EVALUATION OF MECHANISMS OF EXHAUST INTRUSION INTO SCHOOL
BUSES AND FEASIBLE MITIGATION MEASURES
FINAL REPORT
Prepared for the California Air Resources Board
Contract No. 03- 343
Principal Investigator
Dennis R. Fitz
College of Engineering
Center for Environmental Research and Technology
University of California
Riverside, CA 92521
Co- Principal Investigator
Arthur M. Winer, Ph. D.
Environmental Health Sciences Department
Environmental Science and Engineering Program
School of Public Health
University of California
Los Angeles, CA 90095
Participating Researchers
Kathleen Kozawa, Eduardo Behrentz
Environmental Health Sciences Department
Environmental Science and Engineering Program
School of Public Health
University of California
Los Angeles, CA 90095
David Pankratz, David Gemmill
College of Engineering
Center for Environmental Research and Technology
University of California
Riverside, CA 92521
January 11, 2006
i
DISCLAIMER
The statement and conclusions in the Report are those of the contractor and not necessarily those
of the California Air Resources Board. The mention of commercial products, their source, or
their use in connection with material reported herein is not to be construed as actual or implied
endorsement of such products.
ii
ACKNOWLEDGEMENTS
We thank Allen Weber at Savanna River National Laboratory, Professor William Hinds
at UCLA, and Professor Ring Carde at UCR for the use of their instruments on- board the school
buses.
We gratefully acknowledge the staff at Hemet Unified School District for all of their help
and the use of their buses.
We wish to acknowledge Kurt Bumiller and David Cocker and UCR students James
Bristo, Sally Pederson, Adrian Afan, and Cameron Switzer, for their contributions to this
research.
We especially appreciated Cheryl Benson, Lynda Owens- Wolfe, Delvin Lucas, and
Walter “ Bud” Morris of the UCR Transportation Department and Henry Coleman of the UCLA
Transportation Department for their professional driving of the school buses and support of the
project team. We also thank Lance Danks at the UCR Transportation Department for helping to
arrange drivers and a bus for the study.
We gratefully acknowledge support for this research by the California Air Resources
Board. This report was submitted in partial fulfillment of ARB Contract No. 03- 343,
“ Evaluation of Exhaust Intrusion in School Buses and Feasible Mitigation Measures,” by the
University of California, Riverside, College of Engineering, Center for Environmental Research
and Technology and the University of California, Los Angeles, School of Public Health,
Environmental Health Sciences Department, under the sponsorship of the California Air
Resources Board. Work was completed as of June 2006.
iii
TABLE OF CONTENTS
DISCLAIMER..................................................................................................................... .......... i
ACKNOWLEDGEMENTS ......................................................................................................... ii
TABLE OF CONTENTS ............................................................................................................ iii
LIST OF FIGURES ..................................................................................................................... vi
LIST OF TABLES....................................................................................................................... ix
LIST OF PHOTOGRAPHS........................................................................................................ xi
ABSTRACT....................................................................................................................... ......... xii
1.0 EXECUTIVE SUMMARY .................................................................................................... 1
2.0 INTRODUCTION AND BACKGROUND........................................................................... 3
2.1 Introduction................................................................................................................... ...... 3
2.2 Background .......................................................................................................................... 3
2.3 Statement of Problem........................................................................................................... 4
2.4 Previous Vehicle Exhaust Intrusion Studies ........................................................................ 4
2.4.1 Mechanisms of Exhaust Intrusion Studies ................................................................... 5
2.4.2 Other Related Exhaust Intrusion Studies ..................................................................... 6
2.4.2.1 Ventilation Air Flow Patterns Inside Vehicles..................................................... 6
2.4.2.2 Air Exchange Rate Studies ................................................................................... 7
2.5 Objectives..................................................................................................................... ....... 8
2.5.1 Overall Objectives........................................................................................................ 8
2.5.2 Specific Objectives....................................................................................................... 8
2.5.2.1 Pilot Study ............................................................................................................ 8
2.5.2.2 Main Study ........................................................................................................... 8
3.0 PILOT STUDY FINDINGS AND RECOMMENDATIONS ............................................. 9
3.1 Introduction................................................................................................................... ...... 9
3.2 Summary of Pilot Study Findings ........................................................................................ 9
3.2.1 Evaluation of Exhaust System Leaks ........................................................................... 9
3.2.2 Evaluation of Leak Points in the Passenger Cabin and Exhaust Intrusion................... 9
3.2.3 SF6 Tracer Gas Release System................................................................................. 11
3.2.4 Evaluation of Leader Vehicle Exhaust Intrusion ....................................................... 12
3.2.5 Evaluation of Proposed Mitigation Strategies............................................................ 12
3.3 Modifications of Experimental Design for Main Study..................................................... 13
4.0 EXPERIMENTAL METHODS AND STUDY DESIGN.................................................. 14
4.1 Introduction................................................................................................................... .... 14
4.1.1 Vehicle Selection........................................................................................................ 14
4.1.2 Fuel Used in the Test Buses ....................................................................................... 14
4.1.3 Characterization and Justification of Selection of Test Routes.................................. 14
4.1.3.1 Route 1................................................................................................................ 15
4.1.3.2 Route 2................................................................................................................ 16
4.2 Field Sampling Procedures ................................................................................................ 16
4.2.1 Instrument Packaging and Supply.............................................................................. 16
4.2.2 Instrumentation .......................................................................................................... 17
4.2.2.1 SF6 Measurements .............................................................................................. 17
4.2.2.2 Real- Time Particle Phase PAH Measurements .................................................. 17
4.2.2.3 Condensation Particle Counts ( CPC) ................................................................. 18
4.2.2.4 Bus Location....................................................................................................... 18
iv
4.2.2.5 Engine Operating Parameters ............................................................................. 18
4.2.2.6 Propene Measurements....................................................................................... 18
4.2.2.7 Meteorological Measurements ........................................................................... 18
4.2.2.8 Video Camera..................................................................................................... 18
4.2.2.9 Tracer Release Control ....................................................................................... 19
4.2.3 Data Collection........................................................................................................... 19
4.2.4 Experimental Design .................................................................................................. 19
4.2.4.1 Evaluation of Exhaust Train Leaks .................................................................... 19
4.2.4.3 Evaluation of Tailpipe Exhaust Intrusion ( Self- Pollution)................................. 21
4.2.4.3.1 Tracer Gas Release System ........................................................................ 21
4.2.4.3.2 Self- Pollution Runs..................................................................................... 21
4.2.4.4 Evaluation of Exhaust Intrusion from a Leader Vehicle .................................... 25
4.3 Baseline Tracer Gas Measurements ................................................................................... 27
4.4 Data Analysis Methods ...................................................................................................... 27
5.0 RESULTS AND DISCUSSION........................................................................................... 28
5.1 Tracer Gas Release System................................................................................................ 30
5.1.1 Validation ................................................................................................................... 30
5.1.1.1 Pearson’s Correlation Coefficient ...................................................................... 31
5.1.1.2 Tracer Gas Concentration in Exhaust................................................................. 33
5.2 Bus Cabin Leak Potential................................................................................................... 36
5.2.1 Evaluation of Overall Bus Cabin Leak Potential ....................................................... 36
5.2.1.1 Rapid Evaluation of Overall Cabin Leak Rate ................................................... 36
5.2.2 Quantification............................................................................................................. 37
5.3 Exhaust Leaks .................................................................................................................... 37
5.4 Mitigation Measures........................................................................................................... 39
5.4.1 Effect of High Exhaust Position When Driven on the Test Route............................. 39
5.4.1.1 Effect of High Exhaust Position on Self- Pollution When Bus in Motion.......... 39
5.4.1.2 Effect of High Exhaust Position in Leader Bus on Follower Bus ...................... 44
5.4.2 Effect of Power Ventilation ( Blower) When Bus in Motion ..................................... 46
5.4.2.1 Effect of Power Ventilation on Self- Pollution when Bus in Motion, Low
Exhaust Location ............................................................................................................ 46
5.4.2.2 Effect of Power Ventilation in the Test Bus while Following a Leader Bus with
Low Exhaust ................................................................................................................... 48
5.4.3 Combination of High Exhaust and Power Ventilation ( Blower) When Driven on the
Test Route ........................................................................................................................... 49
5.4.3.1 Effect of Combined High Exhaust and Power Ventilation on Self- Pollution.... 49
5.4.3.2 Effect of Combined High Exhaust in Leader Bus and Power Ventilation on
Follower Bus .................................................................................................................. 52
5.6 Stationary Runs .................................................................................................................. 57
5.6.1 Self- Pollution during Stationary Runs ....................................................................... 60
5.7 Comparison to Previous School Bus Study ....................................................................... 69
5.7.1. Comparison of Tracer Gas Release Systems and Tracer Gas Concentrations in
Exhaust ............................................................................................................................... 69
5.7.2. Comparison of Buses and Routes.............................................................................. 70
5.7.3 Effects due to Differences in Run Duration ............................................................... 70
5.7.4 Summary of Comparisons between Studies............................................................... 71
v
6.0 CONCLUSIONS AND RECOMMENDATIONS.............................................................. 73
6.1 Exhaust Leak Potential and Bus Cabin Leak Potential...................................................... 73
6.2 Tracer Gas Release System................................................................................................ 73
6.3 Mitigation Strategies .......................................................................................................... 73
6.3.1 Mobile Runs ............................................................................................................... 73
6.3.2 Stationary Runs .......................................................................................................... 74
6.4 Comparison to Previous Bus Study ................................................................................... 74
6.5 Recommendations .............................................................................................................. 74
7.0 RECOMMENDATIONS FOR FUTURE RESEARCH.................................................... 75
8.0 REFERENCES..................................................................................................................... 78
9.0 INVENTIONS REPORTED AND COPYRIGHTED MATERIALS PRODUCED...... 81
10.0 GLOSSARY OF TERMS, ABBREVIATIONS, AND SYMBOLS................................ 82
vi
LIST OF FIGURES
Figure No. Title Page
3.2.3.1 Tracer gas release system used in main study, using two tracer gases. The system
was controlled by engine intake flow. .................................................................. 11
4.1.3.1.1 Map of Route 1 in Riverside, California............................................................... 15
4.1.3.2.1 Map of Route 2 in Downtown Riverside, California............................................ 16
5.1.1.1.1 Pearson’s correlation coefficient for each day of mobile testing in the main study
............................................................................................................................... 31
5.1.1.1.2 Scatter plots of exhaust flow versus SF6 flow rate for 0427/ Run 44. ................... 32
5.1.1.1.3 Scatter plots of exhaust flow versus propene flow rate for 0427/ Run 44.. ........... 32
5.1.1.2.1( a, b) Time series of one- minute medians of SF6 ( 10 Hz release system) and propene ( 4
Hz release system) exhaust concentrations for 0405( a) and 0412( b).. ............... 325
5.2.1.1.1 Survey of bus cabin pressures using the “ blower door” method. Note the first two
digits of the bus number correspond to the model year. Black bars represent
buses used in the main study............................................................................... 326
5.3.1 Backpressure measurements in psi for the six engine types tested in the exhaust
leak experiment ( n is the number of buses tested for each engine type). Standard
deviations are provided for the Cummins 250 and Cat 3208 engines.. ................ 38
5.4.1.1.1( a) Time series of percent self- pollution for SF6 and propene during mobile runs
conducted on 0427 .............................................................................................. 410
5.4.1.1.1( b) Time series of percent self- pollution for SF6 and propene during mobile runs
conducted on 0504. ............................................................................................... 41
5.4.1.1.1( c) Time series of percent self- pollution for SF6 and propene during mobile runs
conducted on 0510 .............................................................................................. 421
5.4.1.1.1( d) Time series of percent self- pollution for SF6 and propene during mobile runs
conducted on 0511 ................................................................................................ 42
5.4.1.1.2 Time series for SF6 and propene concentrations on 0406. The bus was not flushed
of tracer gas for these runs .................................................................................... 43
5.4.1.2.1 Time series for percent exhaust intrusion by SF6 and propene on 0412 ( leader
exhaust test) .......................................................................................................... 45
vii
5.4.1.2.2 Time series for percent exhaust intrusion by SF6 and propene on 0419 ( leader
exhaust test). ......................................................................................................... 47
5.4.2.1.1 Time series for percent self- pollution by propene on 0427. The boxed areas
represent runs where power ventilation was tested............................................... 47
5.4.2.2.1 Time series for percent exhaust intrusion by SF6 on 0419 during a leader exhaust
test. The boxed areas represent runs when power ventilation was tested............ 48
5.4.3.1.1 Time series of percent self- pollution for SF6 and propene on 0510 examining the
combination of exhaust position and blower operation as a mitigation measure. 51
5.4.3.1.2 Time series of percent self- pollution for SF6 and propene on 0503 examining the
combination of exhaust position and blower operation as a mitigation measure. 51
5.4.3.2.1 Time series of percent exhaust intrusion for SF6 and propene during a leader
exhaust run conducted on 0412............................................................................. 52
5.4.4.1 Average propene tracer gas concentrations ( by exhaust position) in the bus cabin
for windows sealed ( 0407) versus unsealed ( 0406).............................................. 54
5.4.4.2 Time series for SF6 and propene concentrations during window seal test on 0407
............................................................................................................................... 54
5.5.1 Time series of percent self- pollution and exhaust intrusion for SF6 and propene,
respectively, during a leader exhaust- follower exhaust test on 0413 ................... 56
5.5.2 Time series of percent self- pollution and exhaust intrusion for propene during a
leader exhaust- follower exhaust test on 0420....................................................... 56
5.6.1 School bus orientation in relation to wind direction for stationary self- pollution
runs. 58
5.6.2 School bus orientation for leader exhaust testing. Both tracer gases are released
from the leader bus from a high exhaust position and low exhaust position. 58
5.6.1.1( a, b) Time series for SF6 and propene during stationary self- pollution testing on 0426
( a) and 0504 ( b)..................................................................................................... 61
5.6.1.1( c) Time series for SF6 and propene during stationary self- pollution testing on 0510
…………............................................................................................................... 62
5.6.2.1 Time series for SF6 and propene during stationary exhaust intrusion runs
conducted on 0413.. .............................................................................................. 66
viii
5.6.2.2 Time series for SF6 and propene during stationary exhaust intrusion runs on
0420....................................................................................................................... 66
5.7.1 SF6 tracer gas exhaust concentrations ( 1- minute medians) for Bus 982/ TO1 on run
0407 calculated using measured SF6 and exhaust flow rates versus those
calculated with constant 2 lpm SF6 flow rate and measured exhaust flow rate.. . 70
5.7.2 Time series for 0406 run in current study and average self- pollution value
( 0.0275%) from our previous study.. .................................................................... 71
7.1 MapPoint locations where the speed was less than 10 dm/ sec along Route 1
during a test run conducted on 4- 20- 05. 76
7.2 MapPoint locations of where the SF6 was greater than 750 ppt along Route 1
during a test run conducted on 4- 20- 05. 77
7.3 MapPoint locations of SF6 concentrations and speed along Route 1 during a test
run conducted on 4- 20- 05. 77
ix
LIST OF TABLES
Table No. Title Page
1.1 Beneficial or negative effects of different mitigation methods under different run
conditions................................................................................................................ 2
2.4.2.2.1 Results from ventilation test conducted in our previous study ( Fitz et al., 2003) for
selected buses.......................................................................................................... 8
3.2.4.1 SF 6 data ( in ppt) for pilot study leader/ follower test, evaluating effect of window
position and exhaust position in a leader vehicle….............................................. 12
4.2.2.1 Measurement methods utilized in the main study................................................. 17
4.2.4.3.2.1 Flow, as measured by pressure in inches of H 2 O for split exhaust hardware
including ratios for high versus low exhaust flow for each test bus ..................... 23
5.1.1 Characteristics of the test buses. ........................................................................... 28
5.1.2 Description of all mobile runs ( conducted in 2005). ............................................ 29
5.1.3 Meteorological data during mobile runs ( with standard deviations) .................... 30
5.1.1.2.1 Variation of SF6 and propene exhaust concentrations for all mobile runs in the
main study. SP= Self- Pollution, LE= Leader Exhaust, LE- FE= Leader Exhaust-
Follower Exhaust. ................................................................................................. 33
5.4.1.1.1 Percent self- pollution for individual runs examining the effect of the high exhaust
position.................................................................................................................. 42
5.4.2.1.1 Effect of power ventilation ( tracer gases released from low exhaust position) on
self- pollution......................................................................................................... 47
5.4.3.1.1 Percent self- pollution for individual runs examining the effect of the high exhaust
position and blower operation mode..................................................................... 50
5.4.3.1.2 Percent self- pollution for individual runs examining the effect of high versus low
exhaust position when the blower was in operation ............................................ 50
5.6.1 Description of all stationary runs ( conducted in 2005)......................................... 59
5.6.2 Meteorological data during stationary runs ( with standard deviations)................ 60
5.6.1.1 Average percent self- pollution ( relative to low exhaust) and changes in self-
x
pollution for individual stationary runs examining the effect of high versus low
exhaust position. ................................................................................................... 62
5.6.1.2 Average percent self- pollution ( relative to blower off) and changes in self
pollution for individual stationary runs examining the effect of blower operation.
