基于KPCA-LSTM的旋转机械剩余使用寿命预测

Remaining useful life prediction of rotating machinery based on KPCA-LSTM

旋转机械的剩余使用寿命(remaining useful life, RUL)预测对工业设备预测和健康管理的具有重要意义。该文针对多传感器冗余数据导致旋转机械退化信息提取困难、剩余使用寿命预测效果差的问题,提出了一种基于核主成分分析-长短期记忆网络(kernel principal component analysis-long short term memory, KPCA-LSTM)的方法对旋转机械剩余使用寿命预测。首先,分析旋转机械的多维退化数据,选择可以表征旋转机械退化的数据;其次,对退化数据进行(kernel principal component analysis, KPCA)融合及特征提取,将降维融合的特征作为预测模型的输入;然后构建旋转机械的健康指标,并通过多阶微分划分旋转机械的不同健康状态,建立KPCA-LSTM模型对旋转机械的剩余使用寿命进行预测;最后,在实验室搭建的矿用减速器平台上进行了试验验证。试验结果表明:该文所提方法与LSTM、粒子群优化LSTM的方法比较,该方法预测效果优于其他两种模型,并降低模型训练的复杂性,减少预测用时。

The prediction of the remaining useful life(RUL)of rotating machinery is of great significance to the prediction and health management of industrial equipment.This paper addresses the problems of harrowing extraction of degradation information and poor prediction of the RUL of rotating machinery due to redundant data from multiple sensors.This paper proposes a kernel principal component analysis-long short term memory(KPCA-LSTM)based method for predicting the RUL of rotating machines.Firstly,the multi-dimensional degradation data of rotating machinery were analyzed,and the data that can characterize the degradation of rotating machinery were selected.Secondly,KPCA fusion and feature extraction were carried out on the degraded data,and the features of dimensionality reduction fusion were used as the input of the prediction model.Then,the health indicators of rotating machinery were constructed,and a KPCA-LSTM model was established to predict the remaining useful life of rotating machinery by dividing the different health states of rotating machinery by multi-order differentiation.Finally,the proposed method was tested by a mine reducer platform.Experimental results show that compared with LSTM and particle swarm optimization LSTM,the proposed method has better prediction effect than the other two models,and reduces the complexity of model training and the time of prediction.