臺灣博碩士論文加值系統

近年來，太陽能因為有大量可再生能源上的需求，以及在永續發展觀點上無限的潛力，在全世界已逐漸變成很重要的研究主題。因此為了要有效的把太陽能發電產生出的電力和固有電網整合，一個可靠的系統如果能準確預測太陽能電廠的發電量，不僅在電力管理和運用策略上都可以以更有效率的方式完成。然而在太陽能發電的預測上是充滿挑戰的，因為太陽能發電會因為太陽輻射、天氣、以及其他外在因素而產生震盪。現階段的研究上，大多研究都無法同時預測未來多個時間點的太陽能發電量，也沒有分析可能會造成太陽能發電變化的因子的影響，單純把收集到的特徵資料全部送進模型，而沒有考慮到各個特徵資料的特性。因此，這篇論文提供了對太陽能發電量有影響的因子(太陽輻射，天氣...等)，進行完整且全面的分析，同時也提出了多模型、點對點的類神經網路架構，藉由多工學習的方式，同時預測未來多個時間點的太陽能發電量。最後，藉由多工學習運用在異質資料上，我們提出的方法，與現有的方法中相比，可以達到最低的錯誤率(11.83\%在五分鐘的預測上)。

In recent years, solar energy has become a significant field of research across the globe because of the growing demand for renewable energy and its promising potential in sustainability aspects. Therefore, with the increasing integration of photovoltaic (PV) systems to the electrical grid, reliable prediction of the expected production output of PV systems is gaining importance as a basis for management and operation strategies. However, power production of PV systems is highly variable due to its dependence on solar radiance, meteorological conditions and other external factors. Currently, most studies are unable to predict the solar power production at multiple future time points and never analyze the influence of the possible factors on the solar power production but simply feed all features without considering their properties. Therefore, this paper provides a holistic comparison among all factors (e.g., solar radiance, meteorology) affecting the solar power production and also, a multimodal and end-to-end neural networks model is proposed to simultaneously predict the solar power production at multiple future time points by multitask learning. Finally, with multitask learning on heterogeneous (and multimodal) data, the proposed method achieves the lowest error rates (11.83\% for 5-minute prediction) compared to the state of the art.

口試委員會審定書 ii
誌謝 iii
摘要 iv
Abstract v
1 Introduction 1
2 Related work 5
3 Method 7
3.1 Heterogeneous and multimodal features 7
3.2 Architecture of Heterogeneous Multitask Learning 9
3.3 Loss function for multitask learning 11
4 Experiments 13
4.1 Datasets 13
4.2 Evaluation Metrics 14
4.3 Settings 15
4.4 Baselines 15
5 Results 17
5.1 Results based on different features 17
5.2 Results based on different features 18
5.3 T-test on learned models 20
5.4 Comparisonamongdifferentmethods 21
Bibliography 24

