卷積長短期記憶神經網路結合全天空影像特徵之短期日射量預測模型

以作者查詢圖書館館藏

、以作者查詢臺灣博碩士

、以作者查詢全國書目

、勘誤回報

、線上人數：58

、訪客IP：18.188.205.249

姓名

張至妤(Chih-Yu Chang) 查詢紙本館藏

畢業系所

資訊工程學系

論文名稱

卷積長短期記憶神經網路結合全天空影像特徵之短期日射量預測模型
(Very Short-Term Solar Irradiance forecasting using convolution Long-Short-Term Memory Network with All-Sky images Features)

相關論文

★ 影片指定對象臉部置換系統	★ 以單一攝影機實現單指虛擬鍵盤之功能
★ 基於視覺的手寫軌跡注音符號組合辨識系統	★ 利用動態貝氏網路在空照影像中進行車輛偵測
★ 以視訊為基礎之手寫簽名認證	★ 使用膚色與陰影機率高斯混合模型之移動膚色區域偵測
★ 影像中賦予信任等級的群眾切割	★ 航空監控影像之區域切割與分類
★ 在群體人數估計應用中使用不同特徵與回歸方法之分析比較	★ 以視覺為基礎之強韌多指尖偵測與人機介面應用
★ 在夜間受雨滴汙染鏡頭所拍攝的影片下之車流量估計	★ 影像特徵點匹配應用於景點影像檢索
★ 自動感興趣區域切割及遠距交通影像中的軌跡分析	★ 基於回歸模型與利用全天空影像特徵和歷史資訊之短期日射量預測
★ Analysis of the Performance of Different Classifiers for Cloud Detection Application	★ 全天空影像之雲追蹤與太陽遮蔽預測

檔案

[Endnote RIS 格式]

[Bibtex 格式]

[相關文章]

[文章引用]

[完整記錄]

[館藏目錄]

至系統瀏覽論文 (2027-7-15以後開放)

摘要(中)

隨著再生能源的興起，太陽能源的發展日漸重要，臺灣因年均日照長，發展太陽能條件良好，若能有穩定的產電預測，將使整體電力系統更彈性、更能因應劇烈的太陽變化。
太陽能發電量主要受到日射量多寡來轉換，因此本篇論文將以日射量的預測進行研究。而短期的日射量變化容易受到天空狀況影響，雲的狀態將影響當下太陽照射到地面的輻射量，故本篇論文提出以深度學習的方式，使用卷積神經網路與長短期記憶模型，根據歷史的日射量及全天空影像資訊，來預測未來逐分鐘的日射量。先藉由全天空影像判斷不同的天空狀態，再將日射計及全天空影像儀蒐集的資訊當作訓練特徵並結合，根據不同的天空狀態類別訓練出不同的深度學習模型，預測時即能根據當下天空狀態選擇適合的模型得到未來日射量結果。其中訓練模型時，透過全天空影像藉太陽位置演算法得到影像中太陽位置，再分析太陽附近灰階度特徵，藉由太陽附近灰階度特徵及全天空影像特徵來輔助日射量特徵進行未來日射量預測。
整體實驗藉由分月份的方式進行訓練、驗證及測試，所有實驗結果將在RMSE、RMSPE、MAE、MAPE上做評估，並根據預測的時長分開統計，實驗顯示藉由不同天空狀態分模型的方式，並影像及數值特徵結合的方法，能夠以較低的誤差預測出日射量。

摘要(英)

Solar energy is becoming more and more popular and important with the rise of renewable energy. Taiwan experiences long, hot summers, and short, mild winters. This makes Taiwan an ideal location for solar energy development. If there is a stable solar energy power generation, the overall power system will be more flexible and able to respond quickly to high fluctuations in supply and demand.
Solar power generation is mainly converted by the amount of solar irradiance. Short-term changes in solar irradiance are easily affected by sky conditions, and the state of clouds. In this work, a very short-term solar irradiance prediction mechanism based on automatic sky conditions classification is proposed. First, different sky states are judged by the all-sky images. Multiple deep learning models are constructed by different sky states using historical irradiance values and all-sky images as features. The deep learning model, which combines convolutional neural network and long- short-term memory network, can extracted both image and numerical features. Moreover, the solar position in the image is obtained through the solar position algorithm, and then the around pixels’ gray-scale features are analyzed and then input to the training model too.
The overall experiment is trained, validated and tested by monthly. All experimental results will be evaluated on RMSE, RMSPE, MAE, and MAPE. Results have shown that incorporating sky conditions information can capture different characteristics of irradiance variation under different sky states, and the method combined with images and numerical features can predict the irradiance with low error.

