Data Center Modelling and Simulation with EcosimPro

EcosimPro provides a powerful simulation environment for modelling the complete dynamic behaviour of Data Center infrastructure.

From individual CPU/GPU racks and cooling distribution units to facility cooling systems, electrical infrastructure and supervisory controls, EcosimPro enables engineers to build integrated, physics-based models and use them throughout the data center lifecycle — from concept and design to commissioning and operation.

One model. Multiple physical domains. The complete Data Center.

A modern data center is not simply an HVAC system or an electrical network. It is a tightly coupled dynamic system in which IT workloads, cooling equipment, electrical infrastructure and control systems continuously interact.

EcosimPro’s multidisciplinary modelling approach makes it possible to represent these interactions within a single simulation environment for design (classical modelling) and to deploy in operation (Digital Twin).

Existing and dedicated libraries provide modelling capabilities for:

• • IT infrastructure: CPU and GPU racks, power consumption, heat generation and transient workload profiles.
  • Liquid cooling: cold plates, Cooling Distribution Units (CDUs), pumps, valves, pipes and heat exchangers.
  • HVAC and cooling plants: chillers, cooling towers, dry coolers, air handling systems and other heat rejection equipment.
  • Thermal systems: heat transfer, thermal inertia and thermal storage.
  • Electrical systems: grid connections, transformers, UPS, batteries, generators and power distribution.
  • Control systems: temperature and flow control, pump and chiller sequencing, valve control and supervisory strategies.
    
    

The result is an executable virtual representation of the data center that captures not only its nominal operating point, but also how the complete failure modes over time.

Design data centers as integrated systems

During design, EcosimPro allows engineers to investigate alternatives before equipment is selected and the facility is built.

Models can be used to evaluate cooling architectures, size equipment and explore the interaction between IT loads and facility infrastructure under different operating and environmental conditions.

Typical studies include:

• • Liquid cooling system design
  • Equipment sizing
  • Energy and water efficiency
  • Cooling architecture comparison
  • Heat recovery
  • Electrical and cooling integration
    
    

Because EcosimPro uses reusable component libraries and hierarchical models, engineers can move naturally from individual subsystems to complete facilities and reuse validated models across projects.

Understand what happens when conditions change

Steady-state design calculations provide only part of the answer. Many of the most critical data center engineering questions are inherently dynamic.

• • What happens if hundreds of GPU racks suddenly move from partial to full computational load?
  • If a chiller, pump or electrical supply fails, how long can the facility maintain acceptable IT temperatures?
  • What happens to coolant temperatures and flows while backup systems start?
  • How much thermal margin is really available during a power transition?
    
    

From engineering model to operational digital twin

Once a data center enters operation, its instrumentation provides large amounts of information: temperatures, pressures, flow rates, electrical power, pump speeds, valve positions, environmental conditions and IT workload.

EcosimPro models can be integrated with external environments and operational systems, enabling the engineering model developed during design to evolve towards a physics-based digital twin.

The operational model can combine measured data with the physical behaviour represented in the simulation to support questions such as:

• • What will happen if tomorrow’s ambient temperature increases while GPU demand rises?
  • Do we have sufficient cooling margin for an additional IT workload?
  • Why is a CDU consuming more power than expected?
  • What is the most efficient combination of pumps and chillers for the current operating condition?
  • How will the facility respond if a particular piece of equipment becomes unavailable?
    
    

Instead of using measured data only to understand what has already happened, physics-based simulation makes it possible to investigate what is likely to happen next.

EcosimPro provides connectivity with external environments and standards such as Python, FMI and OPC UA, allowing simulation models to participate in broader engineering, optimization, data analytics and digital-twin architectures.

Python integration also opens the possibility of combining physics-based simulation with data analytics, optimization and machine-learning techniques.