以深度學習數據驅動方法對短中期太陽能發電預測之研究

簡易檢索 / 詳目顯示

回結果列表

研究生：	洪雅玟 Ya-Wen Hung
論文名稱：	以深度學習數據驅動方法對短中期太陽能發電預測之研究 Deep Learning Based Data Driven Approach for Short and Mid Term Solar Power Generation Prediction
指導教授：	周碩彥 Shuo-Yan Chou 郭伯勳 Po-Hsun Kuo
口試委員:	游慧光 Hui-Kuang Yu
學位類別：	碩士 Master
系所名稱：	管理學院 - 工業管理系 Department of Industrial Management
論文出版年：	2023
畢業學年度：	111
語文別：	英文
論文頁數：	78
中文關鍵詞：	時間序列預測、太陽輻射量預測、太陽光伏預測、序列分解、encoder- decoder 、預測資料集
外文關鍵詞：	Solar radiation prediction, Solar photovoltaic forecasting, predictive datasets
相關次數：	點閱：210 下載：3
分享至:	分享至facebook 分享至twitter

查詢本校圖書館目錄查詢臺灣博碩士論文知識加值系統勘誤回報

由於氣候變化的日益嚴重和全球溫室氣體排放的影響，減少碳排放已成為全球關注的焦點。各國紛紛制定具體的政策和目標，旨在實現近零排碳。在這一努力中，發展可再生能源扮演著至關重要的角色，而太陽光電發電被認為是最為穩定和迅速發展的可再生能源形式。
太陽光電作為可再生能源的代表，不僅具備清潔、可再生和無污染的特點，而且其穩定的發電量有助於提升區域電網的穩定性。為了更好地了解未來的發電量並促進電廠的有效運作，天氣和輻射量對太陽光伏發電具重要影響。本研究利用分析天氣的多種特徵，找出對發電量產生影響的特徵，實現對輻射量和發電量的預測。在數據預處理階段，我們使用了決策樹(Decision Tree)、隨機森林(Random Forest)、梯度提升決策樹(Gradient Boosting Decision Tree)和eXtreme Gradient Boosting (XGBoost)方法，對外部數據集與台灣中央氣象局的數據進行比較，以利後續使用。特徵工程方面，將時間、風向和風速進行轉換，並利用遞歸特徵消除（Recursive Feature Elimination, RFE）進行特徵選擇。在模型建構階段，比較多種序列分解方法，選擇合適的方法進行序列分解。同時採用編碼器-解碼器模型，加入預測數據集，以改善模型的表現。本研究的特點在於加入預測數據集，以預測未來1、3、6、12和24小時的發電量。
最終，本研究提出了一種基於encoder-decoder預測太陽能光伏模型。編碼器階段利用歷史天氣特徵，並通過機器學習進行特徵選擇以預測輻射量，加入預測數據集並進入解碼器階段以預測發電量。本研究收集台灣中部和南部各一個電廠的數據，以比較模型的表現。實驗結果顯示，本研究提出的多變量預測模型在加入天氣特徵和預測數據集後較不包含預測天氣資料集的模型或是僅加入預測天氣資料集但沒有先預測輻射量的模型表現佳。

Due to climate change and global greenhouse gas emissions, reducing carbon emissions has become a global focus. Countries are implementing specific policies and goals to achieve near-zero carbon emissions. Developing renewable energy, particularly solar photovoltaic power, is crucial.
Solar photovoltaic power is clean, renewable, and stable, making it an ideal form of renewable energy. Its reliable power output helps enhance the stability of regional power grids. Weather conditions and solar radiation play significant roles in solar power generation. This study analyzes various weather features to identify factors affecting power generation and predict solar radiation and power output.
Data preprocessing involves comparing external datasets with data from the Taiwan Central Weather Bureau using decision trees, random forests, gradient-boosting trees, and XGBoost methods. Feature engineering includes transforming time, wind direction, and wind speed, and using recursive feature elimination for feature selection. The model construction phase compares multiple sequence decomposition methods and employs an encoder-decoder model with predictive datasets for accurate power output forecasting in different time intervals.
In conclusion, this study introduces an encoder-decoder-based solar photovoltaic prediction model. The decoder stage uses the predictive datasets to forecast power output. Experimental results using data from solar power plants in Taiwan's central and southern regions demonstrate improved performance when integrating weather features and predictive datasets.
Reducing carbon emissions and achieving near-zero carbon goals are global priorities. Developing renewable energy, particularly solar photovoltaic power, is crucial in this effort. By analyzing weather features and incorporating predictive datasets, accurate and reliable power output forecasts can be achieved, contributing to the effective operation of solar power plants.

