Research On Key Technology Of Fault Prognostics And Health Management For EMU

In recent years, the research of Optimize complex equipment maintenance strategy is being paid more and more attention. Prognostic and Health Management and Condition based maintenance are the hot spots of research. By means of advanced sensor technology, a large number of complex equipment operating data information can be collected and stored. However, the use of these data to guide the maintenance of equipment and the development of advanced maintenance strategy is an important research content of the current maintenance work. With the rapid development of high speed railway, railway-related operation and maintenance departments are facing the great challenge, which is to ensure the safety and reliability of the operation.The maintenance mode of high speed railway in China is transforming from traditional "plan repair" to "state repair". Maintaining equipment when there is data to prove the device to fail is possible to improve maintenance efficiency, save maintenance costs and increase safety and reliability of the equipment. Economic development and improvement of maintenance strategy requires an accurate grasp of the equipment failure time, health status assessment, and the remaining useful life (RUL). So, failure prediction is core research of high-speed rail maintenance strategy optimization. This paper focuses on the failure prediction technique on EMU high-speed rail, and the main contents include:(1) In order to optimize the maintenance strategy of complex equipment, this paper analyzes the current mainstream fault prediction algorithm, and studies the failure mechanism of the traction system of the key equipment of high speed train,4 kinds of degenerate states are divided. On the basis of this, the HSMM prediction model of the traction system based on the degraded state is proposed.(2) Due to the limitations of HSMM, this paper uses the PSO algorithm to optimize the HSMM parameters estimation. On this basis, we improved the PSO algorithm and give the new characteristics of the particle. Thus, a fission particle swarm optimization (FPSO) algorithm is proposed to optimize the parameters of the HSMM model. FPSO-HSMM prediction model is established.(3) The improved algorithm is applied to EMU traction system, Evaluation of the health state of traction system and estimation of remaining life. Assess the health status of the traction system and estimate the remaining life successfully.Experimental results show that the proposed FPSO-HMM prediction model can be a good predictor of the key components of EMU. Compared with the traditional prediction models, the prediction accuracy has been greatly improved.