A data center emergency power checklist is a structured set of operational and compliance actions that confirms every critical power system will activate, transfer load, and sustain operations without interruption. NFPA 110 Type 10, Level 1 requires emergency power supply systems (EPSS) to assume full data center load within 10 seconds of utility grid failure. That 10-second window covers engine cranking, voltage stabilization, and frequency lock under load. The checklist is the tool that proves your system can meet that standard every time. Key components covered include diesel generators, uninterruptible power supplies (UPS), automatic transfer switches (ATS), and power distribution units (PDUs).
1. verify EPSS transfer time and startup sequence
The first item on any data center emergency power checklist is confirming the EPSS transfers load within the 10-second NFPA 110 requirement. This is not a soft target. Inspectors from the Authority Having Jurisdiction (AHJ) treat it as a hard pass/fail criterion. Test the full startup sequence: engine crank, governor response, voltage rise, frequency stabilization, and ATS transfer. Log every phase with timestamps. A single missed transfer during an audit can trigger a compliance deficiency notice.
2. size generators to handle critical load and UPS recharge
Generator sizing is one of the most frequently miscalculated items in emergency power plans. Generators should be sized at 125% of total critical load, or 150% when integrated with UPS systems. The extra capacity exists for a specific reason: UPS battery recharge after an extended outage can draw up to 20% of UPS rating for several hours. That recharge surge hits the generator minutes after startup, when thermal and electrical loads are already climbing. Undersized generators trip under this secondary spike, leaving your facility dark during recovery.
Pro Tip: Add a recharge load simulation to your annual test protocol. Run the generator under normal critical load, then introduce a simulated UPS recharge load to confirm the unit holds frequency and voltage within spec.
3. confirm UPS runtime is calibrated for bridging, not backup
UPS systems provide approximately 5 minutes of battery runtime, designed only to bridge the gap during generator startup. Operators who treat UPS as extended backup increase the risk of data corruption and hardware damage. Your contingency checklist must document the actual measured runtime for each UPS string under current load, not the nameplate estimate. Battery capacity degrades over time. A UPS that delivered 6 minutes of runtime two years ago may now deliver 3.5 minutes under the same load conditions.
Battery Energy Storage Systems (BESS) can complement diesel generators by absorbing short outages, but they do not replace diesel for multi-hour or multi-day autonomy. Diesel remains the gold standard for extended runtime in mission-critical facilities.
4. conduct monthly load testing and prevent wet stacking
Monthly generator testing is mandatory under NFPA 110, and the engine must sustain at least 30% of nameplate load during each test. Running a generator at low or no load causes wet stacking, a condition where unburned fuel accumulates in the exhaust system and degrades engine performance. If your facility cannot achieve 30% load during monthly tests, NFPA 110 requires an annual supplemental load bank test at 100% capacity for a minimum of 2 hours. Load bank testing is the most reliable method to verify true generator performance under real emergency conditions.
Pro Tip: Schedule load bank tests during planned maintenance windows and coordinate with your fuel supplier in advance. A 2-hour full-load test on a large generator consumes significant diesel volume. Running low on fuel mid-test is a compliance and operational failure.
5. derate generator capacity for ambient conditions
Generator nameplate capacity must be derated for ambient temperature, elevation, and fuel quality to avoid thermal shutdowns during actual emergencies. A generator rated at 2 MW at sea level and 77°F may deliver only 1.7 MW at high elevation or during a summer heat event. This is not a theoretical concern. Cooling system compatibility failures create what engineers call “second-order failures,” where backup power is technically present but IT loads shut down because the generator cannot sustain output. Your checklist must include a derating calculation verified against your site’s worst-case ambient conditions.
