A dynamic model of engine using neural network description

Automotive engines frequently run under dynamic conditions. Due to the inner complex physical and chemical changes in the course of engine operation, the precision of model obtained by means of mechanism modeling or curve approximation is limited. To meet the requirement of making automatic shift schedule, in this paper the determinant factors of dynamic characteristics of an engine are analyzed, the dynamic test of the Santana 2000 EFI engine is completed, and the dynamic model of the engine based on neural network identification is built through the learning of a great deal of test data, which leads the optimal match of the cooperation between the engine and powertrain system into reality.