BESS commissioning punch list for data center energy storage

BESS commissioning, cold to hot

A BESS, a battery energy storage system, is a packaged plant that stores electrical energy in battery cells and pushes it back out through an inverter on command. Commissioning a BESS is the staged process of proving that plant is installed right and works right, and it runs in a hard sequence: cold first, then hot. Cold means de-energized, the static checks you do with the system dead. Hot means energized, the functional and performance tests you do with the system live and moving real power.

The arc from cold to hot is not a suggestion. You do not energize a BESS with open cold deficiencies, because the things you skip cold are the things that hurt people hot. A reversed DC polarity, an under-torqued bus joint, a fire panel left in bypass, a coolant loop full of air. None of those announce themselves until current flows, and by then the cheap fix has become an incident.

The punch list is the spine of all of it. It is the tracked list of every deficiency found, sorted by whether it blocks energization or blocks turnover, and it is what the owner signs against at acceptance. A BESS that was energized and cycled but never had its punch list closed and witnessed was not commissioned. It was just turned on.

What is inside a BESS: cells, BMS, PCS, transformer, EMS

Before you can commission a BESS you have to know what is in the box, because each piece has its own checks. The energy lives in cells, the smallest unit, grouped into modules, which are stacked into racks or strings. A string of modules in series builds the DC voltage, commonly somewhere in the hundreds of volts up past 1000 V DC on utility-scale gear, and that voltage exists the moment the cells are charged, switch or no switch.

The battery management system, the BMS, watches every module and usually every cell: voltage, temperature, current, and state of charge. It balances cells, it trips the battery on a limit, and it reports up to the controller. The power conversion system, the PCS, is the bidirectional inverter that turns DC into grid-synchronized AC to discharge and AC back into DC to charge. A step-up transformer often sits between the PCS and the medium-voltage gear.

The energy management system or site controller, the EMS, dispatches the PCS and coordinates the racks. Thermal management keeps the cells in their operating band, by forced air on older or smaller systems and by a pumped liquid loop, usually water-glycol, on most modern high-density enclosures. And the life-safety stack sits over all of it: smoke and heat detection, off-gas and combustible-gas detection, suppression, and deflagration venting or explosion control for the enclosure.

What is cold commissioning of a BESS?

Cold commissioning is everything you verify with the BESS de-energized, before any controlled energization. It is the static phase, and it is where most of the real defects get caught, because once the system is hot the cost and risk of fixing anything climbs fast.

The cold scope is broad: mechanical install verification, torque of every power connection, DC polarity, DC bus insulation and continuity, grounding and bonding, BMS communication and addressing, the thermal loop, the fire and gas detection and suppression systems, enclosure sealing, signage, and spill containment. Each of those is its own punch item with its own sign-off.

One rule governs the phase. Treat the battery as live the entire time, because it is. Cold commissioning is de-energized at the PCS and the AC side, but the cells hold full string voltage at the DC terminals throughout. Cold means the system is not converting power. It does not mean the DC bus is safe to touch.

Mechanical install, torque, and DC polarity

Start with the mechanical install, because the electrical checks assume it is right. Confirm racks and enclosures are anchored to the pad or structure per the drawings and the seismic detail, that modules are fully seated and latched, that working clearances and egress around the enclosure match the listing and the code, and that nothing shipped loose is still loose. A module that is not fully seated reads fine on the BMS and then arcs at the connector under load.

Then torque every power connection to the manufacturer's value and mark it. DC bus bars, module interconnects, PCS terminals, transformer terminations, and the AC landing. Under-torqued joints are the classic BESS hot spot. They run hot, they oxidize, and on a DC bus there is no zero crossing to help an arc self-extinguish, so a loose DC joint that starts to arc keeps arcing. Use a calibrated torque wrench, follow the sequence the maker specifies, and put a witness mark on each joint so a thermal scan later can tell you if one moved.

Verify DC polarity before anything gets connected to the PCS. Reversed polarity on a battery string into an inverter is not a nuisance. It destroys hardware and can start a fire. Meter every string for correct polarity and expected open-circuit voltage, and confirm the polarity at the PCS DC input matches. This is a two-person check on a system you cannot turn off, so treat the leads as live and the voltage as real.

Grounding, bonding, and DC bus checks

Grounding and bonding is the check that the fault path actually exists. Confirm the enclosure, the racks, the PCS, and the transformer are bonded to the equipment grounding system and to the grounding electrode per the design, and that the bonding jumpers are landed and tight, not just present. The ground is what carries fault current long enough to trip protection and what holds touchable metal near earth potential. An open bond on a BESS enclosure is a shock and arc hazard hiding in plain sight.

