Vibration Predictive Modeling And Parameter Optimization In Hardened Steel Precision Turning Process

Precision cutting of hardened steel has become the key technology of manufacturing industry, Hardened steel turning instead of grinding is widely used because of its environmental protection and quality process. Meanwhile, high speed cutting and high machining accuracy are the features of modern machining. With the increase of turning speed, the vibration in the machining process becomes more severe, which leads to the decrease of the workpiece surface quality. That the amplitude becomes larger will make the workpiece surface quality worse and the tool wear more serious. With the background of Precision Turning of Hardened Steel, this paper researches the effects of cutting parameters on the amplitude in the process of processing, Through the experiment, the influence factors of amplitude is analyzed, the prediction model of vibration amplitude is established. cutting parameters, which Considering the influences of amplitude are optimized in this paper.Firstly,By orthogonal analysis method to combine cutting speed, feed rate, depth of cut, corner radius, etc. Conduct experiments of cutting hardened steel in CAK4085 NC lathe with PCBN tool, the amplitudes of the tool relative to the workpiece are collected under different cutting parameters, the influences of the parameters on the amplitude are analyzed. Secondly, the vibration amplitude prediction model is established, by using the BP neural network established by MATLAB. By comparing the error of the network model, we can choose the optimal learning algorithm, the training function and the learning function. We pre-process the amplitude data that has been collected and train BP neural network to build the prediction model of hardened steel turning amplitude, the model can satisfy the practical application requirement through the error validation prediction model. In order to simplify the operation of practical application, the amplitude prediction system based on MATLAB human- computer interaction interface is established. Thirdly,based on the theory of this method, Seeking the extreme value of multivariate function, Based on the theory of method for finding the extreme value of multivariate function, With the uniform design method, Parameters of thecutting test are combined and the mathematical regression model of multi – objecti optimization based on cutting parameters is established, the cutting parameters are optimized, with the minimum surface roughness, the minimum tool vibration amplitude, the minimum white layer thickness, the minimum residual stress and the maximum tool life. it can provide a theoretical reference for the increase of hardened steel turning surface quality and tool life.