Validation of Controller Inputs in Electronically Controlled Engines

The control strategy in an electronically controlled engine critically relies on the proper functioning of the measuring components; sensors measuring the inputs to the controller are given a margin of error no more than a few percentages in order to achieve the performance constraints of the vehicle. For example, a 10% calibration error in the Mass Air Flow sensor will no longer provide the measurement that is accurate enough to satisfy the emission constraints of the vehicle. Diagnosis of sensors in an electronically controlled engine is a relatively new effort, for which some versions have been implemented to provide simpler tasks like identifying open or short circuits. The main purpose of this paper is to develop an analytical method for, and to demonstrate the feasibility of the failure detection for the sensors measuring the critical inputs to the engine controller. Unlike the previous attempts, however, the approach taken in this paper will employ a model-based method, refenred to as the Beard-Jones Detection Filter (BJDE), and will emphasize the capability of detecting a failure in its earliest stage, i.e., an incipient failure.