Data centre design is not defined by how much equipment is installed. It’s defined by how the electrical system performs under worst-case conditions. Fault contribution, protection discrimination, voltage stability, and standby transition behaviour determine whether redundancy operates as intended or collapses during a disturbance.
System resilience must be verified through in-depth studies and analysis, especially in high-capacity environments. Power system modelling and validation provide the necessary evidence that supports investment decisions, planning approvals, and operational confidence.
Why Power System Resilience Determines Data Centre Performance
Modern data centres operate within tightly controlled availability targets. Whether the architecture follows N, N+1, or 2N principles, the integrity of the design depends on an accurate assessment of system behaviour during abnormal operating conditions.
When a system experiences a fault or transfers to standby power, electrical conditions shift rapidly across the network, directly influencing how protection devices and critical loads respond. Without early-stage assessment, this risk carries into commissioning and energisation.
Effective data centre design, therefore, requires early-stage power system analysis to quantify:
- Prospective short circuit levels per IEC 60909 methodology
- Protection grading and discrimination margins
- Arc Flash incident energy calculation allowing for early mitigation during design.
- Rapid Voltage change during fault, energisation and switching events.
- Voltage dip performance during fault clearance
- Transient stability and flicker performance associated with large or rapid GPU load ramping.
- Loading requirements and equipment rating compliance under contingency conditions.
These studies show how the system will respond during faults and power transfers, helping to ensure protection operates correctly, equipment stays within its limits, and redundancy works as intended before equipment is purchased or installed.
Power System Modelling as the Foundation of Data Centre Design
Electrical modelling provides a digital representation of the entire power architecture, from grid interface to final distribution boards. It allows engineers to simulate credible fault scenarios and switching events before construction begins.
In data centre design, modelling is used to:
- Confirm switchgear interrupting capacity against calculated peak and breaking currents
- Validate discrimination between upstream and downstream protective devices
- Assess transformer impedance tolerances and their impact on fault levels
- Evaluate the effect of additional generation or UPS modules on system stability
- Identify constraints that could limit future expansion
Without this validation, configuration changes can introduce hidden non-compliance or compromise selectivity.
Short-Circuit Levels and Equipment Duty
Expanding capacity in a data centre often increases fault contribution. Adding standby generators, installing additional UPS load, or reinforcing the grid connection can increase potential fault current beyond original design assumptions.
If switchgear ratings are exceeded, protective devices may not operate within their certified limits. This exposes the facility to equipment damage, extended outage duration, and safety risk.
Accurate short circuit analysis ensures equipment can withstand the highest possible fault levels and that protection settings are coordinated so faults are cleared without affecting healthy parts of the system.
Protection Coordination and System Selectivity
Redundancy in data centre design only works if protection is correctly coordinated. During a fault, the protective device closest to the fault should operate, isolating only the affected section of the network. If discrimination is lost, upstream devices may also trip, disconnecting healthy supplies and undermining the resilience architecture.
Protection coordination studies assess grading margins across HV and LV systems to ensure time–current characteristics are aligned. This is particularly critical in multi-source configurations where grid and generator supplies operate in parallel, increasing fault contribution and complexity.
Effective discrimination protects uptime by limiting faults to the smallest possible zone and preserving redundant supply paths.
Arc Flash and Operational Safety
Data centre electrical rooms can contain significant fault energy. Incident energy analysis, commonly assessed using IEEE 1584 methodology, quantifies the thermal hazard associated with a potential arc event.
- Arc flash assessment within data centre design informs:
- Selection of suitable switchgear configurations
- Verification of protection operating times
- Definition of safe working distances and PPE requirements
- Labelling and compliance with applicable standards
Reducing fault clearance time through coordinated protection lowers incident energy levels. This improves personnel safety and supports operational continuity.
Transient Stability and Dynamic Load Response
Resilience in data centre power systems depends not only on redundancy, but on the ability of the electrical network to remain stable under rapid and substantial load variation. High-density GPU clusters can introduce large step changes in demand within milliseconds. If the upstream network, transformers or on-site generation cannot respond dynamically, voltage excursions and frequency deviations may occur, potentially leading to protective trips or equipment malfunction.
In most data centres, essential load is supplied via static UPS systems with integral battery back-up as part of a typical double-conversion design. During upstream disturbances, the UPS inverter is required to maintain output voltage and frequency within its defined limits. Transient stability studies, therefore, assess not only network performance, but also UPS control response, short-term overload capability and battery autonomy during severe load ramping and source transfer events. This confirms that rapid GPU load changes do not force the UPS outside its operating tolerances, initiate bypass operation or compromise critical load support.
