Comparative Study on Combustion Features Extraction Methods in IC Engines Using Neural Networks Models

Abstract

The in-cylinder pressure analysis provides important information on the combustion process and stands as an invaluable tool in the internal combustion engine research & development. Its implementation in a combustion control algorithms appears as a promising solution for attaining optimal combustion control in IC Engines. The pressure sensor durability, accuracy and price, along with increased demand for the processing power of Engine ECU, seems to be the main obstacles for putting these concept in IC engines on serial production line. This paper deals with the potentials of Artificial Neural Networks (ANN) and their application in combustion features extraction, based on the crankshaft angular speed measurements. High speed processing capabilities and acceptable accuracy of ANN make them good candidates to become a core component of the future combustion control algorithms. An radial basis function (RBF) ANN and a local linear Neuro-fuzzy model (LLNFM) are compared in order to gain some conclusions on optimal network topology, best suited for job of extracting crucial combustion features on common ECU platforms.