As data centres support increasingly demanding digital workloads, power quality and thermal testing are becoming essential to ensuring operational reliability, protecting critical equipment and improving energy efficiency.
Every time someone joins a video call, makes a digital payment, streams a film or interacts with an artificial intelligence tool, a network of servers and supporting infrastructure works behind the scenes to make that experience possible. At the heart of this digital ecosystem are data centres, facilities designed to keep information accessible, applications responsive and digital services running around the clock.
However, maintaining this level of reliability involves more than installing powerful servers and connecting them to a stable electricity supply. Data centres must manage two closely connected challenges: ensuring that electrical power remains within acceptable operating conditions and keeping equipment temperatures under control.
Even small deviations in either area can have wider consequences. Electrical disturbances may affect sensitive equipment, while uneven cooling can create hotspots, increase energy consumption and shorten hardware life. This makes power quality and thermal testing an important part of data centre design, commissioning and ongoing maintenance.
As artificial intelligence, cloud computing and high-performance computing continue to expand, these tests are becoming increasingly important to the reliability of modern digital infrastructure.
Understanding Power Quality in Data Centres
Data centres depend on a complex electrical infrastructure that includes utility connections, transformers, switchgear, uninterruptible power supply (UPS) systems, backup generators, power distribution units and server-level power supplies. Each component must work effectively within the wider electrical system.
Power quality refers to the characteristics of the electricity supplied to equipment, including voltage, frequency, waveform distortion and continuity of supply. Although a facility may receive electricity without a complete outage, disturbances can still occur and affect equipment performance.
Voltage sags, voltage surges, harmonic distortion, frequency variations and phase imbalances are among the conditions that electrical engineers monitor. Harmonics, for example, can arise from nonlinear loads such as power electronic converters and other equipment. If not properly managed, they can contribute to additional heating, equipment stress and interference with electrical systems.
The IEEE 519-2022 standard establishes guidance for controlling harmonic distortion in electrical power systems, while IEEE 1159-2019 addresses the monitoring and interpretation of power quality phenomena. Applying the appropriate requirements helps engineers assess electrical performance and identify potential sources of disruption.
For data centre operators, the objective is not simply to confirm that electricity is available. It is to establish whether the electrical infrastructure can deliver power within the required operating conditions, respond appropriately to disturbances and support the facility’s expected load.
Why Power Quality Testing Matters
A data centre may have backup generators and redundant power supplies, but redundancy alone does not eliminate every electrical risk. Problems can occur at different points in the distribution network, and a disturbance can affect multiple connected systems if the underlying design or protection arrangements are inadequate.
Power quality testing helps identify these vulnerabilities before they develop into operational problems.
During testing, engineers may measure voltage and current, monitor frequency, assess harmonic distortion, evaluate phase balance and examine electrical behaviour under different load conditions. Testing may also assess the performance of UPS systems, switching arrangements and backup power infrastructure, depending on the scope of commissioning or maintenance.
Load-bank testing is another important activity. It allows generators and associated power systems to be evaluated under controlled electrical loads, helping confirm whether they can perform as expected when required. Integrated systems testing can go further by examining how the electrical infrastructure responds to simulated failures, transfers between power sources and recovery scenarios.
The findings help operators understand whether their systems are performing as designed and whether corrective action is needed.
Importantly, testing should be planned around the facility’s design, applicable standards and operational requirements. Some assessments can be performed during normal operation, while others require controlled test conditions and careful coordination to avoid introducing unnecessary risks.
Thermal Testing: Managing Heat Before It Becomes a Problem
Electricity powers the servers, storage systems and networking equipment inside a data centre, but much of that electrical energy ultimately becomes heat. As workloads increase, removing that heat efficiently becomes essential to maintaining stable operating conditions.
Thermal management involves more than keeping the room cool. Airflow distribution, rack configuration, equipment density, cooling capacity and environmental conditions all influence how effectively heat is removed.
A cooling system may appear to be functioning correctly while individual racks experience higher inlet temperatures or uneven airflow. Cold air may bypass equipment that needs it, while hot exhaust air may recirculate into server intakes. These conditions can create localised hotspots even when average room temperatures appear acceptable.
Thermal testing helps uncover these issues.
Engineers can use temperature sensors, airflow measurements, environmental monitoring and infrared thermography to assess conditions across the facility. Measurements at rack inlets are particularly useful because room-level readings alone may not reveal the temperatures experienced by individual servers.
Infrared thermal imaging can also help identify abnormal heating in electrical connections, distribution equipment and other components. When used alongside appropriate electrical inspections, it can reveal potential problems such as overheating connections or uneven thermal loading.
The US Environmental Protection Agency’s ENERGY STAR programme highlights server-inlet temperature monitoring and thermal imaging as practical ways to identify hotspots and improve cooling management.
