How to Prepare a Protection Coordination Study for a Data Center

Why a Protection Coordination Study Matters for Your Program

A protection coordination study determines how every protective device in your data center’s electrical distribution system responds to a fault. It defines which device trips first, how fast it operates, and whether upstream devices hold off long enough to let the closest device clear the problem. Get this wrong, and the consequences show up during commissioning — failed witness tests, cascading trips that shouldn’t happen, and schedule days spent troubleshooting settings that should have been right from the start.

For AE, MEP, and EPC prime contractors managing Tier III/IV data center programs, the coordination study sits at the front of a dependency chain that gates your energization schedule:

Protection coordination study → relay settings → arc flash calculations → labels → energization clearance.

Every delay in the upstream steps pushes the entire chain. This guide walks through the process from initial data collection through commissioning-ready deliverables.

Step 1: Assemble the Single-Line and Input Data

The coordination study starts with complete input data. Missing or inaccurate inputs force conservative assumptions that inflate clearing times, which cascade into overly restrictive arc flash numbers and PPE requirements.

What you need:

Single-line diagram — as-designed for new construction, as-built for brownfield. This is the foundation for the power system model.
Utility contribution — short-circuit data at the point of common coupling (PCC) from the serving utility.
Transformer data — impedance, tap settings, and winding configuration for every transformer in the distribution path.
Generator data — subtransient reactance and decrement curves for each generator. For Tier IV 2(N+1) configurations, this includes every unit in every distribution path.
Cable data — lengths, sizes, and impedance values for feeder and distribution cables.
Protective device data — type, rating, and time-current characteristics for every breaker, fuse, relay, and controller in the system.

Where data stalls: In multi-vendor environments, input data arrives from different manufacturers in different formats and at different times. Generator controller parameters from the gen-set OEM frequently arrive late in procurement. Flag these early — the coordination engineer cannot finalize the model without them.

Step 2: Build the Power System Model

With input data assembled, the next step is building a power system model in software — typically ETAP, SKM, or PowerFactory. The choice of tool matters less than the fidelity of the model.

Every protective device in the distribution system needs to be modeled with its actual characteristics, not generic library defaults. The model represents the real electrical behavior of the system under fault conditions.

The model produces two foundational analyses:

Load flow study — confirms that the system can carry its intended loads under normal and contingency operating conditions.
Short-circuit study — calculates maximum and minimum fault currents at each bus in the distribution system. Both extremes matter: maximum fault current determines interrupting duty and device ratings; minimum fault current determines whether protective devices can reliably detect and clear faults.

For Tier III/IV data centers with 2(N+1) redundancy architecture, the model must account for every operating configuration — all sources online, N-1 contingency, maintenance bypass, and islanded operation. Each configuration produces different fault current magnitudes, and the protection settings must work correctly across all of them. A distributed control architecture adds operating mode transitions that change the protection topology dynamically.

Step 3: Set the Protection Coordination Criteria

With fault currents calculated, the engineer defines the coordination criteria — the rules that govern how protective devices interact during a fault.

Selectivity is the primary objective: the device closest to the fault should clear it, without tripping upstream devices that protect other parts of the system. When selectivity works, a fault on one distribution path doesn’t take down adjacent paths. When it doesn’t, a single fault cascades into a broader outage — exactly the scenario Tier III/IV designs are built to prevent.

Coordination criteria vary by device pair. Relay-to-relay coordination uses time-current curve separation with defined margins. Relay-to-fuse coordination accounts for fuse total clearing time against relay pickup. Each pair has different margin requirements per IEEE 242 (the “Buff Book”), which defines the methodology for protective device coordination.

The selectivity-versus-speed trade-off: Faster fault clearing reduces arc flash incident energy — a direct safety benefit. But faster settings can sacrifice selectivity, causing upstream devices to trip unnecessarily. The engineer must decide where to accept this trade-off and where to resolve it.

Zone-selective interlocking (ZSI) resolves this tension for many device pairs. ZSI uses relay-to-relay communication (typically over IEC 61850 GOOSE messaging) to allow instantaneous clearing at the faulted zone without sacrificing selectivity upstream. Where ZSI is implemented, you get both fast clearing and selective coordination.

