Exhibitor Press Releases

18 Feb 2026

Data centre cooling: why process and proof matter

Eurovent Certita Certification Stand: F90
Jennifer Preston
Data centre cooling: why process and proof matter
Data centre cooling: why process and proof matter
Data centres place some of the most demanding requirements on cooling systems of any building type. High heat densities, continuous operation, strict uptime expectations and increasing pressure on energy efficiency mean that cooling equipment must perform exactly as specified. For data centre operators, this makes the process of selecting and validating cooling equipment as important as the equipment itself.

Two elements are critical in reducing risk:

1)      A design process based on accurate performance data

2)      The factory witness test – a procedure used to verify performance before equipment ever reaches site.

This article explores the cooling equipment selection process for data centres, the importance of the factory witness test and the role of product certification in reducing risk of project delays, spiralling costs and future system problems.

The supply process for data centres

Unlike comfort cooling applications, data centre HVAC systems must support mission-critical infrastructure. Additionally, the demand for data centres outstrips the speed at which they can be built, meaning that any delays in construction programmes can have serious financial penalties. Therefore, when it comes to specifying cooling equipment, data centre operators are not just choosing a product, they are choosing a partner. As a result, the supply process needs to be rigorous, transparent and seamlessly aligned with operational requirements from the outset.

The process usually begins with a clear definition of the data centre’s needs, including:

  • IT load and heat density
  • Redundancy strategy (e.g. N, N+1, 2N)
  • Environmental conditions and allowable temperature bands
  • Power Usage Effectiveness (PUE) targets and sustainability goals
  • Failure scenarios
  • System type e.g. CRAC, CRAH with chilled water, free cooling, liquid cooling, etc.
  • Planning and compliance considerations
  • Plus, any location and infrastructure constraints.

An SLA (Service Level Agreement) is used to define Data Centre performance. SLAs are incredibly strict and failure to meet them incurs huge financial consequences. Any ambiguity in requirements at this stage increases the risk of performance shortfalls later.

Once operational requirements are defined, they must be translated into precise technical specifications for HVACR equipment. This may include:

  • Cooling capacity under realistic operating conditions over 365 days a year performance
  • Peak and part-load performance (with the required redundancy and resilience)
  • Airflow and control strategies
  • Pipework requirements
  • Integration with monitoring and control systems
  • Acoustic and spatial constraints.

Manufacturers respond with proposed solutions, which are likely to incorporate detailed pre-procurement and technical information such as data sheets, selection sheets, piping and instrumentation diagrams (P&IDs), general arrangement (GA) drawings, control sequences, installation and operation manuals (IOM), and integration details for BMS and EPMS.

For data centre operators, this pre-procurement stage is critical. It is often the final opportunity to validate that performance data, assumptions and interfaces are fully understood before contractual commitments are made. Manufacturer performance claims must be based on verified, independently supported data rather than theoretical or best-case values, as any inaccuracies at this stage can directly affect equipment selection, system design, project cost and long-term operational resilience.

Next the design team, procurement team and client will review the manufacturer offer for compliance with the specification (capacity, redundancy, ambient range, footprint, acoustics, efficiency, codes/standards etc.). Clarifications and revisions are common during this phase to ensure technical and contractual alignment.

Once product selection has been confirmed and procurement approval granted, the manufacturer will proceed with building the equipment. Before shipment, the products will usually undergo a Factory Witness Test (FWT) to verify that the supplied units meet the agreed technical and performance requirements prior to delivery to site.

The factory witness test

Factory witness testing allows operators, consultants and stakeholders to observe and verify equipment performance under controlled conditions before delivery. For data centres, this step is increasingly seen as essential rather than an optional extra.

FWTs involve testing HVAC equipment at the manufacturer’s facility, while representatives from the client or an independent body witness the process on the client’s behalf. It will also usually include a review of all drawings, P&IDs, wiring diagrams, control sequences, and test procedures against the specification. The goal is to confirm that the unit:

  • Meets specified performance criteria
  • Operates as expected across defined load conditions
  • Behaves as expected during start up, shutdown and simulated fault conditions
  • Passes all safety checks
  • Functions correctly before installation and commissioning.

If the unit fails to match the performance on the selection sheet, three things will happen:

  1. Investigations must take place
  2. The issue must be resolved
  3. The unit must be retested.

Depending on the issue and ease of remedy, this can take weeks to months to complete, delaying the construction programme. While the financial implications can be severe, a worse consequence is loss of credibility - especially in a market where partnership and confidence are key.

The FWT is vital as once HVAC equipment is installed in a live data centre, correcting performance issues can be very costly and disruptive. The FWT shifts this risk upstream by identifying potential problems early, such as:

  • Capacity shortfalls
  • Control or sequencing issues
  • Deviations from declared performance
  • Component or assembly issues.

