基于贝叶斯神经网络短期负荷预测模型

中国管理科学 ›› 2012, Vol. ›› Issue (4): 118-124.

基于贝叶斯神经网络短期负荷预测模型

史会峰¹, 牛东晓², 卢艳霞¹

1. 华北电力大学数理学院, 河北保定 071003;
2. 华北电力大学工商管理学院, 北京 102206

收稿日期:2011-03-16 修回日期:2012-03-11 出版日期:2012-08-29 发布日期:2012-08-29
基金资助:
国家自然科学基金资助项目(70671039)

The Short-Term Load Forecasting Model Based on Bayesian Neural Network

SHI Hui-feng¹, NIU Dong-xiao², Lu Yan-xia¹

1. School of Mathematics and Physics, North China Electric Power University, Baoding 071003, China;
2. School of Business Administration, North China Electric Power University, Beijing 102206, China

Received:2011-03-16 Revised:2012-03-11 Online:2012-08-29 Published:2012-08-29

摘要/Abstract

摘要： 本文提出了基于贝叶斯神经网络(BNN)短期负荷预测模型。根据气象影响因素和电力负荷的样本数据,针对权向量参数的先验分布分别为正态分布和柯西分布两种情况,应用混合蒙特卡洛(HMC)算法学习了BNN的权向量参数。由HMC算法和Laplace算法学习的贝叶斯神经网络以及BP算法学习的传统神经网络分别对4月 (春)、8月 (夏)、10月 (秋)和1月(冬)每月25天的每个整点时刻的负荷进行了预测。这些神经网络的输入层有11个节点,它们分别与每个整点时刻和的气象因素、上一个整点时刻的气象因素和时间变量相对应,输出层只有一个节点,它与负荷变量对应。试验结果表明HMC算法学习的BNN的预测结果的百分比平均绝对误差( MAPE)和平方根平均误差( RSME )取值远远小于由Laplace 算法学习的BNN和BP算法学习的人工神经网络的 MAPE和RMSE。而且,HMC算法学习的BNN在测试集和训练集上的预测误差MAPE和RMSE的相差很小。实验结果充分说明HMC算法学习的BNN具有较高的预测精度和较强的泛化能力。

关键词: 贝叶斯神经网络, 短期负荷预测, 蒙特卡洛算法, 先验概率分布, 汉密尔顿动力系统

Abstract: A short term load forecasting model based on Bayesian neural network learned by the Hybrid Monte Carlo (HMC) algorithm is presented in this paper. The weight vector parameter of the Bayesian neural network is considered as multi-dimensional random variables. Using the weather factors and load recorders in training set, HMC algorithm is used to learn the weight vector parameter with respect to normal prior distribution and Cauchy prior distribution respectively. Two Bayesian neural networks learned by Laplace algorithm and HMC algorithm and the artificial neural network learned by the BP algorithm are used to forecast the hourly load of 25 days of April(spring), August(summer), October(autumn) and January(winter) respectively. There are eleven nodes in input layer, ten nodes representing the ten weather factor variables of current hour and the previous hour and one hour variable. There is one node in output layer, corresponding to the load on each hour. The experimental result shows that the roots mean squared error (RMSE) and the mean absolute percent errors (MAPE) of the Bayesian neural network learned by hybrid Monte Carlo algorithm both are much smaller than those of the neural networks learned by Laplace algorithm and BP algorithm. Hence, the forecasting model based on BNN learned by the HMC algorithm has higher forecasting precision, and can be used to short-term load forecasting.

Key words: Bayesian neural network, short term load forecasting, Monte Carlo algorithm, prior probability distribution, Hamiltonian dynamical system

中图分类号:

TM74

史会峰, 牛东晓, 卢艳霞. 基于贝叶斯神经网络短期负荷预测模型[J]. 中国管理科学, 2012, (4): 118-124.

SHI Hui-feng, NIU Dong-xiao, Lu Yan-xia. The Short-Term Load Forecasting Model Based on Bayesian Neural Network[J]. Chinese Journal of Management Science, 2012, (4): 118-124.

参考文献

[1] Pitt B. Applications of data mining techniques to electric load profiling. Manchester: University of Man chester Institute of Science and Technology, 2000.
[2] Douglas A P, Breipohl Arthur M, Lee F N,et al. The impacts of temperature forecast uncertainty on bayesian load forecasting [J]. IEEE Transactions on Power System, 1998, 13(4): 1507-1513.
[3] Buizza T R. Neural network load forecasting with weather ensemble predictions [J]. IEEE Transactions on Power Systems, 2002, 17(2): 626-632.
[4] 秦海潮, 王玮,周晖,等. 人体舒适度指数在负荷预测中的应用 [J]. 电力系统及其自动化学报, 2006, 18(2): 63-66.
[5] Peng T M, Hubele N F, Karagy G G. An adaptive neural network approach to one-week ahead Load forecasting [J]. IEEE Transaction. Power Systems, 1993, 8(3): 1195-1203.
[6] EI-Sharkawi M A, Marks R J, II Atlas L E, et al. Electric load forecasting using an artificial neural network [J]. IEEE Transactions on Power System, 1991, 6(2): 442-449.
[7] 牛东晓, 赵磊, 张博, 王海峰. 粒子群优化灰色模型在负荷预测中的应用 [J]. 中国管理科学, 2007, (01): 69-73.
[8] 牛东晓, 乞建勋, 邢棉. 二重趋势性季节型电力负荷预测组合灰色神经网络模型 [J]. 中国管理科学, 2001,(06): 16-20.
[9] 刘旭, 罗滇生, 姚建刚,等. 基于负荷分解和实时气象因素的短期负荷预测[J]. 电网技术, 2009, 33(12): 94-100.
[10] 谢宏, 陈志业, 牛东晓,等. 基于小波分解与气象因素影响的电力系统日负荷预测模型研究 [J]. 中国电机工程学报, 2001, 21(5): 5-10.
[11] Hippert H S, Pedreira C E, Soza R C. Neural network for short-term load forecasting: A review and evaluation [J]. IEEE Transactions on Power System, 2001, 16(1): 44-55.
[12] Mandal P, Senjyu T, Funabashi T. Nerual networks approach to forecast several hour ahead electricity price and loads in deregulated market [J]. Energy Convers Manage, 2005, 47: 2128-2142.
[13] Buntine W L,Weigend A S. Bayesian back-propagation [J]. Complex Systems, 1991, 5(6): 603-643.
[14] MacKay D J C. A practical Bayesian framework for back-propagation networks [J]. Neural Computation, 1992, 4(3): 448-472.
[15] Lampinem J,Vehtari A. Bayesian approach for neural networks-review and case studies [J]. Neural Networks, 2001, 14: 257-274.
[16] Bishop C M. Pattern recognition and machine learning [M]. Singapore: Springer, 2006.
[17] Neal R M. Bayesian learning for neural networks [M]. New York:Springer Verlag, 1996.
[18] Liang Faming. Bayesian neural networks for nonlinear time series forecasting [J]. Statistics and Computing, 2005,15:13-29.
[19] Lauret P, Fock E, Randrianarivony R N, et al. Bayesian neural network approach to short time load forecasting [J]. Energy Conversion and Management, 2008, 49:1156-1166.
[20] Hastings W K. Monte Carlo sampling methods using Markov chains and their applications [J]. Biometrika, 1970, 57: 97-109.