摘要    i
Abstract    ii
Acknowledgment    iii
Table of Contents    iv
List of Figures    vi
List of Tables    vii
   Chapter 1  Introduction    1
1    Background    1
2    Research Purpose    5
3    Research Limitation    5
4    Organization of the Thesis    5
   Chapter 2  Literature Review    7
1    Renewable Energy Issues    7
2    Solar Energy Issues    7
3    Solar Photovoltaic Prediction    8
   Chapter 3  Methodology    12
1    Pre-analysis    12
1.1    Data Pre-processing    13
1.2    Similarity of Datasets    14
1.3    Data Visualization    15
1.4    Time Series Analysis    15
1.5    Causality analysis    16
2    Machine Learning of Missing Values Imputation    17
2.1    Cross Validation    17
2.2    Decision Tree    17
2.3    Random Forest Regression    18
2.4    eXtreme Gradient Boosting    18
3    Feature Engineering Methods    18
3.1    Transformation of Wind    19
3.2    Cyclical Features of Time    19
4    Feature Selection    19
5    Prediction Method    20
5.1    Multi-step Forecasting Methods    20
5.2    Complete Ensemble Empirical Mode Decomposition with Adaptive Noise    21
5.3    Empirical Wavelet Transform    22
5.4    LSTM    24
5.5    Encoder-decoder    25
5.6    Forecasting Scenario    26
6    Result Evaluation    27
   Chapter 4  Result and discussion    29
1    Data Description    29
1.1    Feature Correlation    35
1.2    Time Series Analysis    37
1.3    Causality analysis    43
2    Feature Engineering    47
2.1    Missing Value Imputation of Irradiance    47
2.2    Missing Value Imputation of Photovoltaic    49
2.3    Transformation of Wind Vectors    50
2.4    Time Cyclical Features    51
2.5    Recursive Feature Elimination    52
3    Forecasting Result    53
3.1    Results of univariate analysis    53
3.2    Results of multivariate analysis    55
   Chapter 5  Conclusion And Future Research    60
1    Research Conclusion    61
2    Future Work    62
References    63

                                

