PREDICTIVE DIAGNOSIS METHOD AND SYSTEM OF NUCLEAR POWER PLANT EQUIPMENT

The present invention provides a predictive diagnosis method of nuclear power plant facilities which performs machine learning based on information acquired from nuclear power plant facilities to automatically predict and diagnose nuclear power plant facilities, and a system therefor. The method com...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: SHIN YOU SOO, LEE WON KYU, JEON I SEUL, KIM DAE WOONG, KIM HEE CHAN, KIM MIN HO
Format: Patent
Sprache:eng ; kor
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:The present invention provides a predictive diagnosis method of nuclear power plant facilities which performs machine learning based on information acquired from nuclear power plant facilities to automatically predict and diagnose nuclear power plant facilities, and a system therefor. The method comprises the steps of: monitoring a state of nuclear power plant facilities based on data provided from the nuclear power plant facilities; diagnosing a first fault value for the nuclear power plant facilities by performing rule-based machine learning based on the data; performing statistics-based machine learning based on the data to diagnose a second fault value for the nuclear power plant facilities; diagnosing the second fault value for the nuclear power plant facilities by performing the statistics-based machine learning based on the data; and performing predictive diagnosis for the nuclear power plant facilities based on the first and second fault values. Accordingly, the rule-based machine learning and the statistics-based machine learning are performed together to increase accuracy, and information on a combined value is visualized to allow an inspector to intuitively determine presence or absence of abnormalities. 본 발명은 원전설비로부터 취득된 정보를 기반으로 머신러닝을 수행하여 원전설비에 대한 자동 예측 진단을 수행하는 원전설비 예측 진단 방법 및 시스템을 제공하기 위하여, 원전설비로부터 제공되는 데이터를 기반으로 상기 원전설비의 상태를 감시하는 단계 및 상기 데이터를 기반으로 규칙 기반(Rule Based) 머신러닝을 수행하여 상기 원전설비에 대한 제1 결함 값을 진단하는 단계 및 상기 데이터를 기반으로 통계학 기반 머신러닝을 수행하여 상기 원전설비에 대한 제2 결함 값을 진단하는 단계 및 상기 제1 및 제2 결함 값을 기반으로 상기 원전설비에 대한 예측 진단을 수행하는 단계를 포함한다. 이에, 본 발명은 Rule Based 기반 머신러닝 및 통계학 기반 머신러닝을 함께 수행하여 결함 값에 대한 정확성을 높이고, 결합 값에 대한 정보를 시각화하여 검사자가 직관적인 이상 유무 판단을 수행할 수 있게 하는 효과가 있다.