基于最小二乘支持向量机超声电机摩擦材料寿命的预测

Life Prediction of Friction Material Used in Ultrasonic Motor Using LS-SVM

超声电机是一个典型的机电耦合系统,其定子与转子的摩擦过程和压电陶瓷与定子压电耦合过程之间具有紧密的联系,该联系具有较强的非线性。超声电机摩擦材料寿命的预测一直是超声电机研究领域尚未很好解决的难题。针对该问题,提出了基于最小二乘支持向量机(LS-SVM)的超声电机摩擦材料寿命的预测模型。该模型以电信号特征参数作为输入量,摩擦材料的厚度为输出量。电信号的检测方便快捷,无需复杂的数据处理。通过对LSSVM模型参数的优化,提高了该模型的预测精度。测试结果表明摩擦材料厚度的预测值与试验值有很好的一致性。

The friction process and electromechanical coupling process in ultrasonic motor （USM） possess strong nonlinearity characteristics, life prediction of friction material used in USM is a problem which have not been solved in the research field of USM. In order to solve this problem, a life prediction model based on least squares support vector machine （LS-SVM） of friction material was proposed. The characteristic param- eters of the electrical signals were the input vector and the thicknesses of friction material were the output vector. The electrical signals are easy to measure and need no further data processing. After optimizing the parameters of LS-SVM model, the prediction accuracy of the model was improved. The prediction results of examples proved that prediction values and experimental validity of proposed model. values were in good agreement which confirmed the