Research On Remote Intelligent Health Assessment System For Power Supply Vehicles

Health assessment is the basis for implementation of condition-based maintenance.At present,there is no effective health assessment method and systematic implementation scheme for power supply vehicles.Based on this,considering the mobility of power supply vehicles,this thesis starts with the construction of an intelligent health assessment system framework for power supply vehicles based on local-remote fusion.Under this framework,the intelligent health assessment method for power supply vehicles is studied,which mainly includes fault diagnosis,fault prediction and system-level health assessment.The main work is as follows:1)Construction of Intelligent Health Assessment System Framework for Power Supply Vehicles Based on Local-remote FusionConsidering the actual requirement of the health assessment for power supply vehicles and the mobility of power supply vehicles,and with the help of the advantages of network technology,this thesis constructs an intelligent health assessment system framework for power supply vehicles based on local-remote fusion.The framework makes full use of the real-time signal processing capability of local DSP and the large data analysis capability of remote Spark platform,and provides data storage and platform support for the joint intelligent health assessment of power supply vehicles based on cloud edge collaboration.2)Research on the Fault Diagnosis Method of Power Supply Vehicles Based on LSTM-SPRTFault diagnosis is the beginning of health assessment for power supply vehicles.For the current fault diagnosis methods based on deep learning,most of them still have the defects of low accuracy and poor reliability due to inadequate extraction of fault features and judgment based on isolated single sample.In this thesis,a fault diagnosis method for power supply vehicles based on LSTM-SPRT is proposed.The simulation results show that the method effectively improves the accuracy and reliability of fault diagnosis for power supply vehicles.3)Research on the Fault Prediction Method for Power Supply Vehicles Based on Multi-state Time Series Prediction LearningFault prediction is an important part of the health assessment of power supply vehicles.Aiming at the difficulty of applying the existing fault prediction methods to power supply vehicles,In this thesis,a fault prediction method for power supply vehicles based on multi-state time series prediction learning is proposed.In the method,a real-time prediction model for status trend of power supply vehicles based on LSTM network is established,and the improved kNN algorithm is applied to predict the fault of power supply vehicles.The simulation results show that the method has high fault prediction accuracy.4)Research on System-level Health Assessment Method of Power Supply Vehicles Based on Deep Learning Reconstruction ModelIn order to evaluate the overall health status of the power supply vehicles,considering the different effects of different components of the power supply vehicles on the system-level health status,a system-level health assessment method based on the deep learning reconstruction model is proposed.This method establishes a multistate reconstruction model of power supply vehicles based on LSTM network,and builds the system-level health index by using the prediction residual of the reconstruction model,achieves the goal of quantitative evaluation of system-level health status of the power supply vehicles.5)Design and Development of Remote Intelligent Health Assessment System for Power Supply VehiclesBased on the verification of the above research methods on the simulation platform of power supply vehicles,this thesis integrates the current advanced technology framework fully and implements the remote intelligent health assessment system for power supply vehicles.The testing of functional and performance shows that the system has excellent performance,friendly interface,flexible operation and strong practicability.