ABSTRACT
This chapter provides a gentle and accessible introduction to modeling and reasoning with Bayesian networks, with the domain of systems health management in mind. It provides a brief overview of the types of algorithms, developed in the artificial intelligence community, available for reasoning in Bayesian networks. Systems health management is often needed due to uncertainty regarding a system’s operation as well as in the system’s environment. There are several classes of algorithms for reasoning in Bayesian networks, computing marginal probabilities, and finding explanations. Our society relies on electrical power in many ways. Obviously, there is the electrical power grid, which is undergoing a major upgrading effort in the context of creating a smart grid. The chapter shows that the ProDiagnose diagnostic algorithm combined with arithmetic circuits compiled from Bayesian networks provides high-quality diagnostic results for a real-world electrical power system called Advanced Diagnostics and Prognostics Testbed.
