LPG Task Group

Meeting Summary * July 28, 1998

I. Introduction

This meeting was held to update members on the results from the emissions and octane tests to date and to discuss the fuel selection process for continued testing. The following issues were discussed at the meeting:

Test Program Status

- Emissions Test Results

- Octane Test Results

Performance Testing Protocol

Budget Update

Meeting materials are enclosed for your reference. As a reminder, the nominal fuel properties for the test fuels are listed below in Table 1:

Table 1

Test Fuel Properties

Fuel ID Propane Propene Butane
Base fuel 94% 4% 2%
Fuel 1 85% 10% 5%
Fuel 2 80% 15% 5%
Fuel 3 80% 10% 10%
Fuel 4 76% 4% 20%
Fuel 5* 77% 21% 2%

* Added by the LPG Task Group during the meeting (discussed below)

II. Test Program Status

During this meeting, members agreed to conduct emissions tests on a fifth test fuel (Fuel 5) at the request of Tosco Corporation. The additional tests should not change the time line for the project.

The medium duty emissions tests on the Cummins B5.9 LPG engine began in late May and are nearly complete. The emissions tests included duplicate tests with 1 cold cycle and three hot cycles for each test. Duplicate tests were originally planned; however, if the results from the first and second test varied significantly above what was expected for any pollutant a third test was required (please refer to the February 1997 LPG Test Protocol for more details). The emissions test results for the Cummins B5.9 LPG engine is shown below in Table 1.

Table 1

Preliminary Emissions Test Results*

for the Cummins B5.9L Engine


Fuel NMHC+NOx THC CO Ozone
Base Fuel 3.71 0.841 0.411 0.718
Fuel 1 3.85 0.702 0.407 0.829
Fuel 2 3.88 0.674 0.529 0.896
Fuel 3 4.05 0.856 0.618 0.893
Fuel 4 3.76 0.782 0.816 0.788

* For data reported as of July 24, 1998

The light duty emissions tests on the Ford F150 began in late May. Several of the initial tests were run with the incorrect dynamometer coefficients and were redone. Because of this delay, results for test fuel 4 had not been conducted by the date of this meeting. Also, due to the lag time in obtaining the NMOG and ozone forming potential results, the results were not available for the meeting. The results available to date are shown below in Table 2.

Table 2

Preliminary Emissions Test Results1

for the Ford F150 LPG/Gasoline Pickup


Base Fuel 0.030 0.052 1.89 0.033 0.043
Fuel 1 0.039 0.048 1.832 0.029 0.039
Fuel 2 0.043 0.057 2.08 0.039 0.068
Fuel 3 0.046 0.053 1.492 --- ---
Fuel 4 --- --- --- --- ---

1. For data collected as of July 15, 1998

2. Suspect data under review

The variability seen in emissions testing in many cases was higher than originally expected. Because of the higher than expected variability, it is unclear if the criteria originally outlined in the test protocol are the most appropriate. For illustrative purposes, ARB staff showed different ways to compare the emissions results. The comparisons were based on the methodology outlined in the test protocol developed by the Task Group. The equation in the test protocol for comparing fuels is the following:

Xc less than or equal to Xr (1+2 delta)

Where: Xc= Average composite emissions during testing with the candidate fuel.

Xr = Average composite emissions during testing with the reference fuel.

Delta = Coefficient of variability.

Staff's presentation showed how the results would change if the coefficient of variability (delta) were calculated based on data available prior to the test program, on the current test data only, or on the combination of both. It is important to remember that testing is not complete. The medium-duty emissions tests are nearly done but the light duty tests are only half finished. In either case, with the data gathered to date, the comparisons show that Fuel 3 and Fuel 4 would not meet the criteria no matter which approach is used to calculate the variability. Fuel 1 is the most likely to pass and Fuel 2 is borderline. In other testing, the octane tests results were largely within expectations compared to calculated estimates. The octane and emissions results are enclosed.

III. Performance Testing Protocol

Performance testing is the next phase of the testing and will be conducted by Southwest Research Institute (SwRI). Bids were received from Cummins, ORTECH and SwRI. The Technical Assessment Committee met twice on the selection and recommended that Cummins be selected. The LPG Task Group members did not agree with the recommendation for various reasons. Ultimately, SwRI was selected with a 5 to 2 vote over Cummins. Members also agreed that if Fuel 3 and Fuel 4 did not meet the emissions criteria then the first fuel to be tested for performance should be Fuel 2. The basis for the suggestion is that if Fuel 2 does well on performance then it follows that Fuel 1 would also do well since it is more similar to the base fuel.

IV. Budget Update

To date we have raised $414,000 for conducting the test program; however, the currently estimated project cost is about $521,000. Another $107,000 needs to be raised to complete the test program as planned. Additional detail is available in the enclosed presentation.


1. Meeting agenda

2. ARB Staff presentation