Risk based engineering for design, material selection and maintenance of power plants

Abstract

Risk based engineering (RBE) provides a measure of decision making through risk assessment of design, manufacturing, operation and maintenance. The application study of RBE to fossil power plants is presented here especially focusing on steam turbines. Typical damage modes and mechanisms are presented, and failure scenarios are expressed in terms of event trees. The comprehensive risk assessment approach is expressed in terms of the conventional risk matrix method for prioritizing the important events and the probabilistic approach for more quantitative evaluation. To make more rational decision making, a ‘Bayesian’ network approach is introduced as an effective tool of casual structure determination in the probabilistic risk assessment framework. The Bayesian network approach for determining the casual structure of risk scenario is shown to be a potential tool for constructing the risk scenario automatically in the data-driven manner even in the case of insufficient available data. Finally a material selection and design map is presented for the creep-fatigue damage condition of turbine components using a two-parameter probability distribution function to make a reliable initial design of high temperature structures.