Fatigue life prediction of turbine blades with geometric imperfections made of stainless steel

Abstract

This research addresses critical challenges faced by steam turbine blades, particularly in low-pressure (LP) turbines, where premature failures are common due to stress concentrations at the blade root area. The study introduces a numerical methodology aimed at predicting the life of mistuned steam turbine blades, with a focus on variations in blade geometry which have received limited exploration in existing literature. A simplified, scaled-down mistuned steam turbine bladed disc model was developed using Abaqus finite element software. Acquisition of steady-state stress response of the disc models was performed through finite element analysis (FEA). Thereafter, numerical stress distributions were obtained. Subsequently, within Companion software, a Monte Carlo simulation-based probabilistic approach was applied to evaluate and quantify uncertainties in fatigue life for 17 cases. This analysis considered an accepted manufacturing percentage scatter of ± 5 % for the steam turbine bladed disc. That was conducted by selecting mistuning (geometry variation) percentages as the random variables. The methodology demonstrated reliability, correlating well with literature-based and discrete fatigue life results. This study establishes the potential for accurately predicting the fatigue life of mistuned steam turbine blades using the developed methodology.