....................................................................................................................... 63
5.6.1.3( a) Average percent self- pollution ( relative to low exhaust) for individual stationary
runs examining the effect of high versus low exhaust position when the blower
was in operation. ................................................................................................... 63
5.6.1.3( b) Average percent self- pollution ( relative to blower off) and changes in self-pollution
for individual stationary runs examining the effect of blower operation
when the exhaust position was high...................................................................... 63
5.6.2.1 Average percent exhaust intrusion ( relative to low exhaust) and percent change in
exhaust intrusion from a leader bus measured in the follower bus during stationary
leader exhaust runs examining the effect of high versus low exhaust position on a
leader bus. ............................................................................................................. 67
5.6.2.2 Average percent exhaust intrusion ( relative to blower off) and percent change in
exhaust intrusion from a leader bus measured in the follower bus during stationary
leader exhaust runs examining the effectiveness of using the blower in the
follower bus in preventing exhaust intrusion from the low exhaust position of the
leader bus. ............................................................................................................. 67
5.6.2.3( a) Average percent exhaust intrusion ( relative to low exhaust) and percent change in
exhaust intrusion from a leader bus measured in the follower bus during stationary
leader exhaust runs examining the effectiveness of high versus low exhaust
position on the leader bus in preventing exhaust intrusion in the follower bus
while operating the blower in the follower bus. ................................................... 67
5.6.2.3( b) Average percent exhaust intrusion ( relative to blower off) and change in percent
exhaust intrusion from a leader bus measured in the follower bus during stationary
leader exhaust runs examining the effectiveness of blower operation in the
follower bus and high exhaust position on the leader bus in preventing exhaust
intrusion on the follower bus. ............................................................................... 68
xi
LIST OF PHOTOGRAPHS
Photograph No. Title Page
3.2.2.1 A leak at the bottom of a bus door............................................................ 10
4.2.4.1.1 Backpressure method to evaluate exhaust system leaks using a silicon
stopper and magnehelic............................................................................. 20
4.2.4.2.1 Sealing passenger windows with plastic sheeting to assess contribution of
window leaks to overall leaks on the bus.................................................. 20
4.2.4.3.1.1( a, b) Tracer Gas Release System: ( a) toggle switch to reverse tracer gas release
position and ( b) mass flow controllers to control tracer gas release into the
tailpipe....................................................................................................... 22
4.2.4.3.2.1 Split exhaust configuration for self- pollution experiments. One tracer gas
was released from each exhaust branch.................................................... 22
4.2.4.3.2.2( a- d) Blower inlet positions on Bus 982 ( a- b) and Bus 872 ( c- d) ..................... 24
4.2.4.4.1 Exhaust configuration for leader vehicle during leader exhaust tests to
evaluate impact of following a bus with high versus low exhaust ........... 25
4.2.4.4.2( a, b) Exhaust configurations for follower bus ( a) and leader bus ( b) during
leader exhaust/ follower exhaust tests to assess impact of exhaust intrusion
from a leader vehicle versus self- pollution............................................... 26
xii
ABSTRACT
“ Self- pollution,” the intrusion of a bus’s own exhaust into the bus cabin, leads under
some conditions to very high exposures. This study attempted to elucidate how and where self-pollution
occurs, and to test various methods to mitigate this phenomenon. The mechanism of
self- pollution was investigated by evaluating the magnitude of exhaust system leaks, searching
for exhaust entry points using a tracer gas, and determining the overall leak rate of the bus cabin.
Comprehensive detection of leaks in the exhaust system using SO2 from the exhaust as a tracer
gas and a survey of leak potential using back pressure measurements showed that exhaust system
leaks in a well- maintained system were insignificant. However, identifying specific exhaust
entry points into the passenger compartment using tracer gas was found to be infeasible due to
the large number of potential entry points. To quantify overall air tightness of cabins, the leak
rate of 17 buses was evaluated by pressurizing them with an air blower with a constant flow rate
and measuring the pressure differential between the inside and outside of the bus (“ blower door
method”). Pressure differentials ranged over a factor of five, but in general, newer buses showed
lower leak rates.
The primary self- pollution mitigation methods evaluated consisted of elevating the
exhaust outlet, power ventilating the cabin, or a combination of the two methods. Because
following other buses is also a major source of high bus cabin concentrations, these methods
were evaluated for their efficacy in reducing not only self- pollution but also pollution from a
leader bus. Comparisons were made both in stationary mode and while driving a prescribed
route, using four test buses representative of the current in- use school bus fleet. Exhaust
intrusion into the cabin was measured using a dual tracer gas approach to allow for a direct
comparison between the mitigated and unmitigated scenarios. Two separate, non- interfering
tracer gases were metered into the exhaust in proportion to engine intake flow rates to maintain
near- constant tracer gas concentrations in the exhaust. Real- time analyzers were used to monitor
the concentration of each tracer gas inside the cabin of the test bus. The concentration data were
used to calculate the volumetric fraction of air inside the bus that originated from each tracer-labeled
exhaust.
Evaluation of the high- exhaust mitigation strategy used a split exhaust ( half of the flow
released above the roof and half released at the normal low position) with a separate tracer gas
metered into each half. When evaluating exhaust intrusion from a leader bus, both tracers were
similarly released on a leader bus while measurements were taken on a follower bus. A second
set of leader- follower experiments involved metering one tracer gas in the leader bus exhaust and
metering the other tracer gas in the follower bus exhaust. This allowed comparing the magnitude
of self- pollution versus exhaust intrusion from a leader vehicle. The effects of power ventilation
were evaluated by comparing the above test outcomes with the blower on versus off. While
results showed the blower reduced the exposure to self- pollution and leader- pollution most of the
time, occasionally exhaust plumes reached the blower inlet at low speeds or during idling,
causing high peak concentrations that largely negated the benefits of the power ventilation.
Using an elevated exhaust outlet significantly reduced the exposure due to self- pollution, but
resulted in only modest reductions in leader- vehicle pollution. Our overall recommendations are
to employ elevated exhaust outlets on school buses and to minimize exposure to leader vehicle
exhaust by avoiding close caravanning of diesel school buses.
1
1.0 EXECUTIVE SUMMARY
Background: Previous studies have shown “ self- pollution” of school bus cabins is a
significant source of pollutant exposure and the pollution from a leading diesel vehicle leads to even
greater exposure. The objective of this study was to identify and evaluate reasonable feasible
mitigation measures to reduce the exposure in the school bus micro- environment. The measures
evaluated included the repair of exhaust system leaks, better sealing of the bus cabin, power
ventilation of the bus cabin, raising the exhaust release point, and a combination of the last two
methods.
Methods: Leaks in the exhaust system itself were evaluated by two approaches: probing with
the inlet of a real- time sulfur dioxide detector and by inducing and measuring backpressure in the
exhaust system. To identify tailpipe exhaust entry points in the bus cabin, SF6 tracer gas was
metered into the exhaust and a real- time SF6 analyzer was used to probe for entry points. The
overall cabin leak rate was evaluated by the blower door approach: pressurizing the bus with a
blower with a constant flow rate and measuring the pressure differential between the inside and
outside of the test bus. To measure self- pollution and leader- pollution, SF6 and/ or propene tracer gas
was added to the exhaust of one or both vehicles, in both stationary and mobile modes. The follower
or self- pollution test bus was equipped with real- time SF6 and propene analyzers. The concentration
data were used to calculate the volumetric fraction of air inside the bus that originated from the
tracer- labeled exhaust ( percent intrusion), which is the ratio between the concentration of tracer gas
in the bus cabin and the concentration of tracer gas in the exhaust. The effectiveness of the raised
exhaust position to mitigate pollutant intrusion was determined by adding exhaust piping to split the
flow evenly between the normal bumper position outlet and a position above the bus body. SF6 was
metered to one path while propene was metered to the other. This approach allowed continuous
comparison under identical conditions. The effectiveness of power ventilation to mitigate pollutant
intrusion was determined by alternating tests with the blower on or off.
Results- Pilot Study: A pilot study utilizing a single instrumented test bus was conducted to
demonstrate the study design feasibility. A real- time sulfur analyzer was used to probe for exhaust
leaks on a single older bus ( 1985 Thomas Coach) using a sulfur- enhanced fuel. No significant leaks
were found and the method was found to be impractical for testing a large number of buses, since
sulfur needed to be added to the fuel. Probing the exterior of the bus while pressurizing the cabin
using a blower whose output was dosed with propene indicated leaks were present throughout the
bus. The blower output of 34 m3/ min resulted in a pressure differential of 0.18 inches of water
column, indicating widespread leakage. Individual leaks could not be pinpointed by adding tracer
gas to the exhaust and probing the inside of the cabin due to ubiquitous leak locations which resulted
in elevated concentrations throughout the cabin. Some leaks, however, allowed tracer- free ambient
air to enter the cabin.
A tracer gas release system that varied the flow of tracer gas in relationship to the engine’s
air processing flow rate to maintain a constant concentration in the exhaust was designed, built and
evaluated. The initial leader- follower tests showed little difference between the high and low
exhaust release points on the leader vehicle. We concluded a split exhaust system using separate
tracers injected at each position was needed so that results could be compared directly.
The combined results of the pilot study suggested that the main study should focus on the
mitigation measures of raising the exhaust and ventilating the cabin.
