A diagnosis technique for the identification of misfire in a converted-diesel HCCI engine

Abstract

The homogeneous charge compression ignition (HCCI) and the exploitation of ethanol as alternative fuel is one way to explore new frontier of internal combustion engines and towards maintaining its sustainability. Misfire and partial burning can cause several serious problems in the HCCI engines fuelled with ethanol. Novel methods for identification of misfire play important roles in HCCI engine operation and exhaust after-treatment system. In this study, a 0.3 liter single-cylinder direct-injection diesel engine was modified to operate on HCCI. The diagnosis technique is introduced for misfire identification. An artificial neural network (ANN) model were designed based on in-cylinder pressures at 0, 5, 10, 15 and 20 CAD aTDC as main inputs to predict indicated mean effective pressure (IMEP). The model was tested with the experimental data, indicating an average error less than 0.1 bar between predicted and measured IMEP and the results indicated that the model can predict IMEP as well as misfire.