LAUSR

Abstract A hybrid method combining the stochastic pseudo excitation method (SPEM) and the response surface method (RSM) was presented to handle the effect of both random characteristics of power spectral density of excitation and uncertain parameters. Based on basic principle of pseudo excitation method (PEM), an optimized algorithm was proposed to be SPEM. The effects of the uncertainties of structure and excitations could be considered through SPEM, thus directly acquiring the whole sampling results of random vibration of uncertain structural system. To reduce the cost of computing the total number of required finite element analysis simulations, RSM was adopted to hold calculated sampling results to gain the probability density function (PDF) of random vibration. The proposed method can consider both the effects of random characteristics of power spectrum of random excitations and the uncertainties of target system. Neverthless, Since the proposed method is based on the pseudo excitation method, the SPEM-RSM can handle the stochastic vibration analysis of determine and uncertain structure. In order to verify the accuracy and efficiency of proposed method, taking a single degree of freedom linear uncertain oscillator and a pier construction as two verification cases, the PDF of responses computed by the hybrid method was verified by Monte Carlo method (MCM), and the efficiency of the proposed method was presented. Then, stochastic responses of the railway continuous bridge subjected to earthquake were calculated, several uncertainties are involved in the computation. In addition, sensitivity analysis was solved for uncertain parameters.