As energy infrastructure operators around the world increasingly turn to automation to boost safety and efficiency, a leading Chinese power utility based in southern China’s Guangdong province is scaling up research and deployment of intelligent robots designed to handle high-stakes grid operations. Guangdong Power Grid Co., a regional subsidiary of China Southern Power Grid Co., has been actively developing and training a fleet of specialized robotic systems to take on core tasks across its vast power network, with some solutions already active in daily operations.
The company’s in-house power robotics laboratory centers its research on cutting-edge capabilities that allow robots to operate reliably in the challenging, variable conditions common to power grid work. Key areas of innovation include multi-modal environmental perception, fully autonomous navigation and dynamic obstacle avoidance, adaptive operation in complex terrain and weather, and AI-powered accurate identification of hidden equipment defects that could threaten grid stability.
Automated systems are already delivering measurable performance gains across the company’s operations. To date, Guangdong Power Grid has deployed more than 10,000 power inspection drones across its service area. These unmanned systems complete roughly 500,000 independent inspection flight missions every year, and the company records that their inspection efficiency is 2.5 times higher than traditional manual inspection methods.
Recent upgrades to drone operations have unlocked even greater time savings, driven by the integration of new digital tools. Earlier this year, the company’s Jiangmen Power Supply Bureau rolled out a new AI-powered digital employee system that acts as a centralized “super brain” for the bureau’s drone fleet. According to Sun Tiancheng, a technician with the Jiangmen branch, the digital tool automatically generates customized flight plans and optimizes inspection routes in real time. This technological upgrade has cut pre-flight preparation time dramatically, from an average of 30 minutes per mission down to just one minute, streamlining workflows and allowing inspectors to respond faster to potential grid issues.