Functional Link Neural Network Prediction on Composite Regeneration Time of Diesel Particulate Filter for Vehicle Based on Fuzzy Adaptive Variable Weight Algorithm

Abstract

In order to enhance the precision of prediction on composite regeneration time of diesel particulate filter for vehicle, different predictive values of each prediction model are selected as the primeval input values of functional link neural network, and the functional link neural network prediction model of composite regeneration time based on fuzzy adaptive variable weight algorithm is established after the necessary and sufficient conditions for fitting of functional link neural network are analyzed. The application result of the model shows that the absolute value |emax| of maximum relative error of functional link neural network prediction model on composite regeneration time of diesel particulate filter for vehicle based on fuzzy adaptive variable weight algorithm is less than 0.86%, indicating the high accuracy of the prediction model. Moreover, the result draws that the factors influencing composite regeneration time prediction of diesel particulate filter for vehicle, influence degree of which is from big to small, are exhaust oxygen concentration, exhaust mass flow, microwave power, exhaust temperature and the amount of cerium-based additive.