Active Fault Diagnosis with Sensor Selection in a Diesel Engine Air Handling System

This paper describes a methodology to improve the precision of FDI for a diesel engine air handling system. Specifically, we present a technique that manipulates admissible inputs to determine conditions that maximize and separate the evidence of persistent faults in sensed outputs. The method developed was applied on a simulated diesel engine air handling system to improve the detectability and isolability of numerous component and actuator faults. In addition, the available sensors of the air handling system were examined to determine the best set of sensors for improving fault diagnostics. Afterwards, a simple FDI test using fixed thresholds was implemented to evaluate the impact of the test design optimization.