Predication For Surface Roughness And Optimization Of Cutting Parameters In Precision Turning

In the process of machining, the roughness of surface is one of the key factor of judging the quality. How to choose reasonable cutting parameters to control the surface roughness is an increasing concern. The studying of this issue not only has great practical significance, but also has certain theoretical significance.Firstly, this paper sorted out situation of surface roughness prediction study in the process of machining, on the base of studying the formation of the workpiece surface and discussing the prediction given by theoretical formula, it was proposed that using artificial neural network techniques to build surface roughness prediction model. Prediction model adopted three-layer feed-forward radial basis function (RBF) neural network, used clustering method to determine the center of radial basis function in the first phase of neural network training, used LMS method to determine the connection weights between the hidden layer and output layer in the second phase of neural network training.Secondly, prediction model was applied to surface roughness prediction in ultra-precision turning. Matlab is used for compiling the neural network working procedure. Select eighteen groups of experimental data as training sample and testing sample of neural network. The testing sample is used for examination of the trained neural network prediction accuracy. The test results show that the model based on neural network prediction is valid. Through prediction model based on Neural network, we develop a theory of the roughness of the surface can be varied according to the speed and the feed of cutting.Thirdly, this paper got further study of cutting parameters optimization based on Genetic algorithm. In the process of using surface roughness as the objective function, accuracy of the theoretical roughness formula is low, in this thesis, the form of index relationship between machining data and roughness has been applied, what's more, genetic algorithm has been used as a distinguishment tool for the data. The optimized cutting parameters can not only ensure the surface quality, but also can improve the processing efficiency.