Automated decision-analytic diagnosis of thermal performance in gas turbines

We have developed an expert system for diagno-sis of e–ciency problems for large gas turbines. The system relies on a model-based approach that combines an expert’s probabilistic assess-ments with statistical data and thermodynamic analysis. The system employs a causal prob-abilistic graph, called a belief network, to up-date the likelihoods of alternative faults given information about diverse classes of informa-tion. In response to any subset of flndings or reported observations, the system suggests the most cost-efiective tests to perform to determine the source of a performance problem. We discuss the decision-analytic methodology that underlie