SAN FRANCISCO MUNICIPAL RAILWAY FLEET ENGINEERING

ALTERNATIVE FUEL PILOT PROGRAM

INITIAL 6 MONTH EVALUATION RESULTS

for

2 CNG 40-FOOT BUSES

2 HYBRID/ELECTRIC 40-FOOT BUSES

2 CLEAN DIESEL W/ PM FILTER 40-FOOT BUSES

2 CLEAN DIESEL CONTROL 40-FOOT BUSES

Working Draft

May 2002

ABSTRACT

The San Francisco Municipal Railway (Muni) is committed to reducing air pollution in San Francisco.  One of the most significant ways that Muni helps to reduce pollution is to provide safe, reliable public transportation.  With more people on transit, fewer cars on the road means less pollution.  Muni also demonstrates its commitment to clean air by operating low-pollution vehicles.  Fifty-three percent of Muni's active revenue fleet is pollution-free, including 203 rail vehicles and 332 electric trolley buses that generate zero pollution.  Forty percent of Muni's buses are zero-emissions buses.  Almost all of the electricity on which those vehicles run is generated by hydropower, a clean source of energy.  Most of the diesel buses in Muni's fleet are new, and thus generate lower emissions.  By December 31, 2002, the remainder of the diesel buses will be rebuilt with kits that reduce soot emissions.  The diesel buses which have not been rebuilt will be retired from the active fleet.  Muni operates 18 vehicles that are powered by compressed natural gas (CNG) in its non-revenue fleet.  Muni is continuously seeking new ways to reduce pollution.

Some of Muni's emissions-reduction programs were borne out of an agreement between the San Francisco County Transportation Authority (SFCTA) and Muni.  In March 2001, the SFCTA attached a set of 11 conditions to TA Resolution 01-08, which authorized Muni's purchase of 95 diesel motor coaches.  Those conditions are included in the Muni Bus Purchase Proposal, commonly referred to as the "11-Point Agreement."  A copy of this agreement is included in Appendix A.  The highlight of the 11-point agreement is this 6-month review of Muni's Alternative Fuels Pilot Program (AFPP).

The AFPP is designed to allow Muni to identify the most appropriate low emissions bus technology to work within the constraints of San Francisco's unique terrain and duty cycle performance demands.  Additionally, the program will allow Muni to identify associated facility support requirements.

Data will be collected for 24 months on eight (8) buses representing four (4) categories of transit bus technology.  This data will be examined to determine the impact of the various characteristics of each technology to Muni, such as performance, emissions, reliability factors, cost per mile, capital costs for the buses and facilities improvements, any safety concerns, and operator and passenger feedback.  Results of similar studies may also be integrated in Muni's analysis and recommendations.

This 6-month preliminary report and update will feature an overview and summary of the pilot project, with summary discussions and graphical data representation.  Additionally, each data collection category will contain an evaluation procedure documenting the technical details related to the collection and analysis of the data gathered during the AFPP.  Once the study is finished, Muni can then determine the viability and impact of each technology.  These results will impact future bus procurements. The AFPP is designed to ensure that Muni will become experienced at adapting to and operating with the newest and least polluting bus technologies while posing the least risk to Muni's ability to meet service standards required by Proposition E.

TABLE OF CONTENTS

EXECUTIVE SUMMARY

INTRODUCTION

EMISSIONS

TAILPIPE EMISSIONS

EXHAUST ODOR

EXHAUST OPACITY

VEHICLE NOISE LEVELS

PERFORMANCE

ACCELERATION AND TOP SPEED

OPERATION TIMES

FUEL ECONOMY

RANGE

OPERATIONS

OPERATOR FEEDBACK

PASSENGER FEEDBACK

MAINTENANCE

RELIABILITY

20-HOUR PERFORMANCE

MAINTENANCE FEEDBACK

COST

OPERATING COST

CAPITAL COST

FACILITY COMPLIANCE

VEHICLE SAFETY CONCERNS

APPENDIX A

APPENDIX B

APPENDIX C

INDEX OF TABLES

TABLE 1:  AFPP VEHICLE SPECIFICATIONS

TABLE 2:  Exhaust Opacity Levels

TABLE 3:  Exterior Noise Level Summary (Graph)

TABLE 4:  Exterior Noise Level Overview (Graph)

TABLE 5:  Interior Noise Level Summary (Graph)

TABLE 6:  Interior Noise Level Overview (Graph)

TABLE 7:  Top Speed Summary (Graph)

TABLE 8:  Relative Top Speed (Graph)

TABLE 9:  Acceleration Times Summary (Graph)

TABLE 10: Relative Acceleration (Graph)

TABLE 11: Average Time Between Each Stop (Graph)

TABLE 12: Average Total Trip Time Between Stops (Graph)

TABLE 13: Fuel Economy

TABLE 14: Vehicle Range

TABLE 15: Driving Characteristics

TABLE 16: Reliability Rates

TABLE 17: Maintenance Survey - Relative Reliability Rates

TABLE 18: Fuel Cost Per Mile

TABLE 19: Maintenance Cost Per Mile

TABLE 20: Total Cost Per Mile

TABLE 21: New Bus Cost Summary

TABLE 22: Infrastructure Requirements Comparison

TABLE 23: Incremental Facility Costs Comparison

EXECUTIVE SUMMARY

The Alternative Fuel Pilot Program (AFPP) has finished evaluating Compressed Natural Gas (CNG), Hybrid, and exhaust after-treatment (PM filter) vehicle technologies during the initial six months in revenue service at Muni.  This program was designed to provide recommendations regarding future Muni motor coach procurements.  San Francisco's unique and challenging transit environment, featuring the City's famous hills, commitment to "transit first" principles, and potential for emergency service requirements in the event of another earthquake, requires a site-specific evaluation of alternative transit technologies.  The test buses were compared to unmodified conventional diesel buses in 19 different categories.  These categories are grouped into the general areas of:  EMISSIONS; PERFORMANCE; OPERATIONS; MAINTENANCE; COST; and SAFETY CONCERNS.

The University of California at Davis - Institute of Transportation Studies (ITS) was contracted by Muni to compare the emissions and energy efficiency of the different AFPP technologies.  This comparison was primarily made with a chassis dynamometer^[1] using an ITS-designed San Francisco simulation which, for the first time, tested heavy duty vehicle emissions on hills.  Results of this testing are found in the UC Davis-ITS Preliminary Chassis Dynamometer Test Results for Four Types of Advanced Technology Buses presentation, attached at Appendix C.  The detailed UC Davis-ITS Preliminary Chassis Dynamometer Test Results for Four Types of Advanced Technology Buses Report will be available from UC Davis through Muni fleet engineering on June 1, 2002.

ITS and Muni evaluated CNG, hybrid, and exhaust particulate (PM) filter vehicle technologies during their initial six months in revenue service.  San Francisco's unique and challenging transit environment, including the City's famous hills, and commitment to "transit-first" principles required an on-sight evaluation as well as "laboratory" testing on the chassis dynamometer. 

The test buses were compared to unmodified conventional diesel buses in 19 different categories.  These categories are further grouped below into the general areas of:  emissions; performance; operations; maintenance; cost; and safety concerns.

Emissions

 Tailpipe emissions

Researchers at the Institute of Transportation Studies, University of California Davis analyzed the Phase One emissions testing which took place over a period of 6 months.  Eighteen months of testing will continue until the full program concludes, in July 2003.  ITS-Davis staff explained that testing of advanced technology buses is a complex and evolving science.  The tests conducted for Muni thus far were intended to 1) confirm that results were consistent with those reported by other advanced bus technology testing programs and 2) facilitate development of new tests suited to San Francisco's unique terrain and driving conditions. With these caveats, UC Davis offered the following preliminary results of based on the standard emissions tests:

a) the magnitude of emissions were generally consistent with those found in other studies of advanced technology buses operating, 

b)   as is typical of emissions testing, there was a large range of emissions values for each bus,

c) the hybrid, CNG, and diesel buses with traps offered greatly reduced PM emissions compared to conventional diesel buses,

d)   PM emission of CNG, hybrid, and diesel bus with trap were similar and sometimes at the detectable limits of the analyzers, 

e) the CNG buses demonstrated the lowest NOx emissions followed by the hybrid and diesel buses,

f)   CNG buses had the highest CO emissions of any bus,

g) the diesel buses exhibited the best fuel economy and the CNG bus the poorest.  Unlike other testing, the hybrid bus had poorer fuel economy than the diesel.  This can be attributed to the diesel's unexpectedly high fuel economy.

However, these results may not be indicative of the in-use emissions in San Francisco.  San Francisco's unique grade and operations characteristics (such as average speed and idling duration) are not captured in the traditional test cycles discussed above.  ITS-Davis created a preliminary SF specific test cycle using bus speed and road grade information collected on three diverse Muni routes.  Limited testing on the SF Cycle indicates that the fuel economy differences between buses may be less in SF than predicted by the conventional cycles. 

There were several test protocol difficulties that need to be addressed with California Truck Testing Services, the contracted emissions testing facility, prior to the next round of testing. Diesel bus emissions and fuel economy in particular were inconsistent, and there were difficulties with hybrid bus testing.  The testing that was conducted in the first six months of Phase One is insufficient for staff to recommend one type of alternative fuel over another.  The staff of ITS-Davis agrees that more testing is required over the remaining 18 months of the pilot program, in order to achieve more meaningful results. ITS-Davis recommends that they actively participate in the data collection and emissions testing in this next phase.  Comprehensive bus operations and topography data collection could commence in July 2002, and the next round of emissions testing would occur in October 2002.

Visual quality of exhaust

Vehicle exhaust odor was not noticeable by passengers on any of the AFPP buses.  Exhaust opacity, or visual quality, was outstanding on all of the technologies.

