摘要
利用灰色理论和神经网络技术,提出了基于灰色关联度分析的灰色神经网络短期负荷预测短期分析新方法并建立了灰色关联—神经网络模型应用到电力短期负荷预测分析中,由于该模型优化了输入层因子,并克服了BP算法确定隐含层节点数的困难,所以提高了学习效率。最后结合某市电网负荷特点,在输入因子中重点考虑了某些扰动因子来进行优化,最后的预测结果以及与其他方法的预测误差进行比较,表明了该预测方法的正确性、高效性和实用性。
A new gray-connection neural-network load forecast model is set up and employed in short-term load forecast based on grey theory and neural network technology. Because this paper derives excellent turn of input factors of the model, and at the same time the model overcomes the difficulty of how to certain factors of middle layer in the way of BP calculation, the model＇s study efficiency is improved greatly. At last, the paper takes some city＇s load as an example to apply the model to forecast. In order to optimize the model , the paper uses three diffent methods to forecast the city＇s load respectively. With the comparison of forecast error margin of three methods ,the model is proved to be correct, high-efficient and practical.
出处
《继电器》
CSCD
北大核心
2005年第19期36-40,共5页
Relay
关键词
灰色关联
神经网络
模型
电力负荷
短期预测
gray-connection
neural-network
model
power load
short-term forecast
