Advanced monitoring and advice integrating a comprehensive sensor network for improved operational availability

Abstract

This research broadens the prime concern of nuclear power plant operations from safe performance to both economic and safe performance through monitoring and advice. First, rotating machines such as turbine generators and reactor coolant pumps are identified as main contributors for the lost availability through the review of pressurized water reactor (PWR) forced outage records. The integrated architecture utilizes comprehensive sensor networks incorporating modern signal processing systems, advisory systems for sensor validation, and advisory systems for the intelligent diagnosis and maintenance (D&M). For the development of comprehensive sensor networks for complex target systems, an integrated method incorporating a structural system hierarchy and a functional system hierarchy, a fault-symptom matrix, sensor selection criteria, a sensor installation feasibility study, and advanced instrumentation techniques is formulated. Such advanced instrumentation techniques reflect the state of the art in advancement of data acquisition, data processing, and data integration techniques. Once the sensor types and locations are selected definitively, they are incorporated into drawings using a computer aided design tool (e.g. AutoCAD) program in order to make sure that it would be possible to install the comprehensive set of recommended sensors on each specific component studied. The second major part of this study is the development of an intelligent D&M advisory system integrating a comprehensive sensor network. This advisory system employs a Bayesian Belief Network (BBN) as a high level reasoning tool for incorporating inherent uncertainty for use in probabilistic inference. It is demonstrated that a rule-based knowledge representation is simply a special case of a general BBN by showing how the general BBN can be reduced to a rule-based representation. The presented major steps for constructing the BBN based generic inference algorithms are applied to systematic elicitation and synthesis of various levels of experts' knowledge. Prototype D&M algorithms are represented explicitly through topological symbols and links between them in a causal direction. This D&M advisory system is set up with an easy-to-learn, user-friendly, man-machine interface and modern graphics for efficient operator interactions. As new pieces of evidence from sensor networks developed are entered into this system, it provides operational advice concerning both availability and safety so that the operator is able to determine the likely failure modes, diagnose the system state, locate root causes, and take the most advantageous action. Thereby, the comprehensive monitoring supported advice improves operational availability.