Identification of Engine Heat Management Models Using Artificial Neural Network

Abstract

This study aims to get a foothold in the practical application of an ANN (artificial neural networks) aided parameter identification method. Using measurement data obtained from a real engine of commercially available cars, the authors tried parameter identifications for a zero-dimensional theoretical heat balance model of the engine system with the ANN-aided identification method. The identification results show that each heat conductance is a function of coolant and oil follow rates, and heat capacity is also a function of those flow rates. The numerical simulation of the representative temperature of the engine system using the theoretical model with identified parameters shows pretty good predictions.