ANALYSIS AND TRAINING OF AN ARTIFICIAL NEURAL NETWORK NETWORK NUCLEAR POWER PLANTACCIDENT DIAGNOSTIC ADVISOR

THE OBJECTIVE OF THE PROPOSED RESEARCH IS TO DEVELOP AN ARTIFICIAL-INTELLIGENCE FAULT-DIAGNOSTIC SYSTEM THAT WILL CLASSIFY, IN REAL-TIME, NUCLEAR POWER PLANT TRANSIENTS AND PROVIDE ERROR BOUNDS ON ITS DIAGNOSES. THE GOAL IS TO DEVELOP AN ARTIFICIAL NEURAL NETWORK (ANN) PARADIGM THAT IS CAPABLE OF NOT ONLY PROVIDING A PROPER CLASSIFICATION BUT ALSO GIVING A FIGURE OF MERIT FOR ITS ANSWER. THE PROPOSED DIAGNOSTIC ADVISOR WOULD BE INTEGRATED INTO THE PROCESS COMPUTER OR SAFETY PARAMETER DISPLAY SYSTEM. THE ADVISOR'S OBJECTIVE WOULD BE TO WARN AND INFORM PLANT OPERATORS AND ENGINEERS OF PLANT COMPONENT FAILURES IN A TIMELY MANNER. THE OBJECTIVE OF THE PROPOSED RESEARCH IS TO DEVELOP AN ARTIFICIAL-INTELLIGENCE FAULT-DIAGNOSTIC SYSTEM THAT WILL CLASSIFY, IN REAL-TIME, NUCLEAR POWER PLANT TRANSIENTS AND PROVIDE ERROR BOUNDS ON ITS DIAGNOSES. THE GOAL IS TO DEVELOP AN ARTIFICIAL NEURAL NETWORK (ANN) PARADIGM THAT IS CAPABLE OF NOT ONLY PROVIDING A PROPER CLASSIFICATION BUT ALSO GIVING A FIGURE OF MERIT FOR ITS ANSWER. THE PROPOSED DIAGNOSTIC ADVISOR WOULD BE INTEGRATED INTO THE PROCESS COMPUTER OR SAFETY PARAMETER DISPLAY SYSTEM. THE ADVISOR'S OBJECTIVE WOULD BE TO WARN AND INFORM PLANT OPERATORS AND ENGINEERS OF PLANT COMPONENT FAILURES IN A TIMELY MANNER. THE OBJECTIVE OF THE PROPOSED RESEARCH IS TO DEVELOP AN ARTIFICIAL-INTELLIGENCE FAULT-DIAGNOSTIC SYSTEM THAT WILL CLASSIFY, IN REAL-TIME, NUCLEAR POWER PLANT TRANSIENTS AND PROVIDE ERROR BOUNDS ON ITS DIAGNOSES. THE GOAL IS TO DEVELOP AN ARTIFICIAL NEURAL NETWORK (ANN) PARADIGM THAT IS CAPABLE OF NOT ONLY PROVIDING A PROPER CLASSIFICATION BUT ALSO GIVING A FIGURE OF MERIT FOR ITS ANSWER. THE PROPOSED DIAGNOSTIC ADVISOR WOULD BE INTEGRATED INTO THE PROCESS COMPUTER OR SAFETY PARAMETER DISPLAY SYSTEM. THE ADVISOR'S OBJECTIVE WOULD BE TO WARN AND INFORM PLANT OPERATORS AND ENGINEERS OF PLANT COMPONENT FAILURES IN A TIMELY MANNER.