麻雀搜索算法优化的RF-BILSTM短期电力负荷预测

Short term load forecasting of RF-BILSTM optimized by SparrowSearch Algorithm

电力负荷的准确预测是电网安全运行以及社会正常生产的重要保障,但负荷数据因自身的非线性以及众多影响因素的不确定性大大降低预测精度的准确性,因此,为了提高短期电力负荷的预测精度,提出一种基于麻雀搜索算法(SSA)优化随机森林(RF)-双向长短期记忆神经网络(BILSTM)的预测模型。首先,针对电力负荷数据特征构造等问题,对数据进行量化以及标准化预处理操作,便于后续模型的输入。其次,利用RF算法对电力负荷的众多影响因素进行重要性排序,保留重要因素并将其与历史负荷数据进一步结合,从而构成神经网络的最终输入。最后,采用SSA算法对BILSTM模型的部分超参数进行优化选取,解决人工选取困难的问题。通过与其他模型对比,验证了该模型具有较高的预测精度。

Accurate forecasting of power load is an important guarantee for the safe operation of power grid and normal production of society.However,the accuracy of forecasting accuracy is greatly reduced due to the nonlinearity of load data and the uncertainty of many influencing factors.Therefore,in order to improve the accuracy of short-term power load forecasting,a forecasting model based on Sparrow Search Algorithm(SSA)optimized Random Forest(RF)-Bidirectional Long Short-term Memory Neural Network(BILSTM)is proposed.Firstly,in response to issues such as the construction of power load data features,the data is quantified and standardized for preprocessing operations to facilitate subsequent model input.Secondly,the RF algorithm is used to rank the importance of numerous influencing factors of power load,retain important factors,and further combine them with historical load data to form the final input of the neural network.Finally,SSA algorithm is used to optimize the selection of some hyperparameter of BILSTM model,which solves the problem of manual selection.By comparing with other models,it is verified that this model has high prediction accuracy.