關鍵字(中)

★ 深度學習
★ 卷積神經網路
★ 長短期記憶
★ 太陽輻射量
★ 全天空影像
★ 太陽位置演算法

關鍵字(英)

★ Deep learning
★ Convolutional neural network
★ Long short-term memory
★ Solar Irradiance
★ All-sky images
★ Solar Position Algorithm

論文目次

摘要 I
Abstract II
致謝 III
目錄 IV
圖目錄 VII
表目錄 IX
第一章緒論 1
1.1 研究背景與動機 1
1.2 研究目的 2
1.3 論文架構 3
第二章相關研究 4
2.1 過往研究 4
2.1.1 統計方法與機器學習方法 4
2.1.2 神經網路方法 5
2.1.3 全天空影像研究 7
2.2 重要的神經網路架構 10
2.2.1 卷積層(Convolution layer) 10
2.2.2 池化層(Pooling layer) 11
2.2.3 全連接層(Fully connected layer) 12
2.2.4 長短期記憶-門控循環單元(Gated Recurrent Unit) 12
第三章研究方法 14
3.1 模型架構 14
3.2 資料收集 16
3.3 研究方法流程 19
3.3.1 資料前處理流程 20
3.3.2 太陽位置演算法 21
3.3.3 日射量預測實驗方法 22
3.3.4 雲種之天空狀況分類 24
3.3.4.1 雲種介紹 24
3.3.4.2 天空分類模型架構 25
3.3.4.3 天空狀況分類模型訓練方法與結果 26
3.3.5 雙模型日射量預測實驗方法 27
3.3.6 訓練與評估方法 28
第四章實驗結果 29
4.1 設備環境設定 29
4.2 案場之實驗結果 30
4.3 模型超參數選用實驗 35
4.3.1 影像模型卷積區塊數實驗 36
4.3.2 影像及數值模型池化層使用實驗 37
4.3.3 數值模型卷積核長度與數量實驗 38
4.4 消融實驗 39
4.4.1 數值輸入特徵比較實驗 39
4.4.2 影像輸入特徵比較實驗 41
4.4.3 特徵結合比較實驗 43
4.4.4 無影像特徵比較實驗 45
4.5 訓練時間長度比較實驗 47
4.6 其他模型結果比較 49
4.6.1 替換影像特徵模型比較實驗 49
4.6.2 其他日射量模型比較實驗 51
4.7 模型結果可視化 53
4.8 預測時長實驗 57
4.9 數據平滑化之長時間預測實驗 61
第五章結論與未來研究方向 62
參考文獻 63
附錄一.影像模型卷積區塊數實驗(4.3.1) 71
附錄二.影像及數值模型池化層使用實驗(4.3.2) 72
附錄三.數值模型卷積核長度與數量實驗(4.3.3) 73
附錄四.數值輸入特徵比較實驗(4.4.1) 74
附錄五.影像輸入特徵比較實驗(4.4.2) 78
附錄六.特徵結合比較實驗(4.4.3) 82
附錄七.無影像特徵比較實驗(4.4.4) 86
附錄八.訓練時間長度比較實驗(4.5) 90
附錄九.替換影像特徵模型比較實驗(4.6.1) 94
附錄十.其他日射量模型比較實驗(4.6.2) 94
附錄十一.預測時長實驗(4.8) 96
附錄十二.數據平滑化之長時間模型實驗(4.9) 99

參考文獻

[1] 綠能科技產業推動中心, “累計至2021年12月份再生能源裝置容量及預估年發電量”, Internet: https://www.geipc.tw/LiveEnergy.aspx.

