Research On The Security State Prediction And Lathing Strategy Optimization For The Wheelset Of Urban Rail Train

ABSTRACT:With the rapid growth of urban population, urban rail trains, who have the characteristic of large volume and high speed, have been increasingly significant for improving the safety of people’s normal travel and further solving urban traffic problems. As the key component of urban rail train’s running gear, the wheelset bears the rail force directly, which naturally leads to some damages such as the wheelset abrasion and the wheel diameter difference, and further accelerates the wheel-rail relationships’changes because of the changes in tread and its surroundings. Therefore, in order to ensure the safe operation of urban rail trains, mastering the state of the wheelset and taking appropriate treatment measures are necessary. So far, the costs of urban rail train wheelset for purchase have been extremely high in China. However, the costs of urban rail train wheelset for lathing are still rising even faster by the improper lathing strategy. Therefore, the development of rational lathing strategy is of great significance to improve the safety and reduce operating costs of urban rail train.First, in this paper, the fuzzy and the security domains theory were introduced to acquire the wheelset state accurately. Then the fuzzy security domains was proposed on the basis of the above two theories, and used to identify the wheelset state as the state of low risk space, medium risk space or high risk space.Next, in order to predict the future state of wheelset, the model for predicting the size of wheelset was studied. After the introduction of basic knowledge and modeling steps of BP neural networks and support vector regression (SVR), BP neural network prediction model for wheelset’s size was established using the traversal method, and GA-SVR prediction model for wheelset’s size was also built by using genetic algorithms (GA). The comparative analysis of simulation results proved that GA-SVR prediction model for wheelset’s size is more superior and effective.Finally, the lathing strategy optimization strategy was studied. According to the wheelset maintenance data of a certain type of urban rail train, the optimal recovery threshold of flange thickness was acquired through the liner regression analysis of flange wear and flange thickness, which was obtained taking advantage of SPSS. The wheel tread geometry is determined by the rim thickness, wheel diameter and the wheel diameter difference. So the lathing strategy, which took consideration of flange thickness and diameter difference, was studied, and even certified by the repair data from the field.