[1] K. Sarıca, G. U. Harputlugil, G. İnaner, and E. T. Kollugil, "Building sector emission reduction assessment from a developing European economy: A bottom-up modelling approach," Energy Policy, vol. 174, p. 113429, 2023/03/01/ 2023, doi: https://doi.org/10.1016/j.enpol.2023.113429.
[2] G. Iyer et al., "The path to 1.5 °C requires ratcheting of climate pledges," Nature Climate Change, vol. 12, no. 12, pp. 1092-1093, 2022/12/01 2022, doi: 10.1038/s41558-022-01517-z.
[3] D. Gibb et al., "Renewables 2022 global status report+ Renewable energy data in perspective+ Press releases+ Regional fact sheets+ Country fact sheets," 2022.
[4] D. Gibb et al., "Renewables 2022 global status report + Renewable energy data in perspective + Press releases + Regional fact sheets + Country fact sheets," France, 978-3-948393-04-5, 2022. [Online]. Available: http://inis.iaea.org/search/search.aspx?orig_q=RN:53066645
[5] P. Lusis, K. R. Khalilpour, L. Andrew, and A. Liebman, "Short-term residential load forecasting: Impact of calendar effects and forecast granularity," Applied Energy, vol. 205, pp. 654-669, 2017/11/01/ 2017, doi: https://doi.org/10.1016/j.apenergy.2017.07.114.
[6] A. Q. Al-Shetwi, M. A. Hannan, K. P. Jern, M. Mansur, and T. M. I. Mahlia, "Grid-connected renewable energy sources: Review of the recent integration requirements and control methods," Journal of Cleaner Production, vol. 253, p. 119831, 2020/04/20/ 2020, doi: https://doi.org/10.1016/j.jclepro.2019.119831.
[7] Q. Hou, N. Zhang, E. Du, M. Miao, F. Peng, and C. Kang, "Probabilistic duck curve in high PV penetration power system: Concept, modeling, and empirical analysis in China," Applied Energy, vol. 242, pp. 205-215, 2019/05/15/ 2019, doi: https://doi.org/10.1016/j.apenergy.2019.03.067.
[8] T. P. Company, 追風逐日擁抱綠電, 2020, p. 7. [Online]. Available: https://tpcjournal.taipower.com.tw/TPMM/2020/691/#p=8.
[9] N. Mostafa, H. S. M. Ramadan, and O. Elfarouk, "Renewable energy management in smart grids by using big data analytics and machine learning," Machine Learning with Applications, vol. 9, p. 100363, 2022/09/15/ 2022, doi: https://doi.org/10.1016/j.mlwa.2022.100363.
[10] M. O. Moreira, P. P. Balestrassi, A. P. Paiva, P. F. Ribeiro, and B. D. Bonatto, "Design of experiments using artificial neural network ensemble for photovoltaic generation forecasting," Renewable and Sustainable Energy Reviews, vol. 135, p. 110450, 2021/01/01/ 2021, doi: https://doi.org/10.1016/j.rser.2020.110450.
[11] F. Wang, Z. Xuan, Z. Zhen, K. Li, T. Wang, and M. Shi, "A day-ahead PV power forecasting method based on LSTM-RNN model and time correlation modification under partial daily pattern prediction framework," Energy Conversion and Management, vol. 212, p. 112766, 2020/05/15/ 2020, doi: https://doi.org/10.1016/j.enconman.2020.112766.
[12] Y. Jung, J. Jung, B. Kim, and S. Han, "Long short-term memory recurrent neural network for modeling temporal patterns in long-term power forecasting for solar PV facilities: Case study of South Korea," Journal of Cleaner Production, vol. 250, p. 119476, 2020/03/20/ 2020, doi: https://doi.org/10.1016/j.jclepro.2019.119476.
[13] Z. H. Osman, M. L. Awad, and T. K. Mahmoud, "Neural network based approach for short-term load forecasting," in 2009 IEEE/PES Power Systems Conference and Exposition, 15-18 March 2009 2009, pp. 1-8, doi: 10.1109/PSCE.2009.4840035.
[14] A. S. Brouwer, M. van den Broek, A. Seebregts, and A. Faaij, "Impacts of large-scale Intermittent Renewable Energy Sources on electricity systems, and how these can be modeled," Renewable and Sustainable Energy Reviews, vol. 33, pp. 443-466, 2014/05/01/ 2014, doi: https://doi.org/10.1016/j.rser.2014.01.076.
[15] N. Zhang, Z. Hu, B. Shen, S. Dang, J. Zhang, and Y. Zhou, "A source–grid–load coordinated power planning model considering the integration of wind power generation," Applied Energy, vol. 168, pp. 13-24, 2016/04/15/ 2016, doi: https://doi.org/10.1016/j.apenergy.2016.01.086.
[16] I. Staffell and S. Pfenninger, "The increasing impact of weather on electricity supply and demand," Energy, vol. 145, pp. 65-78, 2018/02/15/ 2018, doi: https://doi.org/10.1016/j.energy.2017.12.051.
[17] Q. Wang and S. Wang, "Is energy transition promoting the decoupling economic growth from emission growth? Evidence from the 186 countries," Journal of Cleaner Production, vol. 260, p. 120768, 2020/07/01/ 2020, doi: https://doi.org/10.1016/j.jclepro.2020.120768.
