基于振动的滚珠丝杠预紧力失效预测

Research on Failure Prediction of Ball Screw Preload Based on Vibration

滚珠丝杠预紧力随着摩擦加剧逐渐减小直至失效,预紧力的失效将严重影响机床的进给精度和加工质量,因此提出一种基于变分模态分解和多尺度熵理论的预紧力失效诊断方法。基于变分模态分解理论分解滚珠丝杠振动信号,得到一系列本征模态分量,依据相似性测度筛选敏感分量信号进行振动信号重构,实现降噪处理;接着提取重构后振动信号的时域、频域特征及多尺度熵值构成滚珠丝杠副预紧力失效诊断特征向量集;最后,使用提取的特征训练神经网络,应用训练好的BP神经网络实现对滚珠丝杠副预紧力失效的诊断。实验结果表明,结合变分模态分解和多尺度熵的方法可以有效实现滚珠丝杠副预紧力失效的诊断。

As the friction intensifies,the preload of the ball screw gradually decreases until it fails.The failure of the preload will seriously affect the feed accuracy and processing quality of the machine tool.Therefore,a preload failure diagnosis method based on the variational mode decomposition and multi-scale entropy theory is proposed.Based on the variational modal decomposition theory,the vibration signal of the ball screw is decomposed,and a series of intrinsic modal components are obtained.The sensitive component signals are filtered according to the similarity measure to reconstruct the vibration signal,so as to achieve noise reduction;Secondly,the time-domain,frequency-domain characteristics and multi-scale entropy values of the reconstructed vibration signals are extracted to form a set of fault diagnosis feature vectors of ball screw preload;Finally,the extracted features are used to train the neural network,and the trained BP neural network is used to diagnose the preload failure of the ball screw pair.The experimental results show that the method combining variational mode decomposition and multi-scale entropy can effectively diagnose the preload failure of ball screw pairs.