基于相空间重构理论的电力负荷预测

研究电力负荷预测问题。针对电力负荷既受非线性变量的影响又受自身混沌性质的影响,单一预测方法无法同时完善地映射其复杂关系,导致预测精度较低。为进一步提高电力负荷预测准确度,融合相空间重构理论与支持向量机,提出了一种人工智能的新电力负荷预测方法(PSR-SVM)。首先以对数线性平稳法对初始电力负荷数据进行去趋势处理。然后基于粒子群算法对相空间重构参数与支持向量机参数进行同步优化,以最优延迟时间τ和嵌入维m进行电力负荷序列相空间重构,并以支持向量机及其最优参数构建非线性电力负荷预测模型。最后以某电力公司1978～1998的电力负荷进行仿真,结果表明,新模型预测精度明显高于参比模型,是一种高精度、可行的电力负荷预测方法。

Power Load Forecasting Based on Phase Space Reconstruction

Research power load forecasting problems.The power load is affected by nonlinear as well as chaotic features,and simple methods can not map the complex relationships perfectly.Based on phase space reconstruction theory and support vector machines,a new high-precision power load forecasting method PSR-SVM was proposed in this paper,to improve the accuracy of power load forecasting.Firstly,initial data was smoothly processed.Secondly,the parameters of τ,m,c,σ were optimized by particle swarm optimization.Lastly,the model was built.The simulation results of power load data from 1978 to1998 in Fujian province show that PSR-SVM has the highest prediction accuracy of all reference models,therefore,is a high precision and practical power load forecasting method.