Vehicle noise levels

Exterior noise was significantly lower on the Hybrids.  In general, interior noise was lowest on the Hybrids and CNG buses.  PM filters lowered the interior and exterior noise levels of the conventional diesels.

Performance

Performance indicators included top speed and acceleration on different grades, average time between stops, fuel economy, and fuel range.  Grades of up to 21% were used in these performance evaluations.^[2]  The following performance indicators were observed:

1) Hybrid buses accelerated faster than any of the other technologies, regardless of grade.  In general, the PM filter buses have the best top speed performance on all grades, followed by the CNG buses.  Preliminary results indicate that PM filters did not degrade the performance of conventional diesel buses. 

2) In challenging operating conditions hybrids gained time on each line trip, while the CNG buses lost time compared to the diesel buses.^[3] 

3) The PM filter buses have the best fuel economy and range.  The CNG buses are the least fuel efficient, and hybrid buses have the shortest fuel range.^[4],^[5]

Operations

Both operators and passengers were surveyed informally for their opinions regarding the alternative technologies.

1) Operators of the AFPP buses preferred the PM filter buses.  However, hybrids were said to be smoother while performing just as well.  Both CNG and hybrid buses were said to have better braking.  In general, operators did not feel that CNG buses were well-suited for hills or high passenger loads.

2) Riders want dependable service first and foremost, but would prefer to ride clean-fuel buses whenever possible.  Riders are generally concerned with all kinds of emissions, but don't view Muni as a major source of pollution.  Most riders feel that a different fuel source probably means a cleaner burning engine.^[6]  

Maintenance

Muni moves the equivalent of San Francisco's entire population every day.  Vehicle reliability is crucial. 

1) During their initial 6 months in operation, the PM filter buses were four (4) times more reliable than CNG buses, and 10 times more reliable than the hybrids.^[7]  The PM filter buses accumulated more than three (3) times the mileage during their first 6 months compared to the CNG and hybrid buses, partly due to the relatively low reliability rates found so far in the alternative technologies.  Six months is not a sufficient time period in which to test reliability.

2) Mechanics working on the AFPP felt that many maintenance issues could be resolved as their knowledge of the new technology improved. 

Cost

Ongoing operational costs and onetime capital costs were analyzed.  Operational costs include fuel^[8] and technology-specific maintenance.^[9]  For the purposes of this report both are calculated in terms of cost per mile.  Capital costs include the incremental cost of the vehicles and necessary fueling and facility modifications.

1) Fuel only cost per mile is highest for the CNG buses, and lowest for the PM filter buses.  Technology-specific maintenance costs were also highest for the CNG buses and significantly lowest for the PM filter buses.  Combining fuel and maintenance costs produce the same results.

2) Muni conventional diesel buses with PM filters cost approximately $330,000.^[10]  Hybrids currently cost about $120-170,000 more per bus than a Muni diesel bus with PM trap.  CNG buses cost approximately $45-60,000 more per bus than Muni diesel buses with PM traps.  In order to meet CARB emissions standards in 2004, diesel buses must be equipped with aftertreatment technology that will reduce NOx.  The aftertreatment technology that Muni will be testing (EGR) costs approximately $10,000 for each bus.^[11]  The cost of CNG buses has dropped as more buses are being manufactured.  An order of 15 hybrid buses would cost approximately $450-500,000 at this time.  However, the price is expected to decline as the technology gains popularity.^[12]  Currently, there are 34 hybrid buses in operation in the United States.  In the state of California, eight (8) are in operation, including the two (2) that Muni owns.  In January 2002, Muni requested a federal earmark of $1.5 million for the purchase of alternative fuel buses in the FY03 Congressional appropriations process.  Two other sources of funding are the Program Manager and Regional Funds from the Transportation Fund for Clean Air.  Any public agency can apply for funds from these sources.  The Program Manager fund contains approximately $950,000.  Muni received $500,000 from this fund in FY02 for CNG facilities.  Regional funds total approximately $8 million, but individual annual grants cannot exceed $1 million.  Muni will be submitting an application for funding from the Regional account in May/June 2002.^[13] 

3) None of Muni's current facilities are safety compliant with CNG buses.  Fueling and maintenance facilities must be specifically tailored for CNG.  To support a fleet larger than 15 CNG buses, Muni would need to spend roughly $7 million for capital costs associated with one facility.  For additional capacity, the cost to Muni would be an additional $5 million, bringing the total cost to $12 million for capital costs associated with modifications to two facilities.  It is noted, however, that one facility could provide enough capacity for a large purchase of CNG buses.  Capital costs associated with facility modifications for a large purchase of hybrid buses would be about $1.05 million. 

Safety

The safety category is primarily concerned with the hazards of CNG vehicle operation on bridges, through tunnels, under electric trolley contact wires, and inside terminals.  Secondary concerns include threats from intentional harm, and earthquake conditions.

Hazards

In general, the hazard potential from CNG buses during operation is extremely low.  However, it is strongly recommended that certain areas be evaluated further.  Other areas include streets with low trolley wires, where the possibility of an explosion exists in which an errant wire may strike a CNG tank on the top of a passing bus.^[14]  Hazards in long tunnels such as Broadway and Stockton present other areas for further evaluation, and should be avoided in the meantime.  A comprehensive safety evaluation and certification must be in place before the procurement of CNG buses or modification of facilities. 

Terrorism

Muni buses operate around City Hall, the Federal Building, the United Nations plaza, the TransAmerica Building, and on the Golden Gate Bridge, which may be considered terrorist targets for various reasons.  Given the volatility of CNG relative to diesel fuel, CNG buses could be considered a target for terrorism.^[15] 

Earthquakes

The US Geological Survey states that there is a 67% probability that San Francisco will one day experience another Richter-magnitude 7 earthquake.  During the 1989 Loma Prieta earthquake, electricity and natural gas supplies were unreliable in San Francisco.  Fifty-three percent of Muni's entire fleet is currently electric powered.  In the event of an emergency, with gas and electric supplies cut or reduced, a reliable fuel source will be crucial.  The Independent Oversight Committee (IOC) has suggested that the use of liquefied natural gas (LNG) would obviate the risk of service interruption in the event of an earthquake.  San Francisco's solid waste hauler, NorCal Waste Systems is in the process of converting its fleet to LNG.

Summary

The preliminary results of the AFPP indicate that Muni should continue its testing program.  The past six months of testing has yielded preliminary, and not reliable, emissions results.^[16]  Muni may also consider the testing of new technologies, such as LNG, the use of oxidation catalysts, and other aftertreatments in Phase Two of testing.

Muni should also consider the impact of regulations on its future procurements.  California is mandating new emissions standards in 2004-2006.  Even more stringent regulations will be imposed in 2007 by the federal Environmental Protection Agency.  Natural gas buses do meet the 2004 CARB mandate.  Currently, neither conventional nor hybrid diesel buses meet the 2004 mandate although aftertreatment technology and new hybrid certification standards may enable diesel buses to meet the new standards.  Diesel, hybrid, and CNG currently do not meet 2007 standards.

INTRODUCTION

BACKGROUND

The mission of the Alternative Fuel Pilot Program (AFPP) is to objectively evaluate alternative fuel technologies so that Muni can make an informed decision when considering propulsion technologies for future vehicle procurements.

This program is designed to determine what additional considerations must be addressed as well as considering which alternative fuel technology is most appropriate for use in San Francisco's unique transit environment.  The AFPP will last for a total of 24 months.  An Independent Oversight Committee, headed by personnel from the City and County of San Francisco Clean Cities Program, acts in the interest of the many citizen groups and environmental organizations that worked to develop the fundamentals of the AFPP.  Preliminary and final recommendations developed by Muni, as well as contingency plans if testing or project scheduling falls behind, will be addressed by this oversight committee.  Should results from similar studies (performed by other transit agencies) be incorporated into Muni's analysis, these studies will be current and evaluated for appropriateness by the Independent Oversight Committee (IOC) prior to inclusion.^[17]

PROGRAM MISSION

The goal of this program is to objectively evaluate the performance, reliability, emissions, safety, operating and capital costs of compressed natural gas (CNG) bus, hybrid bus, and exhaust particulate matter (PM) filter bus technologies.  These technologies will be evaluated for 24 months, using conventional diesel buses as a control/base measurement.

PROGRAM OBJECTIVES

The program is set up using the following guidelines:

  Evaluations of new, alternative technology will provide San Francisco specific information so as to lay the foundation for future vehicle procurements.^[18]

  Alternative technologies being evaluated include: two (2) CNG buses, two (2) hybrid buses, two (2) exhaust PM filter technology clean-diesel buses, and two (2) conventional clean-diesel control group buses.

  Evaluations of supplemental emissions technology will provide information on the impact of adding California Air Resources Board (CARB) mandated PM filters to clean-diesel buses in the fleet so that they too will exhibit dramatically reduced emissions.  Note that Muni is going ahead with this retrofit on the entire clean-diesel fleet four (4) years ahead of CARB's schedule.

  Fleet engineering will produce periodic and quarterly reports, as well as project summaries/recommendations after six (6) months, 12 months, and after 24 months of testing.

APPROACH

The eight (8) AFPP buses are stored and operated out of Muni's Kirkland and Woods divisions, and maintained at Muni's Marin Street division.  The diesel buses are fueled at Kirkland and Woods, while the CNG buses are fueled at an off-site CNG station.^[19]  Three (3) maintenance personnel and up to 50 operators have been specifically trained on the AFPP buses, as well as in the record keeping techniques required for the testing.

