Forecastability Measures that Describe the Complexity of a Site for Deep Learning Wind Predictions

Authors

  • Jaume Manero Technical University Catalonia
  • Javier Béjar

DOI:

https://doi.org/10.14529/jsfi210102

Abstract

The application of deep learning to wind time series for multi-step prediction obtains good results at short horizons. The accuracy of a wind forecast is highly dependent on the specific structure of wind in the specific location, as many local features influence wind behaviour. The characterization of the complexity of a site for wind prediction is defined as forecastability or predictability and can be obtained from the inner structure of the meteorological time series observations from a site. We analyze the time series structure searching for properties that have a high correlation with the prediction result, properties that can create measures that have the potential to describe the forecastability of a site. The best measures will show a high correlation with the accuracy of the predictions. In this work, we analyze wind time series from 126,692 wind locations in the US, where we apply several deep learning methods first, and then we verify several forecastability descriptors with the accuracy deep learning results. We require High-Performance Computing (HPC) resources for this task as the deep learning algorithms have sensible resource requirements and are applied to a large set of data. The measures defined and explored in this work are based on several techniques that decompose or transform the wind time-series. By combining several of these measures, we can obtain better predictors of the site complexity, which will allow us to evaluate the future error of a prediction on this site. Forecastability measures can contribute to a wind site multi-dimensional description, becoming a valuable tool for wind resource analysts and wind forecasters.

References

Bastian, J., Jinxiang Zhu, Banunarayanan, V., Mukerji, R.: Forecasting energy prices in a competitive market. IEEE Computer Applications in Power 12(3), 40–45 (1999), DOI: 10.1109/67.773811

Ben Taieb, S., Bontempi, G., Atiya, A.F., Sorjamaa, A.: A review and comparison of strategies for multi-step ahead time series forecasting based on the NN5 forecasting competition. Expert Systems with Applications 39(8), 7067–7083 (2012), DOI: 10.1016/j.eswa.2012.01.039

Cao, Q., Ewing, B.T., Thompson, M.A.: Forecasting wind speed with recurrent neural networks. European Journal of Operational Research 221(1), 148–154 (2012), DOI: 10.1016/j.ejor.2012.02.042

Chollet, F.: Xception: Deep learning with depthwise separable convolutions. In: 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 21-26 July 2017, Honolulu, HI, USA. pp. 1800–1807. IEEE Computer Society, Los Alamitos, CA, USA (2017), DOI: 10.1109/CVPR.2017.195

Cleveland, W.S.: Robust locally weighted regression and smoothing scatterplots. Journal of the American Statistical Association 74(368), 829–836 (1979), DOI: 10.1080/01621459.1979.10481038

Draxl, C., Clifton, A., Hodge, B.M., McCaa, J.: The Wind Integration National Dataset (WIND) Toolkit. Applied Energy 151, 355–366 (2015), DOI: 10.1016/j.apenergy.2015.03.121

Feng, C., Chartan, E.K., Hodge, B.M., Zhang, J.: Characterizing time series data diversity for wind forecasting. In: Proceedings of the Fourth IEEE/ACM International Conference on Big Data Computing, Applications and Technologies, 5-8 Dec. 2017, Austin, Texas, USA. pp. 113–119. Association for Computing Machinery, New York, NY, USA (2017), DOI: 10.1145/3148055.3148065

Feng, C., Sun, M., Cui, M., Chartan, E.K., Hodge, B.M., Zhang, J.: Characterizing forecastability of wind sites in the United States. Renewable Energy 133, 1352–1365 (2019), DOI: 10.1016/j.renene.2018.08.085

Girard, R., Laquaine, K., Kariniotakis, G.: Assessment of wind power predictability as a decision factor in the investment phase of wind farms. Applied Energy 101, 609–617 (2013), DOI: 10.1016/j.apenergy.2012.06.064

