基于改进广义回归神经网络的日负荷曲线预测  

Daily Load Curve Forecasting Based on Revised General Regression

作　　者：边居政 

机构地区：[1]山东陵县供电公司,山东陵县253500

出　　处：《山东电力技术》2013年第6期35-38,共4页Shandong Electric Power

摘　　要：提出了基于改进广义回归神经网络(GRNN)的日负荷曲线预测模型。对GRNN模型的输入元素进行分析筛选,并同时输出一天48点的负荷预测值,在保证预测精度的基础上大大提高了预测速度。采用烟台某变电站实际负荷进行预测分析,结果证明了该方法的有效性。A new short-term load forecasting method based on revised general regression neural network is proposed.The input elements are analyzed and 48 load forecasting results can be got from the model at the same time.This method can improve the speed of the forecast and keep high accuracy.The load from a real station in Yantai is taken as an example for the load forecasting method proposed in the paper,and the results prove the effectiveness of the proposed method.

关 键 词：日负荷曲线预测 广义回归神经网络 平滑参数 

分 类 号：TM715[电气工程—电力系统及自动化]