Data centres require a reliable power supply around the clock. Servers, network equipment, storage systems and cooling systems must be supplied with power continuously. Even brief interruptions can cause IT systems to fail or applications to become inaccessible. That is why the power supply for a modern data centre consists of more than just a simple mains connection. It is structured as a multi-stage system that distributes electrical energy, monitors it and safeguards it in the event of a failure in the public supply.

A data centre’s power infrastructure includes, amongst other things, transformers, switchgear, uninterruptible power supplies, power distribution units, busbar systems and automatic transfer switches. This infrastructure is often supplemented by emergency generators and energy storage systems, which step in the event of a mains failure.

To ensure these systems function reliably, various operating parameters must be monitored. In addition to electrical measurements such as current and voltage, temperature and pressure also play an important role.

A data center continuously processes data. Many services must be available 24 hours a day, seven days a week.

These include, for example, cloud applications, enterprise software, web services, databases, communication systems, artificial intelligence, production systems, and digital platforms.

If the power supply fails, these systems can become unavailable within a very short time. For this reason, data centers typically have multiple levels of backup power. In the event of a failure in the public power grid, the power supply must be switched over to these backup systems with as little interruption as possible.

The exact configuration varies depending on size and security requirements. Generally, however, the electrical power passes through several stages. First, the electricity enters the data center via the utility connection. Transformers adjust the voltage to the required levels. Switchgear then distributes the power to different areas.

Uninterruptible power supplies (UPS) bridge short-term outages and the time it takes, for example, for an emergency generator to start up. From there, the electricity travels through power distribution units and busbar systems to the individual server racks. Inside the rack, power distribution units handle the further distribution to the individual servers.

In the event of a failure in the public power grid, a data center’s power supply must be maintained with as little interruption as possible. A UPS (uninterruptible power supply) can bridge short outages, often with the help of battery storage. For longer-term power supply, additional power generation systems such as diesel or gas generators and fuel cell systems are used. Pressure and temperature measurements play an important role here—for example, in monitoring oil pressure in internal combustion engines or hydrogen pressure in fuel cell systems.

With sensor solutions such as VSP, IPS, and HPS, i2s offers suitable pressure sensors for a variety of media and operating conditions.

Transformers, power electronics, batteries, busbars, power supplies, and switchgear can heat up during operation. A certain amount of heat generation is normal. However, excessively high temperatures may indicate a heavy load, insufficient cooling, or a technical fault.

Temperature measurements can therefore help to:

• detect unusual heating
• monitor cooling systems
• record thermal loads
• protect components from overheating
• identify maintenance needs early on

Power electronics systems can generate significant amounts of heat. Depending on their power rating and design, these systems are cooled by air or liquid. In liquid cooling systems, pressure and temperature sensors can provide important information about the condition of the coolant. Temperature sensors measure how hot the coolant is. Pressure sensors can monitor whether the required operating pressure is present and whether the flow conditions are changing.

Transformers adjust the electrical voltage to suit different areas of the data center. This process generates heat. Depending on their size and design, transformers can be air- or liquid-cooled.

In liquid-filled transformers, for example, the temperature and condition of the insulating or cooling medium can provide important operational information. Switchgear distributes and switches electrical power. Here, too, high currents can cause conductors, contacts, and connection points to heat up. Temperature monitoring can help detect unusually hot areas at an early stage.

Specially designed temperature or infrared sensors are used for direct measurements on electrical conductors. i2s, on the other hand, focuses primarily on sensor technology for the associated fluidic and thermal systems.

Power Distribution Units, or PDUs for short, distribute electrical power within the data center. Depending on the model, they are installed at the room level or directly in the server rack. There, they supply power to the individual IT components and can also perform important monitoring and control functions.

Modern rack PDUs enable, for example:

• Power consumption monitoring
• Measurement of individual outlets
• Remote switching
• Overload alerts
• Network communication
• Temperature monitoring

This makes PDUs a key component of power distribution and supports power supply monitoring down to the rack level.

An Automatic Transfer Switch (ATS) automatically switches between different power sources. It monitors the regular grid supply and, in the event of a failure, can switch to an emergency power generator, for example.

This ensures that the data center’s power supply is maintained even during a temporary disruption of the public grid. The ATS is therefore an important component of a redundant and fail-safe power supply.

Emergency power generators provide an important safeguard against prolonged power outages. In data centers, stationary generator sets are often used for this purpose.

A generator consists of more than just the electrical generator itself. A so-called genset typically includes an internal combustion engine, a generator, a cooling system, a fuel supply system, a lubrication system, an exhaust system, and a control system. This creates an environment where pressure and temperature measurements are particularly important.

Sensors within a data center must operate reliably over long periods of time. Emergency power generators present additional challenges.

These include:

• Vibrations
• High ambient temperatures
• Oil and fuel
• Changing operating conditions
• Pressure spikes
• Long periods of downtime
• Short-term full load operation

Therefore, a sensor should not be selected based solely on its measurement range.

Media Compatibility
A pressure or temperature sensor may come into direct contact with the respective medium. Therefore, all materials in contact with the medium must be suitable for that purpose. Depending on the application, these may include, for example:

engine oil, transmission oil, coolant, water-glycol mixtures, air, nitrogen, refrigerants, hydrogen.

Robustness and Vibration Resistance
A running internal combustion engine generates vibrations. Sensors mounted directly on the engine or on adjacent components must withstand these mechanical stresses.

Measuring Range and Accuracy
The required measuring range depends heavily on the application. An engine oil circuit operates under different conditions than a cooling system or a hydrogen supply system. The sensor should therefore be designed for the normal operating range and any possible pressure spikes. The required accuracy also depends on the measurement task. Simple limit monitoring poses different requirements than precise control.

Fast Response Time
In certain applications, a fast response is critical. For example, if the coolant temperature of an engine or power electronics rises rapidly, the control system should detect this change as early as possible.

Long-Term Stability
A data center’s power supply systems are designed for long operating times. Even an emergency generator must remain reliably operational for many years. Sensors should therefore provide stable measurements throughout their service life. This is particularly important when measurement data is also used for long-term condition analysis.

Measured values can be used for more than just immediate control. If pressure and temperature readings are continuously recorded, changes can be monitored over longer periods of time.

Examples include:

• a slow drop in oil pressure
• rising coolant temperatures
• increasing pressure losses
• changes in temperature trends
• unusual pressure fluctuations

Such changes can indicate wear or the need for maintenance. This allows some maintenance tasks to be better tailored to the actual condition of the system.

Data centers often use redundant systems. This means that critical components are present in multiple instances. If, for example, a power supply unit fails, a second unit can take over. This principle can be applied at various levels:

• multiple network connections
• multiple transformers
• redundant UPS systems
• multiple power paths
• multiple emergency power generators
• redundant cooling

Sensors help monitor the status of individual systems. This makes it possible not only to detect whether a system has failed completely, but also whether its operating status is slowly changing.