Main Study: Seventeen buses were screened for exhaust leaks using the backpressure
2
approach and for cabin “ tightness” using the “ blower door” approach, with a squirrel- cage blower
mounted on the bus door opening to pressurize the bus. Using engine backpressure to evaluate
exhaust leakage, leakage rates appeared to be dependent on the engine make and model. Six
different engine types were employed. We concluded lower pressures within an engine make and
model may be indicative of a leak, although physical examination did not indicate any buses had
substantial exhausts leaks. The pressure drop for the “ blower door” tightness test ranged from 0.04”
to 0.25” of water column. Newer buses were generally tighter.
The mitigation measures of raising the exhaust outlet and power ventilating the cabin were
evaluated using four different instrumented test buses covering a range of manufacturers and model
years to be representative of those most commonly used in California. The four buses chosen for
testing were a 1987 Blue Bird, a 1993 Carpenter SPT- 3908, a 1998 Thomas Saf- T- Liner and a 2002
Thomas Saf- T- Liner. A total of 54 mobile test runs and 32 stationary ( with the bus’s exhaust
pointed into the wind) test runs were conducted. Table 1.1 summarizes the results for self- pollution
and leader exhaust intrusion when the bus was on a test route ( mobile) and when stationary. The
high exhaust release location consistently reduced the amount of self- pollution. Using power
ventilation gave less consistent results, and at times it appeared the test bus’s own exhaust was
pulled into the blower inlet with relatively little dilution. This was particularly noticeable when the
exhaust was discharged in the high position. Similar results, although of a more qualitative nature,
were obtained for tracer added to the exhaust of a leading vehicle. Exhaust intrusion in the test
follower bus from leader vehicles was typically twice that of self- pollution on the test bus.
Table 1.1 Beneficial or negative effects of different mitigation methods under different run
conditions. “++” ( or “--“) indicates consistent and sizeable reductions ( or increases) in
exhaust intrusion; “+” ( or “-“) indicates frequent but less sizeable increases ( or
decreases); and “+/--” indicates mixed effects with sometimes large increases in exhaust
intrusion if “--” included.
MITIGATION METHOD Exhaust High Blower On
TEST CONDITION Blower Off Exhaust Low Exhaust High
RUN TYPE
Self- Pollution, Mobile ++/- ++ +/--
Self- Pollution, Stationary ++/- +/-- +/--
Leader- Follower, Mobile + +/- -
Leader- Follower, Stationary ++ +/-- --
Conclusions: The blower door method was an effective method to determine the overall
tightness of a bus and should be used as a diagnostic test to ensure tightness is maintained as buses
age. Bus exhaust system leaks in well- maintained buses were found to be insignificant. Children’s
exposure to exhaust, particularly from self- pollution could be significantly reduced by placing the
exhaust outlet above the bus. Exhaust from a leading vehicle can be more significant than self-pollution
and therefore close caravanning of school buses should be avoided, and buses should also
avoid following other diesel- powered vehicles closely, further reinforcing this recommendation
made in our previous school bus exposure study ( Fitz et al., 2003). Results from the current study,
however, are not directly comparable to our previous school bus study due to differences such as the
time of day in which the tests were conducted, bus types, and routes.
3
2.0 INTRODUCTION AND BACKGROUND
2.1 Introduction
Children’s health has been the focus of intense interest in California across all levels of
government, as well as in academic research, the advocacy community, and Federal health and
environmental agencies. California Senate Bill 25 ( Escutia 1999) required the California Air
Resources Board ( ARB) to identify areas where exposure of infants and children to air pollutants
were not adequately measured by the current fixed- site monitoring network and to conduct enhanced
monitoring. Among the greatest concerns has been growing evidence of the impacts of air pollution
on children’s respiratory function and other health indicators. Children are especially susceptible to
air pollution because of their high inhalation rates relative to body mass, high activity rates, greater
time spent outdoors, narrower lung airways, immature immune systems and rapid growth ( Lipsett,
1989; Pope, 1989; Phillips et al., 1991; Wiley et. al., 1991; U. S. EPA, 1996). The ARB has been
particularly concerned with exposures resulting from the amount of time children spend during
school bus commutes, since one million children are transported by public school buses each day
( California Department of Education, 2002). About 70% of the 26,000 school buses in California
remain powered by diesel engines ( Long, 2000), which emit exhaust particulate the ARB has
declared to be a Toxic Air Contaminant.
2.2 Background
Concern about this issue led the ARB to fund a recently- completed study by the present
research team, designed to characterize the range of children’s pollutant exposure during school bus
commutes ( Fitz et al., 2003; Sabin et al., 2005a, b). Following a pilot study to demonstrate
feasibility of the study design and measurement protocols, real- time and integrated measurements of
a wide range of gaseous and particulate pollutants were conducted while driving several distinct
school bus routes in Los Angeles with eight different school bus and fuel/ emission control
technology combinations. Across the pilot and main studies, three key microenvironments were
investigated: bus stops, loading/ unloading zones, and school bus interiors during commutes. It was
shown that children’s typical urban commute times were far more important as a determinant of
exposure than typical times spent in either school loading/ unloading zones or at bus stops ( Behrentz
et al., 2005).
The key variables affecting children’s exposure on school buses were identified ( Sabin et al.,
2005a, b) and included the degree of exhaust intrusion (“ self pollution”), window position, nearby
diesel vehicles ( especially other diesel school buses), and roadway type. Due to self pollution,
directly emitted, vehicle- related pollutants such as black carbon and particle- bound, polycyclic
aromatic hydrocarbons ( PAHs) were higher with windows closed than open. In addition, these same
pollutants were much higher on urban routes compared with a rural/ suburban route ( Sabin et al.,
2005a). Additional findings were that higher exposures to pollutants such as nitrogen dioxide ( NO2),
benzene, 1,3- butadiene and a range of aldehydes and ketones occurred during children’s commutes
than indicated by measurements at nearby central sites.
High commute exposures resulted in part from expected causes, including the high
concentrations of pollutants already present on roadways, especially in heavy traffic, and the direct
influence of other vehicles being followed. A critical and novel finding from this ARB- sponsored
study was that self- pollution could contribute as much to the high exposures that children experience
during school bus commutes as the surrounding traffic itself. This phenomenon of self- pollution
was unambiguously demonstrated through the use of a tracer gas, sulfur hexafluoride ( SF 6 ), injected
4
into the exhaust of each bus tested during commutes ( Behrentz et al., 2004).
In general, higher concentrations of diesel- related pollutants ( i. e., more than double) were
observed when the windows were closed, and older buses had greater intrusion of their own exhaust
into the cabin compared with newer buses. For conditions such as idling at bus stops with the wind
coming from the rear, we observed SF6 tracer gas outside at the front of the buses for all commutes
although the mean concentrations were much lower than the SF6 concentrations inside the bus.
The extent of self- pollution we identified through the tracer technique was dramatic: every
bus we tested exhibited some degree of self- pollution during every bus commute. Moreover, about
25% of the variance in black carbon within- cabin concentrations could be explained by intrusion of
the bus’s own exhaust. In a sample calculation for one of the “ representative” buses in our study,
approximately half of the mean black carbon concentration during the one- hour commute could be
accounted for by self- pollution.
Although this earlier investigation of children’s exposure in school bus commute- related
microenvironments appears to be the most definitive study of its kind, investigation of mitigation
measures for “ self- pollution” or “ leader/ follower” pollution was beyond the scope of that study.
Therefore an investigation of the various mechanisms of intrusion of exhaust into school bus cabins,
and feasible mitigation measures, was needed.
2.3 Statement of Problem
There are at least four possible mechanisms for exhaust gases to enter the cabin of school
buses. First, leaks from the engine’s compartment can enter into the cabin, possibly as the result of
leaks in the exhaust train. Second, the exhaust plume exiting from the tailpipe can travel from the
rear of the bus and enter the cabin through the windows ( if open, or perhaps when closed as well) or
through other entry points in the cabin. Third, the exhaust from a “ leader” vehicle can enter the
cabin of a following bus through windows or the cabin. Fourth, crankcase emissions can also enter
the bus cabin; however at the time of this study, crankcase emissions were not recognized as an
important source of self- pollution. While it was beyond the scope and resources of our previous
study to directly investigate any of these mechanisms in detail, the results conclusively demonstrated
the importance of self- pollution due to the intrusion of tailpipe exhaust into the bus cabin ( Behrentz
et al., 2004) and the impacts of leader vehicle exhaust, especially the exhaust of other diesel vehicles
( Sabin et al., 2005a, b).
In our previous project report ( Fitz et al., 2003), we made a number of policy
recommendations designed to mitigate the impacts of exhaust from diesel vehicles being followed
by a school bus. For example, we recommended reducing or eliminating the “ caravanning” of buses
( presently a common practice) and attempting to minimize following other heavy- duty diesel
vehicles. However, we could not recommend or design specific mitigation measures for “ self-pollution”
and “ leader/ follower” pollution without a thorough investigation of the mechanisms of
these phenomena.
2.4 Previous Vehicle Exhaust Intrusion Studies
Chan et al. ( 1991) determined the penetration of volatile organic compounds ( VOC), carbon
monoxide ( CO), and NO2 from a car’s exterior into the car’s cabin by simultaneously measuring the
pollutants inside and outside of two experimental vehicles. The median inside/ outside ratio was
approximately 1.1 for the three pollutants, suggesting a slight but measurable contribution of tailpipe
and engine running loss emissions into the passenger compartment. In- vehicle VOC concentrations
were lower with the air conditioner on and higher when the vent was open with the fan on.
Fletcher and Saunders ( 1994) determined the infiltration rate of a gas into stationary motor
5
vehicles for different wind speeds and directions. Measurements were made on five vehicles under
both positive and negative pressures to determine their leak characteristics. A tracer gas method was
then used to determine the air exchange rates in the vehicles for different wind speeds and directions.
Measurement of air exchanges per hour were also made on a vehicle driven at constant speed and
while moving through a cloud of contaminant.
Clifford et al. ( 1997) analyzed the local aspects of vehicular pollution using 1: 10 scale
models placed in a low- velocity wind tunnel with tracer gases ( SF6 and nitrous oxide) injected into
the airflow. Measurements showed the exhaust gases are entrained in the wake of the vehicle from
which they are emitted, and are dispersed mainly by the movement of such wakes. Thus, the wake
itself may be a self- pollution source, depending on its contact with the bus and its pressure relative to
the bus interior.
Wu et al. ( 1998) reported the use of an iridium tracer to determine soot exposure of high
school students commuting to and from school in passenger cars, and on diesel public transit buses
in Baltimore. During this study a portion of the Baltimore municipal fuel supply was tagged with
iridium traces and exposure was monitored during commutes with personal aerosol monitors. The
tracer was undetectable in personal samples collected by the students commuting in passenger cars
when the windows were closed, but comparable to the samples collected on transit buses when the
vehicle windows were open during the commute.
Chan et al. ( 2000) evaluated in- vehicle and out- vehicle CO concentrations during different
driving microenvironments including tunnels and highways. In- vehicle CO levels were highest in
urban residential, rural districts and on some highways; varied with different land uses; and were
found to be influenced by pollutant levels outside the vehicle. The results suggested the penetration
of emissions from outside sources, ( through leaks, joints, or the ventilation system) were occurring
during commutes.
Behrentz et al. ( 2004) developed a method to evaluate the fraction of a bus’s own exhaust
that entered the cabin of several in- use school buses over a range of roadway types, fuels, and
emission control technologies. The percentage of intrusion of the bus’s own exhaust into the cabin,
or self- pollution, was found to be a function of bus type, age, and window position ( i. e., open or
closed). Older buses exhibited a larger amount of self- pollution compared to newer buses with up to
0.3% of the bus’s own exhaust entering the cabin. Also, 25% of the within- cabin black carbon
concentration variance could be explained by the buses’ self pollution. For all buses tested, the
amount of self- pollution was highest while windows were closed compared to when windows were
open.
Fitz et al. ( 2003) evaluated the impact of a leader bus on a follower bus ( with windows open)
using tracer gas released into the exhaust of the leader and driving the same route that we describe as
Route 1 later in this report. The concentration of tracer gas was found to be approximately five times
higher than when the tracer gas was released from the follower bus ( self- pollution) in separate test
runs using this route.
2.4.1 Mechanisms of Exhaust Intrusion Studies
Although several studies on exhaust intrusion have been conducted, there have been far
fewer studies conducted that have evaluated mechanisms of exhaust intrusion.
In 1981, Ziskind et al. conducted a study investigating the intrusion of carbon monoxide
( CO) into sustained- use vehicles. These vehicles included taxicabs, police cruisers, and school
buses. The main sources of CO were from leaks at the rear of the exhaust system or from tailpipe
exhaust. In vehicles with excessive interior CO levels, the sources and intrusion pathways were
identified using a sulfur SF6 detection system. For school buses, large leaks were most often
6
observed at the rear emergency exit door seal, heater or windshield washer water hoses, and along
the exhaust system. The study also found the greatest potential for CO accumulation occurred when
vehicle windows, doors, and vents were closed.
