On-line predicting diesel engine EGR rate based on Chaos-Neural Networks

To counter the influencing emission of the diesel engine by the EGR rate, the forecasting model of the diesel engine EGR rate was set up by combining Radial Basis Function neural network (RBFNN) with the Chaos-Neural Networks (CNN) theory. Based on the nonlinear and effective modeling capabilities of the chaos system, the emissions performance and the EGR rate of the diesel engine were on-line predicted by the evolution on approximating to phase points of the RBF neural network; the predicting results were compared with the gray prediction results. The tests on the EGR rate and the diesel engine emission were made on the whole vehicle by the chaos-neural networks synergetic strategy. The test result showed that the forecasting model of the diesel engine EGR rate was reasonable; the chaos-neural networks forecasting strategy had the good resolving power and could be much fitted for the on-line forecast of the EGR rate.