Speed and Power Regressions for Quality Control of Heavy Duty Vehicle Chassis Dynamometer Research 1999-01-0614

When performing a transient test on a heavy-duty engine as outlined in the Code of Federal Regulations (CFR), defined regression values of engine speed, torque and power must meet specific tolerances for the test to be considered valid. Regression of actual engine feedback data against target points from a schedule defined from an engine map is performed using the method of least squares to determine the slope, intercept, coefficient of regression and standard error of the estimate. To minimize the biasing effects of time lag between actual and schedule data, shifting of the data in the time domain prior to analysis and certain point deletions are permitted. There are presently no regression criteria available for heavy duty chassis testing. This leaves facilities performing these chassis tests with no suitable guidelines to validate individual tests. This study applies the regression analysis used in engine testing to chassis testing and examines the difficulties encountered. The test cycles examined are the Central Business District (CBD) cycle, the West Virginia University (WVU) 5 peak cycle, and the EPA Heavy Duty Vehicle Driving Schedule (Test D). Regressions in the speed domain were found to be useful, but regressions in power were poor due to the speed-time nature of the schedules and large instantaneous power fluctuations caused by gear changes and pedal dithering.