[18] T. Kober, H. W. Schiffer, M. Densing, and E. Panos, "Global energy perspectives to 2060 – WEC's World Energy Scenarios 2019," Energy Strategy Reviews, vol. 31, p. 100523, 2020/09/01/ 2020, doi: https://doi.org/10.1016/j.esr.2020.100523.
[19] A. M.-Z. Gao, C.-T. Fan, and C.-N. Liao, "Application of German energy transition in Taiwan: A critical review of unique electricity liberalisation as a core strategy to achieve renewable energy growth," Energy Policy, vol. 120, pp. 644-654, 2018/09/01/ 2018, doi: https://doi.org/10.1016/j.enpol.2018.01.010.
[20] I.-S. Tsay and P.-H. Chen, "A dual market structure design for the reform of an independent power grid system—The case of Taiwan," Energy Reports, vol. 5, pp. 1603-1615, 2019/11/01/ 2019, doi: https://doi.org/10.1016/j.egyr.2019.11.001.
[21] C.-T. Chang and H.-C. Lee, "Taiwan's renewable energy strategy and energy-intensive industrial policy," Renewable and Sustainable Energy Reviews, vol. 64, pp. 456-465, 2016/10/01/ 2016, doi: https://doi.org/10.1016/j.rser.2016.06.052.
[22] H.-C. Lee and C.-T. Chang, "Comparative analysis of MCDM methods for ranking renewable energy sources in Taiwan," Renewable and Sustainable Energy Reviews, vol. 92, pp. 883-896, 2018/09/01/ 2018, doi: https://doi.org/10.1016/j.rser.2018.05.007.
[23] T. J. Hammons, "Integrating renewable energy sources into European grids," International Journal of Electrical Power & Energy Systems, vol. 30, no. 8, pp. 462-475, 2008/10/01/ 2008, doi: https://doi.org/10.1016/j.ijepes.2008.04.010.
[24] A. Zahedi, "Maximizing solar PV energy penetration using energy storage technology," Renewable and Sustainable Energy Reviews, vol. 15, no. 1, pp. 866-870, 2011/01/01/ 2011, doi: https://doi.org/10.1016/j.rser.2010.09.011.
[25] A. Strzalka, N. Alam, E. Duminil, V. Coors, and U. Eicker, "Large scale integration of photovoltaics in cities," Applied Energy, vol. 93, pp. 413-421, 2012/05/01/ 2012, doi: https://doi.org/10.1016/j.apenergy.2011.12.033.
[26] U. K. Das et al., "Forecasting of photovoltaic power generation and model optimization: A review," Renewable and Sustainable Energy Reviews, vol. 81, pp. 912-928, 2018/01/01/ 2018, doi: https://doi.org/10.1016/j.rser.2017.08.017.
[27] A. Murata, H. Ohtake, and T. Oozeki, "Modeling of uncertainty of solar irradiance forecasts on numerical weather predictions with the estimation of multiple confidence intervals," Renewable Energy, vol. 117, pp. 193-201, 2018/03/01/ 2018, doi: https://doi.org/10.1016/j.renene.2017.10.043.
[28] J. Alonso-Montesinos and F. J. Batlles, "Solar radiation forecasting in the short- and medium-term under all sky conditions," Energy, vol. 83, pp. 387-393, 2015/04/01/ 2015, doi: https://doi.org/10.1016/j.energy.2015.02.036.
[29] R. Prasad, M. Ali, Y. Xiang, and H. Khan, "A double decomposition-based modelling approach to forecast weekly solar radiation," Renewable Energy, vol. 152, pp. 9-22, 2020/06/01/ 2020, doi: https://doi.org/10.1016/j.renene.2020.01.005.
[30] S. Monjoly, M. André, R. Calif, and T. Soubdhan, "Hourly forecasting of global solar radiation based on multiscale decomposition methods: A hybrid approach," Energy, vol. 119, pp. 288-298, 2017/01/15/ 2017, doi: https://doi.org/10.1016/j.energy.2016.11.061.
[31] B. Gao, X. Huang, J. Shi, Y. Tai, and J. Zhang, "Hourly forecasting of solar irradiance based on CEEMDAN and multi-strategy CNN-LSTM neural networks," Renewable Energy, vol. 162, pp. 1665-1683, 2020/12/01/ 2020, doi: https://doi.org/10.1016/j.renene.2020.09.141.
[32] M. Mishra, P. Byomakesha Dash, J. Nayak, B. Naik, and S. Kumar Swain, "Deep learning and wavelet transform integrated approach for short-term solar PV power prediction," Measurement, vol. 166, p. 108250, 2020/12/15/ 2020, doi: https://doi.org/10.1016/j.measurement.2020.108250.
[33] T. Peng, C. Zhang, J. Zhou, and M. S. Nazir, "An integrated framework of Bi-directional long-short term memory (BiLSTM) based on sine cosine algorithm for hourly solar radiation forecasting," Energy, vol. 221, p. 119887, 2021/04/15/ 2021, doi: https://doi.org/10.1016/j.energy.2021.119887.
[34] J. Qu, Z. Qian, and Y. Pei, "Day-ahead hourly photovoltaic power forecasting using attention-based CNN-LSTM neural network embedded with multiple relevant and target variables prediction pattern," Energy, vol. 232, p. 120996, 2021/10/01/ 2021, doi: https://doi.org/10.1016/j.energy.2021.120996.
