Research On State Evaluation Of Rolling Bearing Based On Similarity Measurement

With the introduction of intelligent manufacturing,China has made remarkable achievements in both the production and manufacturing of precision instruments and large equipment.And accompanied by the rise of manufacturing and the development of national economy,the growing demand for machinery and equipment,also more and more demanding,and for the safety testing and control of the mechanical equipment is especially important,once the mechanical equipment appears wear and aging and other faults in production and operation,if it cannot be found in time and necessary protective measures are taken,it will cause major accidents.It will not only bring about huge economic losses,but even endanger people’s lives.If the fault can be found out accurately and timely and the correct decision can be made,the occurrence of major catastrophic accidents can be largely avoided.Gearboxes and bearings are important components of mechanical transmission systems,affecting the critical performance of the equipment.Therefore,it is necessary to carry out relevant life prediction and corresponding state assessment to ensure the safe operation of mechanical equipment and achieve healthy management of mechanical equipment.In this paper,the state monitoring of bench rolling bearing and field fan gearbox bearing data is the main research object.The similarity measurement method is applied to the time-frequency distribution of mechanical vibration signals and the principal curves extracted from high-dimensional characteristic space.Based on the similarity measure,the state evaluation method of mechanical equipment realizes the monitoring of the running state of mechanical equipment,and based on this,an early warning analysis of early bearing faults is carried out.Finally,the software system is designed to meet its application in subsequent engineering demand.This paper mainly focuses on the following aspects:(1)Firstly,based on the traditional signal analysis,several time-frequency analysis methods,such as Short-time Fourier transform(STFT),Wavelet transform,S transform,Hilbert-Huang transform and Wigner-ville distribution,are compared and studied.Then point out the shortcomings and advantages of various time-frequency analysis methods..Secondly,this paper introduces the methods to measure similarity between two images,such as the Mean square error(MSE)and Peak signal-to-noise ratio(PSNR)which only consider the similarity measure method of grey value,such as Structural similarity index(SSIM)of integrated image brightness,contrast and structure information of,Feature similarity(FSIM)which combined with Phase consistency(PC)and gradient magnitude(GM)similarity measure methods of characteristics ect.Finally,combining Wigner-ville distribution and Feature similarity(FSIM),a mechanical equipment state evaluation method for time-frequency distribution similarity measurement is proposed.The effectiveness of the method is verified by the whole life experiment of rolling bearing.(2)Aiming at the insensitivity and instability of single features to early faults,a high-dimensional feature set combining time domain,frequency domain and wavelet packet energy characteristics is proposed,and the problem of high-dimensional feature set information redundancy,computational complexity and low operational efficiency is solved.The method of Laplacian Eigenmaps(LE)and soft-K principal curves extraction for the original high-dimensional feature set is studied,and a comparative verification experiment is carried out.Finally,the curve similarity is evaluated,and the state evaluation curve is made by using classical Euclidean distance,Dynamic time warping(DTW),Discrete Frechet distance measurement and Hausdorff distance.The experimental analysis of the rolling bearing life test shows that the method can effectively identify the early faults in the running process of the rolling bearing and provide the basis for the maintenance decision of the mechanical equipment.The method is compared with the state evaluation methods such as HMM and DBN,and the method is more sensitive to the early failure of the bearing.(3)Combined with the advantages of C# in interface development and port operation,and MATLAB in data processing and graphics,C# language and MATLAB platform mixed programing(based on.NET component technology)were applied to jointly develop the mechanical equipment state assessment software system.The vibration data of wind turbine gearbox bearing is analyzed.The results show that the software system can effectively evaluate the running state of the gearbox and make timely warning of the fault,thus proving the effectiveness of the software system.