Using Multiple-Model Adaptive Estimation and System Identification for Fault Detection in Nuclear Power Plants

Abstract

One challenge in nuclear power plant operation is the detection and identification of system faults and plant transients. Timely and accurate identification will reduce operational costs and increase plant safety. This paper describes a combined model-based and data-driven approach to identifying faults in nuclear power plants. Faults are detected for a GSES Generic Pressurized Water Reactor simulator using the multiple-model adaptive estimation (MMAE) technique. In this technique, multiple input-output system models are used that represent different operating conditions. The models predict sensor measurements for both normal and faulted operating conditions simultaneously. The predicted measurements are then compared to the sensor measurements to determine the most likely operating condition. The system models are obtained using system identification techniques for a specific set of faulted conditions. This technique uses sensor measurements from the simulation to identify appropriate parameters for the system models. The MMAE technique is then used to detect similar faults using the identified model. This combination of model-based and data-driven techniques can ultimately be used to create robust fault models that take advantage of both the models created during the design and validation process and real plant data.