摘要
小波变换在时域和频域都具有良好的局部化性质,在电力系统负荷预测中得到了广泛应用。小波变换的实质是卷积运算,在负荷预测过程中存在边界效应,降低了预测的精度。本文采用阈值处理和差分补偿数据延拓方法对原始数据进行处理,然后采用Mallat分解算法对处理后的负荷序列进行分解,针对电力负荷为随机序列的特点,利用时间序列法的随机模型对小波子序列分别进行预测,最后采用Mallat重构算法对预测结果进行重构,提出了一种基于Mallat算法的负荷预测实用方法。算例结果表明该方法有效地减小了边界效应对预测结果的影响,针对具有随机序列特点的电力负荷的预测具有良好的计算精度。
Wavelet transform has good local characteristics in time domain and frequency domain. It has been widely used in power load forecasting. Wavelet transform is convolution operation essentially and its boundary effect in decomposition and reconstruction process reduces the accuracy. In this paper, the original data is processed by difference compensation data extension and threshold value method. Then the processed data is decomposed by Mallat decomposition algorithm. Considering the power load is stochastic sequence, the time series stochastic model is chosen to forecast every wavelet sequence. Finally the forecasting results is reconstructed by Mallat reconstruction algorithm to compose daily load forecasting data. The example shows that the method in this paper can reduce boundary effect clearly and has good accuracy.
出处
《电力系统保护与控制》
EI
CSCD
北大核心
2009年第20期89-92,共4页
Power System Protection and Control