All rapid voltage changes must also be assessed against applicable local grid operator and transmission system operator requirements at the point of connection, such as ENA P28, EN 50160, and 61000-3-7. Assessment against these frameworks ensures that internal load dynamics do not adversely impact the public network or breach connection agreements
Effective transient analysis protects uptime by confirming that the combined grid, generator and UPS system remains electrically stable during extreme but credible load transitions, preserving both supply integrity and resilience architecture.
Earthing and Lightning Protection
Earthing systems in data centres must safely dissipate high fault currents while maintaining touch and step voltages within the limits defined by ENA 41-24.
As facilities expand, increases in fault level or changes in soil conditions can alter earth grid performance. Earth potential rise during high magnitude faults must be reassessed to confirm that safety limits and equipment performance remain within acceptable parameters.
Detailed earthing modelling, supported by site-testing where required, ensures the grounding system continues to perform throughout a facility’s lifespan.
A Practical Failure Scenario
Consider a facility that increases transformer capacity to support additional IT load growth. The higher-rated transformers increase prospective fault current at downstream distribution boards. Under normal operation, a double-conversion UPS inherently limits downstream short-circuit current to a defined multiple of rated current for a short duration, due to inverter current-limiting characteristics.
However, when the UPS is placed into maintenance bypass, this current-limiting effect is removed, and the downstream system is exposed directly to the upstream network fault level. In this condition, the low-voltage assembly may be subjected to significantly higher prospective short-circuit current than under inverter-supported operation.
If short-circuit levels and protection settings are not recalculated following the capacity upgrade, circuit breaker interrupting ratings may be exceeded and discrimination margins reduced.
Designing for Scalable Growth Without Performance Drift
Data centres are rarely static. Capacity upgrades, transformer replacement, additional generation and revised grid connection requirements are common over the asset lifecycle.
Scalable data centre design requires verification that each modification preserves system integrity. Expansion should prompt reassessment of:
- Changes in available fault current across the network
- Preservation of selective coordination between protective devices
- Transformer loading and thermal capability
- Voltage performance under contingency conditions
- Grounding system effectiveness under revised fault levels
Independent modelling before expansion protects capital investment and ensures that resilience architecture continues to operate as intended.
How EPS Delivers Resilient Data Centre Design
Electrical resilience is embedded through coordinated studies and independent verification at each stage of the project lifecycle.
EPS supports data centre developers, operators and principal contractors through:
Power System Modelling and Validation – Short circuit analysis to IEC 60909, protection coordination across HV and LV systems, load flow and voltage stability assessment, and generator and UPS fault contribution modelling to confirm system behaviour under worst-case conditions.
Grid Code Compliance and Connection Studies -Fault level impact assessments, protection alignment with DNO requirements and technical support for connection approvals, ensuring reinforced or embedded generation schemes meet network operator expectations.
Arc Flash Risk Assessment – Incident energy analysis and protection optimisation to reduce clearance times and improve operational safety.
Earthing Modelling and Verification – Earth grid modelling to BS 7430 limits, earth potential rise assessment and site verification following reinforcement or modification.
Independent Technical Assurance – Third-party review of electrical design, validation of redundancy architecture and pre-energisation compliance checks.
These services ensure that data centre design decisions are validated before procurement, energisation and expansion.
Notable Projects:
Client: AI Data Centre Developer
Location: Norway
Services: Power System & Arc Flash Studies
Date: June 2025
EPS delivered a fully integrated power system analysis package at the FEED stage for a new AI data centre, validating MV/LV network performance, fault duties and protection grading before construction. Load flow modelling confirmed voltage compliance and transformer headroom under peak demand, while short circuit analysis verified that switchgear remained within rated breaking capacities. Building on the same system model, EPS assessed protection coordination constraints and quantified arc flash incident energy across multiple UPS operating modes, identifying maintenance bypass as the most onerous scenario for LV exposure. The combined studies enabled the client to de-risk equipment selection, refine protection philosophy and implement an informed arc flash labelling strategy before energisation.
The Engineering Standard for Modern Data Centre Design
Resilient data centre design is achieved through verification, not assumption. Availability targets must be supported by evidence that the electrical system performs correctly under credible fault and switching conditions.
For developers, operators and engineering leaders, early-stage system studies prevent redesign, equipment replacement and regulatory challenge later in the project lifecycle.
If you are planning a new data centre project, expanding existing capacity or reviewing the resilience of a live facility, EPS can support your design team with validated engineering studies and independent technical assurance.