Most utility-scale BESS run an ungrounded or high-resistance-grounded DC bus monitored by an insulation monitoring device, so the grounding scheme is not the same as a solidly grounded AC system. Confirm the actual grounding topology against the design before you assume anything, because what you bond and how depends on whether the DC system is grounded, ungrounded, or resistance-grounded.

On the DC bus itself, run insulation-resistance tests to ground on the de-energized sections you can isolate, using the test voltage and method the manufacturer allows, and check continuity across the bus. Be careful here. A battery string is a charged source you cannot meg the way you would a dead cable, so insulation testing on a BESS follows the manufacturer's procedure, not habit carried over from switchgear work. NETA acceptance testing practice and IEEE 43 cover insulation-resistance methods for the electrical gear, and recent NETA editions added coverage specific to battery energy storage.

BMS communication, addressing, and trip checks

The BMS gets commissioned before the battery does any work, because the BMS is what protects the battery and what tells you whether anything else is wrong. Confirm every module is addressed correctly, with no duplicate or missing addresses on the communication bus. A duplicated address makes one module invisible, and the BMS will happily report a system that looks complete while a module sits unmonitored.

Walk the communication path end to end: module to rack controller, rack controller to system BMS, system BMS to the EMS or site controller. Confirm the telemetry is sane and not just present. Cell voltages should sit in a tight band at the shipped state of charge, temperatures should read close to ambient and close to each other, and any reading that is wildly off is either a bad sensor or a real problem. You find out which one now, cold, not during a discharge.

Verify the BMS alarm and trip thresholds are set to the values the manufacturer specifies, and that a simulated or forced limit actually opens the contactor it is supposed to open. A BMS whose trips were never proven is a protection system you are taking on faith, and on a battery the BMS is the protection that matters most.

Thermal loop: coolant, flow, and leak check

Thermal management keeps cells inside the temperature window where they are safe and where they make their rated capacity, so the loop gets commissioned cold before the cells generate any heat of their own. On a liquid-cooled system, fill with the specified coolant, usually a water-glycol mix, confirm the glycol concentration with a refractometer against the freeze and heat-transfer spec, and bleed the loop. Air trapped in a liquid loop is the quiet killer: it blocks flow to a module, that module runs hot under load, and the BMS derates or trips.

Pressure-test or leak-check the loop, confirm pump rotation and flow at each circuit, and verify the coolant distribution unit or chiller starts, holds setpoint, and reports to the controls. A coolant leak inside a battery enclosure is both a thermal problem and an electrical one, so the leak check is not optional, and the containment for a leak is part of the same scope.

On an air-cooled system the checks are simpler but the same in spirit: confirm fans run the right direction, filters are clean and installed, dampers and louvers move, and the airflow path is not blocked or short-circuiting hot discharge back into the intake. Either way, prove the thermal system holds setpoint cold, so that when the cells start working you are watching the battery, not chasing the cooling.

Fire detection, gas detection, suppression, and enclosure integrity

The life-safety systems get proven cold, and this is the section crews are most tempted to wave through and most likely to regret. Confirm smoke, heat, and off-gas or combustible-gas detection respond to a test stimulus and report to the panel and to the building fire alarm where required. Gas detection on a BESS matters because a failing lithium cell vents flammable and toxic gas before it ignites, and the gas detectors are what give early warning and trigger ventilation or shutdown. Bump-test or calibrate the gas sensors against the manufacturer's procedure, because a gas detector that was never proven is a blank gauge.

Verify the suppression system is installed, charged, and correctly armed. Confirm it is not accidentally left active during work, where it could discharge on a crew, and not left in bypass, where it will sit dead when it is needed. Clean-agent, water, or aerosol systems each have their own checks; confirm agent cylinder pressure, nozzle coverage, releasing-panel logic, and any interlocks to the HVAC and the PCS. Deflagration venting or explosion control for the enclosure is part of the listing, so confirm vents and any explosion-prevention ventilation are installed as the system was tested and listed, and verify the gas-detection-to-ventilation interlock actually runs.

Enclosure integrity ties into all of it. A suppression system that relies on holding an agent concentration needs the enclosure sealed, so check gaskets, door seals, and cable and conduit entries for a real seal, not daylight. While you are at it, confirm the required signage and labels are in place: the ESS placard, DC voltage and arc-flash warnings, emergency shutoff identification, and the disconnect labeling the code and the AHJ require. And confirm spill and containment provisions, an electrolyte spill kit for the chemistry on site and secondary containment for liquid systems, are present before the system goes live.

Why is a BESS dangerous even when it is switched off?

A BESS is dangerous de-energized because the energy is chemical, and you cannot switch chemistry off. Open every breaker and every disconnect, lock them out, and the battery strings still sit at full DC voltage at their terminals, because the cells are charged. There is no position of any switch that brings the cell stack to zero volts. That single fact is what makes a BESS different from almost everything else an electrician locks out.