2.4.2 Other Related Exhaust Intrusion Studies
2.4.2.1 Ventilation Air Flow Patterns Inside Vehicles
The ASHRAE applications handbook ( 1999) discusses how to optimize the air- flow within
the cabin of a bus. The handbook states it is necessary to position air inlets and outlets based on the
pressure gradient distribution in the cabin. Most of the pressure is positive on the front surface with
the stagnation point located at about 1/ 3 of the height of the bus. The pressure is strongly negative at
the top and side leading edges, due to localized high velocities. Behind the recirculation bubble in
the front, the pressure on the roof is nearly zero and in the rear, the pressure coefficient is always
slightly negative. Thus, the best location for inlets is the lower part of the front surface. The areas
with strong negative pressure coefficients in the side panels just behind the front are the best
locations for the outlets because vehicle movement drives the flow.
A number of workers have studied flow distribution within vehicles using tracers. For
example, Komoriya ( 1989) used kerosene smoke to study the effect of air changes per hour ( ACH)
on the conditions inside a vehicle compartment and demonstrated that a numerical method could be
used to qualitatively simulate ventilation experiments.
Ishihara et al. ( 1991) determined the flow velocity distribution inside a vehicle by combining
a particle- tracking technique with a pulsed- laser- light- sheet technique. By using a 1: 4 scale vehicle
model and water as the flow medium, flow velocity distributions were determined. Lasers were
directed toward the flow to visualize paths of distinctive particles. The authors suggested that a
similar methodology could be used to measure flows from the exhaust system into the cabin and
external flows around the vehicle body, although neither of these topics was investigated in that
study.
Conceicao et al. ( 1997a) installed a “ removal” duct in a commuter bus to improve ventilation
rate and modeled the airflow with a simple, uni- dimensional flow model, predicting the air exchange
rate as a function of the vehicle velocity. In addition, tracer gas experiments were performed to
demonstrate the adequacy of the model and the efficacy of an air removal duct. Conceicao et al.
( 1997b) also mapped the flow field of the zone occupied by passengers, in terms of mean air
velocity, turbulence intensity, and temperature. A full- scale bus section was used in the laboratory
tests, with the passenger presence simulated by thermally- regulated mannequins. Measurements
were performed with and without “ passengers” seated in the windows seat and in the aisle seat. Air
velocity and turbulence in the vehicle were not affected by the presence of passengers, but did
increase temperatures for certain test conditions when passengers were in the vehicle.
Lee et al. ( 1998) measured, simultaneously, the temperature and velocity field variations of
the ventilation flow inside a vehicle cabin by using a digital image processing technique. In this
study, micro- encapsulated TLC ( thermochromic liquid crystal) particles were also used as a tracer
for temperature and velocity measurement inside a 1/ 10 scale vehicle. The measured temperature
and velocity fields exhibited a close relationship and a high degree of correlation. The simultaneous
use of the two techniques can give reliable information on ventilation flow in the passenger
compartment.
Aroussi and Aghil ( 2001) investigated the ventilation flow inside a 1: 5 scale model of a
typical mid- size passenger compartment with a driver present. Water was used as the fluid medium
seeded with neutrally buoyant particle tracers. The fluid measurements used a particle image
7
velocimetry technique to acquire the velocity distribution. The prediction of velocity distributions
showed this methodology could be useful for studying ventilation performance.
Oshio et al. ( 2001) studied the pressure levels observed in the ventilation ducts by making
modifications to the ventilation system to understand how the shape and configuration of the air
ventilation system determines the ventilation performance in a vehicle. These methods were able to
predict the ventilation characteristics without the use of vehicle prototypes ( passenger cars).
2.4.2.2 Air Exchange Rate Studies
Air exchange rate ( AER) can play a significant role in determining the magnitude of self-pollution.
A number of studies have reported measured AERs for a wide variety of vehicles and
conditions, and have found in general that AER is a strong function of vehicle speed and is much
higher if windows are open.
During a study to measure the exposure to emissions from gasoline within automobile cabins,
Weisel et al. ( 1992) showed the concentration of volatile organic compounds ( VOC) inside the cabin
of vehicles being driven on a suburban route in New Jersey, and on a commute to New York City,
were inversely related to driving speed and wind speed relative to roadway air, although wind
direction was not considered.
Ott et al. ( 1992) measured the air exchange rate, or air changes per hour ( ACH), of an
automobile ( station wagon) moving at 20 miles per hour, and reported ACHs of 13 h- 1 for windows
closed and 121 h- 1 for windows open. The ACH was calculated using a box mass balance model that
is generally defined by the following relationship:
Q ( F ) qC t qC t kQ t S t out in ¶ = 1 - ¶ - ¶ - ¶ + ¶ ( 2.1)
where Q is the mass of indoor contaminant; F is the fraction of the contaminant removed from the
entering air; q is the volumetric air flow rate in and out the automobile; V is the interior volume of
the automobile; t is the time; k is the rate of decay, settling, and removal; S represents the emissions
from the internal source; Cout is the contaminant outside concentration; and Cin is the contaminant
within- vehicle concentration.
Using carbon dioxide ( CO2) as the tracer gas, Park et al. ( 1998) measured ACHs under four
different wind conditions and four ventilation situations in three stationary vehicles. The initial CO2
concentration was approximately 3000 ppm at the start of each test run and the decay in CO2
concentrations was used to calculate the ACH, which ranged between 1.0 and 3.0 h- 1 with windows
closed and no mechanical ventilation to 36 and 48 h- 1 for windows closed with the fan set on fresh
air. ACHs for windows closed with no mechanical ventilation were higher for older automobiles
than for newer vehicles. This study only used stationary vehicles since idling is a major component
of a typical commute in heavy- traffic urban areas.
Brauer et al. ( 2000) estimated average ACHs, using CO as an internal tracer gas, in two
buses being driven in urban British Columbia during real school bus runs while under normal
occupancy loads. ACHs ranged between 10.3 h- 1 and 13.5 h- 1 for two buses tested with windows
closed while on the bus route.
In our previous study, Fitz et al. ( 2003), we measured air exchange rates with the windows
open and the windows closed in seven different buses at speeds of 0, 20, and 40 mph. Air exchange
rates inside the buses were measured by releasing an SF6 tracer gas inside the cabin and monitoring
the gas concentration over time. The results of the ventilation tests are presented in Table 2.4.2.2.1,
which shows the time constant, or the time required for 63% of the bus air to be exchanged. The
8
time for essentially complete exchange is three times ( i. e. 95% exchange) to five times ( i. e. 99%
exchange) longer with windows closed versus windows open; the shorter the time for air to
exchange, the higher the ventilation rate.
Table 2.4.2.2.1 Results from ventilation test conducted in our previous study ( Fitz et al., 2003) for
selected buses.
BUS HE3 RE1 TO1 CNG
Response Time Response Time Response Time Response Time
TEST CONDITION ( mm: ss) ( mm: ss) ( mm: ss) ( mm: ss)
Windows closed 0 mph 09: 47 > 30 min > 15 min > 42 min
Windows open 0 mph 03: 16 03: 57 02: 18 07: 00
Windows closed 20 mph 01: 52 01: 56 04: 38 02: 00
Windows closed 40 mph 00: 38 01: 05 01: 22 01: 21
Windows open 20 mph 00: 58 00: 48 00: 23 00: 26
Windows open 40 mph 00: 29 00: 17 00: 12 00: 23
2.5 Objectives
2.5.1 Overall Objectives
The overall objectives of this study were to determine mechanisms of exhaust intrusion into
school buses, and determine methods to economically reduce children’s pollutant exposure during
school bus commutes.
2.5.2 Specific Objectives
2.5.2.1 Pilot Study
The objectives of the pilot study were to:
1. Determine a method to systematically characterize school bus exhaust system leaks.
2. Identify and characterize intrusion mechanisms and locations for the bus’s own exhaust.
3. Determine the intrusion potential of the exhaust from a vehicle being followed.
4. Evaluate the effectiveness of changes in ventilation and exhaust hardware in reducing exhaust
intrusion into the bus.
2.5.2.2 Main Study
The objectives of the main study were to:
1. Determine a method to rapidly evaluate exhaust system leaks and survey a number of buses to
characterize exhaust leaks.
2. Determine a method to rapidly evaluate bus cabin sealing and survey a number of buses to
characterize cabin leak potential.
3. Evaluate the effectiveness of raising the exhaust outlet to a high position in reducing self-pollution
and pollution from a leading vehicle with a high exhaust.
4. Evaluate the effectiveness of a centrifugal blower ( i. e., power ventilation) in reducing self-pollution
and pollution from a leading vehicle.
9
3.0 PILOT STUDY FINDINGS AND RECOMMENDATIONS
3.1 Introduction
The pilot study was conducted to develop effective investigative methods since little specific
and relevant background information was available. All testing was conducted using a 1985 Thomas
Coach, an 84- passenger school bus currently in use by a school district, and studied in our previous
school bus study. The main study used the methods developed during the pilot study on a wider
variety of buses.
3.2 Summary of Pilot Study Findings
For the pilot study, we conducted four experiments: evaluation of exhaust system leaks;
evaluation of leak points in the bus cabin ( for self- pollution); testing of a tracer gas release system to
help better quantify self- pollution; and evaluation of a leader vehicle. Further detail is found in Fitz
et al. ( 2004).
3.2.1 Evaluation of Exhaust System Leaks
A Meloy SA 285 real- time SO2 ( sulfur dioxide) analyzer was used to probe for leaks in the
bus exhaust system. The bus’s fuel was spiked to 1000 ppm using an organic sulfide blend. Three
small leaks were found in the pilot study bus. By probing an artificial SO2 leak of known leak rate,
we were able to quantify the leaks in the exhaust system as being in the range of 50 ml/ min. This
leak rate would represent less than 0.01% of the exhaust flow at idle. Based on these results, and the
close proximity of the leaks to the exhaust outlet ( 2 meters), exhaust leaks in a well- maintained
system ( such as in the pilot study bus) were considered to be insignificant contributors to self
pollution compared to exhaust rates from the tailpipe.
While gross exhaust leaks could be identified by traditional methods ( visible carbon residue
streaking, noise of escaping gas), it was difficult to evaluate the magnitude of such leaks. Although
in principle, leaks could be quantified by this SO2 method, it was not a practical method for
surveying a large number of buses, primarily because it was necessary to add significant organic
sulfur to the fuel ( buses are routinely operated on low- or non- sulfur fuel and may have exhaust
system catalysts that are poisoned by sulfur) to make quantitative measurements.
3.2.2 Evaluation of Leak Points in the Passenger Cabin and Exhaust Intrusion
The potential for exhaust intrusion into the bus’s cabin was evaluated in two steps. First, we
determined the location of leaks along the outside of the bus’s cabin using propene tracer gas while
the bus was stationary ( engine off) and windows were closed. Tracer gas was introduced into a
blower used to pressurize the bus cabin. A PID ( photoionization detector) instrument was then used
to search for leaks on the exterior of the bus. Leaks were found all over the bus, particularly around
the windows and the front door. An example of a door leak is shown in Photograph 3.2.2.1. This
amount of leakage eliminated the possibility of significantly reducing self- pollution by sealing the
bus cabin.
10
Photograph 3.2.2.1 A leak at the bottom of a bus door.
We determined that the overall leak rate for the pilot study bus was 34 m3/ min at a pressure
differential of 0.18 inches of water column.
The second step to evaluate exhaust intrusion into the bus’s cabin was to determine the
location and magnitude of intrusion points using a tracer gas injected into the test bus’s exhaust and
measuring tracer gas in the bus cabin at potential leak points while both stationary and mobile. For
the stationary test, the bus was parked so the tailpipe was upwind of the cabin.
Windows were closed for both the stationary and mobile tests. The SF6 release system as
discussed in the next section ( Section 3.2.3) was utilized in both tests to maintain a constant
concentration of tracer gas in the bus’s exhaust. We were unable to pinpoint leak locations within
the cabin due to elevated and variable tracer gas concentrations found throughout the cabin. This
was likely due to numerous gross leak points all over the bus as found in the first test, and a rapid
overall accumulation of tracer due to self- pollution.
11
3.2.3 SF6 Tracer Gas Release System
A schematic of the tracer gas release system employed in the pilot ( and main) study is
illustrated in Figure 3.2.3.1.
Figure 3.2.3.1 Tracer gas release system used in main study, using two tracer gases. The system
was controlled by engine intake flow.
Note the figure shows the use of two tracer gases. A 1% SF6 cylinder was used for the pilot
study. We used the remainder of the contents of the 1% SF6 cylinder for the first runs of the main
study then switched to a 5% SF6 cylinder for the remainder of the main study runs. A second tracer
gas ( propene) was added in the main study as discussed in Section 4.2.4.3.1. For the pilot study, SF6
alone was used in the release system.
The purpose of the release system was to maintain a constant concentration of the tracer gas
in the exhaust to more accurately quantify self- pollution. First, we determined the pilot bus’s
exhaust flow by approximating exhaust flow with engine air intake flow. Second, based on the
SF6 Mass
Flow
Controller
C3H6 Mass
Flow
Controller
Flow Rate
Set Points
Flow Rates
Measured
Pressure Transducer
Campbell Data
Logger/ Controller
C3H6
Cylinder
Gas
SF6
Cylinder
Gas
Intake Engine Split Exhaust
Switch
12
intake flow, a mass flow controller was adjusted to release the appropriate amount of tracer gas into
the tailpipe as to maintain a constant concentration of tracer gas in the exhaust. This method was
evaluated while stationary ( with varying rpm) and while traveling on a test route by measuring the
concentration of SF6 tracer gas directly from the exhaust.
