摘要
由于中长期电力系统负荷数据相对较少,影响因素广泛,常用的负荷预测方法很难达到满意的精度,所以提出一种结合小波软阈值和广义回归神经网络的综合预测模型,该模型充分考虑了长期负荷数据相对较少和影响因素广泛的特点,把负荷预测看作一个信号序列,应用小波去噪原理,将信号与噪声分离,然后利用广义回归神经网络进行预测。实际应用证明,预测结果是令人满意的。
Mid-and long-term load data of power system is comparatively inadequate and influenced by many factors. Hence it is difficult to achieve satisfactory precision by use of common load forecast methods. This paper presents a comprehensive forecast model integrating wavelet soft-threshold with general regression neural network (GRNN). With full consideration of the characteristics of long-term load data, the load forecast is viewed as a signal sequence and the signal is denoised with wavelet denoising theory. Then GRNN is used to forecast the load. Practical applications show that the forecast precision is satisfying.
出处
《广东电力》
2007年第12期5-8,29,共5页
Guangdong Electric Power
关键词
广义回归神经网络
中长期负荷预测
小波软阈值
小波去噪原理
general regression neural network CGRNN)
mid- and long-term load forecast
wavelet soft-threshold
waveletdenoising theory