基于分层稀疏编码的轴承剩余寿命预测方法

A RUL prediction approach for rolling bearing based on hierarchical sparse coding

剩余寿命预测技术是装备智能维护与智能制造的关键技术。滚动轴承作为旋转机械装备关键零/部件之一,对其进行剩余寿命预测具有重要工程与实际意义,因此提出一种基于分层稀疏编码的滚动轴承剩余寿命预测方法。该方法作为深度学习模型的一种,克服了传统机器学习模型需要大量训练、标签学习以及鲁棒性差的缺点,有效提高了轴承剩余寿命预测精度。实验结果表明该方法具有更高的预测精度和更好的鲁棒性。

Remaining Useful Life(RUL) prediction is very important for intelligent maintenance of equipment and intelligent manufacturing.Roller bearings are among the most frequently used components in the majority of rotating machines,thus RUL prediction of roller bearing is also very significant for practical applications.A novel RUL prediction approach is proposed for roller bearing based on Hierarchical Sparse Coding(HSC).HSC is one of deep learning techniques and it overcomes the shortcomings of traditional machine learning models which require trainings,labels and have poor robustness.This proposed method can effectively improve RUL prediction accuracy of roller bearings.Experimental validation are conducted to well demonstrate its accuracy and robustness.