Combustion Torque Estimation and Misfire Detection for Calibration of Combustion Engines by Parametric Kalman Filtering

In this paper, a new estimation method for the inner combustion torque of internal combustion engines and a misfire detection algorithm are presented. The torque estimate is based on a parametric Kalman filter using the measurements available on commercial engine testbeds. The combustion torque is modeled as a superposition of specially designed basis functions. Excellent noise attenuation, phase-free estimation, and robust parametrization result from the chosen structure. Misfire detection is achieved by an interacting multiple model approach, where two dedicated Kalman filters for each cylinder represent regular operation and misfire, respectively. This approach enables reliable detection of the actual cylinder state early in the combustion phase. The functionality and performance of the new method are illustrated by means of simulation examples and test-bed results.