基于粒子群算法优化参数的VMD-GRU短期电力负荷预测模型

VMD-GRU Short-term Power Load Forecasting Model Based on Optimized Parameters of Partical Swarm Algorithm

为了更有效提取电力负荷数据中的特征信息,从原始含噪声较多的负荷序列中提取包含丰富特征信息的信号分量,提高电力负荷预测精度。针对变分模态分解(variational mode decomposition,VMD)参数设定经验与主观性较强,提出一种基于粒子群算法(particle swarm algorithm,PSO)优化参数的变分模态分解和门控循环单元(gated recurrent unit,GRU)的组合模型短期电力负荷预测方法,先通过粒子群算法对VMD最佳影响参数组合进行搜寻,得到最佳效果的分解子序列,减少不同趋势信息对预测精度影响。然后运用GRU网络,针对各子序列分量建立基于GRU的预测模型。最后叠加各子序列预测结果得到短期电力负荷的最终预测值。实验结果表明,相对于相对于BP神经网络(Back Propagation Neural Network)、支持向量机(support vector machine,SVM)、GRU模型和EMD-GRU模型以及未经优化VMD-GRU模型,此模型具有更高的负荷预测精度。

In order to extract the feature information more effectively in the power load data,the signal components containing rich feature information are extracted from the original noisy load sequence to improve the accuracy of power load prediction.For the fact that,there are empirical and subjective parameter settings in variational mode decomposition(VMD),we proposed a short-term power load predication method of variational modal decomposition and gated recurrent unit(GRU)based on particle swarm algorithm(PSO)optimization parameters.In the prediction method,the particle swarm algorithm was used to search for the best influencing parameter combination of VMD to obtain the decomposed sub-sequence with the best effect and reduce the influence of different trend information on the prediction accuracy.Then we used the GRU network to establish a GRU-based prediction model for each sub-sequence component.Finally,the prediction results of each sub-sequence were superimposed to obtain the final prediction value of the short-term power load.The experimental results show that this model has a higher accuracy in load predication than the Back Propagation Neural Network(Back Propagation Neural Network),supports vector machine(SVM),GRU model,EMD-GRU model,and unoptimized VMD-GRU model.