Study On The Wind-Induced Safety Of The High-Speed Train Based On The Reliability

With the improvement of the train speed, the impact of the airflow on the train becomes more and more obvious. The aerodynamic problem of the high-speed train becomes an important engineering issue against the operational safety and speed improvement. The strong crosswinds in nature seriously deteriorate the aerodynamic performance of the high-speed train, and pose a grave threat to the safe operation of the high-speed train. To investigate the strong crosswind stability of the high-speed train, the computational methods of the unsteady aerodynamic loads for the high-speed train exposed to the longitudinal fluctuating winds were set up to study the unsteady aerodynamic characteristics of the high-speed train. The reliability assessment and reliability sensitivity methods for the operational safety of the high-speed train were proposed to study the operational safety characteristics of the high-speed train. The multi-objective optimization and reliability based multi-objective optimization models for the kinetic parameters of the vehicle were set up to perform the wind resistance design of the high-speed train. The computational methods for the unsteady aerodynamic loads of the high-speed train were set up when the longitudinal and lateral components of the stochastic winds were taken into account.The computational fluid dynamics technology was used to set up the numerical methods for the steady aerodynamic load coefficients of the high-speed train exposed to crosswinds. In a crosswind scenario, the fluctuating winds of a moving point that was stationary relative to the train were calculated in this paper based on Cooper theory and harmonic superposition method, and the effects of train speeds and average wind speeds on the dimensionless power spectral density were analyzed. In the cross stochastic wind scenario, four methods, which were the quasi-steady steady assumption, weighting function, improved quasi-steady and improved weighting function, were used to analyze the unsteady aerodynamic loads and their probability distributions. Moreover, in a sidewind scenario, the fluctuating winds of a moving point at the train position were calculated based on Cooper theory and harmonic superposition method, and the effects of the wind angle on the dimensionless power spectral density and the fluctuating winds were studied. The computational method for the unsteady aerodynamic loads of the high-speed train exposed to any wind direction was set up, and the statistics for the aerodynamic loads were studied.The vehicle system dynamic model was built based on the multi-body system dynamic theory. The unsteady aerodynamic loads were dealt with as external loads acting on the vehicle system dynamic model to compute the safety indexes, which are related to the performance function of the system. The stochastic winds, side force coefficient, lift force coefficient, roll moment coefficient, yaw moment coefficient and pitch moment coefficient were dealt with as basic random variables, and the semi-analytical methods and Monte Carlo simulation were used to calculate the random reliability and reliability sensitivity of the high-speed train exposed to stochastic winds. This finally led to the probabilistic characteristic wind curve, which could effectively assess the overturning probability of the high-speed train at a mean wind speed and train speed. Compared with the characteristic wind curve (safety region curve) computed by the conventional deterministic method, it is found that the operational safety region computed by the traditional deterministic method is too conservative, and a more reasonable safety region can be obtained using the method based on the reliability theory.Due to the fuzziness of the limit state of the structure, the failure of the structure was considered as a fuzzy random event. The importance sampling for computing the structural reliability and reliability sensitivity of the fuzzy random event was built and introduced to the high-speed train in order to study the fuzzy random reliability and reliability sensitivity of the high-speed train exposed to stochastic winds. Considering the fuzziness of the failure criteria would lead to an increase of the probability of failure. Considering the fuzzy limit state as the deterministic limit state would lead to an underestimation of the probability of failure. On account of the uncertainty of the distribution parameters for the random variables in the system, the computational method of the structural reliability was set up for the distribution parameters with random uncertainty. This method was applied to the high-speed train to investigate how the random uncertainty of the mean and standard deviation for the random variables affected the operational safety of the high-speed train. Further, the fuzziness of the failure state and random uncertainty of the distribution parameter were taken into account, and the importance sampling formula and the fast approximate method were set up to compute the probability of failure.To improve the wind resistance of the high-speed train, the multi-objective optimization design model of the kinetic parameters was established. The load reduction factor and wheel lateral force were taken as optimization objectives, and the multi-objective genetic algorithm NSGA-Ⅱ was used for the automatic optimization design of the kinetic parameters of the high-speed train. Through the multi-objective optimization of the kinetic parameters of the high-speed train, the values of the load reduction factor and wheel lateral force for each case were reduced significantly. Meanwhile, the Sperling comfort and riding comfort of the train before and after optimization were simulated. It is found that, the lateral vibration acceleration, lateral Sperling index and comfort index were significantly reduced after optimization. The vertical vibration acceleration and the vertical Sperling index were increased, but still far less than their limit, which proves that the optimized kinetic parameters will improve the wind stability without deteriorating other dynamic performance. Further more, the reliability optimization design method of the kinetic parameters of the high-speed train exposed to stochastic winds was established. The probability of failure and the wheelset lateral force were taken as optimization objectives and the multi-objective genetic algorithm NSGA-II was used for the automatic optimization. In the simulation, the train speed was 300km/h, and the mean wind speed was 30m/s. After optimization, the probability of failure is reduced from 0.4884 to 0.1406, and the wheelset lateral force is reduced from 45.13kN to 43.01kN.The algorithms for calculating the unsteady aerodynamic loads of the high-speed train in cross stochastic wind scenario and side stochastic wind scenario were set up, when the longitudinal and lateral components of the stochastic wind speed were taken into account. The characteristics of the standard deviations and maximum of the unsteady aerodynamic loads were studied. In addition, the difference of the unsteady aerodynamic loads when the lateral fluctuating component was included and eliminated was analyzed. When the wind angle was 90°, which is the cross stochastic wind scenario, the change for the statistical properties of the unsteady aerodynamic loads is small if the lateral component of the wind speed was further taken into account. However, in the side stochastic wind scenario, when the lateral component of the wind speed was further taken into account, the change for the statistical properties of the unsteady aerodynamic loads was highly related to the wind angle. The closer to the cross wind direction the mean wind is, the smaller the change is. The farther away from the cross wind direction the mean wind is, the larger the change is.