The initial tracer gas release system worked well in achieving a relatively constant
concentration in the bus’s exhaust while stationary or at steady speeds. During the stationary testing
we found that there was a delay in the mass flow controller’s response in metering the correct
amount of tracer gas into the exhaust. When the bus had a change in the exhaust flow rate, there was
a one second delay before the mass flow meter received the updated set point information. This was
because the controller operated at 1 Hz. Speeding the controller up to 10 Hz was sufficient to solve
this problem.
3.2.4 Evaluation of Leader Vehicle Exhaust Intrusion
Exhaust intrusion from a leader vehicle was studied with the bus windows open and closed,
and a tracer gas ( SF6) released from the leader vehicle, a small moving truck. SF6 was released
either 0.5 m above the ground ( low exhaust) or 0.5 m above the height of the leader vehicle ( high
exhaust). Four runs were conducted around the UCR campus ( Route 1). For this combination of
buses, exhaust position did appear to have somewhat of an effect on in- cabin concentrations of SF6,
which originated from the leader vehicle when windows were closed. Window position also
appeared to have an effect on in- cabin SF6 concentrations with higher concentrations observed when
windows were open. Results for this test are shown in Table 3.2.4.1. This experiment showed a
second, different tracer needed to be released simultaneously at the other exhaust position to
properly assess the impact of exhaust position due to variability between runs. A second tracer was
utilized in the main study.
Table 3.2.4.1 Mean SF6 data ( in ppt) for pilot study leader/ follower test, evaluating effect of
window position and exhaust position in a leader vehicle.
Window Position
Follower Bus
Exhaust Position
Leader Bus Open Closed
High 3200 2800
Low 3000 3400
3.2.5 Evaluation of Proposed Mitigation Strategies
The mitigation methods we proposed included repairing exhaust leaks, sealing leaks in the
bus cabin, improving bus cabin ventilation, pressurizing the cabin, and raising the exhaust outlet so it
extended above the height of the bus. As noted above, exhaust system leaks were shown to be
insignificant in our pilot study test bus. Cabin leaks were found to be too extensive to seal and it
was not clear indiscriminate and incomplete sealing would be useful. Some of the leaks were so
large they allowed significant amounts outside air to enter the cabin, improving cabin ventilation and
causing tracer gas concentrations to decrease near these leak points in the cabin, especially while
moving.
13
3.3 Modifications of Experimental Design for Main Study
As noted, the purpose of the pilot study was to develop methods to test mitigation strategies
for reducing exposure to be used in the main study. Over the course of the pilot study, some
methods proved to be time consuming or inadequate and modifications were needed. Based on our
observations in the pilot study, several recommendations were made for the main study:
· We recommended the first bus of the main study be reasonably representative of California’s
school bus fleet, and that this first bus be used to further evaluate and refine all test
procedures before testing additional buses. This was important to ensure the quality of the
data collected from subsequent buses.
· The method for evaluating exhaust system leaks with a tracer gas proved to be too
cumbersome considering the small impact exhaust leaks had on self- pollution. A more
convenient method was needed to survey exhaust system leaks. We subsequently developed
and employed a backpressure method for this purpose as described in Section 4.2.4.1.
· A straightforward and fairly rapid procedure needed to be developed to test overall bus leak
rates due to the difficulty we found in the pilot study of isolating individual leaks in the cabin
( in the main study no attempts were made to pinpoint individual leaks in the cabin).
Development of this procedure would also allow for a survey of overall leak rates in buses in
the in- use fleet from which we were recruiting test buses. To accomplish this task a
centrifugal blower and the methods described in Section 4.2.4.2 were used.
· Improvements in the SF6 tracer release system were needed for the main study. This was
accomplished largely by changing the recording and speeding up the control rate of the data
logger.
· The precision of both SF6 and hydrocarbon analyzers needed to be fully documented.
· Methods and testing for mitigation measures were focused on raising the exhaust outlet for
both the test and leader vehicles, and increasing the ventilation rate from front to rear using a
blower and/ or establishing a positive pressure in the bus cabin. When evaluating raised
exhaust, we recommended the exhaust be evenly split between upper and lower outlet
locations in a “ T” shape, with SF6 injected in one outlet and propene in the other, alternating
the two tracers between bus commutes over the test route ( see Section 4.2.4.3.2).
· Meteorological guidelines under which to conduct tests, especially wind speed, needed to be
established to control for meteorological effects.
14
4.0 EXPERIMENTAL METHODS AND STUDY DESIGN
4.1 Introduction
This study was designed to identify cost- effective methods to reduce exhaust pollutant
concentrations in the passenger cabin of school buses by evaluating routes of pollutant intrusion and
methods for reducing such intrusions. The assessment strategy consisted of the following:
1. Development and documentation of a rapid method for the identification and ranking of
exhaust leaks between the engine and the tailpipe ( i. e., leaks in the exhaust system) and a
survey of exhaust system leaks in buses within a district’s fleet.
2. Development and documentation of a rapid method for the identification and ranking of the
leak potential of a bus’s cabin and a survey of buses within a district’s fleet.
3. Evaluation of the effectiveness of mitigation measures such as raising the exhaust outlet
and/ or use of power ventilation in the bus’s cabin in reducing self- pollution, for both
stationary and mobile configurations, including driving under realistic traffic conditions.
4. Performance of leader/ follower experiments in both stationary and mobile configurations and
includes driving conditions to evaluate the effectiveness of mitigation measures such as high
exhaust in the leader vehicle or power ventilation in the follower vehicle.
5. Providing a robust data set and distinguishing spatial and temporal factors contributing to in-cabin
pollution, two different tracer gases and one to three tracer gas analyzers for each of
these two tracer gases were utilized simultaneously during the driving tests.
4.1.1 Vehicle Selection
This study called for the use of 4- 8 school buses. The exact number was subsequently
determined by the effort needed to fully evaluate mitigation methods after testing the first bus. The
goal was to utilize buses found to be representative of California’s current bus fleet. In our previous
study ( Fitz et al., 2003), a 1999 California motor vehicle database was used to plot the distribution of
buses by model year and manufacturer to aid in bus selection. One bus was chosen to be
representative of older vehicles. The other three buses were chosen to represent both the ends and
the middle of the distribution in the most recent 15 model years. When appropriate, buses from this
previous study were used or additional buses were recruited from the same school district. If
available, we chose Thomas and Blue Bird manufacturers since they accounted for approximately
35% of the fleet in southern California. After these vehicles were obtained and prepared, the
monitoring instrumentation was installed in the bus on plywood sheets in a manner similar to that
used in our earlier study.
4.1.2 Fuel Used in the Test Buses
The fuel used in all diesel buses tested ( except for the leader bus) was Arco Emission Control
Diesel ( ECD- 1). This fuel, or “ green” diesel, has ultra- low sulfur content (< 15ppm), low aromatics,
and a high cetane number. Ultra- low sulfur fuel must be used for after- treatment emissions control
technologies ( e. g. particle trap catalysts) to function properly. The leader bus used diesel fuel
meeting California regulations for sulfur in fuel (< 500 ppm).
4.1.3 Characterization and Justification of Selection of Test Routes
The two routes described below were selected based on relative traffic density. Route 1 had
relatively low traffic density and fewer stoplights and stop signs, a desirable characteristic for the
leader/ follower experiments, where the focus was evaluating the impact of the leader vehicle. Route
15
2 was characterized by increased traffic density and several stoplights and stop signs, conditions that
would promote self- pollution.
4.1.3.1 Route 1
Route 1, mapped in Figure 4.1.3.1.1 was utilized in all leader/ follower runs and was also
used for all runs in the pilot study. The start point was the intersection of Spruce Street and Iowa
Avenue in the city of Riverside. The route traveled east on Spruce Street to Watkins Drive then
southeast on Watkins Drive to State Highway 60, where the street name changed to Central Avenue,
and then curved to the southeast. At Chicago Avenue a right turn was made and the route continued
north to Spruce Street, where another right turn was made. The route was mostly free of other
diesel- powered vehicles, which was optimum for our leader/ follower tests, in order to better evaluate
the effect of the leader bus only. The total length of the test route was approximately 10 miles and
required 20- 25 minutes to complete depending on traffic signals and congestion, which was
generally light depending on the time of day. During morning and afternoon commutes, congestion
added 10- 15 minutes to the driving time.
Figure 4.1.3.1.1 Map of Route 1 in Riverside, California.
16
4.1.3.2 Route 2
Route 2, mapped in Figure 4.1.3.2.1 was utilized in all self- pollution runs. The route began
at the intersection of Market and 3rd Streets in downtown Riverside, then headed south on Market
Street, which changed its street name to Magnolia Avenue. At Central Avenue a left turn was made,
and the route headed east toward Riverside Avenue. A left turn was made at Riverside Avenue and
another left at Jurupa Blvd. At Magnolia a right turn was made and the route headed back north
toward 3rd
Street. This route was characterized by moderate to heavy traffic with several stops ( e. g.
for stop lights, train tracks, or stop signs), creating conditions to promote self- pollution, for example,
the exhaust plume of the bus being blown over the bus when stopping. The route was about 8- 10
miles long taking approximately 30 minutes to complete a single loop. For the first 8 runs conducted
in this study, Route 2 included a residential area. This section of the run was cut for all subsequent
runs conducted on Route 2. Figure 4.1.3.2.1 does not include this residential area.
Figure 4.1.3.2.1 Map of Route 2 in downtown Riverside, California.
4.2 Field Sampling Procedures
4.2.1 Instrument Packaging and Supply
Power for all vehicle- mounted instruments was provided by 12V automotive batteries. A
sine wave inverter was used to generate 110VAC for the instruments requiring AC power.
Jurupa
Riverside
17
4.2.2 Instrumentation
Table 4.2.2.1 summarizes the measurement methods used in the main study. This section
provides detail on these methods and their purpose.
Concurrent with the preparation of the school buses, measurement instruments were
assembled, configured, and tested at University of California, Riverside’s College of Engineering,
Center for Environmental Research and Technology, including all necessary calibration and data
logging equipment. The instruments were tested for proper operation and proper interfacing with
their respective calibration and data logging systems. After these tests were successfully completed,
the instruments were installed in the bus and further tested.
Table 4.2.2.1 Measurement methods utilized in the main study.
Species/ Measurement Instrument/ Model Detection Limit
Sulfur Hexafluoride ( SF 6) AeroVironment CTA 1000 10 ppt
Particle Bound PAH ( inside and
outside bus)
EcoChem PAS 2000 0.01 μg/ m
3
Particulate Matter Number> 7nm
Thermo Systems Inc.
Model 3022 1 particle/ cm
3
Total Hydrocarbons ppb RAE 0.05ppm
Bus Location Garmin Map 76 GPS 3 m
Bus Engine rpm Engine Alternator Signal Single pulse
Temperature & Relative Humidity
( inside bus)
Rotronics PM101A 0.5° C/ 5% RH
Exhaust Gas Flow Rate
Omega PX274 Pressure
Transducer
0.00” H 2 O
Wind Speed, Wind Direction, and
Temperature
Climatronics F460
0.1 m/ s, 2 deg WD,
0.1° C
4.2.2.1 SF6 Measurements
SF6 ( sulfur hexafluoride), one of the tracer gases used in the main study and also used in the
pilot study, was measured with two AeroVironment Model CTA 1000 analyzers. This instrument
uses electron capture detection after water and oxygen are removed from the sampled air. The
instrument was developed for operation on a moving platform and had a sensitivity of approximately
10 ppt with a response time of twenty seconds.
4.2.2.2 Real- Time Particle Phase PAH Measurements
Two EcoChem Model PAS 2000 analyzers were used to measure concentrations of particle-bound
PAH inside the cabin and outside the cabin ( roadway concentrations). This instrument uses a
UV lamp to photoionize PAH components of particles. An electric field is then applied to remove
negatively charged particles. The positively charged particles are collected on a filter and the total
charge collected is measured with an electrometer; the charge collected is proportional to the
concentration of particle- bound PAH. The sensitivity of the instrument is approximately 10 ng/ m3.
18
4.2.2.3 Condensation Particle Counts ( CPC)
A Thermo Systems Incorporated Model 3022 Condensation Particle Counter was used to
determine the number concentration of particles. This device uses butanol to grow particles and
light scattering to detect them. It detects particulates starting at 3 nanometers in diameter with the
measurement efficiency increasing with size ( 50% of particles 7 nanometers in diameter) at
concentrations up to 107 particles/ cc. The response time is 13 seconds for a 95% response to a step
change.
4.2.2.4 Bus Location
Location was monitored with a Garmin GPS Map76 global positioning system with WAAS
( Wide Area Augmentation System) capability. Position was determined to within three meters. In
addition to horizontal position ( e. g., latitude and longitude or UTM coordinates), the system also
provided elevation and bus velocity data. These data were displayed on a liquid crystal display on
the GPS with a digital output ( RS232) for data logging along with the air quality data. The GPS unit
was used as the time reference during this study. The clocks on all other devices were set to the GPS
time before each run.
4.2.2.5 Engine Operating Parameters
Each bus was operated at several different engine speeds to obtain a relationship between
engine rpm, manifold vacuum and exhaust flow rate. An Omega model PX274 differential pressure
transducer was used to monitor manifold vacuum in real time during bus operations. Near- constant
exhaust tracer gas concentration was obtained by using a data logger/ controller programmed with the
manifold vacuum exhaust flow rate relationship and using the vacuum signal from the pressure
transducer to control the tracer gas flow set point of mass flow controllers for introducing tracer gas
into the bus’s exhaust.