[35] A. Agga, A. Abbou, M. Labbadi, and Y. El Houm, "Short-term self consumption PV plant power production forecasts based on hybrid CNN-LSTM, ConvLSTM models," Renewable Energy, vol. 177, pp. 101-112, 2021/11/01/ 2021, doi: https://doi.org/10.1016/j.renene.2021.05.095.
[36] Y. Gao, S. Miyata, and Y. Akashi, "Multi-step solar irradiation prediction based on weather forecast and generative deep learning model," Renewable Energy, vol. 188, pp. 637-650, 2022/04/01/ 2022, doi: https://doi.org/10.1016/j.renene.2022.02.051.
[37] S. Almaghrabi, M. Rana, M. Hamilton, and M. S. Rahaman, "Solar power time series forecasting utilising wavelet coefficients," Neurocomputing, vol. 508, pp. 182-207, 2022/10/07/ 2022, doi: https://doi.org/10.1016/j.neucom.2022.08.016.
[38] J. Tong, L. Xie, S. Fang, W. Yang, and K. Zhang, "Hourly solar irradiance forecasting based on encoder–decoder model using series decomposition and dynamic error compensation," Energy Conversion and Management, vol. 270, p. 116049, 2022/10/15/ 2022, doi: https://doi.org/10.1016/j.enconman.2022.116049.
[39] T. Limouni, R. Yaagoubi, K. Bouziane, K. Guissi, and E. H. Baali, "Accurate one step and multistep forecasting of very short-term PV power using LSTM-TCN model," Renewable Energy, vol. 205, pp. 1010-1024, 2023/03/01/ 2023, doi: https://doi.org/10.1016/j.renene.2023.01.118.
[40] Z. Xiao, X. Huang, J. Liu, C. Li, and Y. Tai, "A novel method based on time series ensemble model for hourly photovoltaic power prediction," Energy, vol. 276, p. 127542, 2023/08/01/ 2023, doi: https://doi.org/10.1016/j.energy.2023.127542.
[41] B. Ratner, "The correlation coefficient: Its values range between +1/−1, or do they?," Journal of Targeting, Measurement and Analysis for Marketing, vol. 17, no. 2, pp. 139-142, 2009/06/01 2009, doi: 10.1057/jt.2009.5.
[42] G. Qian, S. Sural, Y. Gu, and S. Pramanik, "Similarity between Euclidean and cosine angle distance for nearest neighbor queries," in Proceedings of the 2004 ACM symposium on Applied computing, 2004, pp. 1232-1237.
[43] I. Cohen et al., "Pearson correlation coefficient," Noise reduction in speech processing, pp. 1-4, 2009.
[44] J. Hauke and T. Kossowski, "Comparison of values of Pearson's and Spearman's correlation coefficients on the same sets of data," Quaestiones geographicae, vol. 30, no. 2, pp. 87-93, 2011.
[45] Y.-W. Cheung and K. S. Lai, "Lag order and critical values of the augmented Dickey–Fuller test," Journal of Business & Economic Statistics, vol. 13, no. 3, pp. 277-280, 1995.
[46] C. F. Baum, "sts15: Tests for stationarity of a time series," Stata Technical Bulletin, vol. 57, pp. 36-39, 2000.
[47] M. C. McGraw and E. A. Barnes, "Memory matters: A case for Granger causality in climate variability studies," Journal of climate, vol. 31, no. 8, pp. 3289-3300, 2018.
[48] S. Arlot and A. Celisse, "A survey of cross-validation procedures for model selection," 2010.
[49] S. R. Safavian and D. Landgrebe, "A survey of decision tree classifier methodology," IEEE transactions on systems, man, and cybernetics, vol. 21, no. 3, pp. 660-674, 1991.
[50] G. Biau and E. Scornet, "A random forest guided tour," TEST, vol. 25, no. 2, pp. 197-227, 2016/06/01 2016, doi: 10.1007/s11749-016-0481-7.
[51] T. Chen and C. Guestrin, "Xgboost: A scalable tree boosting system," in Proceedings of the 22nd acm sigkdd international conference on knowledge discovery and data mining, 2016, pp. 785-794.
[52] D. O. Afanasyev and E. A. Fedorova, "The long-term trends on the electricity markets: Comparison of empirical mode and wavelet decompositions," Energy Economics, vol. 56, pp. 432-442, 2016/05/01/ 2016, doi: https://doi.org/10.1016/j.eneco.2016.04.009.
[53] D. Shu-wen, L. Lu-jun, W. Qing-qu, W. Kang-le, and C. Peng-zhan, "Fiber optic gyro noise reduction based on hybrid CEEMDAN-LWT method," Measurement, vol. 161, p. 107865, 2020/09/01/ 2020, doi: https://doi.org/10.1016/j.measurement.2020.107865.
[54] M. E. Torres, M. A. Colominas, G. Schlotthauer, and P. Flandrin, "A complete ensemble empirical mode decomposition with adaptive noise," in 2011 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 22-27 May 2011 2011, pp. 4144-4147, doi: 10.1109/ICASSP.2011.5947265.
[55] O. Singh and R. K. Sunkaria, "ECG signal denoising via empirical wavelet transform," Australasian Physical & Engineering Sciences in Medicine, vol. 40, no. 1, pp. 219-229, 2017/03/01 2017, doi: 10.1007/s13246-016-0510-6.
[56] J. Gilles, "Empirical wavelet transform," IEEE transactions on signal processing, vol. 61, no. 16, pp. 3999-4010, 2013.

簡易檢索 / 詳目顯示

相關論文