A Neural Network Fault Diagnosis Method applied for Faults in Intake System of an SI Engine

One essential part of automated diagnosis systems for SI engines is due to elements of air path system. The diagnosis task is getting more challenging by including Exhaust Gas Recirculation (EGR) which its transient effects on temperament complexity of the air-path system are quite significant. The faults occur in this subsystem can result in deviation in air-fuel ratio, which causes increased emissions, misfire and especially loss of power and drivability problems. In this article, a model-based diagnosis system for airpath of an SI engine with EGR is constructed. In addition, a nonlinear four-state dynamic model of an SI engine with EGR is utilized, and then results are validated by a real engine. In the next step, diagnosis system is designed in the framework of Artificial Neural Network (ANN) classifier. Simulation results show that the constructed diagnosis system for six fault modes considering all three kinds of common faults including actuator, component, and sensor faults is applied successfully. In addition, in this article the Manifold Air Temperature (MAT) sensor fault which comparatively less has been evaluated than other elements are also taken into account.