The output of this step is a selectivity matrix — a table documenting which device pairs are selectively coordinated, which are not, and where intentional non-selectivity has been accepted with engineering justification.

Step 4: Develop the Time-Current Curves and Relay Settings

This is the core engineering work: translating the coordination criteria into specific device settings.

Time-current coordination plots are the visual representation. Each plot shows the pickup curves and time delays of every protective device in a coordination path, from the load-side devices up to the source. The engineer works bottom-up — setting the load-side devices first, then working upstream with defined margins between each pair.

In a multi-vendor environment, this is where platform breadth matters. Each relay manufacturer has different parameter structures, naming conventions, and configuration tools. Settings that make sense in one vendor’s ecosystem need to produce equivalent protection behavior when they interact with devices from another vendor across the distribution system. A data center with protective devices across eight manufacturer platforms has eight different toolchains producing settings that must coordinate as a single system.

Every setting must be traceable to a design decision. “Why is this pickup set to this value?” should have a documented answer rooted in the coordination study — not inherited from a previous project or copied from a manufacturer default. Design-basis documentation turns relay settings from tribal knowledge into auditable engineering.

The deliverable from this step is not just a report. It includes ready-to-load settings files — actual configuration files that can be loaded directly into field IEDs. This eliminates the transcription step where a field engineer re-interprets study recommendations into device configurations, removing a common source of errors.

Planning a Protection Coordination Study?

Get Coordination Settings That Match Your Design Intent

We deliver PE-stamped coordination studies with ready-to-load settings files across every vendor platform in your spec.

Book a Scoping Call See Our Services

Step 5: Recalculate Arc Flash with Final Settings

Once protection settings are finalized, the arc flash incident energy must be recalculated using those actual settings — not the preliminary or default settings from earlier in design.

Clearing time is the dominant variable in IEEE 1584 arc flash calculations. The difference between a relay that trips in 50 ms and one that trips in 300 ms can shift PPE requirements by multiple categories. Labels based on conservative assumptions may over-restrict operations (increasing operational burden for the facility owner) or, if assumptions were optimistic, under-protect workers.

The recalculation uses the same power system model from Step 2, updated with the final protection settings from Step 4. The output is a set of incident energy values and arc flash boundary distances for every piece of equipment — the data that goes onto the arc flash labels required by NFPA 70E before energization.

Step 6: Package Deliverables for Commissioning

The coordination study is complete when five deliverables are ready for your commissioning team:

Protective device coordination study report — PE-stamped, documenting the methodology, criteria, and results.
Selectivity matrices — showing coordination status for every device pair, with justification for any accepted non-selectivity.
Relay settings calculations with design-basis documentation — every setting traceable to a coordination decision.
Ready-to-load settings files — configuration files for every protective device, across all vendor platforms in the system.
Arc flash incident energy recalculations — reflecting the final protection settings, ready for label production.

This package is what your commissioning team and the owner’s commissioning agent verify against during witness testing. The L1-L5 commissioning framework lists protection coordination as a pre-commissioning readiness criterion — without it, the verification sequence cannot proceed properly. Use the commissioning readiness checklist to assess whether your program has the documentation and procedures in place before testing begins.

What Can Go Wrong

Three patterns account for most coordination study failures on data center programs:

Study and settings done by different firms. When one firm produces the coordination study and a different firm programs the relay settings, every translation between the study’s recommendations and the field device configurations is an integration boundary. Settings that look correct on paper may not produce the intended protection behavior when loaded into actual devices — and the gap surfaces during commissioning, not before.

Study based on preliminary data, never updated. Design-phase coordination studies use the best data available at the time. When as-built conditions differ — different transformer impedances, revised cable routes, generator substitutions during procurement — the study must be updated. Settings based on preliminary data do not match field reality, and the mismatch shows up during functional testing.

Arc flash calculated with default settings. When arc flash analysis is performed before the coordination study is finalized, it uses assumed clearing times rather than actual relay settings. The resulting labels may not reflect the real incident energy at each piece of equipment. Recalculation with final settings is not optional — it is the step that connects the coordination engineering to the safety documentation.

The schedule cost of discovering these issues during commissioning is measured in weeks, not days. The coordination study is where you invest the engineering time to prevent that outcome.

For guidance on selecting a firm to perform this work, see How to Evaluate a P&C Engineering Subcontractor.