Performance Modeling of Engine Based on Artificial Neural Networks

In order to further improve the precision and generalization ability of the neural network based performance model of engine, back propagation neural network (BPNN), radial basis function neural network (RBFNN) and generalized regression neural network (GRNN) have been investigated. The topologies and algorithms of these three different types of neural networks have been designed to meet the same goal of convergence, and a same set of testing data have been used to test the trained neural networks. Comparison of the training and testing errors as well as the generalization ability of these neural networks shows that RBFNN is more suitable for modeling the performance of engine than BPNN and GRNN.