Jacobson, M.Z., Delucchi, M.A., Bauer, Z.A., et al.: 100% clean and renewable wind, water, and sunlight all-sector energy roadmaps for 139 countries of the world. Joule 1(1), 108–121 (2017), DOI: 10.1016/j.joule.2017.07.005

Javier, S.R., Frías Paredes, L., Girard, R., et al.: The role of predictability in the investment phase of wind farms. In: Renewable Energy Forecasting: From Models to Applications, chap. 14, pp. 341–357. Woodhead Publishing Series in Energy, Elsevier - Woodhead Publishing (2017), DOI: 10.1016/B978-0-08-100504-0.00014-7

Li, G., Shi, J.: On comparing three artificial neural networks for wind speed forecasting. Applied Energy 87(7), 2313–2320 (2010), DOI: 10.1016/j.apenergy.2009.12.013

Liu, Z., Gao, W., Wan, Y.H., Muljadi, E.: Wind power plant prediction by using neural networks. In: IEEE Energy Conversion Congress and Exposition (ECCE), 15-20 Sept. 2012, Raleigh, NC, USA. pp. 3154–3160. IEEE (2012), DOI: 10.1109/ECCE.2012.6342351

Manero, J.: Deep learning architectures applied to wind time series multi-step forecasting. Ph.D. thesis, Technical University of Catalonia UPC. Department of Computer Science (2020), http://hdl.handle.net/2117/328183

Manero, J., Béjar, J., Cortés, U.: Wind energy forecasting with neural networks. a literature review. Computación y Sistemas 22, 1085–1098 (2018), DOI: 10.13053/CyS-22-4-3081

Manero, J., Béjar, J., Cortés, U.: “Dust in the Wind...”, Deep Learning application to Wind Energy time series forecasting. Energies 12(12), 2385 (2019), DOI: 10.3390/en12122385

Martorell, J.M.: Barcelona Supercomputing Center: Science accelerator and producer of innovation. Contributions to Science 12(1), 5–11 (2016), DOI: 10:2436/20.7010.01.238

Richman, J.S., Moorman, J.R.: Physiological time-series analysis using approximate entropy and sample entropy. American Journal of Physiology-Heart and Circulatory Physiology 278(6), H2039–H2049 (2000), DOI: 10.1152/ajpheart.2000.278.6.H2039

Rogers, A.L., Rogers, J.W., Manwell, J.F.: Comparison of the performance of four measure-correlate-predict algorithms. Journal of Wind Engineering and Industrial Aerodynamics 93(3), 243–264 (2005), DOI: 10.1016/j.jweia.2004.12.002

Shi, J., Guo, J., Zheng, S.: Evaluation of hybrid forecasting approaches for wind speed and power generation time series. Renewable and Sustainable Energy Reviews 16(5), 3471–3480 (2012), DOI: 10.1016/j.rser.2012.02.044

Sun, W., Wang, Y.: Short-term wind speed forecasting based on fast ensemble empirical mode decomposition, phase space reconstruction, sample entropy and improved back-propagation neural network. Energy Conversion and Management 157, 1–12 (2018), DOI: 10.1016/j.enconman.2017.11.067

Walsh, C., Pineda, I.: Wind energy in Europe in 2018 trends and statistics. http://windeurope.org/about-wind/statistics/european/wind-energy-in-europe-in-2018/, accessed: 2021-01-09

Wang, J., Zong, Y., You, S., Trholt, C.: A review of Danish integrated multi-energy system flexibility options for high wind power penetration. Clean Energy 1(1), 23–35 (2017), DOI: 10.1093/ce/zkx002

Wang, X., Smith, K., Hyndman, R.: Characteristic-based clustering for time series data. Data Mining and Knowledge Discovery 13(3), 335–364 (2006), DOI: 10.1007/s10618-005-0039-x

Downloads

Published

2021-05-29

How to Cite

Manero, J., & Béjar, J. (2021). Forecastability Measures that Describe the Complexity of a Site for Deep Learning Wind Predictions. Supercomputing Frontiers and Innovations, 8(1), 8–27. https://doi.org/10.14529/jsfi210102