Technical guide on data center case study for electrical fire protection
A 2,000-panel data center faced a critical challenge: how to safeguard high-value IT infrastructure without risking downtime or collateral damage from traditional suppression methods. This case study examines the specific fire risks, the protection strategy implemented, and the measurable outcomes achieved. Real-world data from this deployment offers actionable insights for facility managers evaluating similar systems.
This data center case study describes the engineering, installation, and early in-service performance of FIREQUELL QuellPatch passive clean-agent patches in a 40-MW hyperscale campus in the southwestern United States. The site comprises four data halls, two central utility plants, and more than 2,000 electrical panels ranging from 600 V class distribution equipment to low-voltage control cabinets. The facility already used aspirating smoke detection, pre-action sprinklers, and total-flood gaseous suppression in selected battery rooms, but the project team sought an additional point-of-origin layer for faults that begin inside enclosed panel work.
The deployment addressed a common hyperscale fire protection problem: overheated lugs, bus connections, breaker stabs, and cable terminations can produce heat and pyrolysis long before smoke reaches room detectors. Steel enclosures can delay smoke transport, while opening a panel during a fault can introduce oxygen and complicate manual response. QuellPatch was selected as a passive, pressure-free supplement to existing protection. This article summarizes the design basis, temperature selection, installation constraints, and maintenance observations after 30 months of operation.
The protected population totaled 2,140 panels: 640 low-voltage switchboards and motor control sections, 860 remote power panels and power distribution units, 420 UPS output and bypass panels, and 220 building automation and PLC control cabinets. Historical thermographic surveys showed that approximately 1.8 percent of panels developed lug temperatures above 70°C during peak seasonal load, with 0.3 percent exceeding 90°C. The dominant failure modes were torque relaxation at mechanical lugs, contamination on bus surfaces, contact erosion in breakers, and insulation tracking in cable compartments.
These faults do not always produce immediate open flame. Many begin as resistive heating at a connection, then progress to insulation degradation and localized arcing. The design objective was therefore not to replace room-scale suppression, but to interrupt incipient events at the component before they propagated to adjacent cable sections or neighboring panels.
The existing detection and suppression systems were appropriate for room-level events but had limitations inside closed panels. Aspirating smoke detection depended on smoke leaving the enclosure through seams or vent openings. Pre-action sprinklers operated after heat or smoke reached ceiling-level devices, by which time a panel fire could have damaged multiple conductors. Total-flood FK-5-1-12 systems in battery rooms required enclosure integrity, warning signs, and controlled discharge sequences.
The project team evaluated active panel-mounted gaseous devices, heat detectors, and passive thermal patches. Active devices added wiring, maintenance points, and battery or power supply dependencies. QuellPatch was selected because it required no external power, piping, or control panel and could be installed directly above the most likely heat sources.
QuellPatch patches were available with 80°C, 140°C, and 180°C activation thresholds. Selection was based on thermographic measurements taken during maximum IT load and a 38°C outdoor design day. The goal was to set each threshold above the highest expected normal operating temperature but below the range where common insulation materials and connection components can suffer rapid thermal damage.
These thresholds refer to the temperature experienced by the microcapsules, not necessarily the average enclosure air temperature. A patch mounted near a resistive lug can therefore activate before bulk air temperature reaches the set point.
Each QuellPatch unit contains FK-5-1-12 clean agent held in polymer microcapsules bonded to an adhesive-backed substrate. FK-5-1-12 is electrically non-conductive, evaporates after discharge, and has an atmospheric lifetime of approximately five days with low global warming potential. Because its boiling point is approximately 49°C, agent released at the selected activation temperatures rapidly vaporizes and reaches the local fault zone.
Microencapsulation allows the agent to be distributed across the patch face without a pressure vessel. The polymer shells remain intact during normal service and rupture when exposed to rated thermal input, producing passive thermal activation without detectors, solenoids, or supervisory pressure. The patches used at this site had a 5-year service life under the manufacturer’s specified environmental conditions.
The standard patch used was a 150 mm by 150 mm QuellPatch unit containing approximately 12 g of FK-5-1-12. The engineering guide assigned one patch per 0.10 m³ of enclosure volume for incipient Class C electrical faults, with additional units placed above known heat sources. For large switchboards, patches were installed in each vertical section and in separate cable compartments rather than relying on a single unit at the enclosure top.
