基于液压马达内泄漏故障的预测研究  

Fault prediction research based on internal leakage fault in hydraulic motor

作　　者：田秋实[1] 赵鹏[1] Qiu-shi TIAN;Peng ZHAO(Civil Aviation Flight University of China,Guanghan 618307,China)

机构地区：[1]中国民用航空飞行学院航空工程学院,四川广汉618307

出　　处：《机床与液压》2020年第18期99-103,151,共6页Machine Tool & Hydraulics

基　　金：四川省教育厅科研重点项目(16ZA0020);中国民航青年飞行基金项目(Q2018-66)。

摘　　要：为了研究液压马达可能出现内泄漏故障,并对液压马达状态进行预测性监控。通过建立液压马达内泄漏故障试验平台,获得液压马达内泄漏的故障数据。在MATLAB中建立基于T-S模糊神经网络的故障预测模型,将实验数据用于模型的训练以及预测结果的验证。对预测结果进行分析后,讨论了不同数量样本用于模型训练对故障预测精度的影响。在分析过程中发现数据波动较大的地方,相对误差较大。研究后发现,通过将实验数据进行拟合后,用相同的模型进行训练和预测,讨论了拟合后不同数量样本的预测模型精度与拟合前的差别。结果表明:虽然模型训练的数据数量越大,预测的精度越高,但数据拟合后只需将较少的数据用于建模,预测就能达到较高的精度,为小数据样本进行故障分析提供了参考。To study the intermal leakage of the hydraulic motor and monitor the state of the hydraulic motor.The fault data of the leakage in the hydraulic motor is obtained by establishing a simulation leakage fault test platform.A fault prediction model based onT-S fuzzy neural network is established in MATLAB,and the experimental data is used for the training of the model and verify the prediction results.After analyzing the prediction results,the influence of different number of samples in model prediction on the accuracy of fault prediction is discussed.Where the data fluctuates greatly during the analysis,the relative error is large.After the study,it was found that by fitting the experimental data,using the same model for training and prediction,the difference between the ac-curacy of the prediction model and the pre-fit difference of different numbers of samples after fitting was dis-cussed.The results show that although the larger the number of data trained by the model,the higher the accuracy of the prediction,but only a small amount of data is used for modeling after the data is fitted,the prediction can achieve higher precision.It provides a reference for fault analysis of small data samples.

关 键 词：T-S模糊神经网络 液压马达 故障预测 数据拟合 

分 类 号：TH137[机械工程—机械制造及自动化]