FTTB场景下基于强化学习的智慧保供电决策方法  

Intelligent power supply decision method based on reinforcement learning in FTTB scene

作　　者：王愚 李佩 李江伟 符华 聂雷刚 WANG Yu;LI Pei;LI Jiang-wei;FU Hua;NIE Lei-gang(Nanning Power Supply Bureau of Guangxi Power Grid Co.,Ltd.,Nanning 530022,China)

机构地区：[1]广西电网有限责任公司南宁供电局,南宁530022

出　　处：《信息技术》2022年第10期147-152,共6页Information Technology

摘　　要：为了提高智慧保供电决策能力,在FTTB场景下提出基于强化学习的智慧保供电决策方法。采用4G/光纤通信和中压载波网络,构建智慧保供电决策系统结构模型;利用车载采集终端与电源车的机组控制器采集智慧保供电参数,实现输出控制;通过光时域反射和光纤Raman散射技术,建立发电机组数据及环境数据分析模型;结合强化学习算法,利用智慧保供电决策函数,实现智慧保供电决策和优化调度。测试结果表明,设计的方法寻优能力较强,稳定性和自适应性较好。In order to improve the decision-making ability of intelligent power supply, a decision-making method of intelligent power supply based on reinforcement learning is proposed in FTTB scene. 4 G/optical fiber communication and medium voltage carrier network are used to build the structural model of intelligent power supply decision-making system. The on-board acquisition terminal and the unit controller of the power vehicle are used to collect intelligent power supply parameters to realize the output control. Through optical time domain reflection and optical fiber Raman scattering technology, the analysis model of generator set data and environmental data are established. Combined with reinforcement learning algorithm, intelligent power supply decision-making function is used to realize intelligent power supply decision-making and optimal dispatching. The test results show that this method has strong optimization ability, and good stability and adaptability.

关 键 词：FTTB场景 强化学习算法 优化调度 决策函数 智慧保供电决策 

分 类 号：TM73[电气工程—电力系统及自动化]