Analysis based on generalized regression neural network to oil atomic emission spectrum data of a type diesel engine

Abstract

In order to deeply mine the information of Oil Atomic Emission Spectrum Data, a simulation model and a prediction model of Cu concentration of a type of six- cylinder diesel engine were established by applying Generalized Regression Neural Network. Seven different working conditions had been set up and sixty-nine oil samples had been taken from engine. The results show that the absolute errors of the simulation value of the 69 samples are within the acceptable accuracy indices and the absolute errors of the prediction value of the 19 samples are lower than the acceptable accuracy indices. It has been proved effective that Cu concentration can be predicted via Generalized Regression Neural Network algorithm. © 2011 Springer-Verlag Berlin Heidelberg.