A multifunctional operator support system for transient identification and operation guidance in nuclear power plants

In this study, a neural network based operator support system was proposed for alarm analysis, computerized procedures presentation and operation validation in nuclear power plants (NPPs). There are many complicated situations, in which operator needs to promptly and accurately access the alarms and execute regular and suitable operations. In order to help operator quickly recognize the situations and regulate and validate their operations, it is necessary to develop an operator support system which includes a transient diagnostic system and computer based procedures with the functions for operation validation. The proposed multifunctional system integrates two subsystems: fault diagnostic system (FDS) and operator guidance system (OGS). The fault diagnostic system is different from the existed operator support systems by using a hierarchical structure with artificial neural network (ANN) aggregation models instead of conventional general purpose ANNs. It not only analyzes the transient and provides information to the operator but also indexes the trained ANN from the database for the OGS. OGS includes two subsystems: computerized procedures system and operation validation system. The computerized procedures system (CPS) was developed as the base for the operation validation system (OVS). As the key function of the integrated system, OVS provides both qualitative and quantitative evaluation for each operator's action and validates their actions