Model-based detection and isolation of faults due to ageing in the air and fuel paths of common-rail direct injection diesel engines equipped with a λ and a nitrogen oxides sensor

Abstract

The air and fuel management of modern diesel engines is based on a feedforward control approach. As a consequence, faults of the air or the fuel path have a direct impact on the emissions that cannot be observed nor compensated. Faults due to the ageing of engine parts are very common; thus engines are designed with tolerances large enough to fulfil the emission standards for their entire lifetime. This paper presents a method to locate and quantify engine faults due to the ageing of parts of modern diesel engines in order to help to reduce design tolerances. A relative air-to-fuel ratio and a nitrogen oxides emission sensor have been applied to a test-bench engine, and control-oriented emission models have been developed. Based thereon, a model-based fault detection and isolation system has been implemented. The sensor signals are compared with the models in order to detect a possible fault, and fault location and size are estimated by means of a bank of extended Kalman filters, one for each fault considered. Results from the New European Driving Cycle show that the developed control-oriented models are accurate, and the faults can be correctly detected and isolated.