Short-Term Power Load Forecasting Based on Improved Wavelet Neural Network

Abstract:

Accurate short-term load forecasting contributes to safe and economic operation of power systems. Due to the shortcomings of traditional wavelet neural network (WNN), which usually has low convergence rate and easily falls into local minimum, an improved wavelet neural network (IWNN) is proposed to modify the algorithm by introducing momentum. Together with the weighted average method (WA) and WNN, these three methods are applied to an example of short-term load forecasting. The results show that compared with the WA method, WNN has obvious advantages of nonlinear fitting and forecasting, and the IWNN method is superior to the others in terms of prediction accuracy and generalization capability, which is helpful to further improve the accuracy of short-term load forecasting.

• Periodical: Advanced Materials Research (Volumes 816-817)
• Pages: 766-769
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.816-817.766
• Online since: September 2013
• Authors: Dong Xiao Niu, Lei Lei Fan, Chun Xiang Liu
• Keywords: Momentum, Short-Term Load Forecasting, Wavelet Neural Network

References

[1] Zhen Zu, Guihong Bi, Li Liu, et al. Research on power system short-term load forecast model based on wavelet neural network[J]. Computer Technology and Development, 2012, 22(10): 237-241.

[2] Zhang Q H, Benveniste A. Wavelet networks[J]. IEEE Transactions on Neural Networks, 1992, 3(11): 889-898.

DOI: 10.1109/72.165591

[3] Daubechies I. Ten Lectures on Wavelets: CBMS-NSF Region Conference Lecture Notes[M]. Philadelphia, PA: Society for Industrial and Applied Mathematics(SIAM), 1992: 87.

[4] Turgay Partal, Kerem Cigizoglu H. Estimation and forecasting of daily suspended sediment data using wavelet neural networks[J]. Journal of Hydrology, 2008, 358: 317-331.

DOI: 10.1016/j.jhydrol.2008.06.013

[5] Chauhan N, Ravi V, Karthik Chandra D. Differential evolution trained wavelet neural networks: application to bankruptcy prediction in banks[J]. Expert Systems with Applications, 2009, 36: 7659-7665.

DOI: 10.1016/j.eswa.2008.09.019

[6] Weigen Chen, Chong Pan, Yuxin Yun, et al. Fault diagnostic method for power transformer based on improved wavelet neural network algorithm[J]. Chinese Journal of Scientific Instrument, 2007, 27(7): 1489-1493.

[7] Xuemin Tian, Qiang Wang, Xiaogang Deng. Soft sensing based on wavelet neural networks with momentum[J]. CIESC Journal, 2011, 62(8): 2238-2242.