In this paper, a prediction method based on least square support vector machine is introduced into the surface roughness prediction model in low-frequency vibration cutting. The model is based on low-frequency vibration cutting experiment, to obtain the corresponding relation between vibration parameters and cutting parameters and the workpiece surface roughness. Construct the training sample set to train regression models of least square support vector machine through experimental data. Identify training sample set to solve the regression parameters a and b. The amplitude of A, vibration frequency f, feeding f1 and spindle speed n as the input variable Xi, obtain predicted values of surface roughness by applying forecasting model; using the value, which equals to the difference between predicted value and its actual measurements value of surface roughness, to evaluate forecasting model. Through examples, and compared with BP neural network and support vector machine method, obtain the following conclusion: under the condition of the same sample, based on least square support vector machine prediction model constructs an order of magnitude faster than SVM method, and the model prediction error is about 29% to support vector machine, the prediction accuracy is an order of magnitude higher than the BP model.
The study is sponsored by the National Special Project of International Cooperation in Science and Technology （S2013HR0021L)）
College of Mechanical Engineering and Automation,Huaqiao University,FuJian Xiamen,361021,College of Mechanical Engineering and Automation,Huaqiao University,FuJian Xiamen,361021,College of Mechanical Engineering and Automation,Huaqiao University,FuJian Xiamen,361021