# Data Center Cooling and Power Outages

Analyze temperature-dependent cooling and the interaction between electrical supply and cooling equipment during grid outages.

# Dynamic Simulation of Data Center Cooling Operation and Power Outages

**Summary**: Data centers need reliable cooling because nearly all the electrical power consumed by IT hardware is ultimately released as heat. Different temperature levels and cooling methods are used depending on the hardware and cooling concepts. Alongside conventional air cooling, direct liquid cooling is becoming increasingly relevant, particularly at high power densities.

A dynamic simulation model can be used to investigate how cooling responds to changing conditions and how a power outage affects operation. The following application example examines both scenarios using a simplified data center.

Free cooling offers an energy-efficient operating option. It releases waste heat to the ambient air without using a vapor-compression chiller. Whether this is possible depends on factors including the outdoor temperature as well as the required cooling water temperatures.

 ![Data center diagram with electrical and thermal energy flows](/images/sites/green-city-library/pages/solutions/data-center-cooling/energy_flows.webp)

*Figure 1: Simplified model of the data center, showing electrical and thermal energy flows.*

## Simulation Model Structure

The model represents the data center as an electrical and a thermal load. Individual servers and airflow within the building are not modeled in detail. The focus is on the interaction between cooling equipment, controls, and the electrical supply.

The cooling system consists of two hydraulic separators, a heat exchanger, a vapor-compression chiller, and dry coolers. The hydraulic separators decouple the cooling circuits, while their water volume also provides thermal storage. The heat exchanger and chiller are arranged in parallel, providing two alternative cooling paths. The dry coolers ultimately release the waste heat to the ambient air.

The data center’s cooling water supply is designed for a supply temperature of 17 °C and a nominal return temperature of 22 °C. Dynamic controls adjust cooling operation to the current temperature conditions.

 ![Data center simulation model with named elements](/images/sites/green-city-library/pages/solutions/data-center-cooling/model_display.webp)

*Figure 2: Data center simulation model, naming the main components .*

A grid connection and a diesel generator are modeled for the electrical supply. This allows both normal grid-connected operation and island operation during a grid outage to be investigated. All components are secured either via the general supply or the emergency supply.

The model was built using the **GreenCity Simulation Library**. The library provides prebuilt components for the dynamic simulation of energy systems. Combining these components makes it possible to represent more complex system configurations with a manageable modeling effort and extend them with custom control strategies.

## Temperature-Dependent Cooling Operation

During normal operation, the cooling water temperature should remain as close as possible to its setpoint despite changes in thermal load and outdoor temperature. Two operating modes are available for this purpose.

At low outdoor temperatures, cooling takes place through the heat exchanger. Waste heat passes from the data center’s cooling circuit to the heat rejection circuit and is released to the ambient air through the dry coolers. A vapor-compression chiller is not required in this mode.

As the outdoor temperature rises, the temperature difference between the two cooling circuits decreases. Once this difference is no longer sufficient for free cooling, the vapor-compression chiller takes over. It removes heat from the first hydraulic separator and raises the temperature level on the heat rejection side. The heat is transferred to the second hydraulic separator so that it can continue to be dissipated through the dry coolers.

 ![Cooling water supply temperature plot](/images/sites/green-city-library/pages/solutions/data-center-cooling/cooling.webp)

*Figure 3: Cooling water supply temperature at the first hydraulic separator over time. Green areas indicate free cooling; red areas indicate operation of the vapor-compression chiller.*

The simulated cooling water supply temperature fluctuates around the 17 °C setpoint. These fluctuations reflect variations in the data center’s thermal load, control hysteresis, thermal inertia, and component response times. During transitions between operating modes, the simulated supply temperature ranges from approximately 16 °C to 19 °C.

The model allows these control processes to be examined over time. Changes to switching conditions or control parameters can be evaluated directly for their effects on the cooling water temperature.

## Grid Outage and Island Operation

In addition to normal cooling operation, the simulation model also represents a grid outage. This can be triggered at any point during the simulation.

In this simulation example, a grid outage reduces the data center’s total electrical demand to approximately 10% of its normal operating level. Loads connected only to the normal supply are disconnected, while selected IT equipment and cooling components remain supplied by the emergency power system. A diesel generator supplies the remaining loads during island operation. Cooling operation continues to serve the reduced thermal load.

 ![Data center power demand plot from the grid and during a grid outage via the emergency power supply.](/images/sites/green-city-library/pages/solutions/data-center-cooling/heat_demand.webp)

*Figure 4: Power draw of the data center from the grid and the emergency power supply over time. The dashed line marks the grid outage.*

The abrupt drop in power shows the transition from normal operation to the emergency supply. The reduction to about 10% load is a simplified assumption for this simulation example and is not a universal design requirement for data centers.

The grid outage serves as a deliberate intervention in the simulation model. This allows exceptional operating conditions to be examined alongside normal system operation, and the interactions between the electrical supply and cooling equipment to be analyzed.

## Beyond Data Center Cooling

The data center example illustrates how the GreenCity Simulation Library connects individual components into a dynamic energy system model. The same approach supports a wide range of applications beyond the scenarios presented here.

Models can combine HVAC systems, electrical networks, energy conversion technologies, and storage to evaluate integrated local energy systems and their interaction with the public grid. This enables investigations into renewable energy integration, sector coupling, peak load management, and demand response. Batteries and combined heat and power (CHP) units can also be evaluated as part of backup power concepts.

Thermal applications include ice, slush ice, and the ground for passive energy storage. Different load scenarios can be used to assess how these options shift energy demand and reduce peak loads. Sustainable cooling concepts and waste heat recovery can also be investigated, including the supply of recovered heat to district heating networks. Building models extend the scope to HVAC control algorithm development and indoor air quality analysis.

 ![Alternativ Data Center setup with redundant cooling system](/images/sites/green-city-library/pages/solutions/data-center-cooling/other_options.webp)

*Figure 5: Example of a redundant cooling system with ice thermal storage and waste heat recovery for district heating.*

Simulation results provide a basis for comparing energy efficiency across alternative system configurations. Combined with investment, operating, and maintenance cost assumptions, they can support techno-economic optimization and life-cycle cost assessments.

The GreenCity Simulation Library helps engineers explore these interactions within a common modeling framework and make informed decisions about the design and operation of interconnected energy systems.

[Are you planning a project? Get in touch to discuss your simulation needs.](/contact/)