Fatigue Analysis of Steam Turbine Blades During Startup Using Probabilistic Concepts

Abstract

The deterministic approach in engineering design and analysis compensates for the uncertainty in specific system parameters through an empirical factor of safety. The probabilistic approach incorporates this randomness or uncertainty and ensures optimal use of resources without compromising safety. This paper presents a probabilistic model to determine the fatigue life of a low-pressure steam turbine blade under the influence of transient loading conditions characteristic of a start-up. The methodology consists of simulating the randomness in the values of structural damping, rotational speed, and the blade material's elastic modulus through Finite Element Analysis (FEA). Monte Carlo simulation is utilized to generate random stress and fatigue cycle values obtained from the results of Finite Element Analysis. The Palmgren-Miner rule is then used to determine the probabilistic fatigue life. Results show that the deterministic approach underestimates fatigue life. Also, it fails to demonstrate the relative impact of these random variables on fatigue life. The probabilistic model developed for this study clearly establishes that randomness in the values of turbine rotational velocity has the most significant effect on the probabilistic fatigue life of a low-pressure steam turbine blade.