Use of High-performance Graphics Processing Units for Power System Demand Forecasting

  • He, Ting (School of ITEE, The University of Queensland) ;
  • Meng, Ke (Department of Electrical Engineering, The Hong Kong Polytechnic University) ;
  • Dong, Zhao-Yang (Department of Electrical Engineering, The Hong Kong Polytechnic University) ;
  • Oh, Yong-Taek (Korea University of Technology and Education) ;
  • Xu, Yan (School of ITEE, The University of Queensland)
  • Received : 2010.01.18
  • Accepted : 2010.05.07
  • Published : 2010.09.01


Load forecasting has always been essential to the operation and planning of power systems in deregulated electricity markets. Various methods have been proposed for load forecasting, and the neural network is one of the most widely accepted and used techniques. However, to obtain more accurate results, more information is needed as input variables, resulting in huge computational costs in the learning process. In this paper, to reduce training time in multi-layer perceptron-based short-term load forecasting, a graphics processing unit (GPU)-based computing method is introduced. The proposed approach is tested using the Korea electricity market historical demand data set. Results show that GPU-based computing greatly reduces computational costs.


Artificial Neural Network;Graphics Processing Unit;Multi-layer Perceptron;Short-term Load Forecasting


  1. P.R.J. Campbell, K. Adamson, “Methodologies for load forecasting,” Int. IEEE Conf. Intell. Syst., pp. 800-806, Sept. 2006.
  2. L. Xu, Z.Y. Dong, and A. Tay, “Time series forecast with Elman neural networks and genetic algorithms,” in K.C. Tan, M.H. Lim, X. Yao, and L.P. Wang (ed): Recent Advances in Simulated Evolution and Learning, World Scientific series, Vol. 2, Advances in Natural Computation, 2004.
  3. Z. Xu, Z.Y. Dong, and W. Liu, “Neural network models for electricity market forecasting,” in D. Wang and N.K. Lee (ed): Neural Networks Applications in Information Technology and Web Engineering, Borneo Publications, 2005.
  4. Z.Y. Dong, “A neural network based framework for electricity load and price forecasting,” Int. Conf. Comput. Intell., Robot. and Auton. Syst., Singapore, pp. 88-93, Nov. 2001.
  5. B.L. Zhang, Z.Y. Dong, “An adaptive neural-wavelet model for short term load forecasting,” Int. J. Electr. Power Syst. Res., Vol. 59, No. 2, pp. 121-129, Sep. 2001.
  6. H.A. Salama, A.F.A. El-Gawad, H.M. Mahmoud, E.A. Mohamed, S.M. Saker, “Short-term load forecasting investigations of Egyptian electrical network using ANNs,” Int. Uni. Power Eng. Conf., Vol. 4-6, pp. 550-555, Sep. 2007.
  7. G.C. Liao, “A novel particle swarm optimization approach combined with fuzzy neural networks for short-term load forecasting,” IEEE Power Engineering Society General Meeting, Jun. 2007.
  8. Y. Xu, Z.Y. Dong, K. Meng, J.H. Zhao, and H.M. Yang “A short-term load forecasting model for Australian NEM,” Chin. Univ. Conf. Power Syst. Automat., Changsha, Oct. 2009.
  9. K. Liu, S. Subbarayan, R.R. Shoults, M.T. Manry, C. Kwan, F.I. Lewis, J. Naccarino, “Comparison of very short-term load forecasting techniques,” IEEE Trans. Power Syst., Vol. 11, No. 2, pp. 877-882, May 1996.
  10. K. Meng, Z.Y. Dong, and K.P. Wong, “Self-adaptive RBF neural network for short-term electricity price forecasting,” IET Gen Trans & Dist., Vol. 3, No. 4, pp 325-335,Apr. 2009.
  11. K. Meng, Z.Y. Dong, H.G. Wang, and Y.Y. Wang, “Comparisons of machine leaning methods for electricity reference price forecasting,” ISNN 2009: Int. Symp. Neural Netw., Wuhan, pp. 827-835, May 2009.
  12. X. Yin, Z.Y. Dong, and P. Zhang, “Fundamentals of emerging techniques,” in Z.Y. Dong, P. Zhang, et al. (edit): Emerging Techniques in Power System Analysis, Springer, pp. 23-44, Jan. 2010.
  13. T. Senjyu, H. Takara, K. Uezato, T. Funabashi, “Onehour-ahead load forecasting using neural networks,” IEEE Trans. Power Syst., Vol. 17, No. 1, pp. 113-118, Feb. 2002.
  14. C. Lursinsap, J.H. Kim, “Parallel learning for backpropagation network in binary field,” IEEE Int. Symp. Circuits Syst., Vol. 3, pp. 1477-1480, Jun. 1991.
  15. H. Schabauer, E. Schikuta, T. Weishaupl, “Solving very large traveling salesman problems by SOM parallelization on cluster architectures,” Int. Conf. Parallel and Distri. Comput., Appl. and Technol., pp. 954-958, Dec. 2005.
  16. I. Buck, “GPU computing: programming a massively parallel processor,” Int. Symp. Code Gen. and Optimiz., pp. 17, Mar. 2007.
  17. K. Levenberg, “A method for the solution of certain non-linear problems in least squares,” Q. J. Appl. Math., Vol. II, No. 2, pp. 164-168, 1944.
  18. NVIDIA CUDA Compute Unified Device Architecture, Programming Guide Version 2.1 Beta, Dec. 2008.
  19. J. Bolz, I. Farmer, E. Grinspun, and P. Schroder, “Sparse matrix solvers on the GPU: conjugate gradients and multigrid,” ACM Trans. Graph., Vol. 22, No. 3, pp. 917-924, Aug. 2003.
  20. I. Buck, T. Foley, D. Horn, J. Sugerman, K. Fatahalian, M. Houston, and P. Hanrahan, “Brook for GPUs: stream computing on graphics hardware,” ACM Trans. Graph., Vol. 23, No. 3, pp. 777-786, Aug. 2004.
  21. H.S. Hippert, C.E. Pedreira, R.C. Souza, “Neural networks for short-term load forecasting: a review and evaluation,” IEEE Trans. Power Syst., Vol. 16, No. 1, pp. 44-55, Feb. 2001.

Cited by

  1. Optimal Allocation of Energy Storage System for Risk Mitigation of DISCOs With High Renewable Penetrations vol.29, pp.1, 2014,