DATA COLLECTION

  Tailpipe emission data

  Exhaust odor

  Exhaust opacity

  Interior noise levels

  Exterior noise levels

  Acceleration performance on grades

  Top speed performance on grades

  20-Hour performance

  Total operational trip time and average time between stops

  Fuel economy

  Fuel range

  Operator feedback

  Passenger feedback

  Maintenance feedback

  Reliability rate

  Capital costs

  Operating costs

  Facility compliance

  Safety concerns

TABLE 1

AFPP VEHICLE SPECIFICATIONS

EMISSIONS

  Tailpipe Emissions

  Exhaust Odor

  Exhaust Opacity

  Vehicle Noise Levels

TAILPIPE EMISSIONS

Category leader:  San Francisco specific results were inconclusive.

SUMMARY

The University of California at Davis - Institute of Transportation Studies (ITS) was contracted by Muni to compare the emissions and energy efficiency of the different AFPP technologies.

ITS-Davis staff and Muni evaluated CNG, hybrid, and PM filter vehicle technologies during their initial six months in revenue service. This comparison was primarily made with a chassis dynamometer^[21] using an ITS-designed San Francisco simulation which, for the first time ever, tested heavy duty vehicle emissions on hills.

Researchers at the ITS-Davis oversaw Phase One emissions testing over a period of 6 months.  Eighteen months of testing will continue until the full program concludes, in July 2003.  ITS-Davis staff explained that testing of advanced technology buses is a complex and evolving science.  The tests conducted for Muni thus far were intended to 1) confirm that results were consistent with those reported by other advanced bus technology testing programs and 2) facilitate development of new tests suited to San Francisco's unique terrain and driving conditions. With these caveats, UC Davis offered the following preliminary results based on the standard emissions tests:

a) the magnitude of emissions were consistent with those found in studies of advanced technology buses operating,

b)   as is typical of emissions testing, there was a large range of emissions values for each bus,

c) the hybrid, CNG, and diesel buses with traps offered greatly reduced PM emissions compared to conventional diesel buses,^[22]

d)   PM emission of CNG, hybrid, and diesel bus with trap were similar and sometimes at the detectable limits of the analyzers,

e) the CNG buses demonstrated the lowest NOx emissions followed by the hybrid and diesel buses,^[23]

f)   CNG buses had the highest CO emissions of any bus,^[24]

g) the diesel buses exhibited the best fuel economy and the CNG bus the poorest.  Unlike other testing, the hybrid bus had poorer fuel economy than the diesel.  This can be attributed to the diesel's unexpectedly high fuel economy.

However, these results may not be indicative of the in-use emissions in San Francisco.  San Francisco's unique grade and operations characteristics (such as average speed and idling duration) are not captured in the traditional test cycles discussed above.  ITS-Davis created a preliminary SF specific test cycle using bus speed and road grade information collected on three diverse Muni routes.  Limited testing on the SF Cycle indicates that the differences between buses may be less in SF than predicted by the conventional cycles.

Other lessons learned:  There were several test protocol difficulties that need to be addressed with California Truck Testing Services (CaTTS), the contracted emissions testing facility, prior to the next round of testing.  The testing that was conducted in the first six months of Phase One is insufficient for staff to recommend one type of alternative fuel over another.  The staff of ITS-Davis agrees that more testing is required over the remaining 18 months of the pilot program, in order to achieve more meaningful results.  ITS-Davis recommends that they actively participate in the data collection and emissions testing in this next phase.  Comprehensive bus operations and topography data collection could commence in July 2002, and the next round of emissions testing would occur in October 2002.

Next steps:  Comprehensive bus operations and topography data collection could commence in July 2002, and the next round of emissions testing would occur in October 2002.  If possible, on-road emissions testing should be used to collect data in addition to chassis dynamometer data.

RESULTS

Results of this testing are found in the UC Davis-ITS Preliminary Chassis Dynamometer Test Results for Four Types of Advanced Technology Buses presentation, attached at Appendix C.  The detailed UC Davis-ITS Preliminary Chassis Dynamometer Test Results for Four Types of Advanced Technology Buses Report will be available from UC Davis through Muni fleet engineering on June 1, 2002.

The objective of ITS-Davis was to measure and compare the mass emissions (carbon monoxide, oxides of nitrogen, hydrocarbons, and particulate matter) and energy efficiency of AFPP buses.

Specific tasks handled by UC Davis included:

  Managing chassis dynamometer-based testing.

  Oversight of emission testing using conventional driving cycle test protocols: the Central Business District (CBD) and New York Bus (NY Bus) Cycles, which have been used extensively in previous research.^[25]

  Comparison of the emission test results to those from other large advanced technology bus testing projects.

  To design a San Francisco specific test, conduct preliminary testing on this cycle, and compare the results of this cycle to the results from other cycles.

Phase 1 Uncertainties:

  Typical variation between tests

  Representativeness of test procedure

  Passenger load

  Correction for State of Charge^[26]

  SF driving cycle inconsistency

  Sulfur interference with trap

  Testing complexity

EXHAUST ODOR

Category leader:  Inconclusive due to confusing survey questionnaire.

SUMMARY

Passengers were surveyed for their opinions regarding each technology's exhaust odor as part of the passenger survey.  This is a quality of life issue similar to exhaust opacity.  Results from this survey question are inconclusive.  In general, none of the technologies seems to have a noticeable exhaust odor.

Next steps:  The exhaust odor question should be refined further, and the survey re-administered to a larger sample of passengers. 

RESULTS

Inconclusive.  According to the preliminary survey designer and administrator the survey language must be revised:  "People who think I'm asking 'can you tell right now' say NO. People who think I'm asking 'could you tell if you put your nose up to (the exhaust)' say YES. The question needs to get better so we can see if we are measuring actual perception, or mental perception."^[27]

EVALUATION CRITERIA

1. EXHAUST ODOR

1.1.   Evaluation:  Data is collected for each alternative bus technology and a qualitative comparison is made using the Muni's conventional clean-diesel bus fleet as the standard.  Passenger feedback surveys are primarily used to evaluate exhaust odor.

1.2.   Test Procedure:  The surveys are verbally given to passengers while the test bus is in revenue service.  Surveys shall be given to a large number of passengers in order to better represent public opinion.

1.3.   Significant Variables:

1.3.1.   Language:  The survey is given in English only, which may prevent many Muni riders from providing feedback.

1.3.2.   Runs:  The survey is only given on a specific runs only; the survey is only given in certain parts of the city at certain times of the day.

1.4.   Test Personnel:  Test personnel who are clearly identified as being employed by Muni administer the surveys.

EXHAUST OPACITY

Category leader:  All.

SUMMARY

Exhaust opacity is a method of evaluating the visual appeal of a vehicle's exhaust.  This is a quality of life indicator similar to exhaust odor.  Although quantified opacity results differ slightly between each technology pair, the difference between the highest and lowest measured exhaust opacity levels on the test buses is insignificant.  None of the test buses produce visible exhaust smoke under normal operating conditions and the hybrid's exhaust opacity could not be measured.

Other lessons learned:  As anticipated, adding a PM filter to a conventional diesel engine lowered the exhaust opacity.  However, due to Muni's recent conversion to ultra low sulfur diesel (ULSD) fuel,^[28] the PM filters in the test program may not be performing to maximum potential with the new fuel.  This because the relatively high sulfur content of the previous fuel tends to partially plug the filter for a period of time.  Once properly cleaned and combined with dedicated use of ULSD, greater opacity reductions are predicted for the diesel powered buses with PM filters (conventional diesel and hybrid).

Next steps:  Since the exhaust gas tester could not measure hybrid exhaust opacity, data should be recollected in order to verify actual levels.  Exhaust particulate filters should all be properly serviced and cleaned by the manufacturers prior to retesting.^[29]

RESULTS

Relatively high exhaust opacity levels are an indication of darker exhaust smoke.  The control group conventional diesel buses have the highest exhaust opacity.  The hybrid bus exhaust opacity is apparently so low that it could not be measured with Muni's opacity measurement equipment.^[30]  The CNG buses and PM filter equipped diesel buses fall in between the control group diesels and the hybrids.  Overall average levels range from 2.6% to less than 1%^[31]

TABLE 2

EXHAUST OPACITY LEVELS:

EVALUATION CRITERIA

2. TAILPIPE EXHAUST OPACITY

2.1.   Evaluation:  Data is collected for each bus technology pair and a relative comparison is made using the unmodified diesel bus pair as the standard.  The average percent opacity for each technology type determines the final result for that pair.

2.2.   Definitions:  The visual quality of a vehicle's tailpipe exhaust is measured by a smokemeter in units of percent visual opacity.

2.2.1.   Zero percent (0%) opacity defines clear.  100% opacity defines black.

2.3.   Test Procedure:

2.3.1.   Prior to Testing:  Warm up and calibrate opacimeter.

2.3.2.   Measurement:  Snap accelerate the engine for three (3) seconds, maintain engine speed for the following two (2) seconds, return the engine to idle speed.  Repeat five (5) times.

2.3.3.   Final result:  The average of five (5) opacity reading determines the final result for that vehicle.

2.4.   Vehicle Conditions:

2.4.1.   Operating Temperature:  Test engine should be allowed to reach operating temperature prior to testing.  For purposes of this test, operating temperature has been reached if engine oil temperature is 140 F or above.

2.4.2.   Engine Inspection:  There are no exhaust leaks or misadjusted engine settings.

2.5.   Data Collection Equipment:  An exhaust gas tester is used to analyze the content of black, white and blue exhaust smoke from the test engines.

VEHICLE NOISE LEVELS

Category leader:  Hybrid.

SUMMARY

There are two sub-categories for vehicle noise in this evaluation:  Exterior noise and interior noise.  An average of the two highest noise levels measured in each category are represented in these results.

Exterior noise levels are lowest on the hybrids, and highest on the CNG buses.  Interior noise levels are lowest on the CNG and hybrid buses, and highest on the conventional diesel buses. 

Some of the results are misleading.  For example, when the CNG buses are in service, the exterior noise levels are not offensive.  Similarly, the measurably low interior noise levels found on the hybrid buses are misleading.  Qualitatively, the interior noise levels for passengers in the very rear of the hybrids are unacceptably loud.