The Growing Challenge of AI and High-Density Computing
The increasing use of AI is changing the way data centres are designed and operated. Training and running advanced AI models can require large numbers of accelerators and other high-performance computing components, often concentrated within densely populated racks.
This creates additional demands on both electrical distribution and cooling infrastructure.
Higher rack power densities can increase current requirements, intensify local heat loads and make airflow management more complex. Traditional cooling arrangements may not be sufficient for every high-density configuration, leading operators to evaluate more advanced approaches, including direct-to-chip liquid cooling and rear-door heat exchangers.
The challenge is not simply to install equipment with a higher rated capacity. Operators must establish whether the complete system including electrical distribution, cooling equipment, monitoring and control systems can perform reliably under the intended workload.
The ASHRAE AI Data Center Energy Performance Framework addresses energy efficiency, thermal management, resilient design, commissioning and operational performance across the data centre lifecycle.
This broader approach is important because power and thermal systems cannot be evaluated entirely in isolation. A change in computing load affects electrical demand and heat generation, while cooling equipment also consumes electricity. Understanding these relationships helps engineers identify constraints and make more informed design and operational decisions.
Testing for Efficiency as Well as Reliability
Reliability remains a central objective of data centre testing, but energy efficiency is becoming equally important as facilities scale and electricity demand increases.
Overcooling, poorly managed airflow and inefficient equipment operation can increase energy consumption without necessarily improving reliability. At the same time, reducing cooling or electrical overhead without adequate validation can create unacceptable operating risks.
The goal is to find the right balance.
Thermal assessments can help operators identify areas where airflow needs adjustment, cooling capacity is poorly distributed or temperature setpoints could be optimised. Power quality measurements and electrical performance assessments can help identify inefficiencies, imbalances and conditions that require corrective action.
The findings can inform maintenance schedules, equipment upgrades and operational changes. When testing results are integrated with building management systems and data centre infrastructure management platforms, operators can also develop a clearer picture of performance over time.
The ASHRAE AI Data Center Energy Performance Framework recommends foundational measures such as effective hot- and cold-aisle management, reduced airflow recirculation, appropriately controlled fan speeds and granular temperature monitoring. These measures can improve cooling performance while reducing unnecessary fan energy.
However, efficiency improvements must remain within the environmental and electrical operating limits of the installed equipment. The appropriate temperature ranges, power quality limits and testing procedures depend on equipment specifications, facility design and applicable standards.
Integrating Testing into the Data Centre Lifecycle
Power quality and thermal testing deliver the greatest value when they are incorporated throughout the data centre lifecycle rather than treated as one-time commissioning activities.
During design and construction, testing helps establish whether installed systems meet their intended performance requirements. During commissioning, integrated assessments can validate how electrical, mechanical and control systems respond to normal operation and defined failure scenarios.
Once a facility becomes operational, periodic inspections, continuous monitoring and condition-based maintenance help identify emerging problems. Testing may also be necessary after significant changes, such as adding high-density computing racks, modifying power distribution or upgrading cooling equipment.
A useful testing strategy brings together several elements: clear acceptance criteria, suitable measurement equipment, documented procedures, trained personnel and a process for investigating abnormal results. Findings should be recorded and tracked so that corrective actions can be verified rather than assumed to be complete.
The approach must also account for operational constraints. Testing live infrastructure requires careful planning, appropriate safety procedures and coordination with facility operations teams. Where an assessment could affect critical loads, the test method and operating conditions must be selected accordingly.
International standards, including the ISO/IEC 22237 series for data centre facilities and infrastructure, provide a broader framework for addressing infrastructure availability, power distribution and energy-efficiency considerations. The applicable parts should be selected according to the facility’s requirements and project scope.
Building More Resilient Digital Infrastructure
As businesses, governments and consumers become increasingly dependent on digital services, the infrastructure supporting those services must be designed to handle growing demand without compromising reliability.
Power quality and thermal testing play a practical role in achieving this objective. They help engineers understand how electrical systems behave under different conditions, how effectively cooling infrastructure removes heat and where potential weaknesses may exist.
The growing demands of AI and high-performance computing make this work even more significant. Higher power densities, changing load profiles and evolving cooling technologies require facilities to validate not only individual components but also how complete systems operate together.
Looking ahead, continuous monitoring, better integration between facility systems and more predictive maintenance practices are likely to strengthen the role of testing in data centre operations. These approaches can help teams identify abnormal conditions earlier, plan interventions more effectively and make decisions based on measured performance.
Ultimately, a reliable data centre is not defined only by the equipment it houses or the amount of computing power it delivers. Its performance also depends on the quality of the electrical supply, the effectiveness of its thermal management and the ability of its infrastructure to respond to changing demands.
By treating power quality and thermal testing as ongoing priorities, operators can build facilities that are better prepared to support the next generation of digital services while balancing reliability, efficiency and long-term operational resilience.














