【摘要】 提出一种基于HHT的配电网短期负荷组合预测方法。该方法利用EMD分解将负荷信号分解为一组IMF分量,通过分析IMF分量对应的边际谱提取周期项、随机项及趋势项。对各个特征量分别采用CPN、SVM及二次指数平滑法进行预测,最后通过直接叠加各分量预测值得到最终预测结果。通过仿真算例,验证该组合模型的精度高于任何一种单一模型和传统线性组合模型,具有较高的预测精度。更多还原

【Abstract】 A distribution network short-term load forecasting model based on HHT is proposed.Through the HHT method of EMD model,the non-stationary and nonlinear distribution short-term load is decomposed into stationary IMF components.Through the analysis of each component corresponding to the Hilbert marginal spectrum of HHT,the periodic,the random and the trend are extracted from original signal,and each component is predicted separately by using CPN,SVM and the secondary exponential smoothing method,the final forecasting result can be obtained by adding each forecasting result.The combined forecasting method in this paper has higher prediction precision,which is confirmed by comparing the results of other methods and each single forecasting method.更多还原