The exterior noise levels of the CNG buses, and the interior noise levels of the hybrids can be significantly reduced in future purchases through cooling system and noise insulation modifications.  However, it is unlikely that any of the other technologies will be able to match the low exterior noise levels produced by the hybrid buses due to the nature of the technology and layout differences between hybrid technology and conventional bus technology.^[32]

Other lessons learned:  Adding PM filters to the conventional diesel buses appears to decrease both interior and exterior noise levels.

Next steps:  Noise level data should be collected while the vehicles are in service in order to supplement data taken under controlled, extreme test conditions.

RESULTS

TABLE 3

TABLE 4

TABLE 5

TABLE 6

Exterior noise levels for each technology differed significantly.  In general, the hybrid buses produce the least amount of noise, and the CNG buses produce the most exterior noise.  The conventional diesel buses fall in between the alternative technologies when it comes to exterior noise, and they are louder than the alternative technologies inside.  The exterior noise level results for the CNG buses are not as representative of actual operating conditions as the results for the other test buses.  In this case, test conditions require the engine's cooling fan to operate continuously, yet in service the CNG engine cooling fan rarely engages.

Interior noise is clearly split between conventional technology and alternative technology.  The CNG and Hybrid buses are considerably quieter for passengers and operator than the conventional diesel buses. 

Differences in vehicle noise, both interior and exterior, can be greatly influenced by vehicle specifications.  For example, interior noise was greatly raised by the ventilation system on the NABI diesel buses. 

EVALUATION CRITERIA

3. NOISE EMISSIONS

3.1.   Evaluation:

3.1.1.   Technology:  Data is collected for each bus technology pair and a relative comparison is made using the unmodified diesel bus pair as the standard.  The highest overall set of vehicle values for each technology type determines the final results.

3.1.2.   Vehicle:  For each vehicle, at least three (3) readings are taken for each specific measurement.  The two (2) highest readings for each measurement location or condition, within one (1) dBA of each other, are averaged to determine the final noise result for each measurement.  For example, if the curbside idle readings are 67.6 dBA, 67.8 dBA, and 68.0 dBA, then the final result will be 67.9 dBA.  Three (3) significant figures are used to express the final result of all subsets.

3.2.   Definitions:

3.2.1.   Set is defined here as all of the results for a single vehicle in each main noise category: interior and exterior.

3.2.2.   Subset is defined here as a singe category within the main categories of internal and external noise.  For example, noise level results for the driver's area is a subset of interior noise levels.

3.3.   Test Procedure:

3.3.1.   Exterior Noise:  Exterior noise levels are measured on both vehicle curbside and vehicle street side.^[33]  The three (3) categories measured are:  Idle; pull-away; and pass-by.  Society of Automotive Engineers (SAE) J366 FEB87 Exterior Sound Level for Trucks and Buses measurement procedure is followed.  In reference to J366 FEB87 section 4.3.3:  The sound level for each side of the vehicle shall be the average of the (2) two highest readings within (1) dBA of each other, rather than within (2) dB of each other.

3.3.2.   Interior Noise:  Interior noise levels are measured at four (4) points in the vehicle:  At the driver area, at the very front seats, at the rear doors, and at the very rear of the vehicle.  All readings are taken at seated head level.  All readings are taken in the middle of the passenger isle, except for the driver area reading.  The driver area reading is taken at head level directly above the operator's seat cushion.  The sound level meter microphone is pointed toward the front of the vehicle during all readings.  Readings are collected as the vehicles are accelerated from zero (0) to 30 mph.  The same city street is used for all measurements.

3.4.   Site Conditions:

3.4.1.   Ambient Noise:  All data is collected in dry weather and still wind conditions.

3.5.   Vehicle Conditions:

3.5.1.   Fans:  Engine cooling fans are configured to provide maximum cooling at all times during testing.  For interior noise testing, all interior fans and heaters are set to provide maximum output.

3.5.2.   Chassis:  All windows, hatches, and doors are closed.

3.5.3.   Passengers:  The vehicle is empty with the exception of test personnel (3.8) and equipment (3.7).

3.6.   Significant Variables:  Every effort is made to eliminate data that may be influenced by an intermittent ambient noise.  If such a noise occurs, it will void the measurement.  Examples of random ambient noise are:  Extremely loud street conditions, a plane flying overhead, or a car horn.

3.7.   Data Collection Equipment:  An SAE J366 approved digital sound level meter is used to determine vehicle noise levels.

3.7.1.   Settings:  Sound meter is set to the "A" scale, fast response.  Meter reading units are decibels (dB), with a resolution of one-tenth of a dB.  The meter's hold feature is used to determine maximum readings.^[34]  Every effort is made to disregard readings which may be the result of a sudden ambient noise condition (3.6).  This may cause the meter's hold feature to display the result of the ambient noise spike.

3.7.2.   Repeatability:  The same sound meter is used for all vehicles.  Sound meter is self-calibrating.  Sound meter calibration is verified with an external sound meter calibration device.

3.8.   Test Personnel:  Personnel consist of one (1) driver and one (1) data collector.  The data collector witnesses the test equipment readings and records test results.

3.8.1.   Driver:  The test driver is given simple, consistent verbal and visual commands.  Upon seeing the command to accelerate, the driver will press the throttle pedal immediately to the floor.  The throttle shall remain fully depressed throughout the acceleration run.  Every effort is made to eliminate performance variation caused by different driving styles

PERFORMANCE

  Acceleration and Top Speed

  Operation Times

  Fuel Economy

  Range

ACCELERATION AND TOP SPEED

Category leader:  Conventional diesel.

Alternative leader:  Hybrid.

SUMMARY

Test buses carrying the equivalent of 85 passengers were tested for top speed and acceleration on 0-21% grades.  The conventional diesel buses have relatively superior top speed performance regardless of grade.  As grade increases, the top speed performance of the hybrid buses improves, while the top speed performance of the CNG buses decreases.

The hybrid buses accelerate faster than any of the other technologies on every grade, with the exception of the 16% grade^[35] where they were nearly identical with the acceleration rate of the diesel buses equipped with PM filters.  CNG buses accelerate at a slower rate than the other test buses in every category.^[36]

Overall, the hybrid buses are ideally configured for hill climbing and acceleration.  The CNG buses are well suited for level ground, high-speed applications.  The conventional diesel buses perform well in all operating conditions, providing operating flexibility.

Other lessons learned:  Adding PM filters to conventional diesel buses seemed to generally improved both top speed and acceleration.

Next steps:  Future procurements should specify certain performance goals based on the results of this testing, since performance can be greatly influenced by vehicle specifications.

RESULTS

TABLE 7

TABLE 8

TABLE 9

TABLE 10

EVALUATION CRITERIA

4. PERFORMANCE

4.1.   Evaluation:  Data is collected for each bus technology pair in each top speed and acceleration subcategory and a relative comparison is made using the unmodified diesel bus pair as the standard.  The best complete set of results between each bus technology pair determines the final result for each technology type.  Grades ranging from 0% to 21% are used to determine performance on flat ground, medium hills, and steep hills.  Muni's maximum grade of 21% is found on the Hyde Street cable car line.^[37]  The second most extreme Muni line grade is found on the 1-California trolley coach line as it leaves Chinatown.^[38]

4.2.   Definitions:

4.2.1.   Set is defined here as all of the top speed and acceleration results for a single vehicle.

4.2.2.   Subset is defined here as a singe category within the main categories of top speed and acceleration.  For example, a performance subset is 0-40 mph acceleration on a 0% grade.

4.2.3.   Full passenger load is defined here as not being able to accept additional passengers onto the bus.  This assumes that all passengers are behind the yellow passenger line in front, and are not standing directly in the rear door area.

4.3.   Test Procedure:

4.3.1.   Top Speed:  Top speed runs on 0% and 2% grades begin with rolling starts.  Both grade locations cover sufficient distance to allow the vehicle to reach terminal speeds.  Top speed runs for all but 0% and 2% grades begin with the vehicle's front bumper at the base of each hill.  Vehicle speed is zero (0) at the start of all but the 0% and 2% grade runs.  For 0% grade, starting speed is 55 mph.^[39]  For the 2% grade, starting speed is 45 mph.^[40]

4.3.2.   Acceleration:  Acceleration runs begin with both axles completely on the test grade; the measured angle of the vehicle's passenger isle is identical to the angle of the grade.

4.3.3.   Driver Commands:  The test driver is given simple, consistent verbal commands.  Upon hearing the command to accelerate, the driver will press the throttle pedal immediately to the floor.  Data recording will not start until the throttle pedal hits the floor.  The throttle remains fully depressed throughout each acceleration run.  Every effort is made to eliminate performance variation caused by different driving styles.

4.3.4.   Determination of Speed:  When the predetermined terminal test speed is attained, either the vehicle speed is taken (during top speed determination), or the stopwatch is stopped (for acceleration times).

4.3.5.   Determination of Subset Results:  Three (3) readings are taken for each performance data subset.  The two (2) best performance readings in that subset are then averaged to determine the final result for that particular subset.  For example, if the top speed readings for a particular subset are 59 mph, 60 mph, and 62 mph, then the final result for that subset will be 61 mph.  Two (2) significant figures are used to express the final result of all subsets, primarily due to measurement resolution.

4.4.   Site Conditions:

4.4.1.   Data is collected in dry weather conditions.  Extreme weather temperatures and wind are avoided during testing.

4.4.2.   There is no interference from traffic or other variables that could effect the consistency of the data collected for all test vehicles.

4.5.   Vehicle Conditions:

4.5.1.   Starting Vehicle Weight:  Each vehicle's fuel supply and other fluids are completely full at the beginning of each performance test set.