4.2.2.6 Propene Measurements
Propene, the second tracer gas used in the main study, was measured using three RAE
Systems ppbRAE hydrocarbon analyzers. This instrument determined the concentration of
hydrocarbons using a 10.6 electron volt photoionization detector ( PID) and has a lower detection
limit for propene of approximately 50 ppb.
4.2.2.7 Meteorological Measurements
Prevailing wind, wind direction, and temperature in the study area were determined using a
system located at a height of 5 meters at CE- CERT.
A Climatronics F460 wind speed and wind direction monitoring system was connected to a
Campbell 10X data logger. This system measured and processed winds into hourly averages and
had an accuracy of +/- 5 degrees for wind direction and +/- 5% wind speed accuracy for winds greater
than 5 m/ s.
4.2.2.8 Video Camera
A Sony DCR- TRV330 video camera was mounted in front of the bus and operated at all
times when the bus was moving. The video records were stored and archived on a computer for
future reference, but were not analyzed as a part of this project.
19
4.2.2.9 Tracer Release Control
A Campbell 23X data logger was used on the follower bus to monitor engine manifold
vacuum and used this signal to control the tracer gas mass flow measurements. A Campbell 21X
data logger was used on the leader bus to perform the same engine manifold vacuum monitoring and
tracer gas release control functions on the leader bus for runs that included the leader bus. The
Campbell 23X data logger was operated at a scanning and logging rate of 10 Hz. The 21X data
logger was operated at 4 Hz, its maximum scanning and logging rate.
4.2.3 Data Collection
Data from the following instruments were collected using a laptop PC with Labview software
and appropriate A/ D cards and RS- 232 multiplexers.
• AeroVironment CTA1000 continuous SF6 analyzers
• EcoChem PAS 2000 particle- phase PAH analyzers
• TSI model 3022 condensation particle counters
• Garmin GPSMAP76 Global Position System
• ISSPRO R8930 magnetic sensor
• SF6 mass flow sensor
• Propene mass flow sensors
• Omega PX274 differential pressure transducer
At the conclusion of each set of tests, all data were transferred to a networked PC for storage
and backup. The PIDs had internal logging capabilities and were downloaded to a PC. The clocks
for all of these instruments were synchronized at the beginning of each test run using the GPS time
as a reference.
4.2.4 Experimental Design
4.2.4.1 Evaluation of Exhaust Train Leaks
Exhaust leaks were evaluated by placing a silicon stopper ( with approximately 2 cm2 hole) in
the tailpipe of several buses. A magnehelic was also attached to the stopper to measure back
pressure. Once the stopper was in place, the back pressure obtained by covering the exhaust flow,
was recorded. Exhaust leaks were qualitatively evaluated by listening for hissing sounds in the
exhaust system and noting any visible carbon streaks on the exterior of the bus ( near the engine
compartment or along the exhaust).
The back pressure method of exhaust leak detection was used on 17 school buses in the fleet
from which we selected the test buses. This method is shown in Photograph 4.2.4.1.1.
20
Photograph 4.2.4.1.1 Back pressure method to evaluate exhaust system leaks using a silicon stopper
and magnehelic.
4.2.4.2 Evaluation of Leak Potential in the Bus’s Cabin
The method for determining the leak potential of bus cabins (“ blower door” method) was the
same method developed to measure building tightness. A centrifugal blower was set to a nominal
flow rate of about 28 m3 and the pressure inside the buses was measured. This method of leak
testing was used to conduct a survey of 17 buses in the district’s fleet. Plastic sheeting was used on
one bus, to seal off window areas to assess the contribution of those locations to overall leaks on the
bus ( see Photograph 4.2.4.2.1).
Photograph 4.2.4.2.1 Sealing passenger windows with plastic sheeting to assess contribution of
window leaks to overall leaks on the bus.
21
4.2.4.3 Evaluation of Tailpipe Exhaust Intrusion ( Self- Pollution)
4.2.4.3.1 Tracer Gas Release System
The tracer release system was initially tested in the pilot study ( Section 3.2.3) and further
improved in the main study as documented in Section 5. Two tracer gases were utilized for the main
study, the purpose of which was to remove variability that may have occurred between subsequent
runs.
Several steps were taken to improve the performance of the release system. Pressure
resulting from the intake flow and tracer gas mass flow rate were continuously monitored and
recorded at 10 Hz during these tests to validate proper performance of the release system. To
preclude cabin air contamination, the SF6 and propene gas cylinders and associated release systems
were either located outside of the passenger cabin or were checked for leaks before each test run
when the bus was stopped ( in these conditions the air exchange rate on the bus was low).
Photograph 4.2.4.3.1.1( a, b) shows the location of the release system and associated parts outside of
the bus cabin. For runs where only a single bus was used ( i. e. non leader- follower runs), the tracer
gas cylinders remained inside the bus.
4.2.4.3.2 Self- Pollution Runs
These tests were conducted while stationary ( exhaust outlet upwind) and on Route 2. Tests
were conducted only if the average wind speed measured at the CE- CERT facility was less than 5
m/ s. Windows were closed for all tests as opening windows would result in the air quality within the
cabin being dependent primarily on roadway pollutant concentrations and impacts of exhaust from
nearby vehicles ( Sabin et al, 2005a, b). The degree of self- pollution in the test bus was determined
with and without the following mitigation strategies: directing the bus exhaust above the height of
the bus ( high exhaust), use of a power ventilation system inside the bus, a combination of the two,
and sealing the window areas.
To evaluate the effectiveness of high exhaust, the flow of exhaust was split with a 4”
diameter “ T” pipe, directing half the exhaust flow upward while the other half of the exhaust flow
was directed out lower the rear of the bus. SF6 was directed into one side of the split and propene
into the other. The two tracers reversed positions between subsequent runs by way of the toggle
switch shown in Photograph 4.2.4.3.1.1a.
22
( a) ( b)
Photograph 4.2.4.3.1.1( a, b) Tracer Gas Release System: ( a) Toggle switch to reverse tracer gas
release position and ( b) Arrow points to mass flow controllers to
control tracer gas release into the tailpipe.
The split exhaust arrangement is shown in Photograph 4.2.4.3.2.1, and allowed for direct
simultaneous comparison of high and low exhaust position.
Photograph 4.2.4.3.2.1 Split exhaust configuration for self- pollution experiments. One tracer gas
was released from each exhaust branch.
Original exhaust
pipe location
23
The flow rates of both tracer gases were regulated by the release system to maintain a near
constant concentration in the exhaust. The exhaust volume split ratio was expected to be nearly even
because each branch had the same number of bends and were the same length. This ratio was
determined for every bus tested by measuring the flow rate from each exhaust branch with a pitot
tube and measuring the exhaust temperature. The results are shown in Table 4.2.4.3.2.1.
Table 4.2.4.3.2.1 Flow, as measured by pressure in inches of water for split exhaust hardware
including ratios for high versus low exhaust flow for each test bus.
Bus No. RPM
High Exhaust
( inches of H2O)
Low Exhaust
( inches of H2O)
Ratio
982 700 0.15 0.15 1.0
2100 0.3 0.25 1.2
2650 0.75 0.75 1.0
872 2000 0.25 0.25 1.0
2850 0.75 0.65 1.2
021 1000 0.15 0 0.0
1500 0.25 0.23 1.1
2000 0.35 0.22 1.6
2400 0.7 0.7 1.0
923 600 0.18 0.2 0.9
1000 0.38 0.45 0.8
1500 0.55 0.55 1.0
2000 1.1 1.1 1.0
2200 1.7 1.8 0.9
Power ventilation was evaluated by placing a centrifugal blower inside the bus cabin, toward
the front of the bus. Ten- inch diameter tubing was attached to the blower’s inlet and the end of the
tubing was positioned to bring outside air into the bus cabin. For three of the four buses tested
( Buses 982, 021, and 923), the tubing was connected to a vent located at the roof of the bus. In Bus
872, the tubing was brought out through a window near the front of the bus and wrapped around to
the roof of the bus. These configurations are pictured below ( Photograph 4.2.4.3.2.2a- d).
Power ventilation ( a centrifugal blower operating at approximately 28 m3 cfm for all tests)
was evaluated by alternately turning the blower on for an entire run and then off for an entire run.
Both primary mitigation strategies could be tested during a single run ( power ventilation, high
exhaust, and combination) as exhaust tracer gas concentration was constantly monitored
simultaneously at both positions and power ventilation could easily be turned on and off.
All tracer gas analyzers sampled from a single location in the bus’s cabin in a typical
breathing location, the middle of the bus approximately 8 cm above the seat back.
24
( a) ( b)
( c) ( d)
Photograph 4.2.4.3.2.2 ( a- d) Blower inlet positions on Bus 982 ( a- b) and Bus 872 ( c- d).
For routine testing, a pair of SF6 tracer gas analyzers and three hydrocarbon analyzers
continuously measured concentrations at the reference point in the bus’s cabin. An EcoChem PAS
2000 analyzer was collocated with a TSI model 3022 condensation particle counter ( CPC) to
monitor particle- bound PAH and concentration of particles, respectively, at the reference point. A
second EcoChem PAS 2000 was used to measure roadway concentrations via a sample line extended
out through a window.
25
For stationary tests, the bus was run at idle and positioned so the exhaust was generally
upwind of the cabin. These tests were conducted during periods of calm winds ( typically in the
mornings) and during periods of increased on- shore winds ( afternoons). As in the mobile runs, the
effectiveness of high exhaust and power ventilation were evaluated. Between each run the engine
was shut off and the cabin was ventilated ( with the blower if necessary) to flush tracer gas from the
cabin. As in the mobile runs, the two PAH and a single CPC instruments were used to collect
measurements during stationary runs.
As an exploratory test, we evaluated the effect of sealing the window areas with heavy plastic
sheeting for one bus. This test was conducted on- road only and not while stationary.
In summary, the following self- pollution tests were conducted on each bus: effect of exhaust
position ( high exhaust versus low exhaust), use of power ventilation ( blower on/ blower off), and a
combination of the two mitigation measures ( blower on and high exhaust). The route used for all
self- pollution tests was Route 2 and were conducted while the test buses were stationary and mobile.
4.2.4.4 Evaluation of Exhaust Intrusion from a Leader Vehicle
The purpose of these tests was to characterize the relative changes in exposure resulting from
mitigation measures when following other heavy- duty diesel vehicles. In the main study, the leader
vehicle was another diesel bus. The two test buses used in these experiments were Bus 982 and Bus
872. The leader bus was a Bluebird diesel bus borrowed from the UCR Transportation Services
fleet.
Two types of leader/ follower experiments were conducted. In both experiments all air
quality instruments were located in the follower bus ( the test bus). The leader bus had a second GPS
for measuring position and speed, tracer gases and release system, manifold vacuum monitoring,
data logging, and tracer gas injection control PC. The leader bus was equipped with release system
controlled by a data logger/ controller operating at its maximum update rate of 4 Hz ( a 4 Hz
controller was the best available to us) while the follower bus was equipped with a release system
controlled by a data logger/ controller operating at an update rate of 10 Hz. In the first type of
leader/ follower experiments, both the SF6 and propene tracer gas release systems were placed in the
leader bus with a split high/ low exhaust “ T” ( Photograph 4.2.4.4.1).
Photograph 4.2.4.4.1 Exhaust configuration for leader vehicle during leader exhaust tests to
evaluate impact of following a bus with high versus low exhaust.
26
In the second approach, one tracer release system was placed on the leader bus and one on
the follower bus, each releasing tracer gas at a low exhaust position ( Photograph 4.2.4.4.2) to
simulate the effects of the buses “ caravanning” ( as observed in our previous study), and also
evaluate the impact of exhaust intrusion versus self- pollution.
( a)
( b)
Photograph 4.2.4.4.2( a, b) Exhaust configurations for follower bus ( a) and leader bus ( b) during
leader exhaust/ follower exhaust tests to assess impact of exhaust
intrusion from a leader vehicle versus self- pollution.
The leader/ follower tests were conducted in both stationary ( with the leader bus upwind of
the follower bus) and mobile configurations with the windows of the follower bus open and closed.
The effect of power ventilation in the follower bus was also evaluated during these tests. The
leader/ follower tests were conducted for multiple runs on two buses in the main study.
27
4.3 Baseline Tracer Gas Measurements
The baseline for each tracer gas analyzer, or the “ zero” instrument response, was determined
before and after each series of up to four consecutive, half- hour runs. The average of the baseline
concentrations from the beginning and end of each series of runs was then subtracted from the
measured tracer gas concentrations. Between runs, the bus was ventilated by opening the windows
and turning on the blower. Between some of the runs, small amounts of residual tracer gas may have
been present, but this was not observed to significantly affect average run concentrations.
4.4 Data Analysis Methods
Because of the dynamic nature of pollution effects aboard moving vehicles, real- time data
collection was emphasized in this project. Therefore, various time- series analysis techniques
including descriptive analyses were employed. Descriptive analyses were also used to study overall
and cyclic patterns as well as to identify outliers and turning points within the time- series.
