SIGNAL COMPONENT ANALYSIS FOR MISFIRE DETECTION

M. Lindemann, D. Filbert

Abstract

The misfire detection plays an essential role for the on-board diagnosis of sparc ignition engines and is prescribed by law in the United States of America since 1989 (OBDII). According to the last revised edition from 1996, the misfire detection rate must equal 100 percent for the 2002 model year under all operating conditions relating to speed and torque. Actual methods for misfire detection in sparc ignition engines base on measuring speed fluctuations. But they don't work reliably even at high speed and low torque. Hence this paper describes a new approach which consists of measuring the vibration at the cylinder walls. The composition of the knock sensor signal is introduced and classified in several elementary parts. All of them carry information about the state of combustion. A new polynomial based signal element decomposition method is proposed to extract features for the misfire detection. Results are given using real knock sensor signals.

Keywords

engine misfire detection, signal component analysis, polynomial filtering

Download

IMEKO-WC-2000-TC10-P268.pdf

DOI

-

IMEKO TC

TC10 - Measurement for Diagnostics, Optimization and Control

Back to the proceedings