Optimizing Of Asymmetric Rail Profile In The Curve Segment Of Heavy Haul Railway

Increase of the axle load is the main development direction of heavy haul railways. As a result, it aggravates the interaction between wheel and rail, speeds up the wear rate of the rail, and shortens the lifecycle of the rail. These problems are even worse in the curve segment. For this reason, it shows a great significance in engineering to optimize the rail profile in curve. The optimized rail profile has the minimum wear rate. It contributes to decreasing the wear and extending the service life of the curved rail. Based on the optimization theory, the numerical method is applied to optimize the asymmetric rail surface of the curve segment of the heavy haul railway.Based on the wheel-rail contact theory and wear model of Archard, the vehicle-track coupling model aimed at analyzing the wear of the rail is established.According to the characteristics of the rail wear in the curve, the abscissa of the rail head in optimized interval is selected between-10 mm and 36 mm. Ordinate of the fourteen points chosen as optimized point is defined as independent variable.Combined with grinding depth of the rail, D-optimal sampling is used to get the sample of rail profile. In the cycle of the given wear analysis, the profiles are simulated by the vehicle-track coupling model so as to obtain the wear data. Rail wear rate is selected as the dependent variable. The goal to get the nonlinear numerical model between the independent and dependent variables is achieved with the help of support vector machine whose parameters including of C, penalty coefficient, and g, nuclear radius, are optimized by using particle swarm optimization.By being evaluated, the numerical model can be used to optimize the rail profile in curve. By being taken the geometric features of the rail profile and characteristics of the model trained, genetic algorithm is used to solve the numerical model. Eight groups of solutions are calculated after the parameters such as the population of the independent variable, genetic operators, design of capable function and termination conditions are determined.The wear of the profiles made up by optimized ordinates are simulated. The predicted wear rate and simulated wear rate of the optimized profiles are smaller than the standard profile. Meanwhile, the total number of vehicle and axle load has been increased. By being simulated and analyzed, in the optimized profiles, Opt5 has the minimum wear rate which is 1.715 62×10-5 mm2/t and decreases by 5.8 percent than the standard profile. Opt3 has the maximum total number of axle load with a value of100.936 Mt which increases by 34 percent than the standard profile. Indicators including derailment criterion, lateral force, rate of wheel load reduction and vibration acceleration are used to evaluate the security and stability of the optimized profile. It can be analyzed that the optimized solutions are reasonable to use.