[1] M. Abadi, A. Agarwal, P. Barham, E. Brevdo, Z. Chen, C. Citro, G. S. Corrado, A. Davis, J. Dean, M. Devin, et al. Tensorflow: Large-scale machine learning on heterogeneous distributed systems. arXiv preprint arXiv:1603.04467, 2016.
[2] R. Aler, R. Martín, J. M. Valls, and I. M. Galván. A study of machine learning techniques for daily solar energy forecasting using numerical weather models. In Intelligent Distributed Computing VIII, pages 269–278. Springer, 2015.
[3] N. S. Altman. An introduction to kernel and nearest-neighbor nonparametric regres- sion. The American Statistician, 46(3):175–185, 1992.
[4] P.Bacher,H.Madsen,andH.A.Nielsen.Onlineshort-termsolarpowerforecasting. Solar Energy, 83(10):1772–1783, 2009.
[5] P. Chakraborty, M. Marwah, M. F. Arlitt, and N. Ramakrishnan. Fine-grained pho- tovoltaic output prediction using a bayesian ensemble. In AAAI. Citeseer, 2012.
[6] J. Chen and C.-W. Ngo. Deep-based ingredient recognition for cooking recipe re- trieval. In Proceedings of the 2016 ACM on Multimedia Conference, pages 32–41. ACM, 2016.
[7] J. Chung, C. Gulcehre, K. Cho, and Y. Bengio. Empirical evaluation of gated re- current neural networks on sequence modeling. arXiv preprint arXiv:1412.3555, 2014.
[8] M. Diagne, M. David, P. Lauret, J. Boland, and N. Schmutz. Review of solar irradi- ance forecasting methods and a proposition for small-scale insular grids. Renewable and Sustainable Energy Reviews, 27:65–76, 2013.
[9] M. Elhoseiny, T. El-Gaaly, A. Bakry, and A. Elgammal. Convolutional models for joint object categorization and pose estimation. arXiv preprint arXiv:1511.05175, 2015.
[10] J. H. Friedman. Greedy function approximation: a gradient boosting machine. An- nals of statistics, pages 1189–1232, 2001.
[11] Y. Gala, A. Fernández, and J. Dorronsoro. Machine learning prediction of global photovoltaic energy in spain. In International Conference on Renewable Energies and Power Quality, number 12, page 278, 2014.
[12] A.GravesandN.Jaitly.Towardsend-to-endspeechrecognitionwithrecurrentneural networks. In ICML, volume 14, pages 1764–1772, 2014.
[13] C.-T. Hsu, R. Korimara, L.-J. Tsai, and T.-J. Cheng. Photovoltaic power gener- ation system modeling using an artificial neural network. In Proceedings of the 3rd International Conference on Intelligent Technologies and Engineering Systems (ICITES2014), pages 365–371. Springer, 2016.
[14] E. Izgi, A. Öztopal, B. Yerli, M. K. Kaymak, and A. D. Şahin. Short–mid-term solar power prediction by using artificial neural networks. Solar Energy, 86(2):725–733, 2012.
[15] S. Jafarzadeh, M. S. Fadali, and C. Y. Evrenosoglu. Solar power prediction using interval type-2 tsk modeling. IEEE Transactions on Sustainable Energy, 4(2):333– 339, 2013.
[16] K. R. Kumar and M. S. Kalavathi. Artificial intelligence based forecast models for predicting solar power generation. 2016.
[17] H. Long, Z. Zhang, and Y. Su. Analysis of daily solar power prediction with data- driven approaches. Applied Energy, 126:29–37, 2014.
[18] E.Lorenz,T.Scheidsteger,J.Hurka,D.Heinemann,andC.Kurz.Regionalpvpower prediction for improved grid integration. Progress in Photovoltaics: Research and Applications, 19(7):757–771, 2011.
[19] R. Martin, R. Aler, J. M. Valls, and I. M. Galván. Machine learning techniques for daily solar energy prediction and interpolation using numerical weather models. Concurrency and Computation: Practice and Experience, 28(4):1261–1274, 2016.
[20] F.Pedregosa,G.Varoquaux,A.Gramfort,V.Michel,B.Thirion,O.Grisel,M.Blon- del, P. Prettenhofer, R. Weiss, V. Dubourg, J. Vanderplas, A. Passos, D. Cournapeau, M. Brucher, M. Perrot, and E. Duchesnay. Scikit-learn: Machine learning in Python. Journal of Machine Learning Research, 12:2825–2830, 2011.
[21] V. P. Singh, K. Vaibhav, and D. Chaturvedi. Solar power forecasting modeling us- ing soft computing approach. In Engineering (NUiCONE), 2012 Nirma University International Conference on, pages 1–5. IEEE, 2012.
[22] S. Venugopalan, M. Rohrbach, J. Donahue, R. Mooney, T. Darrell, and K. Saenko. Sequence to sequence-video to text. In Proceedings of the IEEE International Con- ference on Computer Vision, pages 4534–4542, 2015.
[23] B. Wolff, E. Lorenz, and O. Kramer. Statistical learning for short-term photovoltaic power predictions. In Computational Sustainability, pages 31–45. Springer, 2016.