摘要
电力系统中存在的随机负荷波动现象影响了电力系统的功率稳定性,导致电力系统的利用率和运行效率较低。为此,提出基于改进深度强化学习的电力安全优化调度模型。考虑需求响应条件,建立安全调度目标函数。在保证内部设备功率平衡的基础上,建立分支线路潮流约束和启停约束。设置机组爬坡约束以保证稳定的能效。使用改进深度强化学习,训练历史数据,以完成调度决策和关系映射,并构建负荷预测数据。根据时间序列关系减少调度周期时长,使调度模型能够在更短的时间内完成调度过程,以实现电力系统安全优化调度。试验结果表明,经模型调度分配后,各机组设备的功率变化稳定、能耗较小,对机组功率的上、下限约束宽松。调度后电力系统的利用率和运行效率均较好,灵活性较高。
The phenomenon of stochastic load fluctuation existing in the power system affects the power stability of the power system,resulting in poor utilization and operational efficiency of the power system.For this reason,a power security optimization scheduling model based on improved deep reinforcement learning is proposed.Considering the demand response conditions,the security scheduling objective function is established.Based on ensuring the power balance of internal equipment,branch line current constraints and start-stop constraints are established.Unit climbing constraints are set to ensure stable efficiency.Improved deep reinforcement learning is used to train historical data,so as to complete scheduling decisions and relationship mapping,and construct load forecast data.The length of the scheduling cycle is reduced according to the time series relationship,so that the scheduling model can complete the scheduling process in a shorter period,in order to achieve the power system security and optimization of scheduling.The experimental results show that after the model scheduling allocation,the power change of each unit equipment is stable,the energy consumption is smaller,and the upper and lower constraints on the power of the unit are loose.The utilization rate and operation efficiency of the power system after dispatching are better,and the flexibility is high.
作者
陈渊博
周翔宇
荣发权
范东亮
姚兰波
CHEN Yuanbo;ZHOU Xiangyu;RONG Faquan;FAN Dongliang;YAO Lanbo(Construction Branch,State Grid Anhui Electric Power Co.,Ltd.,Hefei 230000,China;State Grid Anhui Electric Power Co.,Ltd.,Hefei 230000,China)
出处
《自动化仪表》
2025年第7期87-91,共5页
Process Automation Instrumentation
关键词
强化学习
优化调度
需求响应
安全调度
电力安全
潮流约束
启停约束
Reinforcement learning
Optimal scheduling
Demand response
Secure scheduling
Power security
Current constraints
Start-stop constraints
