Resonance reliability and importance measure analysis of multi-span pipe conveying fluid embedded in temperature-varying matrix

Abstract

Pipe system conveying fluid faces the problem of multi-order resonance failure caused by broadband excitation. For solving above problem, the dynamic stiffness method is employed to solve the dynamic equations of multi-span pipes considering the temperature effect. Combining the obtained natural frequency and the rule of resonance failure of pipe system, a multi-order anti-resonance system reliability model is established in this paper. To analyze the effect of input variable uncertainty on the probability of system resonance failure, the variance-based importance measurement index is further established. By introducing the active learning Kriging (ALK) model, the resonance failure probability and importance measurement index can be calculated efficiently. The effects of fluid flow velocity, pressure and temperature on the probability of pipe resonance failure are analyzed in detail, which has significant guidance for the anti-resonance optimization design of pipes.