi = threshold reduction for ith emission specie
emissions to be
identify specific
80
-------
. The average cost per vehicle per inspection may be
constrained to not exceed a given value
~ d. R. < g
J J J-
where g = maximum allowable cost per vehicle
d. = a prorated cost per unit adjustment of the jth parameter
J
The ubi's" are normally expressed in terms of some minimum percentage
emission reduction desired for each exhaust pollutant, e.g., 20 percent CO
reduction for an extensive maintenance procedure. These values are exactly
analogous to the emission reductions program goals employed in earlier
studies. The cost constraint equation is used to relate the cost of adjust-
ing engine parameters with total system costs. The "g" coefficient
represents an assigned program cost for a given inspection/maintenance
procedure that cannot be exceeded.
The above analytical approach yields estimates of the optimal average
adjustment, Xj' for each parameter. These initial values can then be
employed in an interactive process to estimate values of the optimal cut-
points of each parameter for each inspection/maintenance process. This
scheme is shown in Figure 8-8.
The first step computes a set of mean values for the engine parameter
settings for the rejected vehicle population. This is accomplished by
adding the 6P. values obtained from the linear programming algorithm to
J
the original engine parameter reference specifications. Using the basic
engine parameter distributions, a systematic search is made to find the
pass/fail criteria which will yield the calculated mean values. In this
manner the linear programming algorithm provides a mechanism for determin-
ing pass/fail criteria for multiparameter inspections. The same basic
technique is also used in deriving mode emission cutpoints for the emission
inspection strategy.
81
-------
Linear programming
model
maximum rWi~ei
Emission response co-
efficients, aei/aP.
Estimate mean population parameter
changes required for emission
reductions, ~P.Rj
Iteratively solve for
optimal pass/fail criteria
00
= f
c'P'j
~P. R.
J J
p (P. ) dP .
J J
PARAMETER DISTRIBUTION ,P/Pj
FIGURE B-8
Linear Programming Interactive Scheme
82
-------
6.6 Economic-Effectiveness Computer Processor
To accommodate the program developments discussed herein, a substantial
modification of the Phase One Economic-Effectiveness Computer Processor was
required. A schematic diagram of the revised processor is shown in Figure
6-9. The processor is used to simulate the performance of various vehicle
inspection/maintenance procedures over a given time interval. The key in-
put data required by the processor include:
. strategy selection
. regional identification
. system pass/fail criteria
The first two inputs are specified based on the inspection/maintenance
procedure selected for study and the urban area of interest. The system
pass/fail criteria are then estimated utilizing the linear programming
algorithm discussed in the previous section. Having identified the type
of strategy to be examined, e.g., idle parameter inspection, the economic-
effectiveness (E/E) model is used in a simulation mode to predict resultant
vehicle emission levels at the end of each year and to estimate the cost of
the program. In the case of a state lane inspection, it ;s also used to
design and size the basic inspection system. In addition to vehicle emis-
sion levels and program costs, the E/E model computes a figure of merit for
each inspection/maintenance strategy. This figure of merit is used to
rank the overall attractiveness of those inspection/maintenance strategies
under study.
Another mechanism used for screening inspection/maintenance strategies
is to perform a statistical analysis on the projected emission levels.
This operation employs the new statistical confidence model. The emission
levels predicted for the test fleet are contrasted statistically with those
of the baseline fleet. Figure 6-9 depicts this statistical comparison.
A final product of the statistical analysis is the development of confidence
limits on all predicted emission time histories.
83
-------
FIGURE 8-9
Economic-Effectiveness Program Design
CX)
.j:o,
r---------------------------l
I POLICY INPUT DATA I
I 0 I/M PROCEDURES BASELINE ~ B' a B I
I 0 FINANCIAL METHO D I
I I
I I
I ECONOMIC-EFFECTIVENESS STATISTICAL I
PASS/FAIL MODEL CONFIDENCE MODEL 0 SELECTION OF I
I CRITERIA 0 EMISSION PREDICTION 0 NULL HYPOTHESIS OPTIMUM PROCEDURES
I USING L. P. 0 COST ESTIMATION jJ. B - I.l T < X 0 CONFIDENCE LIMITS I
I 0 SYSTEM DESIGN ON EMISSIONS I
AT 95%
I I
I I
I REGIONAL I
CHARACTERIZATION
I I
L_-----------______ECONOMIC-EFFECTIVENESSPROCESSOU
-------
REFERENCES
1.
liThe Economic Effec;tiveness of Mandatory Engine Maintenance for Reducing
Vehicle Exhaust Emissions,1I Vol. II, IIModeling of Inspection/Maintenance
Systems," TRW Systems Report in Support of CRC APRAC Project No. CAPE-13-
68, 1971.
2.
liThe Economic Effectiveness of Mandatory Engine Maintenance for Reducing
Vehicle Exhaust Emissions," Vol. III, "Procedures Development," TRW
Systems Report in Support of CRC APRAC Project No. CAPE-13-68, 1971.
3. A. J. Hocker, "Surveillance of Motor Vehicle Emissions in California, II
Quarterly Progress Report No. 23, California Air Resources Board.
4. W. F. McMichael and A. H. Rose, "A Comparison of Automotive Emissions
in Cities at Low and High Altitudes," June 1965.
5.
IIProfile of Air Pollution Control in Los Angeles County," Air PoTlution
Control Distrtct, Los Angeles County, January 1969.
--New York State Motor Vehicle Emission Program Final Report,1I prepared
by Scott Research Laboratories, Inc., 24 August 1968.
6.
7. IIAutomotive News A1manac,1I Thirty-Fifth Reference Edition, Dana Corpor-
ation, 26 April 1971.
8. L. J. Bintz, "Automotive Fleet Emission Program," Automobile Club of
Southern California Report, June 15, 1968.
9. Vehicle Emissions Surveillance Study, TRW Systems Report in Support of CRC
APRAC Project Number CAPE-l3-68, 20 August 1970.
10. G. D. Ebersole and G. E. Holman, Lubricant-Closed PCV System Relation-
ships Influence Exhaust Emissions, SAE Paper 680113, presented in
Detroit, Michigan, January 1968.
85
------- |