DC arc flash is the second hazard, and it is worse than people raised on AC expect. An AC arc gets help twice a cycle when the current crosses zero. A DC arc has no zero crossing, so once it strikes it wants to sustain, and the available fault current from a battery bank is enormous. Treat the DC side with arc-flash respect, use the boundary and PPE the study calls for, and do not work the bus energized when you have any way not to.

Thermal runaway is the hazard that defines BESS safety. A cell that is overcharged, overheated, internally shorted, or physically damaged can go exothermic, heat its neighbors, and cascade, venting flammable and toxic gas as it goes and reigniting after it looks out. You do not put out a thermal runaway the way you put out a fire. You try to keep it from spreading and you protect people and exposures. That is why the gas detection, the venting, the spacing, and the BMS trips are not paperwork. They are the layers that keep one bad cell from becoming a destroyed enclosure. And lockout has limits here: LOTO can isolate the AC and DC connections, but it does not make the racks safe to touch, so the whole system gets treated as energized until proven otherwise at the point of work.

What is hot commissioning, and how do you energize a BESS safely?

Hot commissioning is everything you do with the system energized, and it starts with a controlled, sequenced energization, not a single switch thrown by one person. The cold punch list has to be closed first. Energizing over open cold deficiencies is the mistake that turns a punch item into an incident.

Energize in stages and in the order the manufacturer specifies, which generally means bringing up the AC side first. Energize the medium-voltage gear and the transformer, bring up control and auxiliary power, and let the PCS and the controls power up and run their self-tests before any battery is connected to the inverter. The DC link in a PCS is usually precharged through a resistor or a soft-start circuit before the main DC contactors close, so the bus comes up controlled instead of slamming the capacitors with battery inrush. Confirm that precharge works, because a failed precharge that gets forced closed is how you weld a contactor or blow a fuse on the first energization.

With the PCS up and the DC connected, verify the things you could only check live. Confirm the EMS reads BMS state of charge, state of health, cell voltages, and temperatures, and that the values agree across the BMS, the PCS, and the EMS, because a mismatch means one of them is reading wrong and you do not want to learn that during a capacity test. Confirm the insulation monitoring device sees the DC bus and actually alarms on a simulated ground fault, since on an ungrounded DC system the device is your only warning that the first ground fault has happened. Then bring the PCS to grid-synchronized operation and run a small, controlled charge and discharge to prove the basics before you commit to a full cycle.

How do you test BESS capacity and round-trip efficiency?

You test capacity by fully charging the BESS to 100 percent state of charge, discharging it at rated power down to the minimum state of charge the manufacturer allows, and measuring the energy that actually came out at the AC terminals. Compare that measured energy, in kWh or MWh, against the nameplate. A common acceptance threshold is on the order of 95 percent or more of rated capacity, but the number is set by the project specification and the supply contract, not by a rule of thumb, and the warranty often ties to it.

Round-trip efficiency comes from the same test. It is the energy you got out on discharge divided by the energy you put in on charge, measured at the same point, usually the AC side. Modern lithium-ion BESS commonly land somewhere around 85 to 92 percent round-trip at the AC terminals, lower than the bare DC cell efficiency because the PCS, the transformer, and the auxiliary loads, including the cooling, all take a cut. Confirm where the measurement point is defined, because AC-to-AC and DC-to-DC efficiency are different numbers and the spec has to say which one it is buying.

Run the test under conditions that match the spec: the right power, the right temperature, the auxiliary loads accounted for, and the metering accurate enough to mean something. The mistake here is accepting the nameplate on faith because the system charged and discharged. A BESS that cycled is not a BESS that proved its capacity. The capacity and efficiency test is the one that finds the weak string, the derated module, and the cooling that eats more than the spec assumed, and it is the test most likely to get cut for schedule. Do not cut it.

Protection, trip, and emergency-stop testing

Protection and trip testing proves the BESS comes off line safely when it has to, and it is tested live but under control. Verify the protective relaying and the PCS protection functions operate at their settings: AC overcurrent and ground fault, DC overcurrent, over and under voltage, over and under frequency, and the anti-islanding function that keeps the system from energizing a dead utility line. Each function gets proven by injection or a controlled condition, not assumed from the settings sheet.

Prove the layers that are specific to the battery. The BMS trips on cell voltage, temperature, and current limits have to actually open the battery contactors, the emergency stop has to drop the system to a safe state from every station that has an E-stop button, and the interlocks between the fire and gas systems and the PCS have to do what the sequence says, shutting down or ventilating on a real signal. A trip that was set but never operated is not protection.

Coordinate the trips so the right device clears the fault. The point of a coordination study is that a fault sheds the smallest piece of the system, not the whole plant, and the commissioning test is where you confirm the real hardware behaves the way the study said. Where the test and the study disagree, the study gets revisited or the settings get corrected, and then you re-test and log it.

