CAJ | 학술논문

为更好地表征滚动轴承性能退化趋势，提出基于流形学习和最小二乘支持向量机的滚动轴承退化趋势预测新方法。提取振动信号的多域特征组成高维特征集，利用局部保持投影算法（LPP）对多域高维特征集进行维数约简，消除各特征指标之间的冗余、冲突等问题。将维数约简后的特征向量作为最小二乘支持向量机的输入，建立退化趋势预测模型，完成退化趋势预测。运用实测的滚动轴承全寿命实验数据进行检验，结果表明该方法能获得准确的预测结果。
위경호지표정곤동축승성능퇴화추세，제출기우류형학습화최소이승지지향량궤적곤동축승퇴화추세예측신방법。제취진동신호적다역특정조성고유특정집，이용국부보지투영산법（LPP）대다역고유특정집진행유수약간，소제각특정지표지간적용여、충돌등문제。장유수약간후적특정향량작위최소이승지지향량궤적수입，건립퇴화추세예측모형，완성퇴화추세예측。운용실측적곤동축승전수명실험수거진행검험，결과표명해방법능획득준학적예측결과。
A new prediction method was proposed based on manifold learning and least squares support vector machine to describe the rolling bearing degradation trend.Time-domain features and features based on information entropy were extracted to construct high-dimensional characteristic sets.The locality preserving projection algorithm was used for dimensionality reduction in order to eliminate the problem of redundancy between each indicators.The characteristic features were input to the least squares support vector machine to train and construct a model,so as to accomplish the trend prediction.The rolling bearing run-to-failure tests were carried out to inspect the prediction model,and the results demonstrate the effectiveness and accurateness of the proposed method.