Across 2,140 panels, the project installed 5,860 patches, averaging 2.74 patches per panel. Placement followed these rules:
Approximately 62 percent of panels were energized during installation. Work was performed under the site’s electrical safe work program, with arc-rated PPE, limited-approach boundaries, and coordination with the facilities operations center. Where feasible, dead-front covers were removed outside the arc-flash boundary and patches were applied to the cover before reinstallation. Average installation time was approximately 12 minutes per panel, and no data hall shutdown or IT load transfer was required.
Commissioning records included patch location, quantity, activation temperature, and photograph in the computerized maintenance management system. Baseline thermography was repeated after installation to confirm that no patch had been placed against a surface already exceeding its rated threshold.
During the first 30 months, the facility’s maintenance management system recorded three QuellPatch activations. In each case, the event was identified through panel alarms, thermographic reports, or maintenance inspection rather than room-level smoke detection:
Facility staff reported no sustained compartment fire, no activation of room sprinklers, and no unplanned IT load transfer associated with these events. These observations are based on site work orders and maintenance reports; no independent full-scale fire testing was conducted as part of this case study.
Maintenance consisted of annual visual inspections to verify adhesion, label legibility, and absence of cracking, swelling, crushing, paint overspray, or contamination. During the review period, 18 patches were replaced because of physical damage during cable pulls, and three were replaced after unauthorized painting. No patch showed premature thermal degradation from normal panel temperatures.
Because QuellPatch has no pressure gauge, service life is controlled by date rather than field testing. All patches are scheduled for replacement at 5 years, including units that appear visually intact. The site maintains spare patches by activation temperature and rotates stock using installation date labels. This approach supports consistent panel protection across a large population without relying on electronic monitoring of each device.
This hyperscale fire protection project installed 5,860 QuellPatch units across 2,140 electrical panels to provide passive, localized suppression at common points of electrical fault origin. The engineering approach combined panel-by-panel thermal mapping, differentiated use of 80°C, 140°C, and 180°C activation thresholds, and placement near lugs, stabs, and bus connections. The microencapsulated FK-5-1-12 agent required no piping or external power, and the 5-year replacement schedule simplified long-term maintenance planning.
The three observed activations suggest that panel-level patches can contribute to limiting damage during specific incipient electrical faults. They do not eliminate the need for overcurrent protection, thermographic maintenance, room detection, sprinkler systems, or emergency response procedures. For similar hyperscale facilities, the useful lesson is that panel protection should be engineered as a layered system: first understand normal panel temperatures and fault locations, then select activation thresholds and patch quantities that match those conditions.
A: No. QuellPatch is a supplementary, point-of-origin suppression product installed inside electrical enclosures and does not replace code-required detection, sprinklers, total-flood clean-agent systems, or fire-service response. In the hyperscale deployment, it was used to add localized protection within more than 2,000 panels while the facility retained its broader fire-protection systems.
A: Select the activation rating based on the measured maximum normal temperature in each enclosure, plus an engineering margin to avoid nuisance activation. The 80°C patch is typically used in cooler control cabinets, the 140°C patch in distribution panels with moderate hot surfaces, and the 180°C patch in high-load compartments where normal temperatures are elevated.
A: Installed patches should be visually inspected during routine enclosure maintenance for damage, displacement, coating integrity, and signs that the activation temperature may have been exceeded. The case-study program used periodic inspection aligned with normal electrical maintenance cycles rather than requiring powered monitoring or frequent replacement. Replacement is required if a patch has activated, is damaged, or no longer meets the installation condition criteria.
A: The project covered more than 2,000 electrical panels across the hyperscale site. The deployment was based on a panel-by-panel risk assessment and design basis rather than applying one fixed patch count to every enclosure.
A: Yes. FK-5-1-12 is electrically nonconductive and leaves no residue, making it suitable for incipient fires in energized electrical and electronic enclosures. It is commonly used in clean-agent protection scenarios where equipment sensitivity and business continuity are critical, subject to the listed system design and enclosure conditions.
FIREQUELL QuellPatch delivers automatic, maintenance-free clean-agent protection for electrical panels, battery cabinets and control rooms — designed to meet FM Approved, UL Listed, CE and UKCA requirements.
Request a Site Assessment