4.5.2.   Test Vehicle Weight:  In addition to the starting vehicle weight (4.5.1), each vehicle carries a simulated full passenger load of 85 passengers, not including the driver, during each test set.  It is assumed that each passenger weighs 150 pounds, and therefore the total passenger load for testing equals 85*150=12,750 pounds.

4.5.3.   Determination of Full Passenger Load:  Full passenger loads were determined by auditing 40-foot bus passenger loads during peak commute hours on the 71-Haight motor coach line.  It was determined that 85 passengers represented an average full passenger load.  Passenger loads as high as 97 people were found on a 40-foot bus operating the 71-Haight line.  However, for this analysis 90+ passenger loads do not fit the theoretical definition of a full passenger load (4.2.3).  In general terms, an 85-person passenger load on a 40-foot Muni motor coach represents a bus with a passenger in every seat, a standee for each seated passenger, plus another five (5) passengers.

4.5.4.   Justification for Exceeding Gross Vehicle Weight (GVW):  GVW is the manufacturers' recommended maximum total vehicle weight including passengers.  This weight is exceeded for all AFPP buses when they are loaded with 12,750 pounds for testing.  Vehicle weights in excess of GVW are common for Muni's 40-foot buses when in revenue service, especially during commute hours and special events.  Of AFPP test vehicles: the conventional diesel bus can carry 78 passenger before reaching GVW; the hybrid bus can carry 73 passengers before reaching GVW; the CNG bus can carry 58 passengers before reaching GVW.^[41]  Rather than evaluate the AFPP buses using an unrealistic standard for San Francisco, it was determined that actual passenger loads be used (4.5.3).  In reality, if a bus has room for additional passengers, Muni patrons expect the bus operator to allow them to board.  In other words, the physical space inside of a bus is what actually dictates passenger loads when a bus is in service, not the weight that the bus is carrying.  Examples of passenger load conditions that potentially exceed even the 85 passengers per bus that defined AFPP test conditions: standard daily commute hours; special event and holiday operation;^[42] and emergency/disaster evacuation.

4.5.5.   Simulating a Passenger Load:  Full passenger loads are simulated by filling the test bus with plastic water containers equal to the weight of 85 passengers (4.5.2).  Both the weight of the water and the container were taken into account when calculating the total simulated passenger load.  The simulated passenger load is equally spaced throughout the interior of the vehicle, including the seats, in an identical manner for each vehicle.  The containers are then secured so that they remain immobile throughout the test.

4.6.   Significant Variables: 4.6.1.   Parasitic Engine Forces:  Vehicle cooling fans are monitored during all test sets in order to establish that parasitic engine forces remained consistent.  In all cases, normal fan operation shall be "off."  Data point outliers have occurred, although not noticeably on the hybrids, if the engine-cooling fan engaged during a particular subset.  If fan activation occurs, the vehicle is allowed to cool to below thermostatically controlled temperatures by idling the vehicle.  Only data collected with the cooling fan off is used in the test analysis.  It should be noted that even when not activated, cooling fans can spin due to mechanical drag or other related forces.  This relatively slow spin is considered to be "fan off" in terms of the elimination of parasitic engine forces.

4.6.2.   Wind:  For all 0% grade evaluations, each of the three (3) data points represents an average of two (2) runs, with the second run being made in the opposite direction of the first.  This provision is made in an attempt to limit the effect of wind force on the test results.

4.7.   Data Collection Equipment:

4.7.1.   Stopwatch:  During acceleration testing, a digital stopwatch is used to determine acceleration times.  The same digital stopwatch is used for all vehicles.  Stopwatch resolution is +/- 0.01 seconds.

4.7.2.   Speedometer:  Analog Speedometer Readings:  Data collector is to make every effort to eliminate parallax for all test readings.  The speedometer reading was verified prior to any testing.  This verification was done by correlating the analog speedometer reading with the onboard digital telemetry reading.  The predetermined speed mark on the speedometer is defined as the center of that mark lined up to the center of the gauge dial indicator.  Combined probability of speedometer error = +/- 1%.

4.7.3.   Protractor:  Test grades were found using a calibrated digital protractor.  The device units are degrees, so a conversion factor was used to determine grade.  Protractor resolution is +/- 0.1 degrees, or +/- 0.175% grade.

4.8.   Test Personnel:  Personnel consistes of one (1) driver and one (1) data collector.  The data collector operates the timing equipment, witnesses the required vehicle speeds, and recorded test results.

OPERATION TIMES

Category leader:  Hybrid.

SUMMARY

This category evaluated the in-service average time between bus stops for each technology type.  Since there is a chassis mix between low-floor buses and high-floor buses, stop dwell times differ significantly.  Therefore, total trip time for each technology is based upon the high-floor control group's average dwell time.  Overall, the hybrid buses had the shortest average time between bus stops, and gain roughly 6% of total trip time when compared to the control group.

CNG buses had the slowest overall line times.  This is likely due to the extreme grades found in certain sections of the 19-Polk line, which hampers the performance of the CNG buses.^[43]

Other lessons learned:  Low-floor buses, like the CNG and Hybrid buses in this test, were found to have much lower stop dwell times than high-floor buses.  In general, low-floor buses, regardless of the propulsion technology, can reduce trip times.

Next steps:  Data from additional routes is necessary in order to make a more accurate analysis of operational times.

RESULTS

TABLE 11

TABLE 12

EVALUATION CRITERIA

5. OPERATION TIMES

5.1.   Evaluation: Data is collected for each bus technology pair, with the exception of the unmodified diesel bus pair, and a relative comparison is made using the PM filer equipped diesel bus pair as the standard.  For purposes of this category, the PM filter and unmodified diesel bus pairs are considered together as the control group.

5.1.1.   Data Collection:  The alternative propulsion vehicles in the AFPP are low-floor buses, while the conventional diesel buses have high floor configurations.  Therefore, operational dwell times are not representative of the different propulsion technologies being evaluated.  Every effort is made to eliminate this unwanted variable by using the following evaluation compensations:

5.1.1.1.   Isolating Dwell Time:  In addition to total trip time, time between stop data is collected.  Therefore, the summation of times between stops, subtracted from the total trip time, equals the total dwell time.

5.1.1.2.   Corrected Total Trip Time:  The control group (high-floor) average dwell time for a particular run is multiplied by a particular test vehicle's number of stops, and added to the same test vehicle's total time between stops to arrive at a "high floor total trip time" for that test vehicle.

5.1.1.3.   Average Time Between Stops:  Average time between stops is considered directly comparable between technologies for any identical run, since there are roughly 200-300 stops per run to be averaged.  These times have nothing to do with the high or low floor configurations.

5.1.2.   Unique Conditions:  Any condition that proves unique to one vehicle during data collection of operation times is either removed from the data set or the particular run is thrown out.  An example of this would be a truck that completely blocks the road during one trip.  In this case, the data collector will note the estimated delay time to the second and subtract this from the recorded time for that segment and total trip.

5.1.3.   Passenger Counts:  Passenger loads are recorded for every designated stop the test bus makes.  These load counts will ensure that comparable passenger loads are used for operation time analysis.

5.2.   Site Conditions:

5.2.1.   Comparable Days:  Every effort is made to collect data on identical runs when the vehicle test conditions are most likely to be similar between vehicle types.  For example, Fridays or days when the there is a Farmer's Market on a particular test run are avoided due to unusually high vehicle and pedestrian traffic on those days.

5.2.2.   Line Compatibility:  Data cannot be collected on any line where the buses have difficulty navigating for any reason.  All AFPP vehicles have proven compatible with every line evaluated during the initial 6 months.

5.3.   Vehicle Conditions:

5.3.1.   Inspection:  A thorough inspection and evaluation is performed of the vehicle's systems prior to testing.  For example, the brakes must function properly.

5.3.2.   Operator Approval:  The operator must feel like the vehicle performs well for its technology type.

5.3.3.   Propulsion System:  There must not be any propulsion system maintenance codes present.  Fuel and all other fluids must be full.

5.4.   Data Collection Equipment:  A digital stopwatch is used to determine time between stops to one-tenth (0.10) of a second.  The stopwatch digital clock is used to determine total trip times to the second.

5.4.1.   The same digital stopwatch and digital clock is used for all vehicles.

5.4.2.   The stopwatch timer has resolution to one-hundredth (0.01)of a second.  Times are rounded to the nearest one-tenth (0.10) of a second.

5.5.   Test Personnel:  Personnel consist of one (1) driver, and two (2) data collectors.  The first data collector operates the timing equipment and records test results.  The second data collector records passenger count data at every designated stop the vehicle makes.

5.6.   Significant variables:  The same operator must drive for every technology type on a given run, in an attempt to reduce driver related variables.

FUEL ECONOMY

Category leader:  Conventional diesel.

Alternative leader:  Hybrid.

SUMMARY

Fuel economy for each technology was evaluated in terms of diesel gallons.  By converting CNG energy to its diesel equivalent, a direct comparison was possible between the different fuels evaluated in this report.

The conventional diesel buses are over 26% more fuel efficient than the hybrid buses.  The hybrid buses are 25% more fuel efficient than the CNG buses.  Overall, conventional diesel technology is over 58% more fuel efficient than the CNG buses.

The relatively low fuel economy result from the hybrids is somewhat surprising.  However, while numbers differ, the general on-road results pattern (conventional diesel leading, followed by hybrid followed by CNG) is consistent with that found during chassis dynamometer emissions testing.^[44]  It should be noted that unlike their passenger car counterparts, which are well known for their superior fuel economy, hybrid transit buses are configured and optimized very differently than these passenger car hybrids.  Furthermore, data collection was interrupted during the hybrid and unmodified conventional diesel evaluations, which potentially compromised the long term fuel economy data for the hybrids.  Therefore, the hybrid data reported here represents only short-term fuel economy for the hybrid buses, and could change during the remaining course of the program.

