Data Center Digital Twin: Real-Time 3D Visualization and AI-Based Expansion Recommendation
Year: 2026
Type: article
Abstract: The rapid growth of global data continuously pushes data center infrastructure toward its physical limits, yet deciding when and how to expand capacity remains a critical challenge for operators. Sudden traffic surges can exceed the throughput of physical infrastructure, while uncontrolled power consumption and overheating threaten densely packed server racks. This paper presents a data center digital twin that unifies real-time telemetry, interactive 3D visualization, and an AI-based expansion recommendation engine in a single web-based system. The architecture follows an eight-block digital twin design: physical facilities, IoT data acquisition, data integration, the virtual twin state, prediction and explainability, optimization and feedback, the user dashboard, and a closed feedback loop. The twin state is maintained as a structured vector covering power infrastructure, cooling, environment, workload distribution, server utilization, network health, and equipment health. An expansion scoring engine computes a vertical score for specification upgrades and a horizontal score for adding nodes or racks from utilization, traffic, power, and cost inputs, and derives a confidence score from the gap between the two. The 3D scene, rendered with WebGL, animates data flow as dynamic particles and maps temperature and load telemetry onto rack heat colors, while a control panel with sliders and quick scenario presets enables zero-risk what-if simulation. A functional prototype with sixteen racks and one hundred twenty-eight servers demonstrates centralized visual monitoring, proactive risk testing through quick scenario presets, and data-driven expansion recommendations with confidence scores. The main contribution is an integrated digital twin architecture that couples real-time 3D visualization with a transparent scoring-based expansion decision engine for data center operators.
Language: English
Subfield: Information Systems
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7225414976