[2] PGE太平洋綠能, “2021太陽能趨勢｜太陽能還有發展性嗎？解析台灣、國際綠能之路”, Internet: https://blog.pgesolar.com.tw/2021/02/08/%E5%A4%AA%E9%99%BD%E8%83%BD%E8%B6%A8%E5%8B%A2.
[3] 台灣智慧型電網產業協會, “認識智慧電網”, Internet:
http://www.smart-grid.org.tw/content/smart_grid/smart_grid.aspx.
[4] G. K. Singh, “Solar power generation by PV (photovoltaic) technology: A review”, Energy, vol. 53, pp. 1-13, 2013.
[5] H. Lund, “Renewable energy strategies for sustainable development”, Energy, vol. 32, pp. 912-919, 2007.
[6] U. Firat, S. N. Engin, M. Saraclar and A. B. Ertuzun, “Wind Speed Forecasting Based on Second Order Blind Identification and Autoregressive Model”, 2010 Ninth International Conference on Machine Learning and Applications, 2010, pp. 686-691, doi: 10.1109/ICMLA.2010.106.
[7] H. Pedro and C. Coimbra, “Assessment of forecasting techniques for solar power production with no exogenous inputs”, Solar Energy, vol. 86, no. 7, pp. 2017-2028, Jul. 2012.
[8] M. Abuella and B. Chowdhury, “Solar Power Forecasting Using Support Vector Regression”, Proceedings of the American Society for Engineering Management 2016 International Annual Conference, 2016.
[9] M. Abuella and B. Chowdhury, “Solar power probabilistic forecasting by using multiple linear regression analysis”, Proc. SoutheastCon, pp. 1-5, 2015.
[10] A. Mellit, S. Kalogirou, S. Shaari, H. Salhi and A. H. Arab, “Methodology for predicting sequences of mean monthly clearness index and daily solar radiation data in remote areas: application for sizing a stand-alone PV system”, Renew. Energy, vol. 33, pp. 1570-1590, 2008.
[11] B. Urquhart, C. W. Chow, D. Nguyen, J. Kleissl, M. Sengupta, J. Blatchford and D. Jeon, “Towards intra-hour solar forecasting using two sky imagers at a large solar power plant”, American Solar Energy Society, 1-6, 2012.
[12] N. Sharma, P. Sharma, D. Irwin and P. Shenoy, “Predicting solar generation from weather forecasts using machine learning”, Proc. 2nd IEEE Int. Conf. Smart Grid Commun., pp. 528-533, Oct. 2011.
[13] H. Zhu, W. Lian, L. Lu et al., “An improved forecasting method for photovoltaic power based on adaptive BP neural network with a scrolling time window”, Energies, vol. 10, no. 10, pp. 1542-1547, Oct. 2017.
[14] F. Wang, Z. Mi, S. Su and H. Zhao, “Short-term solar irradiance forecasting model based on artificial neural network using statistical feature parameters”, Energies, vol. 5, no. 5, pp. 1355-1370, May. 2012.
[15] M. Behrang, E. Assareh, A. Ghanbarzadeh, and A. Noghrehabadi, “The potential of different artificial neural network (ANN) techniques in daily global solar radiation modeling based on meteorological data”, Solar Energy, vol. 84, pp. 1468-1480, 2010.
[16] M. Marquez and C. F. M. Coimbra, “Forecasting of global and direct solar irradiance using stochastic learning methods, ground experiments and the NWS database”, Solar Energy, vol. 85, no. 5, pp. 746-756, 2011.
[17] F. O. Hocaolu, . N. Gerek and M. Kurban, “Hourly solar radiation forecasting using optimal coefficient 2-d linear filters and feed-forward neural networks”, Sol. Energy, vol. 82, no. 8, pp. 714-726, 2008.
[18] A. Mellit, M. Benghanem, and S.A. Kalogirou, “An adaptive wavelet-network model for forecasting daily total solar radiation”, Applied Energy, Vol.83 pp.705-722, 2006.
[19] J.C. Cao and X.C. Lin, “Application of the diagonal recurrent wavelet neural network to solar irradiation forecast assisted with fuzzy technique”, Engineering Applications of Artificial Intelligence, vol. 21, no. 8, pp. 1255-1263, Dec. 2008.
[20] A.C. Tsoi, “Recurrent neural network architectures: an overview”, in Adaptive processing of sequences and data structures, Berlin:Springer-Verlag, pp. 1-26, 1998.
[21] M. Husken, P. Stagge, “Recurrent neural networks for time series classification”, Neurocomputing, 2003, 50(1): 223 – 235.
