摘要
信号不确定性的正确表达是故障诊断的先决条件。然而,在实际情况中,轮对轴承的信号存在各种不确定性,采用传统方法处理这类不确定性,存在信息丢失问题。提出一种基于概率包络的轮对轴承故障诊断方法。对原始信号进行分布类型检验,针对不同分布特点使用不同方法进行概率包络建模。提取概率包络模型的几何形状作为故障特征,并将其输入支持向量机(SVM)训练获得故障诊断模型。以公共数据集及实测数据进行诊断测试,并对诊断结果进行比对验证。实验结果表明,该方法合理有效,提高了诊断精度和效率。
The correct expression of signal uncertainty is a prerequisite for fault diagnosis.However,in practice,there are various uncertainties in the signal of the wheelset bearing.There are several shortcomings including missing useful information when traditional methods are used to deal with those uncertainties.To address those deficiencies,a fault diagnosis method for wheelset bearing based on probability envelope is proposed.Firstly,we test the distribution type of the original signal,and use different methods to model the probability envelope according to the different distribution characteristics.Next,feature extraction is carried out for the geometric shape of the probability envelope model.After that,the feature vectors are taken as input,and support vector machine(SVM)is used to train the model.Finally,experiments on publicly available datasets and real-world datasets are carried out,and the diagnostic results are compared and verified.The experimental results show that the method is reasonable and feasible,and has high diagnostic accuracy and efficiency.
作者
丁家满
原琦
李川
DING Jiaman;YUAN Qi;LI Chuan(Faculty of Information Engineering and Automation,Kunming University of Science and Technology,Kunming 650500,China)
出处
《铁道学报》
EI
CAS
CSCD
北大核心
2020年第4期52-58,共7页
Journal of the China Railway Society
基金
国家自然科学基金(51467007)。
