퍼지-神經網을 利用한 原子力配管의 破壞擧動評價 = Evaluation on fracture behavior of nuclear piping by fuzzy-neural network

다국어 초록 (Multilingual Abstract)

Uncertainties exist in criteria for performing elastic-plastic fracture mechanics analysis of nuclear pipings. In this thesis, fuzzy theory is adopted to deal with the uncertainty of a fracture criterion and neural network theory is used to deal with the discordance of the several fracture criteria. The objective of this thesis is to develop fuzzy-neural network expert system for dealing with such uncertainty.
This thesis is consisted of 4 parts; application of fuzzy theory, application of neural network theory, development of fuzzy -neural network expert system and case studies. The fuzziness at the boundaries of the criteria is quantified by the fuzzy for four fracture criteria of nuclear piping. Evaluation on fracture behavior is performed by the learning technique and the pattern matching of the neural network when the results are different each other. Finally the expert system software was developed by Visual C++, and case studies were performed to demonstrate the usefulness.

목차 (Table of Contents)

• 목차
• * Nomenclature
• * List of figures
• * List of Tables
• 1. 서론 = 1
• 1.1 연구 배경 = 1
• 1.2 연구 목적 = 2
• 1.3 연구 내용 = 3
• 2. 관련이론 = 4
• 2.1 퍼지이론 = 4
• 2.1.1 확률론과 퍼지 이론 = 4
• 2.1.2 퍼지추론 = 4
• 2.2 신경망 이론 = 5
• 2.2.1 신경망 이론 = 5
• 2.2.2 학습 = 6
• 2.3 배관파괴거동 평가기준 = 7
• 2.3.1 DPZP 기준 = 7
• 2.3.2 한계응력기준 = 7
• 2.3.3 ASME 기준 = 8
• 3. 퍼지를 이용한 파괴거동기준 = 12
• 3.1 퍼지를 적용한 DPZP 기준 = 12
• 3.1.1 퍼지화(fuzzification) 및 근사화(approximation) = 12
• 3.1.2 퍼지추론(fuzzy reasoning) = 13
• 3.2 퍼지를 적용한 한계응력기준 = 14
• 3.2.1 퍼지화 및 근사화 = 15
• 3.2.2 퍼지추론 = 15
• 3.3 퍼지를 적용한 ASME 기준 = 16
• 3.3.1 퍼지화 및 근사화 = 17
• 3.3.2 퍼지추론 = 17
• 3.4 경험적 지식의 퍼지화 = 18
• 4. 신경망을 이용한 파괴거동기준 = 24
• 4.1 학습데이터의 포맷(format) = 24
• 4.2 파괴거동기준의 정규화(Normalization) = 24
• 4.3 파괴거동기준의 학습 = 25
• 5. 전문가시스템의 구현 및 사례연구 = 29
• 5.1 퍼지전문가시스템 = 29
• 5.2 신경망전문가시스템 = 29
• 5.3 사례연구 1 = 30
• 5.4 사례연구 2 = 36
• 6. 결론 = 50
• * Reference
• * ABSTRACT