Optimal Topology Optimization of Wireless Networks in Distribution System Substation

As an intelligent and adaptive control technology for distribution systems, distribution system substation play an important role in the stable and reliable operation of the entire power system. This paper proposes a wireless mesh network optimization algorithm based on master-slave multi-agent reinforcement learning to address the optimization problem of delay and stability in multi-hop wireless network transmission links in distribution system substation. Firstly, the problems of wireless mesh networks in distribution system substation are introduced, and a Markov decision process model of multi-agent reinforcement learning is established. Then, a master-slave multi-agent reinforcement learning optimization structure is designed with network stability and delay as rewards. Finally, the master-slave multi-agent reinforcement learning optimization algorithm is used to obtain the optimal transmission power of each wireless mesh node, and then the optimal wireless mesh network topology is obtained. Through experiments on the wireless mesh network in distribution system substation, the average network delay is optimized from around 2.5 seconds to around 1.8 seconds, which is about a 28% improvement compared to the original average delay. The experimental results show that the proposed algorithm effectively reduces transmission delay while ensuring transmission stability.