Computational Structural Reliability Analysis of a Turbine Blade

Abstract

A system reliability methodology has recently been developed to determine accurately and efficiently the reliability of structures with multiple failure modes. This paper explores the computational implementation and application of the methodology through the structural system reliability calculation of a turbine blade subject to randomness in material properties. Failure due to creep, stress overload, and vibration is considered. The methodology consists of a probabilistic system reliability analysis methodology integrated with finite element methods. The failure paths and failure modes are organized using a fault tree. An efficient method for assessing the reliability of a single failure mode, i.e., component reliability, is implemented as well as an efficient adaptive importance sampling method to assess the system reliability. A probabilistic structural analysis program, NESSUS, is used for the calculations.