DOI QR코드

DOI QR Code

A Study on Deep Learning Input Pattern for Summer Power Demand Prediction

하계 전력수요 예측을 위한 딥 러닝 입력 패턴에 관한 연구

  • 신동하 (가천대학교 에너지 IT 학과) ;
  • 김창복 (가천대학교 에너지 IT 학과)
  • Received : 2016.10.17
  • Accepted : 2016.11.10
  • Published : 2016.11.30

Abstract

The machine learning is receiving attention as a new method for energy big data analysis and power demand prediction to be more effectively operating a power system. In this paper, we study input pattern of deep learning for power demand prediction using machine learning package of R and tensorflow. The input pattern and learning rate in deep learning is the most important factor in power demand prediction. However, the input pattern is because humans have directly determined, must determine by repeated experiment. The factor of power demand prediction was used average temperature, sensible temperature, cooling degree hours, discomfort index. As a result, input pattern power demand and average temperature for one week was obtained the best results about power demand prediction. In addition, we were enhanced to more prediction results by adding the sensible temperature and discomfort index elements. As future research, proposed model need to build more suitable network for deep learning, and need to use of meteorological elements Big Data to improve power demand prediction.

기계학습은 전력계통을 효율적으로 운영할 수 있도록, 에너지 빅 데이터 분석과 전력수요예측을 위한 방법으로 관심을 받고 있다. 본 논문은 R과 텐서플로우의 기계학습 패키지를 이용하여, 전력수요예측을 위한 딥러닝 입력패턴에 대해서 연구하였다. 딥 러닝에서 입력패턴과 학습률은 전력수요예측에서 가장 중요한 요소이지만 인간이 직접 결정해야하기 때문에, 반복적인 실험에 의해 결정해야한다. 전력수요 예측요소는 당일 전력수요와 상관관계가 있는 평균온도, 체감온도, 불쾌지수를 이용하였다. 결과로서, 일주일간 전력수요와 평균온도 데이터 입력패턴이 전력수요예측에서 가장 좋은 결과를 나타냈다. 또한, 체감온도, 불쾌지수 등의 예측요소를 추가함으로서 좀 더 예측결과를 향상시킬 수 있었다. 향후 연구과제로서 제안 모델은 전력수요예측 향상을 위해 기상 요소 빅 데이터를 이용하고, 보다 적합한 딥 러닝 네트워크 구축이 필요하다.

Keywords

References

  1. B. G. Koo, H. S Kim, H. S. Lee, and J. Ho Park, "Short-term Electric Load Forecasting for Summer Season using Temperature Data", The Transactions of the Korean Institute of Electrical Engineers, Vol. 64, No. 8, pp. 1137-1144, Aug. 2015. https://doi.org/10.5370/KIEE.2015.64.8.1137
  2. Y. T. Son, "Thoughts on Artificial Intelligence and Power Industry", Korea electric association, Journal of Electrical World Monthly Magazine, pp. 70-70, Apr. 2016.
  3. S. I. Kong, Y. S. Baek, K. B. Song, and J. H. Park, "The Daily Peak Load Forecasting in Summer with the Sensitivity of Temperature", The Transactions of the Korean Institute of Electrical Engineers, Vol. 53(A), No. 6, pp. 538-363, Jun. 2004.
  4. D. H. Ahn, K. H. Song, and E. J. Choi, "The Research for the Change of Load Demand in Wintertime by the Influence of a Climate", Journal of the Korean Institute of Illuminating and Electrical Installation Engineers, Vol. 23, No. 9, pp. 47-54, Sep. 2009. https://doi.org/10.5207/JIEIE.2009.23.9.047
  5. K. S. Kim, G. H. Ryu, and K. B. Song, "Shortterm Load Forecasting of Using Data refine for Temperature Characteristics at Jeju Island", The Transactions of the Korean Institute of Electrical Engineers, Vol. 58, No. 9, pp. 1695-1699, Sep. 2009.
  6. M. K. Kim and C. E. Hong, "The Artificial Neural Network based Electric Power Demand Forecast using a Season and Weather Informations", The Transactions of the Korean Institute of Electrical Engineers, Vol. 53, No. 1, pp. 71-78, Jan. 2016.
  7. S. M. An, "Deep Learning Architectures and Applications", Korea intelligent information system society, intelligent information study, Vol. 22, No. 2, Jun. 2016.
  8. https://cran.r-project.org/, [Accessed: Aug. 13. 2016]
  9. http://www.openstatistics.net/doc/contributed/2013/Eric-R-Data-Mining-20130812_v3.pdf, [Accessed: Aug. 13. 2016]
  10. J. Torres, "First Contact with tensorflow", Hanbit Meadia, 2016.
  11. H, Geoffrey, S. Osindero, and Y. W. Teh. "A fast learning algorithm for deep belief nets", Neural computation, Vol. 18. No. 7, pp. 1527-1554, May 2006. https://doi.org/10.1162/neco.2006.18.7.1527
  12. http://www.swig.org, [Accessed: Sep. 21. 2016]

Cited by

  1. The Implementation of Smart Plug-in Switching Mode Power Supply vol.16, pp.2, 2016, https://doi.org/10.14801/jkiit.2018.16.2.17
  2. The Implementation of Smart Motor Diagnosis System vol.17, pp.4, 2016, https://doi.org/10.14801/jkiit.2019.17.4.27
  3. 인공지능 기반 수요예측 기법의 리뷰 vol.32, pp.6, 2016, https://doi.org/10.5351/kjas.2019.32.6.795
  4. Prediction of Special Day’s Hourly Load Using MLP, SVR and RF vol.18, pp.5, 2020, https://doi.org/10.14801/jkiit.2020.18.5.1
  5. A Study on the Development of Machine-Learning Based Load Transfer Detection Algorithm for Distribution Planning vol.13, pp.17, 2016, https://doi.org/10.3390/en13174358
  6. Deep Learning-Based Smart Meter Wattage Prediction Analysis Platform vol.9, pp.4, 2016, https://doi.org/10.7236/ijasc.2020.9.4.173
  7. 배전계획을 고려한 실데이터 및 기계학습 기반의 배전선로 부하예측 기법에 대한 연구 vol.7, pp.1, 2016, https://doi.org/10.18770/kepco.2021.07.01.171