For data centre operators, this reduces the likelihood of late-stage surprises. However, catching issues early will still cause headaches in the short term. A failed factory witness test will have serious implications on project timelines - costing time and money - as the product will need to be fixed or replaced and retested. 

Reducing risk of factory test failure

Independent, third-party verification - such as that provided through Eurovent Certified Performance programmes - helps to ensure that performance data used at the design and procurement stage is both reliable and comparable.

Eurovent Certified Performance programmes use state of the art evaluation protocols including laboratory tests, factory audits, selection software certification and evaluation of all submitted data (including data sheets and marketing materials) to ensure the data declared is accurate, verified and certified against recognised standards and real-world requirements.

In addition, Eurovent certified IT cooling units and Very High Temperature Process Chillers enter a surveillance process which includes regular factory audits, laboratory tests and software checks. This ensures the products, and their selection software, remain accurate and compliant with certified data – a critical step in guaranteeing product performance.

For data centre operators, this provides:

  • Confidence that equipment performs as stated
  • Greater transparency during equipment selection
  • Trust in software selection tools as they have also been certified
  • A level playing field when comparing different certified solutions.
  • Certified performance data supports better decision-making early in the project lifecycle, reducing the likelihood of underperformance later.

Safeguarding the witness test

Certification’s role doesn’t stop there. It can help give assurance that a certified manufacturer’s in-house testing facilities deliver accurate factory witness test results. In certain Eurovent Certification programmes such as its Chillers and Heat Pump programme (LCP-HP) and its IT Cooling Unit programme (ITCU), Eurovent Certification allows manufacturers to use their own laboratories for Eurovent surveillance tests under what is known as the Participant Laboratory option. This option is available only when the manufacturer’s laboratory meets all requirements of the relevant test standards and methods indicated in programme-specific TCR (technical certification rules) documents.

For surveillance purposes, a certain number of units (determined by the programme rules) is selected for testing. Some of these tests may be conducted in the manufacturer’s laboratory, but always by an accredited independent partner laboratory assigned to the tests by Eurovent Certification. The partner laboratory must use its own calibrated measuring equipment rather than the devices available in the manufacturer’s laboratory and must follow the programme rules and applicable standards.

During these tests, Eurovent Certification may also request manufacturers to take measurements with their own devices across selected tests and share these results. Eurovent Certification then performs an internal comparison between the manufacturer’s measurements and those of the independent laboratory to verify that any deviations remain within the permissible limits defined by the programme-specific TCR document and based on the applicable standards. If deviations fall outside the allowable range, Eurovent Certification requires a root cause analysis and, where necessary, corrective actions.

Eurovent Certification also issues an Approval of Compliance document to manufacturer laboratories approved to conduct Eurovent surveillance tests. While this document is not a formal accreditation or certification of the participant’s laboratory, it confirms that the laboratory meets the relevant standards, demonstrates compliant measurement accuracy, and is therefore approved for Eurovent tests.

For specifiers and data centre operators this provides a triple layer of assurance. Certified data guarantees the truthfulness of the manufacturer’s performance claims, while the factory witness test - in a laboratory with test accuracy verified by Eurovent Certification - confirms that the actual delivered unit meets those claims at the project’s critical operating points.

Conclusion: from assumptions to assurance

As data centres continue to grow in scale, density and criticality, the tolerance for uncertainty in cooling system performance continues to shrink. HVAC equipment is no longer selected solely on the basis of nominal capacity or compliance with a written specification; it must deliver proven, repeatable performance under real operating conditions, from day one and throughout its lifecycle.

A robust supply process, underpinned by accurate performance data and validated through factory witness testing, is central to achieving this. If the manufacturers laboratory has also had its test accuracy independently validated, risk of data deviations can be further reduced. However, factory testing alone cannot compensate for unverified declared performance data at the design and procurement stage. If theoretical performance assumptions are incorrect, the risk of failed factory witness tests, design changes, project delays and spiralling costs increases.

This is why Eurovent Certified Performance is essential for data centre cooling systems. By independently verifying that manufacturers declared performance data is accurate, comparable and repeatable, Eurovent Certification provides a trusted foundation for system design and equipment selection. For operators and specifiers, it reduces uncertainty early in the project lifecycle, supports fair comparison of solutions, and lowers the risk of unexpected outcomes later.

Eurovent Certification provides dedicated certification programmes for data centre cooling equipment including its IT Cooling Unit (ITCU) programme (CRAC/CRAH/Hybrid and rear door units) and its Chillers and Heat Pump (LCP-HP) programme which certifies Very High Temperature Process Chillers with or without free cooling. Eurovent Certification also offers schemes for the full cooling infrastructure including cooling towers, heat exchangers, air handling units and evaporative cooling.

Discover more at www.eurovent-certification.com

View all Exhibitor Press Releases
Loading

Until Tech Show London 2026

Register Now