GRU 모델을 이용한 바닥 조건별 태양광 패널의 발전량 예측 = Power generation forecasting of solar panels under different floor conditions using GRU model

다국어 초록 (Multilingual Abstract)

Solar power generation is usually affected by different meteorological factors, such as rainfall, cloud cover, solar radiation, and temperature. This variability has shown a negative impact on the large-scale integration of solar energy into energy supply systems. In order to smoothly supply solar energy, the amount of power generated from solar panels must be accurately predicted. Problematic solar panels can cause problems that affect the power generation of the entire solar array. These problems result in loss of power production. In this study, we collected, compared and analyzed solar power generation data for each type of solar module and each floor condition, and identified under which floor conditions the power generation of solar modules was high. Solar power generation forecasts were created through modeling using artificial intelligence technology. In the case of deep learning models, this is data obtained from the target power production value in kWh and other characteristics such as weather conditions, solar radiation, and solar radiation. Power generation values from solar panels were collected at time intervals from January 1 to December 31, 2021. To demonstrate the effectiveness of the proposed model, several deep learning approaches GRU models were implemented. Model performance evaluation showed that (MAE) was 0.145, 0.087, 0.082, and 0.078 in the order of cement, snow field, desert, and grass field, respectively. The root mean square error (RMSE) of the model is 0.171, 0.115, 0.111, and 0.106, respectively. The optimal model for a double-sided solar panel with a low error rate was expressed. The proposed model showed efficient performance and was proven to be effective in predictiong time series data.

목차 (Table of Contents)

• 목차 Ⅰ
• Abstract Ⅷ
• 1. 서론 1
• 2. 이론적 배경 4
• 2.1. 태양전지 4
• 2.1.1. 태양전지의 발전 원리 4
• 2.1.2. 태양전지의 종류 7
• 2.1.3. 실리콘계 태양전지 9
• 2.1.4. 단결정질 실리콘 태양전지 10
• 2.1.5. 다결정질 실리콘 태양전지 12
• 2.2. 태양광 모듈 15
• 2.3. 태양광 모듈 소재 17
• 2.3.1. Glass 17
• 2.3.2. 봉지재 19
• 2.3.3. Back sheet 21
• 2.3.4. 정션박스 23
• 2.3.5. 프레임 25
• 2.4. 양면 태양광 모듈 27
• 2.5. 태양광 관리의 문제 29
• 2.6. 머신러닝 31
• 2.7. 순차 데이터 33
• 2.8. 순환 신경망(RNN) 34
• 2.9. 순환 신경망(LSTM) 36
• 2.10. 순환 신경망(GRU) 38
• 3. 태양광 패널의 발전량 비교 40
• 3.1. 태양광 모듈의 발전량 데이터 수집 40
• 3.2. 태양광 모듈의 발전량 데이터 분석 48
• 3.3. 태양광 모듈의 발전량 데이터 비교 51
• 4. 태양광 모듈의 발전량 예측 및 정확도 분석 56
• 4.1. 태양광 모듈 발전량 예측 모델 학습 56
• 4.1.1. 기상 데이터 활용 56
• 4.1.2. 데이터 모델 학습 61
• 4.1.3. 데이터 모델 학습 결과 64
• 4.1.4. 데이터 시각화 결과 66
• 4.2. 태양광 모듈 발전량 예측 모델 정확도 분석 70
• 5. 결론 73
• 6. 참고문헌 74
• 요약 84 Ⅸ