Research On Data-driven Method Of Highway Electromechanical Equipment Performance Degradation Analysis

Condition-based maintenance is the main development direction of highway electromechanical equipment maintenance industry,and analysis of equipment performance degradation is the core content of condition-based maintenance.In this paper,a data-driven method for performance degradation analysis,which combines the data-driven method with the performance degradation analysis of electromechanical equipment in the highway scene,is proposed.This method provides a technical theoretical support for condition-based maintenance of highway electromechanical equipment.First,according to related researches and characteristics of performance degradation analysis of highway electromechanical equipment,the extraction of performance degradation feature and the impact of operating environment on the performance degradation process are regarded as key points of the research.Second,the overall theoretical framework and the research ideas and method principle of each part in the framework are introduced.Then,in the research of degradation characteristics,the equipment monitoring data under different working conditions are used to construct the state characteristic parameters by calculating the characteristic values of different data dimensions.Moreover,based on the correlation between characteristic parameters and performance indicators parameter,the sensitive parameters of the degradation process are screened out.In the research of the influence of environment on the degradation,an influence mechanism model based on the Arrhenius model is build.Considering the time cumulative effect,a cumulative effect model to quantify the impact of environmental factors is further established.And the optimization method is used to look for the hyperparameters of the model.Finally,the sensitive parameters and environmental impact are comprehensively considered in the degradation process modeling.In addition,the RBFNN is used to establish the mapping relationship between the two and the equipment performance,and the RVM is used to fit the timing process of the pseudo equipment performance indicators outputted by RBFNN.Through the combination of RBFNN and RVM,the description of the timing process of equipment performance degradation is realized.Based on the actual experimental data of UPS power supply equipment,LED lighting equipment and board card equipment in the highway electromechanical system,each part of the data-driven method framework is verified.Five sensitive parameters that are strongly related to the degradation process of UPS power equipment are screened out.And the temperature is found to have a significant impact on the degradation process of board card equipment.Taking UPS power supply equipment as an example,the degradation analysis is carried out.The actual monitoring data of similar equipment is used for verification experiments.The experimental results show that the goodness of fit of the model reaches 0.936,and the root mean square error is 0.0719.These values proof that the method can accurately describe the equipment performance degradation process and can provide technical support for conditionbased maintenance of highway electromechanical equipment.