Scholarlink

As urban power grids grow in scale and complexity, rational deployment of sensor networks is critical to ensure operational security and stability. This study proposes a Deep Reinforcement Learning (DRL) framework for sensor network deployment optimization, integrating cluster-based spatial priority scoring. A case study involving 352 transmission towers in Tianhe District, Guangzhou, is presented. The framework incorporates commercial mathematical programming (Gurobi), Genetic Algorithm, and DRL approaches to systematically compare performance under varying deployment budgets. Experimental results demonstrate that the DRL method achieves high coverage and redundancy with superior computational efficiency and adaptability, effectively addressing complex sensor placement challenges in urban environments. This study offers an efficient and scalable decision-support tool for smart grid maintenance with significant theoretical and practical implications.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Deep Reinforcement Learning Framework for Sensor Network Deployment in Urban Power Grid Monitoring Optimization Analytics

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Cheng Su,
  • Dachuan Xu,
  • Shijie Li,
  • Hao Wang,
  • Xiaohan Jiang,
  • Chang Liu,
  • Shaohua Wang

摘要

As urban power grids grow in scale and complexity, rational deployment of sensor networks is critical to ensure operational security and stability. This study proposes a Deep Reinforcement Learning (DRL) framework for sensor network deployment optimization, integrating cluster-based spatial priority scoring. A case study involving 352 transmission towers in Tianhe District, Guangzhou, is presented. The framework incorporates commercial mathematical programming (Gurobi), Genetic Algorithm, and DRL approaches to systematically compare performance under varying deployment budgets. Experimental results demonstrate that the DRL method achieves high coverage and redundancy with superior computational efficiency and adaptability, effectively addressing complex sensor placement challenges in urban environments. This study offers an efficient and scalable decision-support tool for smart grid maintenance with significant theoretical and practical implications.