Non-probabilistic reliability analysis on resonance of fluid-filled pipeline based on Kriging model

To improve the evaluation of reliability, a non-probabilistic reliability method was used to analysis the resonance of fluid-filled pipeline systems. The uncertain parameters of structures were described by both ellipsoidal modes and interval parameters. Kriging model and lating hypercube sampling technique were employed to settle the problem that implicit limit state functions were not easily be solved. The Kriging model, as an approximate model, was introduced to represent the original structure and was update gradually in the computation. Considering the fluid-structure interaction (FSI), and by using the FEM software, a fluid-filled pipeline system was modeled and the system's modal was analyzed. Then, after the anti-resonance reliability theory was presented to process the non-probabilistic reliability of the FSI pipeline system, optimization method was used to calculate the parameter reliability. An example was introduced to demonstrate the validity and efficiency of the method. And it is shown that the proposed method has the capability to improve the anti-resonance design and the reliability estimation of fluid-filled pipelines.