摘要
针对电力负荷预测残差序列具有复杂的非线性状态的特点,运用神经网络模型,对电力负荷的灰色预测残差进行校正;同时对神经网络隐含层的神经元个数进行调整,使得网络模型结构优化,模型参数确定更为合理,进一步提高了预测精度。
To improve the accuracy of grey forecast model of power system, a new forecast method is suggested. According to the complexity and non- linearity of residual error series, the residual error of grey forecast is corrected using neural network model. The number of nerve cell in implicit layer of the neural network is adjusted. The network structure is improved.
出处
《华北电力大学学报(自然科学版)》
CAS
北大核心
2006年第4期93-95,共3页
Journal of North China Electric Power University:Natural Science Edition
基金
国家自然科学基金资助项目(50077007)
高等学校博士点专项基金项目(20040079008).
关键词
电力负荷
GM(1
1)模型
神经网络
残差
结构优化
power load
GM (1,1)model
neural network
residual erroor
structural optimization