Research On Reliability Analysis And Maintenance Decision Optimization Of Mechanical Rolling Bearing

As the most widely used and critical part of mechanical rotating parts,the safety,economy and reliability of rolling bearings have always been the focus of attention,and their operational status is directly related to the safety of the entire mechanical system.Accurate knowledge of the operating condition of rolling bearings and effective maintenance decisions are key elements in ensuring that machinery and equipment can work healthily and efficiently.Therefore,scientific monitoring of the operating condition of bearings is of great significance in reducing the incidence of accidents,improving the efficiency of equipment and reducing maintenance costs.Through real-time monitoring of equipment operating conditions,it provides a basis for bearing reliability analysis and the development of maintenance strategies.This paper conducts research as the following:(1)For the issue that the original vibration signal of a rolling bearing contains a lot of noise,this paper compares various modal decompositions to find the best method to process the signal,and then uses root mean square cliffness and reconstruction error to assess how well each method reduces noise.Comparing the simulated signals demonstrates the superiority of variational modal decomposition;example verification using bearing data from Xi’an Jiaotong University further demonstrates the good effect of variational modal decomposition in the noise reduction of bearing vibration signal decomposition,which provides data assurance for the subsequent reliability analysis and maintenance decisionmaking.(2)The reliability-based Weibull distribution is used to characterize the general law of bearing degradation from the standpoint of ensuring the rolling bearings’ operational dependability.After noise reduction,the data is first processed before the time domain and frequency domain parameter signs are extracted.A principal component analysis method is chosen to reduce the dimensionality of the Govett levy set,and the first three principal elements of the principal components are extracted to create the Weibull proportional risk model.This solution addresses the issue that the high-dimensional feature indexes of the whole-life data cannot completely reflect the operating status of the bearings.The great likelihood technique is used for parameter estimation to address the issue that the parameters of the Weibull proportional risk model are challenging to estimate.The example analysis using the whole-life vibration data of the bearing from the University of Cincinnati shows that this method can more accurately represent the bearing’s degradation state trend.It also yields full reliability results and offers data support for the subsequent bearing maintenance decision.(3)Based on the results of the previous data processing and reliability analysis calculation,this paper further investigates the specific scheme of maintenance decision optimization.From the perspective of high efficiency and high economy,a single objective availability and single objective maintenance cost model is established,as well as a fuzzy multi-objective planning model combined with reliability,and the model constraints are improved by using the hierarchical analysis method proportional scale.Based on the operation status,the online monitoring and offline monitoring maintenance decision scheme is proposed,and finally the superiority of the improved fuzzy multi-objective optimization is proved by example verification,and the maintenance timing and maintenance strategy based on the operation statu s of the bearing is given to realize a series of measures for the scientific maintenance of rolling bearings.