| Derating Factor | Impact on Output | Checklist Action |
|---|---|---|
| Ambient temp above 77°F | Output reduced 1% per 1.8°F above baseline | Verify cooling system capacity and airflow |
| Elevation above 1,000 ft | Output reduced approximately 3.5% per 1,000 ft | Apply manufacturer derating table to nameplate |
| Fuel quality (water or microbial contamination) | Unstable combustion, power fluctuations | Test fuel quarterly; treat or replace as needed |
| Load step sequencing errors | Voltage and frequency instability on startup | Program ATS to stage loads in defined sequence |
6. inspect transfer switches and pdus on a defined schedule
Automatic transfer switches and PDUs are the mechanical and electrical links between your generator and your IT load. They fail silently. An ATS that tests fine under no-load conditions can fail to close under full load during an actual outage. Your power failure strategies must include quarterly ATS inspection: contact condition, timing relay calibration, and manual override function. PDU inspection should verify breaker ratings match current load assignments, since load growth in active facilities frequently causes breaker mismatches over time.
7. validate fuel levels, quality, and delivery logistics
Fuel is the single variable that determines whether a generator runs for 8 hours or 8 days. Your data center contingency checklist must confirm tank levels before every test and after every significant weather event. Diesel stored for more than 12 months degrades through oxidation and microbial growth. Test fuel quarterly using ASTM D975 standards. Beyond tank levels, validate your fuel delivery logistics and confirm your supplier can reach your site during a regional emergency. A full tank at the start of an outage is only as reliable as your resupply plan.
8. maintain comprehensive compliance documentation
Compliance audits require detailed logs of load in kilowatts, voltage, frequency, oil pressure, coolant temperature, ambient conditions, and technician identification for every generator start and test. Without automated, timestamped records, AHJ inspectors routinely reject test results. Manual logbooks are no longer sufficient for facilities operating under NFPA 110 Section 8.4.2. Automated monitoring systems that capture and timestamp every parameter in real time are the current standard. Required documentation includes:
- Monthly test logs with all measured parameters and technician ID
- Annual load bank test report with pre-test and post-test fuel analysis
- Transfer switch inspection records with contact condition ratings
- UPS battery health reports with measured runtime under load
- Fuel delivery receipts and tank level logs with dates and quantities
- Deficiency logs documenting every identified issue and its resolution date
- AHJ inspection reports and any corrective action correspondence
“Incomplete documentation is the leading cause of compliance failures during AHJ inspections. A generator that performs perfectly but lacks a complete paper trail is treated the same as one that failed to start.”
9. plan for fuel storage, secondary containment, and site access
Fuel storage logistics affect site viability in ways that are expensive to fix after construction. Secondary containment requirements, regulatory permits for above-ground storage tanks, and emergency vehicle access routes must all be validated during site selection. Facilities that skip this step during design face costly retrofits or operational compromises when they need fuel most. Your site-specific checklist should address:
- Tank capacity relative to runtime requirements at full load
- Secondary containment compliance with EPA Spill Prevention, Control, and Countermeasure (SPCC) regulations
- Delivery truck access routes, including weight limits on access roads and gate clearance dimensions
- Fuel resupply contracts with guaranteed response times during declared emergencies
- Coordination with local fire marshals on permit requirements for diesel storage volumes
Backup power planning that starts at the site selection phase avoids the most common and costly fuel logistics failures. Facilities that treat fuel access as an afterthought consistently face the worst outcomes during extended outages.
10. build storm readiness into your emergency power plans
Severe weather is the most common trigger for extended grid outages, and 73% of data center outages are preventable with proper preparation. Storm readiness in your emergency power plans means more than a full fuel tank. It means pre-positioning fuel delivery orders before a storm makes roads impassable, confirming generator coolant levels and block heater function before cold weather events, and verifying that outdoor fuel tanks have adequate freeze protection. UPS runtime of approximately 5 minutes provides no meaningful buffer during a multi-day weather event. The generator and its fuel supply are the only systems that matter for extended storm outages.
Pro Tip: Establish a pre-storm checklist that triggers 72 hours before a forecast severe weather event. Include a fuel top-off order, generator test run, and ATS manual transfer verification as standard items.