Other lessons learned:  Due to Muni's early switch from ~120 ppm sulfur CARB-2 diesel fuel to ~15 ppm sulfur ULSD^[45] fuel during the initial evaluation period, fuel economy reported here for diesel fueled buses with PM filters may not be optimized due to excess sulfur in the filter.  Once the PM filters have been properly serviced by the filter manufacturers, fuel economy for those buses is expected to increase.

Next steps:  Hybrid bus data should be recollected to better reflect long term fuel economy.  All PM filters should be factory serviced due to Muni's conversion to ULSD fuel, and data should be recollected on those vehicles.

RESULTS

TABLE 13

FUEL ECONOMY:

EVALUATION CRITERIA

6. FUEL ECONOMY

6.1.   Evaluation:  Data is collected for each bus technology pair, with the exception of the unmodified diesel bus pair, and a relative comparison is made using the PM filter equipped diesel bus pair as the standard.  Fuel records are kept for fuel added and odometer mileage at the time of each fueling.  Total mileage (mi) during measured segment divided by total fuel used during that segment (DGE) equals miles per diesel gallon (mi/DGE).

6.2.   Definitions:  Due to the dissimilar physical states of fuel being evaluated (gas and liquid), energy content is used to compare fuel economy.  Fuel economy results are represented in units of miles per diesel gallon equivalent (mi/DGE).

6.2.1.   DGE:  Diesel fuel energy content is measured in this study in units of diesel gallon equivalents (DGE).  One (1) DGE is defined as 139,000 BTU's.  Diesel fuel referenced here is ultra low sulfur diesel (ULSD).  ULSD is defined as having less than or equal to 15 parts per million (ppm) sulfur content.

6.2.2.   GGE:  Compressed natural gas (CNG) is measured in this study in units of gasoline gallon equivalents (GGE) and Therms.  One (1) GGE is defined here as 128,000 BTU's.  1.086 GGE = 1.000 DGE.  One (1) Therm = 100,000 BTU.  1.39 Therms = 1.00 DGE.

6.3.   Significant Variables:

6.3.1.   Hybrid Diesel-Electric Bus Correction Factor:  Every effort is made to treat the hybrid technology as "transparent" in terms of daily operation.  For example: The hybrids are simply started before going into service, allowed to idle while the operator makes seat/mirror adjustments and performs a pre-trip inspection, and they are then driven into service.  Likewise, at the end of a run, the buses will be fueled (if required) and parked.  Because of this transparent treatment, no provisions are made for battery state of charge (SOC) variations during normal operations.  In keeping with the theory of transparent operation, fuel economy numbers for the hybrid buses are "in-use" numbers, and not necessarily actual fuel economy.  In order to determine actual fuel economy for the hybrids, one would need to correct for battery state of charge.

6.3.2.   Loads and Operating Conditions:  Passenger loads and operating conditions varied during fuel economy data collection.  Load and operating variability can influence fuel economy data.

RANGE

Category leader:  Conventional diesel.

Alternative leader:  CNG.

SUMMARY

Range is the measure of how far a vehicle can travel per full fuel load.  Useful range is clearly less than this maximum range number, since running out of fuel is unacceptable when the bus is in-service.  The conventional diesel buses with PM filers can travel more than 20% farther than the CNG buses.  Hybrid range data collection was interrupted, so the range number reported here is estimated (based on 100 gallon fuel tank and 3.0 mpg fuel economy).  Based on this estimate, the CNG buses can likely travel up to 30% farther than the hybrid buses.^[48]  Overall, the conventional diesel buses with PM filers can travel 22% farther than the CNG buses.

Other lessons learned:  The CNG range data is based on the fuel system pressure dropping to minimum recommended operating levels.  It should be noted that the CNG buses could continue on after this point.^[49]  The hybrid range data is estimated based upon fuel economy data.  Furthermore, true hybrid range data must be calculated once actual range data is corrected with battery state of charge data.

Next steps:  Hybrid range data should be recollected.

RESULTS

TABLE 14

VEHICLE RANGE:

EVALUATION CRITERIA

7. RANGE

7.1.   Evaluation:  Data is collected for each bus technology pair, with the exception of the unmodified diesel bus pair, and a relative comparison is made using the PM filter equipped diesel bus pair as the standard.  Records are kept for odometer mileage at the time of full fueling.  The vehicle then operates at all times either in revenue service or otherwise on surface streets in San Francisco.  Freeway miles accumulated during range evaluation are avoided or are kept to a minimum.  Total mileage (mi) accumulated at the time of fuel starvation equals vehicle range.  Note that the usable range should generally be considered 25 miles less than the actual range.

7.2.   Definitions: 

7.2.1.   Vehicle Range:  Range is defined here as the distance (mi) traveled per full vehicle fuel load at the time the vehicle runs out of fuel.

7.2.2.   Out of Fuel:  Running out of fuel, is further defined as the vehicle's engine running out of fuel; fuel starvation.  Fuel starvation occurs when it is confirmed the engine cannot draw any fuel from the vehicle's gas supply.  This definition is needed for the following reasons:

7.2.2.1.   For CNG, full fueling is defined as 3600 psi on the fueling station pump gauge.

7.2.2.2.   Muni's CNG buses are considered out of fuel in this evaluation if the vehicle's supply fuel pressure reaches 200 psi.  At this point a driver warning alarm is sounded due to the possibility of fuel starvation caused by low primary fuel supply pressure.⁴⁴

7.2.2.3.   The diesel fuel systems in this evaluation may not be able to draw the entire fuel supply; after the point of fuel starvation there may still be residual fuel in the fuel tank.  This lowers the useable volume of the fuel tank to below specified tank capacity, and differentiates actual range from theoretical range.

7.3.   Significant Variables:  Passenger loads and operating conditions may vary during range evaluation.  Load and operating variation can have an influence on range data.

OPERATIONS

  Operator Feedback

  Passenger Feedback

OPERATOR FEEDBACK

Category leader:  Conventional diesel.

Alternative leader:  Hybrid.

SUMMARY

All operators of the test buses were asked to fill out driver surveys following each run (a run is a set of round trips on a specified line).  Seven (7) operators responded, representing roughly 17% of AFPP trained operators.^[51]  The majority of operators that responded prefer the conventional diesel buses that they are familiar with.  Of note was that none of the operators responded as preferring the hybrid buses, however, the majority of operators feel that the hybrid buses perform roughly the same or better than conventional diesel buses.  Hill climbing performance and acceleration from a stop were emphasized as the shortcomings of the CNG buses.  Noise level and braking effort were preferred on both the CNG and hybrid buses, and the overall smoothness of the hybrid buses when accelerating and stopping impressed the operators.^[52] 

Other lessons learned:  Greater run variety is needed in order to obtain more feedback.  Of the standard motor coach fleet, operators prefer the 1999 NABI diesels.

Next steps:  The buses should be evaluated by a larger number of operators, from a larger number of different lines. 

RESULTS

TABLE 15

DRIVING CHARACTERISTICS - COMPARED TO CONVENTIONAL DIESEL BUSES:

The following questions and answers are taken from the CNG section of the written surveys.  All responses are listed:

Q - Any problems related to CNG engine?

A - No CNG specific problems were reported by operators.

Q - Should Muni buy CNG buses?

A - "No."

A - "Not to be recommended."

A - "No."

A - "Not if hybrids available."

General CNG comments from written surveys:^[53]

" will not pull hills." 

"Bus are very low, slow on the take off, bouncing " 

"Please note for the record that this equipment is wholly inadequate.  It is too slow and underpowered for the express line that I had to work today."

"It's a good bus for tour traps.  But not for city lines."  "Unsatisfactory" 

"My assessment is it is a excellent coach.  But not for San Francisco bus lines.  We have lots of hills, which hampers this coach.  Slow acceleration on hills and in traffic." 

"New equipment always seems better because it is new." 

"I don't feel the CNG bus would be good for Muni.  It was very slow, underpowered and heavy to steer.  It might be good for commuter freeway lines but it would add 25% to our running times." 

"Fair." 

"When I drove the bus it had 5,000 miles and felt like it had 200,000 miles.  I noticed that it was losing speed and power from when I tested it and the bus only had 1,000 miles."

The following questions and answers are taken from the hybrid section of the written surveys.  All responses are listed:

Q - Any problems related to the hybrid propulsion system?

A - The only hybrid propulsion system problem that was reported was that the interior noise was unacceptably high.^[54]

Q - Should Muni buy hybrid buses?

A - "Not on hills."

A - "No."

A - "Buy some."

A - "Yes."

A - "Yes."

A - "If the (interior) noise can be minimized then maybe yes."

General hybrid comments from written survey:⁴⁴

"Coach is not for SF operation." 

"Should perform on some lines without grade." 

"The public is very intrigued by a non-polluting electric bus that doesn't need to be on wires." 

"At this stage, it's too early to give comment."

EVALUATION CRITERIA

8. OPERATOR FEEDBACK

8.1.   Evaluation:  Data is collected for each bus alternative technology pair and a relative comparison is made using the Muni's conventional diesel bus fleet as the standard.  Operator feedback is primarily collected using evaluation surveys.  Operators of the test buses are asked to complete the surveys at the completion of each service run in a test bus.  The second form of operator feedback is in written form, such as a letter or email.

8.2.   Significant Variables:  While all operators of test buses are asked to provide feedback, individual time spent driving the test buses differs significantly.  This is generally due to the training schedules, the availability of the test buses, operator schedules, and the different runs that the buses are assigned to for different phases of the testing program.

PASSENGER FEEDBACK

Category leader:  Inconclusive, due to limited data collection.

SUMMARY

In order to gauge passenger opinions regarding the alternative technologies in this program, a preliminary seven (7) question survey was designed.  The survey has been tested briefly in the field.  Based on a very limited sample of 18 surveys and follow up discussions, some potential trends have already been noticed.^[55]

1) Riders are generally concerned with all kinds of emissions but don't view Muni as a major source of urban emissions.

