為了有效利用太陽能發電與確保電力品質，本研究利用Python程式語言，採用日照度、溫度等氣象數據，搭配分散式太陽能發電系統發電成果，配合機器學習中由遞歸神經網路（RNN）衍伸而來的長短期記憶（LSTM）演算法，進行太陽能發電預測模型的訓練與測試。
LSTM透過三個控制記憶的機制，分別為Input、Output與Forget Gate，使機器學習透過太陽能發電紀錄與氣候資料自主學習三個閥門的開關，由此決定長短期記憶的權重並判斷哪些資料是雜訊，特別適合具時序關係之物理量估測。相關演算法特性與邏輯、數據的取得與預處理方式與太陽能發電預測模型的訓練與測試結果將會加以說明，並比較透過不同的參數設定或方法分析其所造成的影響，例如：LSTM方法、RNN方法、輸入與輸出的時間步長、一次性多步預測與漸進式預測…等，最後提供1到10分鐘的短期預估成果並加以討論，最後探討訓練決策方法與實際應用的可行性。

To ensure the effective supply of solar energy and its quality, researchers are looking for better methods to improve the prediction accuracy of a solar energy harvesting system. In this study, weather information such as solar radiation and temperature, together with the experimental results of a distributed solar power harvesting systems were used to train and test by an artificial intelligent algorithm called the long short-term memory (LSTM) method. The LSTM model can assign different weighting coefficients to long-term and short-term memory data, which is particularly suitable for time-series data forecasting. The proposed AI model is able to provide the coming 1 to 10 minutes short-term forecast of the photovoltaic power system. The detail of the method and prediction results as well as potential application of the machine learning algorithm will be discussed.

[1]經濟部能源局。能源統計資料查詢系統。取自 :https://www.moeaboe.gov.tw/wesnq/Views/B01/wFrmB0104.aspx
[2]3S Market「全球智慧科技應用」市場資訊網（2019/5/24）。太陽能 + 住宅儲能微電網助力小鎮輸電。取自:http://3smarket-info.blogspot.com/2019/05/blog-post_71.html
[3]Josep M. Guerrero (2017). MICROGRID TECHNOLOGY RESEARCH AND DEMONSTRATION. From https://www.et.aau.dk/research-programmes/microgrids/activities/microgrid-technology-research-and-demonstration/
[4]台灣電力股份有限公司。歷年太陽光電發電量。取自 https://www.taipower.com.tw/tc/chart_m/b19_%e7%99%bc%e9%9b%bb%e8%b3%87%e8%a8%8a_%e5%86%8d%e7%94%9f%e8%83%bd%e6%ba%90%e7%99%bc%e9%9b%bb%e6%a6%82%e6%b3%81_%e6%ad%b7%e5%b9%b4%e5%a4%aa%e9%99%bd%e5%85%89%e9%9b%bb%e7%99%bc%e9%9b%bb%e9%87%8f.html
[5]林薇如，太陽能電力系統電腦模擬器之開發，國立成功大學系統及船舶機電工程學系碩士論文，2015年。
[6]張智皓，適用於太陽能預測之多步預測即時回饋類神經網路，國立臺北科技大學電機工程系碩士班，2014年。
[7]蔡松諺，以資料探勘方法預測台灣地區太陽能發電量之研究，國立中興大學資訊管理學系碩士論文，2017年。
[8]Sepp Hochreiter, Jürgen Schmidhuber(1997). “Long Short-Term Memory”, Neural Computation 9(8):1735-1780, 1997
[9]Juan Ospina, Alvi Newaz, M. Omar Faruque(2019). “Forecasting of PV plant output using hybrid wavelet-based LSTM-DNN structure model”, IET Renew. Power Gener., 2019, Vol. 13 Iss. 7, pp. 1087-1095
[10]Xiangyun Qing, Yugang Niu (2018). “Hourly day-ahead solar irradiance prediction using weather forecasts by LSTM”, Energy 148 (2018) 461-468
[11]陳煜文，以委託單資料預測當日股價趨勢－LSTM類神經網路模型之應用，輔仁大學金融與國際企業學系金融碩士班，2018年。
[12]M. A. Green, K. Emery, Y. Hishikawa, W. Warta, and E. D. Dunlop, “Solar cell efficiency tables (version 45),” Progress in Photovoltaics: Research and Applications, vol. 23, pp. 1-9, 2015.
[13]K. L. Kennerud, “Analysis of performance degradation in CdS solar cells,” IEEE Transactions on Aerospace and Electronic Systems, vol. AES-5, pp. 912-917, 1969
[14]R. W. Erickson and D. Maksimovic, Fundamentals of Power Electronics, Kluwer Academic Publishers, Norwell, MA, 2001.
[15]R. M. Chao, S. H. Ko, F. S. Pai, I. H. Lin, and C. C. Chang, “Evaluation of a photovoltaic energy mechatronics system with a built-in quadratic maximum power point tracking algorithm,” Solar Energy, vol. 83, pp. 2177-2185, 2009.
[16]F. S. Pai, and R. M. Chao, “A new algorithm to photovoltaic power point tracking problems with quadratic maximization,” IEEE Transactions on Energy Conversion, vol. 25, pp. 262-264, 2010.
[17]S. H. Ko, and R. M. Chao, “Photovoltaic dynamic MPPT on a moving vehicle,” Solar Energy, vol. 86, pp. 1750-1760, 2012.
[18]G. Adinolfi, N. Femia, G. Petrone, G. Spagnuolo, and M. Vitelli, “Energy efficiency effective design of DC/DC converters for DMPPT PV applications,” in Proc. of IEEE 35th Annual Industrial Electronics Conference, pp. 4566-4570, 2009.
[19]H. Patel, and V. Agarwal, “MATLAB-based modeling to study the effects of partial shading on PV array characteristics,” IEEE Transactions on Energy Conversion, vol. 23, pp. 302-310, 2008.
[20]H. Patel, and V. Agarwal, “Maximum power point tracking scheme for PV systems operating under partially shaded conditions,” IEEE Transactions on Industrial Electronics, vol. 55, pp. 1689-1698, 2008.
[21]C. Deline, “Partially shaded operation of multi-string photovoltaic systems,” in Proc. of 35th IEEE Photovoltaic Specialists Conference (PVSC), pp. 394-399, 2010.
[22]IEEE SPECTRUN(2019). The Top Programming Languages 2019.