摘要
电力负荷预测是进行电力能源供需精准化管理的重要手段.然而,许多传统的预测模型只关注预测的准确性,而忽略了预测的稳定性,导致实际应用中表现欠佳.因此,提出能克服传统模型缺陷的新模型是非常有必要的.本文以澳大利亚昆士兰州的电力负荷时间序列数据为例,提出了一种将先进的数据预处理策略、深度神经网络和多目标优化算法相结合的混合模型.其中,数据预处理策略用于对原始数据分解和去噪,深度神经网络用于数据学习和预测,而多目标优化算法用于深度神经网络初始参数的优化.实验结果表明,本文提出的模型具有较高的预测精度和预测稳定性,能较好地运用于电力负荷时间序列的短期预测.
Electricity load forecasting is an important mean of accurate management of electricity energy supply and demand.However,many traditional models only focus on the forecasting accuracy but ignore the forecasting stability,resulting in unsatisfying performance in practical applications.Therefore,it is very necessary to propose a new model to overcome the defects of the traditional models.Taking the time series data of electricity load in Queensland,Australia as an example,we proposes a hybrid model combining advanced data preprocessing strategy,deep neural network and multi-objective optimization algorithm.Among them,the data preprocessing strategy is used to decompose and denoise the original data,the deep neural network is used for data learning and forecasting,and the multi-objective optimization algorithm is used for the optimization of the initial parameters of the deep neural network.The experimental results show that the model proposed in this paper has high forecasting accuracy and stability,and can be applied to the short-term forecasting of electricity load time series well.
作者
倪凯来
柳向东
NI Kailai;LIU Xiangdong(School of Economics,Jinan University,Guangzhou 510632,China)
出处
《中国科技论文在线精品论文》
2021年第1期47-60,共14页
Highlights of Sciencepaper Online
关键词
电气工程
电力负荷预测
混合模型
互补集合经验模态分解
多目标优化
深度置信网络
electrical engineering
electricity load forecasting
hybrid model
complementary ensemble empirical mode decomposition
multi-objective optimization
deep belief network
作者简介
倪凯来(1998—),男,硕士研究生,主要研究方向:金融统计与证券分析、机器学习;通信联系人:柳向东,教授,主要研究方向:金融统计与证券分析、机器学习.E-mail:tliuxd@jnu.edu.cn
