Parameter Optimization Design Of The Magnetic Inspection Process For The Railway Vehicle Wheel Webs

The wheelset is an important moving part of the railway vehicles. During manufacturing and assembly process the wheelsets are subject to residual stresses. Apart from that, the wheelsets experience complex mechanical and thermal loads during service and flaws are likely to emerge. Moreover, the sprag holes in the wheel webs, which are manufactured for easy handling, intend to aggrandize stress concentration and lead the wheels to break. In fact the wheels with sprag holes have led to many accidents recently. Over half of the wheels in service contain sprag holes in China and cannot be totally replaced within 10 years. Thus the sprag hole caused flaws have become a severe threat to the railway safety.Consequently it is a central issue to detect sprag hole caused flaws accurately and effectively and repair them in time. It is necessary to establish standards for magnetic inspection process (MIP) of wheels. Now there are only standards of MIP for shafts. Based on the existing standards, this thesis figures out the accuracy, cost and time as indicator for optimization and perform parameter design (PD) for MIP for wheels. At first we quantify the stress distribution in the wheel webs through mechanical-thermal coupled finite element analysis and determine the circumferential region near sprag hole as inspection focus. Then we perform design of experiment (DoE) with the aid of electromagnetic finite element analysis, administrate field surveys and set up the relationship between definition of the magnetic display and four key factors, namely the magnetizing current, magnetizing time, position of the sprag hole during inspection and spray time of magnetic suspension by using methods of variance analysis and multivariate regression. Taking into account the relationship between the four factors and inspection time and cost, we obtain the optimization solution through multi-response surface techniques. This solution could be the recommendation values for MIP standards for wheels. The field test confirms the result of parameter design finally.The emphasis of the thesis is that, distinctive from other methods for determining MIP parameters for wheels, it is more reasonable to consider not only the demand for accuracy but also other factors like time and cost in order to search for overall optimization, i.e. improving the efficiency and economy while maintaining accuracy. Parameter design based on statistical methodology should be the tool to determine parameter values, instead of mere experience and sense.