Field Damage Data Analysis and Probabilistic Fatigue Life Assessment of Turbine Components

Abstract

A methodology is presented for performing probabilistic life assessment for individual parts based on statistical damage analysis and stochastic damage simulation analysis of actual turbine components. Examples are shown for thermomechanical fatigue (TMF) damage of gas turbine nozzles and steam pipes. TMF damage is multiple cracking and damage parameters are maximum crack length and crack length density. Applying the experimental damage parameters vs. cycle ratio relationships to the field crack data of individual parts, specific crack growth trend curves are derived as well as parts service conditions such as plastic strainrange. Using parts service conditions and stochastic damage evolution law of materials, Monte-Carlo damage simulation analysis is performed. The simulation analysis is the useful tool for the probabilistic life prediction for individual parts and for risk-based maintenance and repair/replace judgement.