2) Most riders feel that a different fuel source probably means a cleaner burning bus; and they can tell that these buses look different than the rest of the fleet.

3) Riders want dependable service first and foremost, but would prefer to ride cleaner buses whenever possible.

Other lessons learned:  These in-field survey tests have revealed that some questions may need to be refined if we are going to measure any significant preference for alternative technology buses compared to conventional diesel buses.

Next steps:  Refine the survey and continue to administer throughout the 24-month AFPP.

RESULTS

Results to date are inconclusive due to limited data.

During the ongoing testing the intent is to survey Muni's ridership in an effort to measure rider opinions in a three main areas:

1) Are Muni riders concerned with emissions from Muni buses and do they view Muni as a significant source of urban emissions?

2) Do Muni riders understand the technical difference between alternative technology buses and diesel buses; can they detect a difference (look, feel, smell)?

3) Do Muni riders have a significant preference for alternative technology buses compared to diesel buses.

EVALUATION CRITERIA

9. PASSENGER FEEDBACK

9.1.   Evaluation:  Data is collected for each bus alternative technology and a relative comparison is made using the Muni's conventional diesel bus fleet as the standard.  Passenger feedback is primarily collected using evaluation surveys.

9.2.   Test Procedure:  The surveys are verbally given to passengers while the test bus is in revenue service.  The goal is to eventually administer the surveys to a large number of passengers in order to better represent public opinion.

9.3.   Significant Variables:

9.3.1.   Language:  The survey is given in English only, which may prevent many Muni riders from providing feedback.

9.3.2.   Runs:  The survey is given on a specific runs only; the survey is only given in certain parts of the city at certain times of the day.

9.4.   Test Personnel:  Test personnel who are clearly identified as Muni employees administer the surveys.

MAINTENANCE

  Reliability

  20-Hour Performance

  Maintenance Feedback

RELIABILITY

Category leader:  Conventional diesel.

Alternative leader:  CNG.

SUMMARY

The average number of miles that a vehicle travels between failures is the measure of reliability that is used in this analysis.  The clear reliability advantage goes to conventional diesel technology.  CNG and hybrid technologies have proven significantly unreliable in service when compared to the control group.  Results are broken down into propulsion system and chassis categories.  

While conventional diesel technology appears over four (4) times more reliable than CNG technology, and over ten (10) times more reliable than hybrid technology, it should be noted that these alternative technologies are new to Muni, and these results were anticipated during the initial 6-month period.  However, Muni's reliability rates for these new technologies are very close to New York City Transit's (NYCT) reported reliability rates for these technology types during their first six (6) months in service.^[56]  Regardless, poor reliability rates directly translate to poor service, it also correlates to low test mileage accumulation, which compounds analysis difficulties.  The PM filters required cleaning one (1) time during the initial 6 months of testing.  Reliability rates for the PM filters will be unknown until additional data has been collected.

Other lessons learned:  Six (6) months is not enough time to build up sufficient vehicle mileage on the alternative technology buses to perform an accurate analysis of reliability.  While this report is primarily concerned with the evaluation of alternative fuels and propulsion technologies, it should be noted that the additional weight added by CNG fuel tanks and hybrid battery loads reduces the number of passengers that a bus can hold before reaching its maximum allowable gross vehicle weight (GVW).  Of AFPP test vehicles: the conventional diesel bus can carry 78 passenger before reaching GVW; the hybrid bus can carry 73 passengers before reaching GVW; the CNG bus can carry 58 passengers before reaching GVW.^[57]  Additional vehicle weight can also increase brake wear rates, however, hybrid brake wear rates seem to be better due to hybrid system's supplemental regenerative (electric) braking.

Next steps:  It is crucial to the analysis of reliability rates that the vehicles build up additional mileage.  Engine oil samples should be analyzed in order to help predict long-term engine wear patterns.

RESULTS

TABLE 16

RELIABILITY RATES DURING INITIAL 6-MONTHS:

EVALUATION CRITERIA

10. RELIABILITY

10.1.   Evaluation:  Data is collected for each bus technology pair and a relative comparison is made using the conventional diesel bus group as the standard.  All records for the first six (6) months in service are evaluated for each technology pair in order to determine the total reliability of each technology.^[58]  PM filters were installed on test buses following their initial six (6) months in service, so the 6-month reliability of the PM filters alone will be evaluated separately.  This study is primarily concerned with the evaluation of different propulsion systems and configurations, so chassis reliability rates are evaluated separately.  A relatively high mileage number result is this study's measure of relatively superior reliability.  In other words, a more reliable vehicle can accumulate relatively more miles between in-service failures.

10.2.   Definitions:

10.2.1.   Reliability:  Mean Distance Between Failure (MDBF) is the measure of reliability used in this study.  Reliability for this study is further defined by two (2) categories:  Propulsion system failures and chassis failures. 

10.2.2.   Propulsion system is defined to include engine, transmission, system generator, traction motor, rear axle, fuel system, batteries, cooling system, and all accessories associated with the normal operation of one of these systems. 

10.2.3.   Chassis failures encompass all non-propulsion system areas of the vehicle, including but not limited to:  Doors, suspension, foundation brakes, windows, tires, and lighting.

10.2.4.   Failure:  Chassis or propulsion system failure is defined as any event or problem that produces a road call or will not allow a vehicle to go into service.

10.3.   Test Procedure:

10.3.1.   Repair Records:  All unscheduled vehicle repairs are recorded.  Vehicle mileage at the time of an unscheduled repair is used to determine MDBF.

10.3.2.   Road Calls:  All road calls are noted in the repair records, and are differentiated from failures that prevent a vehicle from going into service.

10.3.3.   Out of Fuel:  Propulsion system failure includes fuel starvation if there remains sufficient fuel to supply the engine.  However, this is not to be confused with an out of fuel condition, where there is no more fuel available for the engine.  Although an out of fuel condition requires a road call, the event is not included in this evaluation of reliability.  This is due to the fact that each vehicle in this study has sufficient fuel capacity to operate for more than 24 hours in revenue service without refueling.

10.4.   Significant Variables:  There are many factors that contribute to a vehicle's reliability.  Examples of the variables involved include, but are not limited to:  Driving technique; maintenance; passenger loads; grade conditions; weather conditions; and road conditions.  In general, the level of equal distribution of these variables between the test buses is proportional to combined vehicle mileage, assuming all test buses accumulate roughly equal mileage.

20-HOUR PERFORMANCE

Category leader:  Inconclusive, due to non-ideal test conditions.

SUMMARY

This category considers the reliability of the buses as they operate continuously for 20 hours in peak ambient temperature conditions.  Data was not available for this initial 6-month report, primarily due to the relatively mild winter temperatures in San Francisco during this time.

Other lessons learned:  All test buses proved reliable when operated continuously in relatively mild ambient temperature conditions.

Next steps:  Be prepared to collect 20-hour performance data when conditions are ideal for testing.

EVALUATION CRITERIA

11. 20-HOUR PERFORMANCE

11.1.   Evaluation:  Data is collected for each bus technology pair and a relative comparison is made using the unmodified diesel bus pair as the standard.  This is a pass or fail category in the AFPP.  The evaluation consists of operating all test vehicles on a characteristic route, while run continuously in peak ambient temperatures for 20 hours.  At the end of 20 hours in operation, the vehicle must pass the following criteria:

11.1.1.   Operator Evaluation:  The operator must feel like the vehicle performs roughly the same at the beginning and end of the 20-hour performance run.

11.1.2.   Diagnostic Codes:  Propulsion system maintenance codes cannot be present when the vehicle is inspected following the 20-hour performance run.  If there is a code, it could signify a problem with the engine or propulsion equipment.

11.1.3.   Maintenance Inspection:  The vehicle propulsion system must not display any fluid leaks, air leaks, noises, signs of overheating, looseness in mountings or attachments, or other form of wear or problem when the vehicle is inspected following the 20-hour performance run.

11.2.   Test Procedure:

11.2.1.   Operation:  The test will assume normal vehicle operations.

11.2.2.   Terminal Breaks:  The vehicle is only turned off during the operator's terminal breaks.

11.3.   Site Conditions: This test is performed during dry and peak ambient temperatures conditions.

11.4.   Vehicle Conditions:  A thorough inspection and evaluation is performed of the vehicle's propulsion system prior to testing.

11.4.1.   Operator Acceptance:  The operator must feel like the vehicle performs well for its technology type before 20-hour testing begins.

11.4.2.   Diagnostic Codes:  Propulsion system maintenance codes cannot be present when the vehicle is inspected prior to the 20-hour performance run.

11.4.3.   Maintenance Inspection:  The vehicle propulsion system must not display any fluid leaks, air leaks, noises, signs of overheating, looseness in mountings or attachments, or other form of wear or problem when the vehicle is inspected prior to the 20-hour performance run.

11.4.4.   Fluids:  Fuel and all other fluids must be full or at proper levels.

11.5.   Significant Variables:  Passenger loads and operating conditions vary during 20-hour performance data collection.  Load and operating variability can influence reliability data.

11.6.   Data Collection Equipment:  Standard vehicle maintenance and diagnostic equipment is used to diagnose and inspect the vehicle.

MAINTENANCE FEEDBACK

Category leader:  All.

