Design of an intelligent diagnostic architecture to support the condition monitoring of power generation assets

Detailed, timely and accurate condition monitoring and diagnostic capability for generation assets is important for safely operating a power plant. A software system developed to perform this function for a fleet of plants must allow the extraction of important information from a large volume of data while operating over a distributed platform. This paper will focus on the development of a multi-agent system (MAS) to allow the condition monitoring and fault diagnosis of several power plants for a major UK utility. This system will exploit the advantages of a MAS approach while incorporating a range of anomaly detection and fault diagnosis techniques to provide accurate and robust condition monitoring capability. An overview of the User Requirements Specification is provided, as well as an outline of the proposed architecture including appropriate data interpretation algorithms and a definition of the interface and support mechanisms for the engineers.