基于PSO-OMP优化的WD-ASD超短期负荷预测

为提高负荷预测精度,降低电力系统规划决策的保守性,本文提出了一种基于小波-原子稀疏分解(WD-ASD)的超短期负荷预测模型。该模型使用模糊聚类算法提取相似日为历史数据,采用小波分解(WD)作为前置环节,以基于原子表达式的自预测和基于最小二乘支持向量机(LSSVM)的残余分量预测为基础构建原子稀疏分解(ASD)预测模型,分别对负荷的高低频分量进行预测,并将结果相加得到最终预测值。其中ASD分解过程由正弦原子库自适应匹配分解完成,并将粒子群算法(PSO)和正交匹配追踪(OMP)算法相结合以增强原子稀疏分解能力。实际负荷数据算例验证了所提方法的自适应性、快速性及有效性。

Short-term load forecasting based on WD-ASD optimized by PSO-OMP

In order to improve the accuracy of load forecasting and reduce the conservatism in power system plan-ning decisions, this paper proposes a short-term load forecasting model based on the combination of wavelet decom-position and atomic sparse decomposition( WD-ASD) . This model uses fuzzy clustering algorithm to extract the data of a similar day as the historical data. Wavelet decomposition ( WD) is used as the prepositive step. Atomic sparse decomposition prediction model is set up to predict the high frequency and low frequency components of load based on the adaptive prediction which uses atomic expression and the residual component prediction which uses least squares support vector machine (LSSVM). Then the two results are added to get the final predictor. The ASD process is obtained by adaptive matching decomposition from the sine atomic library. In addition, particle swarm optimization ( PSO) and orthogonal matching pursuit ( OMP) algorithm are combined as PSO-OMP to optimize the atomic sparse decomposition. This method enhances the decomposition ability. The actual load example shows the adaptivity, rapidity and validity of this method.