基于进化神经网络的短期电网负荷预测算法  

Algorithm for Predicting Short-Term Load Based on Evolutionary Neural Network

作　　者：刘喜生 LIU Xisheng(Shantou Huadian Power Generation Co.,Ltd.,Shantou 515000,Guangdong,China)

机构地区：[1]汕头华电发电有限公司,广东汕头515000

出　　处：《电力大数据》2023年第4期11-18,共8页Power Systems and Big Data

摘　　要：本文提出了一种基于进化神经网络的短期电网负荷预测算法。该算法使用了改进的人工蜂群算法与BP神经网络融合生成的进化神经网络,并且使用改进的人工蜂群算法对进化神经网络的偏置和权重进行优化。该算法将火电历史负荷数据作为输入,使用进化神经网络训练预测模型,预测未来一段时间内的电网负荷。首先获取历史负荷数据,然后将收集到的数据应用于进化神经网络模型训练。人工蜂群算法作为一种全局搜索算法,可以有效地探索模型参数空间,寻找最佳的模型参数组合以提升预测精度。为了验证所提出的负荷预测方法的有效性,本文使用了火电网负荷数据进行测试。实验结果表明,在短期电网负荷预测方面,本文提出的进化神经网络比传统方法预测结果更加准确可靠。This paper proposes a short-term electricity grid load forecasting algorithm based on an evolutionary neural network.The algorithm combines an improved artificial bee colony algorithm with a backpropagation neural network to generate the evolutionary neural network.Additionally,an improved artificial bee colony algorithm is employed to optimize the biases and weights of the evolutionary neural network.The algorithm takes historical load data from thermal power plants as input and uses the evolutionary neural network to train a prediction model for forecasting the grid load in the future.Firstly,historical load data is obtained and then applied to train the evolutionary neural network model.The artificial bee colony algorithm,as a global search algorithm,effectively explores the parameter space of the model to find the optimal combination of model parameters,thereby improving the prediction accuracy.To validate the effectiveness of the proposed load forecasting approach,the algorithm is tested using load data from a thermal power grid.Experimental results demonstrate that the proposed evolutionary neural network outperforms traditional methods in short-term electricity grid load forecasting,providing more accurate and reliable predictions.

关 键 词：进化神经网络 短期负荷预测 人工蜂群算法 参数优化 

分 类 号：TM715[电气工程—电力系统及自动化]