Techniques included time- series graphs, scatter plots, smoothing ( e. g., moving average), as well as
the estimation of statistical parameters such as arithmetic mean, standard deviation, and median. In
the following section, one minute medians were used to analyze the data from the current study.
28
5.0 RESULTS AND DISCUSSION
In this study, three types of mobile testing experiments were conducted: “ self- pollution”
( tracer gases in test bus exhaust), “ leader- follower” ( tracer gases in exhaust of the leader bus, in front
of the test bus) and “ leader exhaust- follower exhaust” ( tracers in both leader and follower bus
exhaust). Results from these experiments were used to evaluate several strategies for reducing
pollution in school bus cabins as discussed later in this section.
We tested four in- use school buses ranging in age from 3 to 18 years in service. A
description of the buses is shown in Table 5.1.1. Buses were selected to be representative of the
current school bus fleet as noted in section 4.1.1. The make and models chosen depended on the
selection available from the lending District’s fleet. Bus 982 was one of the study buses tested in the
previous school bus study ( Bus TO1, Fitz et al, 2003). Note that the first two numbers of the bus
number correspond to the last two numbers of the model year ( e. g., Bus 982 is a model year 1998
and Bus 021 is a model year 2002).
Table 5.1.1 Characteristics of the test buses.
Bus No. Make/ Model Year Mileage Type
982 1998 Thomas Saf- T- Liner 1998 124,000 Diesel ( with particle trap)
872 1987 Blue Bird 1987 324,000 Diesel
021 2002 Thomas Saf- T- Liner 2002 66,000 Diesel
923 1993 Carpenter SPT- 3908 1992 128,000 Diesel ( converted from CNG)
Table 5.1.2 describes all mobile runs conducted in the study, including test date ( mmdd), bus
number, type of test conducted, route number on which the testing was conducted, window position,
tracer gas release positions, and power ventilation ( blower operation). During mobile testing, our
test buses traveled over one of two selected routes ( discussed earlier) and all such runs were
conducted in the late morning to the late afternoon.
Table 5.1.3 shows meteorological data including wind speed, wind direction, temperature,
relative humidity, and respective standard deviations for most study days. The data in this table
summarize conditions during which mobile tests were conducted. Meteorological conditions were
stable across these test periods. Mean wind speed and direction over all mobile runs were 3 m/ s and
256 degrees, respectively. Temperature and relative humidity for all mobile runs averaged 24 ° C
and 38%, respectively.
29
Table 5.1.2 Description of all mobile runs ( conducted in 2005)
Test
Date
Run
Number
Bus
Number Run Type Route
No.
Window
Position
SF6
Release
Position
Propene
Release
Position
Blower
0405 13 982 Self- Pollution 2 Closed High Low Off
0405 14 982 Self- Pollution 2 Closed Low High Off
0405 15 982 Self- Pollution 2 Closed Low High On
0405 16 982 Self- Pollution 2 Closed High Low On
0406 17 982 Self- Pollution 2 Closed High Low Off
0406 18 982 Self- Pollution 2 Closed Low High Off
0406 19 982 Self- Pollution 2 Closed Low High On
0406 20 982 Self- Pollution 2 Closed High Low On
0407 21 982 Self- Pollution 2 Closed High Low Off
0407 22 982 Self- Pollution 2 Closed Low High Off
0412 23 982 Leader Exhaust 1 Open High Low Off
0412 24 982 Leader Exhaust 1 Open Low High Off
0412 25 982 Leader Exhaust 1 Closed High Low On
0412 26 982 Leader Exhaust 1 Closed Low High On
0412 27 982 Leader Exhaust 1 Closed High Low Off
0412 28 982 Leader Exhaust 1 Closed Low High Off
0413 29 982 Leader Exhaust Follower Exhaust 1 Open Low Low Off
0413 30 982 Leader Exhaust Follower Exhaust 1 Closed Low Low On
0413 31 982 Leader Exhaust Follower Exhaust 1 Closed Low Low Off
0419 32 872 Leader Exhaust 1 Open High Low Off
0419 33 872 Leader Exhaust 1 Open Low High Off
0419 34 872 Leader Exhaust 1 Closed High Low Off
0419 35 872 Leader Exhaust 1 Closed Low High Off
0419 36 872 Leader Exhaust 1 Closed High Low On
0419 37 872 Leader Exhaust 1 Closed Low High On
0420 38 872 Leader Exhaust Follower Exhaust 1 Open Low Low Off
0420 39 872 Leader Exhaust Follower Exhaust 1 Closed Low Low On
0420 40 872 Leader Exhaust Follower Exhaust 1 Closed Low Low Off
0420 41 872 Leader Exhaust Follower Exhaust 1 Open Low Low Off
0420 42 872 Leader Exhaust Follower Exhaust 1 Closed Low Low On
0420 43 872 Leader Exhaust Follower Exhaust 1 Closed Low Low Off
0427 44 872 Self- Pollution 2 Closed High Low Off
0427 45 872 Self- Pollution 2 Closed Low High Off
0427 46 872 Self- Pollution 2 Closed Low High On
0427 47 872 Self- Pollution 2 Closed High Low On
0503 52 021 Self- Pollution 2 Closed Low High Off
0503 53 021 Self- Pollution 2 Closed High Low Off
0503 54 021 Self- Pollution 2 Closed High Low On
0503 55 021 Self- Pollution 2 Closed Low High On
0504 56 021 Self- Pollution 2 Closed Low High Off
0504 57 021 Self- Pollution 2 Closed High Low Off
0504 58 021 Self- Pollution 2 Closed High Low On
0504 59 021 Self- Pollution 2 Closed Low High On
0510 60 923 Self- Pollution 2 Closed Low High Off
0510 61 923 Self- Pollution 2 Closed High Low Off
0510 62 923 Self- Pollution 2 Closed High Low On
0510 63 923 Self- Pollution 2 Closed Low High On
0511 64 923 Self- Pollution 2 Closed Low High Off
0511 65 923 Self- Pollution 2 Closed High Low Off
0511 66 923 Self- Pollution 2 Closed High Low On
30
Table 5.1.3 Meteorological data during mobile runs ( with standard deviations)
Test
Date
( 2005)
Average
Wind Speed
( m/ s)
Average Wind
Direction
( deg)
Temperature
(° C)
Relative
Humidity
(%)
0406 NA NA NA NA
0407 4 ± 0.4 256 ± 5 22 ± 0.7 38 ± 4.2
0412 2.8 ± 0.3 245 ± 11 28 ± 1.2 24 ± 0.5
0413 3 ± 0.3 259 ± 11 26 ± 0.6 34 ± 2.6
0419 3.8 ± 0.4 262 ± 6 19 ± 1.0 46 ± 4.0
0420 2.6 ± 0.4 259 ± 14 23 ± 0.6 32 ± 5.6
0427 3.5 ± 0.4 258 ± 7 19 ± 0.7 49 ± 1.2
0503 2.6 ± 0.6 265 ± 11 26 ± 0.5 45 ± 2.3
0504 2.7 ± 0.2 252 ± 8 26 ± 0.8 50 ± 2.5
0510 3.5 ± 0.4 249 ± 16 22 ± 0.7 32 ± 2.4
0511 3 ± 0.4 255 ± 9 26 ± 0.4 28 ± 1.3
5.1 Tracer Gas Release System
Upon receipt of each bus we did stationary measurements to determine the range and
relationship between engine RPM and both exhaust flow rate and manifold vacuum. Based on the
exhaust flow rate range for each bus, we established a tracer gas flow rate in relation to the exhaust
flow rate. Because the bus engine parameters varied significantly from one another, the tracer gas
flow rate relationship was established independently for each bus to keep the tracer gas at
measurable levels and also to be within the range of our controllers. Therefore the average tracer
flow rates varied from bus to bus because the nominal flow rates were changed to accommodate the
exhaust flow rate range of each bus and the dynamic operating range of the tracer gas flow
controllers and detection range of the tracer analyzers.
5.1.1 Validation
Validation of the tracer gas release system described in Section 3.2.3 was an important
component of this study. As discussed earlier, the purpose of the release system was to control the
amount of tracer gas released into the tailpipe to maintain a constant concentration of tracer gas in
the exhaust. Based on our experience in the pilot study, efforts were made to improve the
performance of the release system in the main study. The following two methods were used to
validate the method of maintaining constant concentration of tracer gas in the bus exhaust.
31
5.1.1.1 Pearson’s Correlation Coefficient
Pearson’s correlation coefficient between the measured bus exhaust flow and the amount of
tracer gas flow released into the tailpipe was calculated. Perfect correlation between these two
variables would indicate the release system was functioning and would correspond to a Pearson’s
correlation coefficient ( r) of + 1.0. The resulting correlation coefficients are shown in Figure
5.1.1.1.1 for each test day ( the aggregate of data for up to six individual runs per day).
When we examined the Pearson’s correlation coefficients for both SF6 and propene tracer
gases for all days ( mobile runs), it appeared the release systems for the two tracer gases performed
well for all mobile runs conducted during the study with an overall average correlation coefficient
for both release systems of 0.97± 0.04. For the SF6 and propene tracer gas release systems, all runs
had coefficients greater than 0.80. The average value of 0.97 is similar to the value obtained in the
pilot study, r= 0.93, which was determined from two runs testing the release system. However, for a
majority of the runs in the main study, the release system performed at least as well, or better than in
the pilot study.
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0405 0406 0407 0412 0413 0419 0420 0427 0503 0504 0510 0511
Test Date ( mmdd)
Pearson's Correlation Coefficient
SF6 C3H6
Figure 5.1.1.1.1 Pearson’s correlation coefficient for each day of mobile testing in the main study.
Both SF6 and propene tracer gases were released from the test bus during self- pollution runs.
The average Pearson’s correlation coefficient for the release system for self pollution runs was
0.94± 0.06 for both tracer gases. Figure 5.1.1.1.2 and 5.1.1.1.3 shows scatter plots of exhaust flow
versus tracer gas flow on a typical run, Run 44 on 4/ 27, for the SF6 and propene tracer gas release
system. Further discussion on calculation of tracer gas concentration in the exhaust is discussed in
Section 5.1.1.2.
32
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90
SF6 Flow Rate ( cc/ s)
Exhaust Flow Rate ( m^ 3/ s)
Figure 5.1.1.1.2 Scatter plots of exhaust flow versus SF6 flow rate for 0427/ Run 44.
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
0 200 400 600 800 1000 1200 1400 1600
Propene Flow Rate ( cc/ s)
Exhaust Flow Rate ( m^ 3/ s)
Figure 5.1.1.1.3 Scatter plots of exhaust flow versus propene flow rate for 0427/ Run 44.
33
5.1.1.2 Tracer Gas Concentration in Exhaust
A second method for evaluating the performance of the release system was to investigate the
concentration of tracer gas in the exhaust. Since conducting direct measurements in the exhaust was
not feasible in this study ( the high concentrations required were far beyond the range of the high
sensitivity analyzers required after dilution), tracer gas concentration in the exhaust was calculated
using Equation 5.1 below, as described in detail in Fitz et al. ( 2003) and Behrentz et al. ( 2004).
Qexh Qcyl
Qcyl
Cexh Ccyl
+
= * ( 5.1)
where:
Qexh = Exhaust flow ≈ Exhaust intake flow
Cexh = Concentration of tracer gas in the exhaust
Qcyl = Tracer gas flow from the compressed gas cylinder
Ccyl = Concentration of tracer gas in the compressed gas cylinder
Table 5.1.1.2.1 shows the average concentration and standard deviation of SF6 and propene
in the exhaust including test type, test day, and bus number.
Table 5.1.1.2.1 Variation of SF6 and propene exhaust concentrations for all mobile runs in the
main study.
SF6 ( ppm) C3H6 ( ppm)
Test Date
Test Type - 2005
Bus
No. Average
Standard
Deviation Average
Standard
Deviation
SP 0405 982 5.8 0.1 7900 150
SP 0406 982 5.8 0.1 7900 160
SP 0407 982 6.3 0.1 8000 150
LE 0412 982 4.6 0.4 7700 710
LE- FE 0413 982 3.0 0.1 3900 330
LE 0419 872 4.6 0.4 7700 720
LE- FE 0420 872 2.0 0.1 3900 360
SP 0427 872 3.3 0.4 8000 1100
SP 0503 21 4.3 1.0 5300 280
SP 0504 21 5.8 0.2 5300 260
SP 0510 923 3.9 0.2 7900 350
SP 0511 923 3.8 0.3 8300 420
34
For self- pollution tests the average percent standard deviation for SF6 and propene
concentrations in the exhaust was 7% and 5% respectively. For leader exhaust runs ( 0412, 0419) the
average percent standard deviation for SF6 and propene exhaust concentrations were both 9%. For
leader exhaust- follower exhaust runs ( 0413, 0420) the average percent standard deviation for SF6
and propene exhaust concentrations were 6% and 11% respectively. The higher average variations
were most likely due to the 4 Hz release system as discussed below.
