No abstract

No abstract

Internal combustion engines exhibit fast pulsating short-time dynamics due to the reciprocating cylinder motion, around mean operating points that change comparatively slow due to inputs such as throttle and load. Comparatively, simple mean value engine models (MVEM) describe the slow changes of the averaged states for automotive control and fault diagnosis. In this paper, a bank of state estimators based on MVEMs is used for fault residual generation. Three faults: 1) throttle mass airflow sensor fault; 2) exhaust gas recirculation valve sensor fault; and 3) exhaust leak fault are considered here. These faults are significant as they affect emission levels. Optimized thresholds for residual classification are derived for minimizing false alarm rates and missed detection rates. The diagnosis logic, based on the principles of structured residuals proposed in literature, is extended here for multiple hypotheses testing. Furthermore, the Dempster-Shafer theory is used to associate a confidence measure with the decision conclusions and this is shown to improve isolation. Performance is demonstrated with automotive engine data obtained from a four-cylinder instantaneous spark-ignition engine (gasoline) system model, developed in the simulation software AMESim.

Mean Value Engine Models (MVEM) have been used to model the averaged dynamics of an automobile engine system for automotive control and fault diagnosis. For these purposes, it is common to estimate states of interest given noisy measurements using state observers. Since the measurements could be noisy and asynchronous, they should be suitably postprocessed before feeding them to a state observer. In this paper, an Unscented Kalman Filter (UKF) was developed for an Extended MVEM and a suitable post-processing algorithm for the measurements has been described.

No abstract