Method to compensate errors in engine speed measurement

Abstract

The invention provides a software method to correct for non-fuel injection induced errors in measured engine speed. This invention provides a means to calculate a least squares regression of the measured engine speed data when the engine is motoring without combustion or load. The motoring test is to avoid the effect of unbalanced cylinder power, thus detecting only the static errors in engine speed measurements. Recognizing that the average cylinder speed should form a straight line within an engine cycle, the lease squares regression of the measured speed data therefore represents the theoretical performance of the engine. A residual between this theoretical performance and the actual measured engine speed can therefore be determined. A table of mean residuals which are averages of individual residuals measured in consecutive engine cycles is constructed for each engine at a reference engine speed. The mean residuals are convertible to other operating speed ranges of the engine and can be subtracted from the engine speed measurements to remove the static errors. This mean residual table is stored as a look-up table in the memory associated with the engine control module (ECM) microprocessor and can be implemented within the software code being executed by the ECM microprocessor.