Key takeaways
A complete data center emergency power checklist must integrate NFPA 110 compliance, proper generator sizing, load testing, fuel logistics, and automated documentation to guarantee uptime during any power failure.
| Point | Details |
|---|---|
| EPSS transfer time | Verify load transfer within 10 seconds per NFPA 110 Type 10 at every scheduled test. |
| Generator sizing | Size at 150% of critical load when UPS is integrated to absorb recharge surges. |
| Monthly load testing | Sustain 30% nameplate load monthly or complete an annual 2-hour load bank test at 100%. |
| Fuel and delivery logistics | Validate tank levels, fuel quality, and resupply contracts before every major weather event. |
| Automated documentation | Use timestamped, automated logs for every test parameter to pass AHJ compliance audits. |
What i’ve learned after years of watching facilities get this wrong
The most common failure I see is not a generator that won’t start. It’s a generator that starts fine but trips 8 minutes later because nobody accounted for the UPS recharge load spike. Facilities spend significant resources on generator procurement and installation, then skip the one test that would expose this vulnerability. Running a recharge simulation during your annual load bank test costs almost nothing extra and catches a failure mode that has taken down more than a few well-funded facilities.
The second pattern I see constantly is backup power planning that starts too late. By the time a facility is in detailed design, the tank location is fixed, the access road is graded, and the fuel delivery logistics are locked in by site geometry. The operators who avoid the worst outcomes are the ones who treat fuel storage and access as a design constraint from day one, not a facilities management problem to solve after the ribbon cutting.
Documentation is the third area where I see facilities consistently underinvest. Automated monitoring systems that capture timestamped logs for every generator start are not expensive relative to the cost of a failed compliance audit. The facilities that get cited are almost always the ones running manual logbooks with incomplete entries. The technology to fix this is mature and widely available. There is no good reason to still be doing this by hand.
My honest recommendation: treat your emergency power checklist as a living document reviewed quarterly, not an annual compliance exercise. The facilities with the best uptime records are the ones where the checklist drives continuous improvement, not just a checkbox before an inspection.
— Justin
Keep your generators fueled and ready when it counts
Datacenterfuels, in partnership with Anytime Fuel Pros, delivers on-site diesel and DEF directly to data centers on a schedule that fits your operations and your compliance requirements. Whether you need routine tank top-offs before monthly load tests or emergency resupply during a regional weather event, the service is built for critical infrastructure.
Anytime Fuel Pros operates 24/7 with guaranteed response times and quality-verified ultra-low sulfur diesel (ULSD) that meets ASTM D975 standards. Fuel quality is not a variable you want to manage during an active outage. Explore on-site diesel delivery for data centers, or use the fuel delivery questions guide to evaluate whether your current supplier meets the standard your facility requires.
FAQ
What is NFPA 110 type 10 for data centers?
NFPA 110 Type 10 requires emergency power supply systems to transfer full critical load within 10 seconds of utility failure. The 10-second window includes engine cranking, voltage stabilization, and frequency lock under load.
How often should data center generators be load tested?
Monthly testing is mandatory under NFPA 110, with the generator sustaining at least 30% of nameplate load. If monthly tests cannot reach that threshold, an annual 2-hour load bank test at 100% capacity is required.
What causes generator tripping during power recovery?
UPS battery recharge after an extended outage draws up to 20% of UPS rating for several hours. This secondary load spike hits the generator during recovery and trips undersized units that were not sized to handle it.
How long does UPS battery runtime last in a data center?
UPS systems are designed to provide approximately 5 minutes of runtime, covering only the generator startup gap. Relying on UPS for extended outages increases the risk of data corruption and hardware failure.
When should fuel logistics be planned for a new data center?
Fuel storage capacity, secondary containment, and delivery access must be validated during site selection. Planning these elements after construction is complete leads to costly retrofits and operational gaps during emergencies.