摘要
电力系统中长期电力负荷数据相对较少,影响因素多且很难预测,其主要由随机性因素和确定性因素决定。根据这些信息中隐藏频率的不同,可利用小波变换原理将其分解到若干频率段上,即得到若干"近似"项和"细节"项,在各项上附加不同的阈值后,应用小波原理对经过阈值处理的各项进行重构,从而达到对原始数据降噪的目的,将去噪后的长期负荷数据作为神经网络的输入进行预测。算例表明,该方法预测结果准确、可靠。
In power system, Mid-term and long-term load data is comparatively absence, is determined by many factors which are mainly decided by determinate and stochastic factors, and so can be viewed the function of time and frequency. According to the frequency difference hiding in the information, the load frequency series can be decomposed into some approximations and details by wavelet analysis. Before reconstruction of the wavelet method, some thresholds were added on the every details, so as to decrease the noise of the origin load data. The denoising long-term data is used as the input of BPNN (back propagation neural network ) for forcast, and it is proved that the forecast precision is exact and reliable.
出处
《电力科学与工程》
2008年第10期18-20,共3页
Electric Power Science and Engineering
关键词
神经网络
电力负荷预测
小波软阈值
去噪
BPNN(back propagation neural network)
electric power load forecast
wavelet soft-threshold
denoising