摘要
为了提高配电网供电机组负荷频率越限预警的精准度,维护配电网供电机组稳定运行,提出基于调控云实时数据平台的配电网供电机组负荷频率越限预警方法。设计调控云实时数据平台架构,充分采集供电机组负荷实时感知数据;在调控云实时数据平台的数据处理层对数据进行压缩抽取、转换重构和汇总融合处理;基于长短期记忆(LSTM)模型结合升压负荷精确度等约束条件构建配电网负荷频率预测模型;利用支持向量机将越限预警问题转换为惩罚函数,建立预警目标模型,并利用模拟退火算法改进支持向量机,实现配电网供电机组负荷频率。实验结果表明,设计的预警方法预警偏差较低,预警精准度较高,且针对连续循环越限的预警完整度较高,预警时延较短。
In order to improve the accuracy of the over-limit warning of the load frequency of the power supply units of the distribution network,and maintain the stable operation of the distribution network power supply units,an over-limit warning method of the load frequency of the power supply units of the distribution network based on the control cloud real-time data platform is proposed.The control cloud real-time data platform architecture is designed to fully collect real-time sensing data of power supply units load,in the data processing layer of the control cloud real-time data platform,data is compressed,extracted,transformed,reconstructed,and summarized for fusion processing.Constructing a distribution network load frequency prediction model based on long short term memory(LSTM)model combined with constraints such as boost load accuracy.Using support vector machine to transform the over-limit warning problem into a penalty function,establishing a warning target model,and improving the support vector machine using simulated annealing algorithm to achieve load frequency of power supply units in the distribution network.The experimental results show that the designed warning method has low deviation,high accuracy,high completeness and short delay for the warning of continuous cycle over-limit.
作者
张小龙
郑立国
肖小苏
张自润
殷寿银
ZHANG Xiaolong;ZHENG Liguo;XIAO Xiaosu;ZHANG Zirun;YIN Shouyin(State Grid Qinghai Electric Power Company Haixi Power Supply Company,Geermu 816099,China)
出处
《微型电脑应用》
2025年第12期265-268,272,共5页
Microcomputer Applications
关键词
供电机组
云数据平台
负荷频率
越限预警
power supply units
cloud data platform
load frequency
over-limit warning
