Research On The Application Of Fault Diagnosis Methods Of The Explosion Relief Valve On The Electric Cars’ Battery Top Cap

Over the past20years, world energy problems become more and more important, the development of new energy automobile such as electric automobile becomes an important proposition researched by the countries of the world. How to make electric automobiles rapidly, high energy and efficiently charge the battery repeatability is the main bottleneck in the current design philosophy of electric automobiles. How to reasonably design the explosion relief valves in the battery components is the necessary issue. Reasonably designing the explosion relief valves is already puzzling the electric vehicles’ development.The explosion relief valve on the electric cars’ battery top cap in the paper are the gland cover safety of the electric automobile battery pack. As core safety components to insure automobiles run normally, researching a detection method on the production line of explosion relief valves becomes rather more important. Therefore, the detection and diagnosis method is the proposition in this paper.In this paper, because the weld of the explosion relief valve is rather more subtle, there are certain difficulties in extraction and identification of the weld defects. In order to obtain more accurate diagnosis of the weld defects, the conductivity measurements of the work pieces are firstly conducted in this thesis. According to the conductivity values, the weld defects locations could be preliminary estimated. And then, with the help of the X-ray image processing, the weld defects would be recognized by using the three-layer forward feedback BP neural network. The work pieces researched in this papers need to bear certain amount of pressure, so that the thesis put forward with taking bearing strength tests on the these pieces after identification. What’s more, combined with the engineering practical application, this paper stated the experimental process of bearing s tests. The thesis stated the testing process for the preliminary decision, meticulous identification and pressure-bearing. The testing process lays a solid foundation for the practical application in the future.This thesis, which combined the characteristic of X-ray image with the conductivity measurements at the same time, employed the efficient image progressing algorithms, proposed a appropriate image progressing sequence, applied the improved three-layer feed-forward BP neural network to obtain the defect identification characteristics parameters, and to realize intelligent identification of the weld defects, and established the system of weld defects parameters measurement and diagnosis identification. In the improved three-layer feed-forward BP neural network structure, the number of the hidden node.momentum coefficient, the error level, step length and other network parameters were obtained by experiments for the optimum value. VC++, MATLAB and Microsoft Office Access were employed to realize system function mixed programming. The establishment of the system and database could lay the foundation of the research on weld parameters and weld image processing.