Data Center Library for Modelon Impact | Faster AI Cooling Design

Data Center Library

The Data Center Library (DCL) is a physics-based simulation solution within Modelon Impact for modeling and evaluating data center cooling systems—from component sizing to full MW-scale facility models with PUE monitoring.

It covers the complete cooling chain end to end: facility-side infrastructure (chillers, cooling towers, pumps, waterside economizers) through coolant distribution to the rack and chip level (CDUs, CRAHs, cold plates), including supervisory control logic. Both hybrid and fully liquid-cooled architectures are supported, reflecting the reality of modern deployments where memory, networking, and other equipment cannot be liquid cooled.

Pre-calibrated CDU vendor models allow teams to simulate exact hardware from leading cooling equipment manufacturers at their specific facility water temperatures, parameterized directly from standard ASHRAE W-class datasheet operating points, without requiring detailed internal geometry. A liquid-cooled AI cluster reference model at 1 MW scale is included out of the box.

DCL is built on the Liquid Cooling Library (LCL) and is compatible with the Vapor Cycle Library (VCL) for chiller modeling. It is a product of Modelon, developed with Modelica open standards, and is exclusive to the Modelon Impact software platform.

The example demonstrates how empirical performance data can be integrated into a comprehensive data center cooling system model to support design analysis, optimization, and control strategy evaluation.

Contact Us

BENEFITS

Understand how the full cooling system behaves under realistic operating conditions, not just isolated subsystem assumptions, enabling better decisions on topology, control strategy, and equipment sizing before deployment.
Simulate exact CDU vendor hardware at your facility’s water temperature conditions before procurement, and compare multiple vendors side by side under identical conditions using pre-calibrated models covering leading CDU manufacturers across different ASHRAE temperature classes (verified accuracy <1.4% error on a 2 MW reference CDU).
Evaluate control strategies, setpoints, and free-cooling utilization at the system level to improve energy efficiency and reduce unnecessary cooling capacity margins.
Accelerate engineering studies using reusable models and reference architectures, including a ready-to-use liquid-cooled AI cluster model, rather than rebuilding each study from scratch.
Aligned to ASHRAE W-class thermal guidelines (W1–W4), with the only documented ASHRAE W1–W4 CDU calibration framework covering the full ASHRAE W1–W4 water temperature class range.
Built on open, Modelica-based foundations, providing transparency and extensibility compared to black-box or proprietary alternatives.

APPLICATION

Data Center Cooling System Design

This library supports end-to-end system design and virtual prototyping for hyperscale, colocation, and enterprise facilities. Components are rapidly assembled to evaluate alternatives, including standard chiller plants, chillers with waterside economizers, multi-chiller configurations, and hybrid rack arrangements. Teams can test different topologies before committing to hardware.

CDU Vendor Evaluation and Procurement Support

Committing to expensive CDU hardware without understanding how it will perform in a specific facility is a key procurement risk. The library’s pre-calibrated vendor records allow engineering teams to simulate exact hardware under actual facility water conditions and compare options side by side before purchase. This approach provides quantitative performance insights that test rigs cannot deliver prior to procurement.

AI Cluster Cooling Optimization

The library includes a reference liquid-cooled AI cluster system model at the 1 MW scale, covering rack-level cold plate models, CDU and chiller loops, and facility-level PUE tracking. The model supports evaluation of supply temperature setpoints, pump sizing, CDU operating ranges, and cooling capacity headroom for GPU-dense workload scenarios.

Control Strategy, Setpoint Optimization, and Sustainability

The library models supervisory plant controls, including chiller staging, waterside economizer mode switching, and CWST-based setpoint logic, within the same environment used for system design. Teams can evaluate free-cooling utilization, energy efficiency, and control behavior under varying IT loads and ambient conditions before deployment or during operational optimization. This provides a quantitative basis for achieving both performance and sustainability goals.