Quality predictive control of gear heat treatment based on Elman

The gears are an important component of industrial machinery; gear heat treatment process plays a very important role on the formation of its final quality. This thesis analyzes the factors affecting the quality of gear heat treatment; the three input layers and an output layer of neurons composes of Elman neural network model are built. According to the actual case data from a car company, through the neural network learning, training and simulation, this thesis applies to the gear of a particular model of heat treatment process quality predictive control. The experimental data shows that the error of the neural network model for simulation is between 3% to 5%, and the control effect of the neural network model is much better, improving the analysis efficiency effectively and achieving control of automation.