사례기반 추론을 통한 냉동 사이클 성능 예측 정확도 향상에 관한 연구 = The Study of Improving the Accuracy of Refrigerant Cycle Simulation using Case-Based Reasoning

다국어 초록 (Multilingual Abstract)

Recently due to the oil and commodity prices, the world countries need to improve the energy efficiency of the system.
Air conditioners, refrigerators, other small household and commercial refrigeration systems are no exception. Each company is trying to get the most effective with less effort and is trying to resolve this problem through using refrigeration cycle simulation. Refrigeration cycle simulation to determine the characteristics of the cycle to reduce effort and cost is a good way. However, the theoretical refrigeration cycle simulation is hard to predict the actual ability, and the difference between the actual test results is occured.

In this study, it shows to reduce difference between the actual experimental results and theoretical simulation results.
Firstly, to extract the main factors affecting these major difference between the actual experimental results and theoretical simulation results, it is used the coefficient of determination and the decision tree method.
Secondly, Case-based reasoning techniques, SVM (Support Vector Machine), MARS (Multivariate Adaptive Regression Splines), Boosting Tree, Multiple Regression, are applied to the main factors and the results of applying each of the cases are compared and analyzed.
Finally, through actual system by implementing the SVM superior to generalization ability, this study shows difference between simulation results and the actual test results and suggest the optimized forecasting model

목차 (Table of Contents)

• 제 1 장 서 론 1
• 1.1 연구배경 1
• 제 2 장 이론적 배경 3
• 2.1 냉동 사이클 시뮬레이션 구조 3
• 2.2 사례 기반 추론 시스템 8
• 제 3 장 주요 인자 선정 10
• 3.1 추진 방안 10
• 3.2 현상 분석 12
• 3.3 변수의 선정 13
• 3.3.1 연속형 데이터 14
• 3.3.2 이산형 데이터 16
• 3.3.3 자료 분석 18
• 3.3.4 주요 인자 분석 19
• 3.3.4.1 결정 계수 분석 21
• 3.3.4.2 결정 계수의 적용 22
• 3.3.4.3 의사결정나무 28
• 3.3.4.4 의사결정나무의 적용 28
• 3.3.5 분석 결과 34
• 제 4 장 사례 기반 예측 모델 36
• 4.1 SVM(Support Vector Machine) 36
• 4.2 MARS(Multivariate Adaptive Regression Splines) 38
• 4.3 Boosting Tree 39
• 4.4 Multiple Regression 40
• 제 5 장 모형별 실험 결과 비교 및 고찰 41
• 5.1 Set 성능 오차 41
• 5.2 Set 소비 전력 오차 46
• 5.3 토출 건구 온도 오차 51
• 5.4 냉매 봉입량 오차 56
• 제 6 장 결론 61
• 참고문헌 62
• Abstract 64
• 감사의 글