Chirag Devendrakumar Parikh, Computer Engineering, California State University, Fullerton, CA, USA
The engineering challenge of continuously complying with safety protocols for high-voltage systems in hyperscale data centers, which deploy AI power management systems to handle high-performance compute workloads, becomes even more difficult. This issue is addressed in the paper by proposing an integrated framework that combines real-time high-voltage hazard alert systems with compliance monitoring and is further enhanced by predictive compliance strategies. The described methods and solutions permit the embedded intelligent control systems to tackle adaptive diagnostics and highly advanced sensor networks and extend their functionality to the power systems to detect early non-conformant conditions and electrical threats to safety. AI-based engines applied to real-time compliance data evaluation enhance the decision processes regarding maintenance, alterations, and updates of the structure in question, and even to the regulations that govern it. Compliance with UL 61010, UL 62368-1, IEC 61010, IEC 62477, and other critical safety standards is also fully observed in the paper, making sure that equipment and processes used will not pose unnecessary hazards. The paper addresses simulations in high-density rack-level power distribution, uninterruptible power supplies, and busway systems, focusing on the application of predictive compliance and high-voltage safety monitoring, reporting reduced operational downtime and enhanced reliability. The intention is to redesign the next generation of data centers by eliminating the traditional approach to risk management and replacing it with an intelligent compliance approach.
Compliance, Product Safety, Data Centers, AI, Global Market Access, Safety
Copyright © ELEN 2025