星火链网络与随机森林算法在电力能源系统调度中的应用  

Application of Spark Chain Network and Random Forest Algorithm in Electric Energy System Dispatching

作　　者：李富鹏 杨明杰 王军 甄鑫 寇小霞 LI Fupeng;YANG Mingjie;WANG Jun;ZHEN Xin;KOU Xiaoxia(Digitalization Division,State Grid Gansu Electric Power Company,Lanzhou 730050,China;Zhangye Power Company of State Grid Gansu Electric Power Company,Zhangye 734000,China)

机构地区：[1]国网甘肃省电力公司,数字化事业部,甘肃兰州730050 [2]国网甘肃省电力公司张掖供电公司,甘肃张掖734000

出　　处：《微型电脑应用》2024年第9期111-113,122,共4页Microcomputer Applications

摘　　要：精准的电力系统负荷预测为能源调度提供强有力的数据支撑,减少执行时间,为此,研究了星火链网络与随机森林算法在电力能源系统调度中的应用。利用星火链网络以及随机森林算法分别预测电力系统短期负荷与长期负荷,以系统长、短期总负荷平衡、常规机组出力、线路潮流等为约束条件,构建电力新能源调度模型,基于蚁群算法求解调度模型,实现电力能源系统最佳调度。实验结果表明:该方法的短期负荷预测误差低于3.19%,长期负荷预测曲线与实际负荷曲线几乎全部吻合,误差不超过50 kW,可准确预测电力能源的短期和长期负荷;该方法调度执行时间低于38005.5 s,有效减少了执行调度任务的时间。In order to provide strong data support for energy dispatching through accurate power system load forecasting and reduce execution time,the application of spark chain network and random forest algorithm in electric energy system dispatching is studied.The spark chain network and random forest algorithm are used to predict the short-term load and long-term load of the power system,respectively.Taking the long-term and short-term total load balance of the system,the output of conventional units and line power flow as constraints,a electric power new energy scheduling model is constructed,and the demodulation degree model is calculated based on ant colony algorithm to realize the optimal scheduling of the power energy system.The experimental results show that the short-term load forecasting error of this method is less than 3.19%,the long-term load forecasting curve is almost consistent with the actual load curve,and the error is not more than 50 kW,which can accurately predict the short-term and long-term load of power energy.The scheduling execution time of this method is less than 38005.5 s,which effectively reduces the time of executing scheduling tasks.

关 键 词：电力能源 能源调度 随机森林 负荷预测 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]