基于经验模态分解和优化BiLSTM的短期负荷预测

Short term Load Forecasting Based on Empirical Modal Decomposition and Optimized BiLSTM

针对电力负荷数据的非线性和不稳定性问题,提出了一种基于经验模态分解改进麻雀搜索算法双向长短期记忆神经网络相结合的EMD ISSA BiLSTM预测模型。首先采用EMD处理非线性负荷数据,将原始负荷数据分解为多个不同尺度的本征模态函数(IMF),引入反向学习策略和Levy飞行策略分别改进麻雀搜索算法(SSA)的收敛速度慢和容易陷入局部最优问题,利用改进麻雀搜索算法(ISSA)对BiLSTM神经网络进行参数寻优。然后再利用优化后的BiLSTM模型对每个分量进行预测,并将各预测结果叠加组合,得到整个负荷序列的预测结果。最后通过实际算例分析,证明该方法相对于传统的预测方法具有更好的预测精度和稳定性,可作为一种有效的短期负荷预测方法。

Aiming at the problem of nonlinearity and instability of load data,an EMD ISSA BiLSTM prediction model based on the combination of empirical modal decomposition improved sparrow search algorithm bidirectional long and short term memory neural network is proposed.Firstly,EMD is used to process the nonlinear load data,and the original load data are decomposed into several different scales of intrinsic modal functions(IMFs)and residuals(Res),and the inverse learning strategy and the Levy flight strategy are introduced to improve the convergence speed and local optimization problem of the Sparrow Search Algorithm(SSA),respectively,and the Improved Sparrow Search Algorithm(ISSA)is utilized to perform BiLSTM neural network parameter search optimization.Then the optimized BiLSTM model is used to predict each component,and the prediction results of each prediction are superimposed and combined to obtain the prediction results of the whole load sequence.Finally,through the analysis of actual cases,it is proved that this method has better prediction accuracy and stability than the traditional prediction methods,and can be used as an effective short term load prediction method.