基于改进型BP神经网络的电网负荷预测

摘要

Since the forecasting accuracy of the traditional linear load forecasting method cannot meet the requirements of the modern power grid management system,the nonlinear BP neural network algorithm suitable for the power grid load predic⁃tion task is used in this paper to build forecasting model. Because the conventional BP neural network is easy to fall into the lo⁃cal optimal solution and has low convergence efficiency,simulated annealing algorithm is used in this paper to optimize the BP neural network weight training algorithm to improve the convergence efficiency and self⁃learning ability of the prediction model. The prediction model studied in this paper is analyzed with an example. The results show that the training times and training time of the improved BP neural network are less than those of the conventional neural network,and it has higher convergence ac⁃curacy,in addition,the prediction error of the improved BP neural network prediction model is obviously reduced.%考虑到传统的线性电网负荷预测方法的预测精度无法满足现代电力电网管理系统的要求，使用更适用于电力电网负荷的预测任务的非线性BP神经网络算法建立预测模型。由于常规的BP神经网络存在容易陷入局部最优解以及收敛效率低等问题，该文使用模拟退火算法对BP神经网络权值训练算法进行优化，提高预测模型的收敛效率和自学习能力。通过实例对所研究的预测模型进行分析，结果表明，所研究的改进型BP神经网络的训练次数和训练耗时均低于常规神经网络，具有更高的收敛精度，同时改进型BP神经网络预测模型的预测误差明显降低，具有较好的工程应用价值。