臺灣博碩士論文加值系統

由於全球暖化、臭氧層破裂等環境議題逐漸浮出檯面，再生能源的產業發展與法規制度儼然已成為各國不得不正視與著手制定的一大政策。其中，太陽能再生能源產業對於位處副熱帶氣候且陽光充沛的台灣來說，具備高度發展前景，因此，準確地預測太陽能案場發電量的技術，將成為發電業者做案場控管與電力分配時的一項重要能力。本研究旨在探討機器學習方法在太陽能發電預測能力與績效之比較，並進一步進行不同時間型態資料集、參數組合、權重設定以及不同案場資料集的結果比對。尤其在不同參數組合中，氣象資料與發電、太陽軌跡變數與發電之間的相互影響，也於本次研究中進行了探討。最後，本研究基於此次實驗結果建立太陽能發電預測系統，期望能應用於實際太陽能案場上。本研究透過類神經網路、最近鄰居法及迴歸分析等機器學習預測方法，對太陽能案場進行發電預測與方法比較。實驗結果顯示，在眾多的影響變數中，太陽輻射值對於案場之發電預測具有最高的影響能力；在方法比較上，類神經網路預測績效普遍優於最近鄰居法的績效結果，但若能取得品值較高之太陽輻射預報資料，可以大幅改善最近鄰居法的預測結果。

The rising importance of solar energy comes along with the attention of solar power forecasting. Accurate forecasting of solar power output is crucial to power distribution and plant management. This study aims to compare performances of different forecasting algorithms coming from machine learning. Comparisons also make on using multiple types of datasets, different sets of input features, different parameter settings, and different solar power plants. In particular, we investigate the influences between the meteorological factors and power outputs, sun path variables and power outputs. At the end, a prototype of solar power forecasting system is constructed and will be deployed on several field plants. In this research, two forecasting algorithms, multiple layer perceptron neural network (MLPNN) and k-nearest neighbors (KNN) are applied to develop the forecasting models. The experimental results indicate that radiation is the most crucial factor for the power output forecasting and MLPNN outperforms KNN algorithm. The performance of KNN can be improved if the predicated values of radiation are stable and fewer outliers.

摘要 i
Abstract ii
Table of Contents iii
List of Figures v
1. Introduction 1
2. Literature Review 3
2.1 Radial Basis Functions 4
2.2 Support Vector Machine 4
3. Methods 6
3.1 Data Description 6
3.1.1 Dataset 6
3.1.2 Parameter Selection 6
3.1.3 Data Normalization 7
3.2 Multiple Layer Perceptron Neural Network 9
3.3 K-Nearest Neighbors 10
3.4 Forecasting Accuracy Evaluation 11
4. Experiment 13
4.1 Result of Daily Dataset 13
4.2 Result of Hourly Dataset 17
4.3 Result of Different Solar Power Plants 19
5. The prototype of solar power forecasting system 22
5.1 Model Training 22
5.2 Solar power forecasting 23
6. Conclusion 25
References 26

M. Oliver, & T. Jackson. (1999). The market for solar photovoltaics. Energy Policy, 27, 371-385.

European Photovoltaic Industry Association. (2014). Global Market Outlook for Photovoltaics 2014-2018. In T. Rowe (Ed.), (pp. 5-8). Brussels, Belgium: European Photovoltaic Industry Association.

BUSINESS SWEDEN TAIPEI. (2015). Opportunities for renewable energy and cleantech industry in Taiwan (pp. 1-15). Taipei, Taiwan: BUSINESS SWEDEN TAIPEI.

Bureau of Energy. (2015). 陽光屋頂政策推動有成 總統宣示支持再生能源發展. from http://web3.moeaboe.gov.tw/ECW/populace/news/News.aspx?kind=1&menu_id=41&news_id=4077

Bureau of Energy. (2013). 推動電力市場自由化. from http://web3.moeaboe.gov.tw/ECW_WEBPAGE/webpage/book4/page2.htm

Patrick Mathiesen, & Jan Kleissl. (2011). Evaluation of numerical weather prediction for intra-day solar forecasting in the continental United States. Solar Energy, 85, 967-977.

M. Zamo, O. Mestre, P. Arbogast, & O. Pannekoucke. (2014). A benchmark of statistical regression methods for short-term forecasting of photovoltaic electricity production, part I: Deterministic forecast of hourly production. Solar Energy, 105, 792-803.

Changsong Chen, Shanxu Duan, Tao Cai, & Bangyin Liu. (2011). Online 24-h solar power forecasting based on weather type classification using artificial neural network. Solar Energy, 85, 2856-2870.

Jie Shi, Wei-Jen Lee, Yongqian Liu, Yongping Yang, & Peng Wang. (2012). Forecasting Power Output of Photovoltaic Systems Based on Weather Classification and Support Vector Machines. INDUSTRY APPLICATIONS, 48, 1064-1069.

Marco Cococcioni, Eleonora D' Andrea, & Beatrice Lazzerini. (2011). 24-hour-ahead forecasting of energy production in solar PV systems. International Conference on Intelligent Systems Design and Applications, 11, 1276-1281.

Bjön Wolff, Elke Lorenz, & Oliver Kramer. (2013). Statistical Learning for Short-Term Photovoltaic Power Predictions. EUROPEAN CONFERENCE ON MACHINE LEARNING AND PRINCIPLES AND PRACTICE OF KNOWLEDGE DISCOVERY IN DATABASES, 1-12.

Huan Long, Zijun Zhang, & Yan Su. (2014). Analysis of daily solar power prediction with data-driven approaches. Applied Energy, 126, 29-37.

Zhaoxuan Li, Mahbobur Rahman, Rolando Vega, & Bing Dong. (2016). A Hierarchical Approach Using Machine Learning Methods in Solar Photovoltaic Energy Production Forecasting. Energies, 9.

A. Mellit , A. M. P. c., V. Lughi,. (2014). Short-term forecasting of power production in a large-scale photovoltaic plant. Solar Energy, 105, 401-413.

Cai Tao, Duan Shanxu, & Chen Changsong. (2010). Forecasting power output for grid-connected photovoltaic power system without using solar radiation measurement Power Electronics for Distributed Generation Systems, 773 - 777 doi: 10.1109/PEDG.2010.5545754

Atsushi Yona, T. S., Ahmed Yousuf Saber, Toshihisa Funabashi, Hideomi Sekine, Chul-Hwan Kim. (2007). Application of Neural Network to One-Day-Ahead 24 hours Generating Power Forecasting for Photovoltaic System. ISAP, 442-447. doi: 10.1109/ISAP.2007.4441657

Navin Sharma, Pranshu Sharma, David Irwin, & Prashant Shenoy. (2011). Predicting solar generation from weather forecasts using machine learning. Smart Grid Communications, 528-533. doi: 10.1109/SmartGridComm.2011.6102379

Hugo T.C. Pedro, & Carlos F.M. Coimbra. (2012). Assessment of forecasting techniques for solar power production with no exogenous inputs. Solar Energy, 86, 2017-2028.

Peder Bacher, Henrik Madsen, & Henrik Aalborg Nielsen. (2009). Online short-term solar power forecasting. Solar Energy, 83, 1772-1783.

徐森雄, 黃光賢, & 廖士吟. (2005). 泥岩坡面日射量之推估. Environment & Bioinformatics, 189, 211-220.