Sensitivity and dynamic analysis of train-bridge coupled system with multiple random factors

The train-bridge system is complex and presents many random factors, including stochastic track irregularities as well as random parameters of bridge and train. Moreover, the entire train-bridge system may contain more than one hundred random variables. Herein, we proposed a new random analysis method that comprehensively considers two random sources of the system. Irregularities are represented using the Karhunen-Loeve expansion and the response of the system is then calculated using the point estimate method based on a two-dimensional reduction technique. The solution of the Monte Carlo simulation is regarded as the exact solution to verify the proposed method. Comparing to the Monte Carlo simulation method, the proposed method demonstrates high precision and efficiency. Moreover, sensitivities of random irregularities and parameters of the bridge and train to the system responses are analyzed. The bridge response demonstrates the highest sensitivity to Young's modulus and mass of the car-body, whereas the responses of the train exhibit the highest sensitivity to stochastic irregularities. Further, the influence of different random source combinations on the system response at different speeds is analyzed. The results suggest that it is sufficient to consider only random track irregularities for the train response, whereas random sources from both the bridge and train must be considered for the bridge response.