[22] G. Li, H. Wang, S. Zhang, J. Xin and H. Liu, “Recurrent neural networks based photovoltaic power forecasting approach”, Energies, vol. 12, no. 13, pp. 2538, 2019.
[23] X. Qing and Y. Niu, “Hourly day-ahead solar irradiance prediction using weather forecasts by LSTM”, Energy, vol. 148, no. 1, pp. 461-468, Apr. 2018.
[24] X. Tang, Y. Dai, T. Wang and Y. Chen, “Short-term power load forecasting based on multi-layer bidirectional recurrent neural network”, IET Gener. Transmiss. Distrib., vol. 13, no. 17, pp. 3847-3854, Sep. 2019.
[25] B. Gao, X. Huang, J. Shi, Y. Tai and R. Xiao, “Predicting day-ahead solar irradiance through gated recurrent unit using weather forecasting data”, J. Renew. Sustain. Energy, vol. 11, no. 4, Jul. 2019.
[26] H. Aprillia, H.-T. Yang and C.-M. Huang, “Short-term photovoltaic power forecasting using a convolutional neural network–salp swarm algorithm”, Energies, vol. 13, no. 8, pp. 1879, 2020.
[27] C. J. Huang and P. H. Kuo, “Multiple-input deep convolutional neural network model for short-term photovoltaic power forecasting”, IEEE Access, vol. 7, pp. 74822-74834, 2019.
[28] Z. Deng, B. Wang, Y. Xu, T. Xu, C. Liu and Z. Zhu, “Multi-scale convolutional neural network with time-cognition for multi-step short-term load forecasting”, IEEE Access, vol. 7, pp. 88058-88071, 2019.
[29] V. Suresh, P. Janik, J. Rezmer and Z. Leonowicz, “Forecasting solar PV output using convolutional neural networks with a sliding window algorithm”, Energies, vol. 13, no. 3, pp. 723, Feb. 2020.
[30] S. Ghimire, R. C. Deo, N. Raj and J. Mi, “Deep solar radiation forecasting with convolutional neural network and long short-term memory network algorithms”, Appl. Energy, vol. 253, Nov. 2019.
[31] Y. Y. Hong, J. J. F. Martinez and A. C. Fajardo, “Day-ahead solar irradiation forecasting utilizing Gramian angular field and convolutional long short-term memory”, IEEE Access, vol. 8, pp. 18741-18753, 2020.
[32] C. N. Long, D. W. Slater and T. P. Tooman, “Total sky Imager model 880 status and testing results”, Pacific Northwest National Laboratory, 2001.
[33] C. N. Long, J. M. Sabburg, J. Calbó and D. Pagès, “Retrieving cloud characteristics from ground-based daytime color all-sky images”, J. Atmos. Ocean. Technol., vol. 23, no. 5, pp. 633-652, May. 2006.
[34] J. E. Shields, R. W. Johnson and T. L. Koehler, “Automated whole sky imaging systems for cloud field assessment”, Proc. 4th Symp. Global Change Studies Amer. Meteorological Soc. Annu. Meeting, 1993.
[35] J. E. Shield, R. W. Johnson, M. E. Karr and J. L. Wertz, “Automated day/night whole sky imagers for field assessment of cloud cover distributions and radiance distributions”, Proc. 10th Symp. Meteorol. Observ. Instrum. Amer. Meteorol. Soc., pp. 11-16, 1998.
[36] Z. Li, M. C. Cribb, F.-L. Chang and A. P. Trischenko, “Validation of MODIS-retrieved cloud fractions using whole sky imager measurements at the three ARM sites”, 14th Atmospheric Radiation Measurement (ARM) Science Team Meeting, Mar. 2004.
[37] M. Kubota, T. Nagatsuma and Y. Murayama, “Evening corotating patches: A new type of aurora observed by high sensitivity all-sky cameras in Alaska”, Geophys. Res. Lett., vol. 30, pp. 1612, 2003.
[38] G. Pfister, R. L. McKenzie, J. B. Liley, A. Thomas, B. W. Forgan and C. N. Long, “Cloud coverage based on all-sky imaging and its impact on surface solar irradiance”, J. Appl. Meteorol., vol. 42, no. 10, pp. 1421-1434, 2003.
[39] J. Calbó and J. Sabburg, “Feature extraction from whole-sky ground-based images for cloud-type recognition”, J. Atmos. Ocean. Technol., vol. 25, no. 1, pp. 3-14, 2008.
[40] A. Heinle, A. Macke and A. Srivastav., “Automatic cloud classification of whole sky images” , Atmos. Meas. Technol., vol. 3, no. 3, pp. 557-567, Jan. 2010.
