摘要
由于钢铁企业电力负荷自身规律复杂,影响因素众多且大多呈现非线性关系,适用于处理非线性数据的智能预测方法成为近年来研究的热点。本文结合以往传统的预测方法,对于钢铁企业目前应用的短期电力负荷预测方法进行了综述。首先,将钢铁企业多样化的电力负荷类型进行归纳,分析了其典型电力负荷类型的相关特性以及影响因素;其次,简述了目前各种预测方法的应用原理及优劣,并以此为依据将目前的预测方法分为单一模型和组合模型两类,并在此基础上进行进一步的细分,同时绐出了几种常见的电力负荷预测评价指标;最后,对于目前钢铁企业电力负荷预测亟需解决的问题和未来可能的发展方向进行了探讨。
Because the law of power load in iron and steel enterprises is complex,there are many influencing factors and most of them show nonlinear relatio nship,the predicti on means apply to address non linear data has become a research hot pints in recent years.Unified with the conventional forecasting methods,power load which owns short period forecasting methods.Combined with the traditional forecasting methods,the short-term power load forecasting methods currently used in iron and steel enterprises are summarized.Firstly,the diversified power load types of iron and steel enterprises are summarized,and the related characteristics and influencing factors of their typical power load types are an a lyzed;Secondly,the brief in troducti on con cer ns the principle and the merits and demerits of the various forecasting methods.On the basis of this,the current forecasting methods are divided into two categories:single model and combined model.Further subdivision is made on the basis of this.Mean while,several comm on in dicators for evaluating the load of power are give n.Fin ally,the problems that need to be solved urge ntly and the possible development directi on in the future are discussed.
作者
祁宇轩
吴定会
Qi Yuxuan;Wu Dinghui
出处
《变频器世界》
2022年第3期49-53,共5页
The World of Inverters
基金
国家高技术研究发展计划资助项目(2020YFB17111002020YFB1711102)。
关键词
电力系统
深度学习
短期负荷预测
Power system
Deep learning
Short term load forecasting