SUMMARY

Three mechanics, with 39 years combined experience with Muni function as the primary maintenance team for the test vehicles.  They were surveyed for their opinions regarding the maintenance of the different technologies represented in the program.  The mechanics all agreed that overall build quality and maintenance practices determine reliability rates.  While acknowledging that the alternative technologies are not yet reliable at the 6-month point in the program, all were in agreement that the technologies would become generally more reliable as they accumulated additional miles.  Each mechanic felt that the most important element in successful maintenance of the alternative vehicles was to have proper training, tools, and maintenance areas provided prior to receiving a new procurement of buses.  Maintenance division supervisors, and Local 1414 have also placed a strong emphasis on this point.^[59]

Other lessons learned:  Reliability comments come from mechanics that have volunteered for the AFPP.  Their optimistic comments do not necessarily represent the opinions of other Muni mechanics that were not interested in applying for the AFPP positions

RESULTS

TABLE 17

MAINTENANCE SURVEY - RELATIVE RELIABILITY RATES:

The following comments are taken from written surveys:

Reliability:

Daily reliability:

Each of the three (3) AFPP mechanics stated that reliability is directly related to the build quality of the entire vehicle.  One comment was that if all motor coaches are maintained identically by Muni, they should therefore be equally reliable in daily service.  Another comment was that the CNG engine and fuel system may not be as reliable on a daily basis as conventional technology.

Long-term reliability:

The AFPP mechanics were split regarding which technology would be most reliable in the long-term.  One stated that conventional diesel technology will have the same reliability as past motor coach fleets.  Two stated that hybrid technology will have the same reliability as past motor coach fleets.  In response to the hybrid bus reliability problems during the initial 6-months, it was thought that hybrid buses will become more reliable as this new technology is developed.

Relative reliability:

AFPP mechanics stated that the CNG buses are generally more reliable than the hybrid buses, but neither CNG or hybrid technology has proven reliable to date.  Furthermore: "We do not have the proper tools, place and safety precautions in place to work on CNG coaches, until that has been done, I believe hybrids are more realistic.  I would also believe fuel cells to be the future of transit."

Safety comments:

CNG:

If appropriate training is provided, Muni mechanics will have every reason to feel comfortable and safe working on CNG buses.  Two (2) AFPP mechanics stated that, in reality, CNG buses may not receive the same level of maintenance attention as conventional diesel buses due to safety concerns - even if appropriate training is provided.

Hybrid:

Each of the three (3) AFPP mechanics stated that if appropriate training is provided, Muni mechanics will have every reason to feel comfortable and safe working on hybrid buses.

Technology specific comments:

CNG:

Two (2) AFPP mechanics stated that whether the buses are different or not makes no difference as long as appropriate training, tools, and infrastructure are provided.

Hybrid:

Two (2) AFPP mechanics stated that Muni should use hybrids to transition its motor coach fleet to be on par with Muni's trolley bus and light rail vehicle (LRV) fleets by using the most modern forms of electric bus propulsion technology currently available.

EVALUATION CRITERIA

12. MAINTENANCE FEEDBACK

12.1.   Evaluation:  Data is collected for each bus technology pair and a relative comparison is made using Muni's conventional diesel fleet as the standard.  Maintenance feedback is primarily collected using evaluation surveys.  Maintainers of the test buses are asked to complete the surveys at six (6) month intervals throughout the program.  The second form of maintenance feedback is in written form, such as a letter or email.

12.2.   Significant Variables:  From two (2) to three (3) maintainers are assigned to the test buses at any one time.  Therefore, feedback from maintenance supervisors familiar with the test buses is encouraged in order to increase the number of responses.

COST

  Operating cost

  Capital cost

OPERATING COST

Category leader:  Conventional diesel.

Alternative leader:  Hybrid.

SUMMARY

Operational cost was split into three (3) subcategories:  Fuel cost per mile, maintenance cost per mile (which includes repair and routine maintenance costs), and total vehicle cost per mile.  Based on fuel alone, the conventional diesels appear to be over 22% less expensive to operate when compared to the estimated fuel cost per mile for the hybrid buses, and 40% less expensive to operate when compared to the CNG bus fuel cost per mile.  Fuel prices are based on the cost that Muni pays for diesel fuel^[60] and what Sacramento RT pays for CNG.

Maintenance cost results here are not representative of life-cycle costs due to the fact that the alternative buses are a new experience for Muni, and the conventional diesel buses represented here are not entirely different than all of Muni's past diesel buses.  In addition, 6 months is not sufficient time to accurately predict life-cycle maintenance costs, and these results are inconclusive.  In general, the conventional diesel buses were the least expensive per mile to maintain, followed by the hybrid buses, and finally the CNG buses.

While no long-term projections have been calculated due to the limited data available after 6 months of testing, it should be noted that the cost to replace hybrid propulsion system batteries, roughly every two (2) years, is about $10,000 per bus, or $60,000 over the 12-year life expectancy of each hybrid bus.^[61]

Other lessons learned:  Fuel prices are constantly changing, and it is therefore difficult to predict long term fuel pricing.  Maintenance data is greatly influenced by the fact that these alternative technologies are new to Muni, and require additional time to establish the repair patterns found with the conventional diesel buses.

Next steps:  Fuel economy data should be recollected for all test vehicles, and fuel cost per mile should be recalculated based on these revised fuel efficiency numbers.  Additional miles are needed on the alternative buses in order to make a meaningful analysis of maintenance costs and life cycle cost estimates.

RESULTS

TABLE 18

FUEL COST PER MILE:

TABEL 19

MAINTENANCE COST PER MILE:

TABLE 20

TOTAL COST PER MILE:

Other operating costs, not included in the primary cost analysis or tables, concerns the effect of vehicle performance differences while in-service, and the cost of special operation and maintenance training.

Limited data concerning the San Francisco operational performance differences between the various bus technologies forced cost calculations based on the annual operational costs for a narrow range of bus performance factors (10% slower, 5% slower, and 6% faster).  The basic analysis sought to quantify how these seemingly minor increases or decreases in bus performance could affect operating costs over the course of a year.  This analysis defined two sets of line assignments: for a fleet of 80 buses on the road at one time; and for a fleet of 60 on the road at one time.  These line selections were based on directives within the 11 Point Agreement.^[65]  One set of routes was determined by environmental justice demographic considerations, and one set was defined by topographic considerations, whereby relatively flat or mild grade routes were selected to better accommodate alternative fuel buses.  Both line assignments demonstrated that a single percentage point in bus performance can cost over $200,000 annually.

One might reasonably expect that a five percent (5%) loss or gain in coach operating performance would not be a significant factor when comparing bus technologies, and that slower buses and the resulting service impacts could be absorbed into the existing schedules.  However, Muni's service is presently very tight and there are very few "slack" periods where slower running times could be absorbed into the current system without reducing service capacity, daily frequencies, peak-hour-headways and hours of service.  Individually a six percent increase in performance or a five percent decrease in performance and running times does not seem significant.  However the detailed analysis estimates the potential, cumulative cost of an 11% difference between the bus technologies.  The analysis indicates that (assuming service capacity and operation costs remain constant) the cumulative annual cost of adopting a bus that is 5% slower and concurrently rejecting a bus that is 6% faster translates into between two (2) and three (3) million dollars annually.

In order to arrive at a baseline cost figure, the annual revenue hours for each motor coach line were determined and multiplied by $91.26 (the motor coach variable cost per hour stated in National Transportation Database).  This figure is the Federal Transit Administration's (FTA's) official measure of bus costs.  It includes most operational and maintenance costs, but excludes administration and overhead (A&O) costs, and it does not differentiate between a standard (40') and articulated (60') coaches. In order to account for a slower buses, with decreased performance measures, the baseline annual revenue hours figure for each motor coach line must be factored upwards by the performance difference. Conversely, in order to account for faster buses with increased performance measures, each motor coach's annual revenue hours figures must be factored downward by the performance difference.  The revised costs were then determined by multiplying the revised hours of service figures with the $91.26 hourly cost.

It is assumed that once better data is collected concerning the performance differences observed while in Muni service, these factors can then be plugged back into this analysis, or the information could be given to Muni scheduling, so that real schedules based on observed run times could be generated.

EVALUATION CRITERIA

13. OPERATING COST

13.1.   Evaluation:  Data is collected for each bus technology pair and a relative comparison is made using the unmodified diesel bus pair as the standard.  Operating cost is evaluated in terms of the cost per mile for each technology pair.  The total cost for each technology pair is divided by the total number of miles for each pair.  Operating cost for test vehicles include, but are not limited to, costs associated with:  fuel used; maintenance labor; maintenance parts; all non-fuel system fluids; body repair labor; and body repair parts.  All records for the first six (6) months in service are evaluated for each technology pair in order to determine the total cost for each technology.  All warranty work is included, even if this cost is not directly paid by Muni.  This study is primarily concerned with the evaluation of different propulsion systems and configurations, so chassis related costs are evaluated separately.

13.2.   Significant Variables:  There are many factors that contribute to operating costs.  Examples of the variables involved include:  Driving technique; maintenance; operating conditions; fuel pricing; market changes; and vehicle age.

CAPITAL COST

Category leader:  Conventional diesel.

Alternative leader:  Hybrid.

SUMMARY

This category considers the incremental cost involved with purchasing alternative technologies as well as the necessary infrastructure costs related to supporting these alternative technologies.  In general, the hybrid buses are 17-52% more expensive to purchase than conventional diesels, but they incur significantly less incremental facility costs at $1.05 million.  The CNG buses are only 14-18% more expensive to procure, but the incremental facility costs are estimated to be approx. $7 million for one facility and $12 million for two facilities.^[66]

Other lessons learned:  The majority of CNG fueling infrastructure cannot be converted in order to support future hydrogen fueling needs, such as those now thought to be required for hydrogen fuel cell powered electric buses.^[67]  Diesel PM filter ash/waste will have to be managed for Muni's entire diesel fleet.^[68]

Next steps:  Further research should be done in order to refine these cost estimates.

RESULTS

The incremental cost of purchasing alternative technology vehicles is due to several factors, including low production volumes, and additional propulsion related equipment and subsystem costs.

TABLE 21

NEW BUS COST SUMMARY⁵⁹