[41] M. Martínez-Chico, F. J. Batlles and J. L. Bosch, “Cloud classification in a mediterranean location using radiation data and sky images”, Energy, vol. 36, no. 7, pp. 4055-4062, Jul. 2011.
[42] H. Huang, S. Yoo, D. Yu, D. Huang and H. Qin, “Correlation and local feature based cloud motion estimation”, Twelfth International Workshop on Multimedia Data Mining, pp. 669-675, 2012.
[43] R. Marquez and C. F. Coimbra, “Intra-hour DNI forecasting based on cloud tracking image analysis”, Solar Energy, vol. 91, pp. 327-336, 2013.
[44] F. Wang, Z. Xuan, Z. Zhen, Y. Li, K. Li, L. Zhao, et al., “A minutely solar irradiance forecasting method based on real-time sky image-irradiance mapping model”, Energy Convers. Manag., vol. 220, Sep. 2020.
[45] T. C. McCandless, S. E. Haupt, and G. S. Young. “A regime-dependent artificial neural network technique for short-range solar irradiance forecasting”, Renewable Energy, 89, 351-359, 2016.
[46] Y. Chu, H. T. Pedro, M. Li and C. F. Coimbra, “Real-time forecasting of solar irradiance ramps with smart image processing”, Solar Energy, vol. 114, pp. 91-104, 2015.
[47] J. Oktavia, T. Lit and T. Suwa, “Sky image-based solar irradiance prediction methodologies using artificial neural networks”, Renewable Energy, vol. 134, pp. 837-845, Apr. 2019.
[48] Q. Paletta, J. Lasenby, “Convolutional Neural Networks Applied to Sky Images for Short-Term Solar Irradiance Forecasting”, In EU PVSEC, 2020.
[49] C. Feng and J. Zhang, “Solarnet: A sky image-based deep convolutional neural network for intra-hour solar forecasting”, Sol. Energy, vol. 204, pp. 71-78, Jul. 2020.
[50] X. Zhao, H. Wei, H. Wang, T. Zhu and K. Zhang, “3D-CNN-based feature extraction of ground-based cloud images for direct normal irradiance prediction”, Sol. Energy, vol. 181, pp. 510-518, Mar. 2019.
[51] H. Yang, L. Wang, C. Huang, and X. Luo, “3D-CNN-Based Sky Image Feature Extraction for Short-Term Global Horizontal Irradiance Forecasting”, Water, vol. 13, no. 13, p. 1773, Jun. 2021.
[52] Q. Paletta, A. Hu, G. Arbod, J. Lasenby, “ECLIPSE: Envisioning Cloud Induced Perturbations in Solar Energy”, arXiv preprint, arXiv:2104.12419, 2021.
[53] K. Fukushima. “Neocognitron: A self-organizing neural network model for a mechanism of pattern recognition unaffected by shift in position”, Biological cybernetics, 36(4), 193-202, 1980.
[54] “池化層示意圖、圖片來源”, Internet: https://www.quora.com/What-is-the-benefit-of-using-average-pooling-rather-than-max-pooling
[55] S. Hochreiter, and J Schmidhuber, “Long short-term memory”, Neural computation, 9(8), 1735-1780, 1997.
[56] K. Cho et al., “Learning phrase representations using RNN encoder-decoder for statistical machine translation”, arXiv preprint, arXiv:1406.1078, 2014.
[57] H.K. Yuen, J. Princen, J. Illingworth, J Kittler, “Comparative Study of Hough Transform Methods for Circle Finding”, Image and Vision Computing, vol. 8 (1), pp. 71-77, 1990.
[58] I. Reda and A. Andreas, “Solar position algorithm for solar radiation applications”, Sol. Energy, 2004. 3, 4.
[59] L. Ye, Z. Cao, Y. Xiao and W. Li, “Ground-based cloud image categorization using deep convolutional visual features”, Proc. IEEE Int. Conf. Image Process. (ICIP), pp. 4808-4812, Sep. 2015.
[60] K. Simonyan and A. Zisserman, “Very deep convolutional networks for large-scale image recognition”, arXiv preprint, arXiv:1409.1556, 2014.
[61] “2020年6月至8月創台灣史上最熱一季”, Internet:
https://news.ltn.com.tw/news/life/breakingnews/3274135
[62] K. He, X. Zhang, S. Ren and J. Sun, “Deep residual learning for image recognition”, Proc. IEEE Conf. Comput. Vis. Pattern Recognit., pp. 770-778, Jun. 2016.
[63] M. Tan and Q. Le, “EfficientNet: Rethinking model scaling for convolutional neural networks”, Proc. 36th Int. Conf. Mach. Learn., pp. 6105-6114, 2019.

指導教授

鄭旭詠(Hsu-Yung Cheng)

審核日期

2022-7-18

推文