The recording speed of the release system ( 10 Hz versus 4 Hz) affected the variability of
tracer gas concentration in the exhaust. The SF6 release system operated at 10 Hz and was always
used on the follower bus, and for all buses tested for self- pollution except during the leader exhaust
experiments. The propene release system operated at 10 Hz for all buses tested for self- pollution
and was always on the leader bus during leader- follower tests; for these runs, both the propene and
SF6 release system operated at 4 Hz. Use of a 4 Hz release system decreased the sensitivity of the
tracer gas release system, but operated well as shown by correlation coefficients for the relationship
between exhaust flow and tracer gas flow for propene of 0.97 ( for all leader- follower experiments)
and 0.96 for SF6 ( for leader exhaust- follower exhaust experiments only).
In Figure 5.1.1.2.1a, one minute medians of SF6 and propene exhaust concentrations ( Cexh)
are plotted against time for a typical day in the main study ( 0405). On 0405, both release systems
operated at 10 Hz using the Campbell 23X controller in the test bus. This time series is contrasted
with a time series from leader- exhaust runs on 0412 ( Figure 5.1.1.2.1b) when both SF6 and propene
release systems were controlled by the Campbell 21X controller operating at its maximum rate ( 4
Hz) on the leader bus. Larger variability in exhaust concentrations for both tracer gases was seen for
the 4 Hz system on 0412. A second 10Hz controller was not available for this study and the 4Hz
controller was the best controller available to us as mentioned previously.
In summary, for the main study, we were able to assess the performance of the tracer gas
release systems and show the system maintained a relatively constant concentration of tracer gas in
the exhaust within 10% of mean tracer gas concentrations ( for one standard deviation).
35
0
1000
2000
3000
4000
5000
6000
7000
8000
9000
11: 38 11: 52 12: 07 12: 21 12: 36 12: 50 13: 04 13: 19 13: 33 13: 48
Time
Propene ( ppm)
0
1
2
3
4
5
6
7
SF6 ( ppm)
Propene in Exhaust SF6 in Exhaust
( a)
0
1000
2000
3000
4000
5000
6000
7000
8000
9000
15: 21 15: 50 16: 19 16: 48 17: 16 17: 45
Time
Propene ( ppm)
0
1
2
3
4
5
6
7
SF6 ( ppm)
Leader Propene Exhaust Leader SF6 Exhaust
( b)
Figure 5.1.1.2.1 ( a, b) Time series of one- minute medians of SF6 ( 10 Hz release system) and
propene ( 4 Hz release system) exhaust concentrations for 0405( a) and
0412( b).
36
5.2 Bus Cabin Leak Potential
5.2.1 Evaluation of Overall Bus Cabin Leak Potential
Cabin leak potential may be defined as the extent of cracks, holes, gaps, or other openings in
a bus cabin. These openings are potential pathways for self- pollution and/ or exhaust intrusion from
other vehicles, as well as intrusion of ambient air ( roadway air). The following sections discuss
methods to evaluate the extent of openings or leak potential in bus cabin.
5.2.1.1 Rapid Evaluation of Overall Cabin Leak Rate
A rapid method for determining a measure of overall cabin leak rate, the “ blower door”
method discussed in Section 4.2.4.2, was evaluated for a total of 17 buses. All buses were diesel
except bus 054, a model year 2005 CNG bus. The “ blower door” method was based on
measurement of the cabin pressure produced by an approximately constant, high volume blower rate.
As seen in Figure 5.2.2.1.1, these data show a wide range of variability in cabin pressure, which is
inversely related to cabin leak potential, for buses of varying ages and types.
0
0.05
0.1
0.15
0.2
0.25
0.3
851 854 855 856 871 872 881 923 926 927 982 983 986 987 003 021 054
Bus Number
Cabin Pressure ( inches water)
Figure 5.2.1.1.1 Survey of bus cabin pressures using the “ blower door” method. Note the first
two digits of the bus number correspond to the model year. Black bars represent
buses used in the main study.
37
Cabin pressures for these buses ranged from 0.04 to 0.25 inches of water. There appeared to
be a trend of newer buses being “ tighter” than older buses, with buses built in the late 1990’ s and
early 2000’ s exhibiting the highest pressures ( or lowest leak rate). Several buses exhibited higher
than expected cabin pressures, based on their ages, including bus 856 and bus 923 ( the latter was one
of the four buses used in the main study). Replicate measurements of cabin pressure for buses 856,
872, and 927 were taken on separate days with differences between the two measurements ranging
from 4 to 10%.
On many buses, the front door carried the greatest potential for leaks as seals were weak
and/ or large gaps were visible around the door area, especially at the bottom of the door in a number
of cases ( see Photograph 3.2.2.1). When we covered the doors, the cabin pressure increased by
about 0.02 inches of water across most buses we tested. From these experiments, we drew two
principal conclusions. First, the observation in our earlier school bus study that older school buses in
general were less well constructed and/ or through age and wear had developed observable openings
in the cabin and around doors and windows, was generally confirmed by the pressure measurements
we made in the present study; a trend of decreasing leak rates was observed from 1985 to 2005
buses. Second, we believe the “ blower door” method developed to measure the pressure ( or leak
rate) in school buses could be employed by school bus maintenance staff to identify the relative leak
rate of buses. This information could be used, in part, to follow one of our previous
recommendations to place the “ cleanest” buses within a school district on the longest routes.
However, it should be noted that self- pollution also depends on other additional factors such as
emission rate.
5.2.2 Quantification
Self- pollution, as described in Behrentz et al. ( 2004), is the percentage of air in the bus cabin
that can be attributed to the bus’s own exhaust, or, in this study, the ratio between the concentration
of tracer gas in the bus cabin and the concentration of tracer gas in the exhaust. To calculate self-pollution,
first, the concentration of tracer gas in the exhaust ( Cexh) is determined from Equation
5.1 as discussed earlier:
Qexh Qcyl
Qcyl
Cexh Ccyl
+
= * ( 5.1)
Second, using direct measurements of SF6 inside the cabin we define percent self- pollution as:
Self- Pollution
Cexh
Ccabin = * 100 ( 5.2)
Using Equations 5.1 and 5.2, we were able to determine the degree of self- pollution across all
test buses. This metric was also used to ( a) assess the potential for exhaust intrusion due to leaks in
the bus cabin, and ( b) investigate the effectiveness of proposed mitigation measures as discussed in
detail below.
5.3 Exhaust Leaks
A systematic method for detecting exhaust train leaks was developed in the pilot study, but
this method was time consuming and was not recommended for use in the main study. Instead,
simple and rapid measures for exhaust system leak detection were developed. These involved both
qualitative measures ( e. g. noise of escaping gas in the exhaust system when a cork was placed in the
tailpipe, or the presence of visible carbon streaks on the outside of the bus near the engine
compartment) and semi- quantitative measures described below.
In general, the qualitative assessments we employed to identify exhaust system leaks failed to
38
reveal any substantial evidence of leaks in the exhaust trains of the 17 buses tested. This was
consistent with results from our earlier study ( Fitz et al., 2003) and our hypothesis that exhaust
system leaks were unlikely to be a prevalent contribution to self- pollution, especially relative to the
impacts of school bus tailpipe emissions.
The same 17 buses ( as in the cabin pressure measurements described in the previous section
and mentioned previously) were tested for semi- quantitative characterization of exhaust leaks using
the method described in Section 4.2.4.1. At the beginning of this study we thought backpressure
measurements might be a good indicator of exhaust leaks; a bus with low exhaust backpressure
might indicate a leaky exhaust train. Backpressures measured using our exhaust restrictor apparatus
ranged from 0.2- 4.0 psi, a range that could not be explained by exhaust system leaks alone. Out of
the 17 buses surveyed, three buses exhibited evidence for possible exhaust leaks. As seen in Figure
5.3.1 the measured backpressures varied by engine type/ manufacturer ( 1- 5 buses were tested per
engine type), with the John Deere CNG- powered buses exhibiting the lowest backpressure while the
Caterpillar diesel- powered buses showed the highest. We conclude backpressure measurements
were dominated by engine type, but it may be possible to use these backpressure measurements by
considering the value expected for each engine type.
0
0.5
1
1.5
2
2.5
3
3.5
4
John Deere
608 CNG
IHC Diesel
Turbo DT 466
Detroit 6 V 92 Cummins 300
8.3 HP
Cummins 250
HP 8.3
Cat 3208
Backpressure ( psi)
n= 3
n= 5
n= 5
n= 1 n= 1
n= 2
SD= 1.2
SD= 1.3
Figure 5.3.1 Backpressure measurements in psi for the six engine types tested in the exhaust leak
experiment ( n is the number of buses tested for each engine type). Standard
deviations are provided for the Cummins 250 and Cat 3208 engines.
39
5.4 Mitigation Measures
As discussed earlier, three major mitigation strategies for reducing self- pollution inside bus
cabins were tested: ( 1) high versus low exhaust position on both the test bus and a leader bus; ( 2)
the use of a blower to pressurize the inside of the test bus cabin ( i. e. power ventilation); and ( 3)
sealing the windows. The first two mitigation measures were tested individually and in combination.
The next several sections discuss the results obtained from testing these three mitigation strategies.
In these analyses, our most quantitative comparisons are based on data obtained from the last
five days ( starting 0427). Earlier runs encountered problems that prevented fully quantitative
comparison. One problem involved residual tracer gas concentrations in the cabin: the cabin was
not adequately flushed of tracer gas between runs up through 0407. The tests before that date can be
compared qualitatively, but overall self- pollution averages cannot be calculated. A second problem
was the blower exhaust port being inadvertently left uncovered until the last five days of the study.
While this may not have significantly affected results, with the blower exhaust port uncovered and
the blower turned off, the bus may have been able to allow outside air to enter the cabin and/ or
create a “ negative” pressure in the bus while it was moving. Therefore, we cannot rule out the
potential for bus exhaust to have entered the bus and to have potentially created the appearance of
higher self- pollution than would have been the case had the blower port been covered. The results
obtained prior to covering the blower still allow direct qualitative comparison of mitigation method
effectiveness, depending on the run type, but not quantitative comparison.
5.4.1 Effect of High Exhaust Position When Driven on the Test Route
Since self- pollution is a phenomenon that primarily occurs when windows are closed
( Behrentz et al., 2004; Sabin et al., 2005a), all self- pollution runs used to test the effects of high
versus low exhaust positions were conducted with windows closed.
In these experiments, SF6 and propene were released simultaneously from a split tailpipe
with one tracer released from the high exhaust position and one tracer released from the low exhaust
position for the duration of one bus loop around Route 2 described in Section 4.1.3.2. The two tracer
gas positions were then switched for the next loop around the test route with up to four consecutive
runs per day conducted in this manner.
5.4.1.1 Effect of High Exhaust Position on Self- Pollution When Bus in Motion
Figures 5.4.1.1.1a- d presents examples of several time series of in- cabin concentrations of
tracer gas during the final five days of testing excluding the first run of 0427 as the blower exhaust
port was not covered for that run. The data include seven runs over 3 buses ( Bus 872, 021 and 923)
and one test route ( Route 2). To eliminate confounders such as differences in meteorology and other
experimental conditions between runs, we compared the effect of high versus low exhaust ( SF6 high
and propene low or SF6 low and propene high) within a single run, taking advantage of our use of a
split tailpipe with dual tracer release.
Examining the first run on 0504, 0510, 0511 ( Figures 5.4.1.1.1b- d), and the second run on
0504, 0510, and 0511, and Run 45 ( 0427) ( Figures 5.4.1.1.1a- d), we found within each run, the high
exhaust position consistently resulted in lower self- pollution compared to the low exhaust position.
This observation is summarized by data in Table 5.4.1.1.1, which shows the percent self- pollution
for individual runs. For all runs but one, the high exhaust position resulted in 35- 95% decrease in
self- pollution compared to the low exhaust position. In Run 61, a 112% increase in self- pollution
was observed. Overall, however, the high exhaust position appears to be a promising approach to
40
reducing self- pollution in school buses. ( Note: During Run 52 and 53, butanol was detected in the
bus, originating from a broken lead on the CPC instrument. Butanol, having an ionization energy of
9.99 eV, was detected by our PIDs which employ a 10.6 eV lamp. Increased concentrations of
butanol in the cabin led to higher PID readings creating the appearance of propene tracer intrusion.
As a result, these runs were discarded from our propene analyses).
Propene data from Runs 62 and 63 from 0510 were also excluded from our analyses due to
diminishing supply of propene gas during these two runs.
0
0.005
0.01
0.015
0.02
0.025
0.03
15: 50: 24 16: 12: 00 16: 33: 36 16: 55: 12 17: 16: 48 17: 38: 24
Percent Self- Pollution
Bus 872
Runs 45
SF6 Low
Propene High
Blower Off
Run 45
Figure 5.4.1.1.1 ( a) Time series of percent self- pollution for SF6 and propene during mobile runs
conducted on 0427.
41
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
0.04
14: 52: 48 15: 14: 24 15: 36: 00 15: 57: 36 16: 19: 12 16: 40: 48 17: 02: 24 17: 24: 00
Percent Self- Pollution
SF6 Propene
SF6 Low
Propene
High
Blower Off
Run 56
SF6 High
Propene
Low
Blower Off
Run 57
Bus 021
Runs 56- 59
Figure 5.4.1.1.1 ( b) Time series of percent self- pollution for SF6 and propene during mobile runs
conducted on 0504.
0.00
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
15: 14 15: 21 15: 28 15: 36 15: 43 15: 50 15: 57 16: 04 16: 12 16: 19
Time
Percent Self- Pollution
SF6 Propene
Propene Low
SF6 High
B