基于向量加权平均算法优化的轴承剩余寿命预测OA
Remaining life prediction of bearings based on INFO optimization
针对轴承振动信号复杂度高的问题,提出基于向量加权平均算法-变分模态分解(INFO-VMD)的特征提取方法.另外,由于轴承振动信号特征差异性较大,因此提出多特征筛选的深度极限学习机预测模型(MFDELM),从而提高预测的准确度.首先,利用INFO-VMD方法寻找最优层数和惩罚系数;然后,对模态分量分别提取时域和频域特征;最后,将特征集合输入到MFDELM预测模型中,计算出轴承剩余使用寿命.计算机仿真实验结果表明,文章预测模型得分为0.47,比基于长短期记忆网络(LSTM)模型得分提高了0.16,同时比基于门控递归单元(GRU)模型得分提高了0.21.通过轴承全寿命实验验证了提出方法的有效性.
Aiming at the problem of high complexity of bearing vibration signals,a feature extraction method based on the weighted mean of vectors-variational mode decomposition-variational modal de-composition(INFO-VMD)was proposed.In addition,due to the high variability of bearing vibration signal characteristics,the Deep Extreme Learning Machine Prediction Model with Multi-feature Filte-ring(MFDELM)was proposed to improve the accuracy of prediction.Firstly,the INFO-VMD meth-od was used to find the optimal number of layers and the penalty coefficient,and then the time domain and frequency domain features were extracted from the modal components.Finally,the feature set was input into the MFDELM prediction model to calculate the remaining service life of the bearing.The results of the computer simulations show that the prediction model score is 0.47,which is 0.16 better than the based on LSTM model and 0.21 better than the based on GRU model,and the effec-tiveness of the proposed method is verified through full life Experiment of Bearing.
周靖诺;郇战;陈瑛;朱学勤
常州大学 王铮微电子学院,江苏 常州 213164常州大学 王铮微电子学院,江苏 常州 213164常州大学 王铮微电子学院,江苏 常州 213164江苏立达电梯有限公司,江苏 常州 213164
机械制造
向量加权平均算法变分模态分解滚动轴承深度极限学习机寿命预测
weighted mean of vectors algorithmvariational mode decompositionrolling bearingsdeep extreme learning machinelife prediction
《常州大学学报（自然科学版）》 2026 (1)
66-73,8
国家自然科学基金资助项目(62201093).
评论