一种改进极限学习机的电力负荷预测方法

A Power Load Forecasting Method Based on Improved Extreme Learning Machine

为了提高电力负荷预测的准确度,需要研究影响电力负荷预测的各个特征变量,笔者提出一种改进极限学习机的电力负荷预测方法。首先,使用箱形图找出原始电力负荷数据的异常值并剔除;然后,使用探索性数据分析和皮尔逊相关性分析确定电力负荷与环境温度、大气压力、相对湿度和排气蒸汽压强等变量的相关性,并确定预测模型的特征输入;最后,使用蜻蜓算法优化极限学习机中的输入层和隐藏层的连接权值、隐层单元的偏置值。实验结果表明,该预测模型的决定系数可达到0.93,预测效果理想,给电力负荷预测提供了借鉴思路。

In order to improve the accuracy of power load forecasting,it is necessary to study the various characteristic variables that affect the power load forecasting.The author proposes an improved power load forecasting method based on extreme learning machine.Firstly,the box diagram is used to find out the abnormal values of the original power load data and eliminate them.Then,exploratory data analysis and Pearson correlation analysis are used to determine the correlation between power load and environmental temperature,atmospheric pressure,relative humidity and exhaust steam pressure,and to determine the characteristic input of the prediction model.Finally,the dragonfly algorithm is used to optimize the connection weights of the input layer and the hidden layer in the extreme learning machine and the offset value of the hidden layer unit.The experimental results show that the coefficient of determination of the prediction model can reach 0.93,and the prediction effect is ideal,which provides a reference for power load forecasting.