Interconnection functional tests and grid support functions

Where the BESS ties to the utility or a microgrid, the interconnection gets its own functional tests, and in North America the governing framework is IEEE 1547, the standard for interconnecting distributed energy resources with the electric power system. The 2018 revision is the current basis, and its companion test standard, IEEE 1547.1, gives the procedures used to verify a resource complies. The utility interconnection agreement and the local jurisdiction set which requirements and settings actually apply, so confirm those before you test.

The functions to verify by topic include voltage and frequency trip settings and timing, voltage and frequency ride-through, which keeps the system connected through temporary sags, swells, and frequency excursions instead of dropping off and making a disturbance worse, and the grid-support functions the interconnection requires, such as volt-VAR and frequency-watt response and a controlled return to service after a trip. IEEE 1547-2018 sorts ride-through into performance categories, with higher categories staying connected through deeper and longer disturbances, so confirm which category the interconnection requires.

Anti-islanding deserves a specific mention because it is a safety function, not just a power-quality one. The BESS must not keep a section of utility line energized after the utility has opened, both for line-worker safety and to allow reclosing. Prove it operates, prove it operates within the required time, and log the result against the setting the utility specified. Do not infer it from the data sheet.

Integration with the data center power chain and the IST

In a data center the BESS rarely stands alone. It sits inside the power chain with the utility, the generators, the UPS, and the switchgear, so the last and most demanding test is the integrated systems test that proves the whole chain behaves together. This is the same IST that closes out generator and UPS commissioning, and the BESS is one more source that has to transfer, ride through, and coordinate when the script forces a failure.

How the BESS is used decides what integrated success looks like. A behind-the-meter or BESS-as-UPS arrangement uses the battery and PCS to carry critical load through the gap when the utility drops and the generators start, so the integrated test has to prove the BESS picks up the load fast enough and holds it until the generator accepts the block, with the transfer happening in the right order and inside the voltage and frequency the load tolerates. A grid-services or peak-shaving BESS has different success criteria, set by its dispatch role and the interconnection.

The failures the IST is built to find are the timing and coordination faults that each piece passing alone will hide. The PCS rides through fine on the bench, the generator starts fine on its own load bank, and then under the real transfer the sequence is a half second off and the load drops. You only catch that with the live, loaded, scripted test, and the script comes from the basis of design and the commissioning plan. A thin script produces a clean report that proves very little, which is exactly the cross-link to the power-QA scope of the project.

What are the acceptance criteria for BESS commissioning?

Acceptance is judged against measured performance and proven safety, not against the fact that the system turned on and cycled. The BESS has to meet its rated capacity and round-trip efficiency, respond to dispatch commands within the required time, report state of charge accurately, hold its protection and trip settings, keep cells inside the thermal window under load, and pass the fire and life-safety verification, all to the numbers in the project specification and the manufacturer's data.

The values in the table below are typical commissioning ranges to know what to watch and roughly where the lines fall. They are not a standard you can cite. The capacity threshold, the efficiency floor, the response-time target, the SOC accuracy, and the regulation bands all come from the contract documents, the equipment data sheets, and the interconnection requirements, and a data center carrying critical load may hold tighter than a grid-services site. Hold to the contract numbers, and where the spec is silent, fall back to the manufacturer's published values and get the gap resolved in writing before turnover.

One criterion gets forgotten in the rush to call a number: the system has to hold. A capacity figure hit once on a perfect afternoon is not the same as a system that makes its rating cycle after cycle at the temperatures the site will actually see. Where the spec allows, prove repeatability, not a single lucky run.

Fire and life-safety acceptance: NFPA 855, UL 9540, and UL 9540A

The fire and life-safety acceptance is its own gate, and on a BESS it carries the same weight as the electrical performance, because this is the chemistry that can burn. NFPA 855, the standard for the installation of stationary energy storage systems, is the installation framework most jurisdictions point to, and it pulls in fire detection, gas detection, suppression, spacing and separation, ventilation and explosion control, and the hazard mitigation analysis for the installation. NFPA 855 runs on a three-year cycle and the 2026 edition expanded its scope and testing expectations, so confirm which edition the AHJ has adopted.

The system itself is generally expected to be listed to UL 9540, the safety standard for energy storage systems and equipment, which evaluates the battery, the PCS, the controls, and the enclosure as a system. Separate from that listing is UL 9540A, which is a test method, not a listing, for evaluating thermal-runaway fire propagation at the cell, module, unit, and installation levels. The UL 9540A test data is what drives the real installation decisions: how far apart units sit, what separation and barriers are needed, and what the ventilation and explosion control has to do. Confirm the project has the UL 9540A data and that the installation matches what that data and the listing assumed.

Verify the life-safety pieces by topic and to the project's fire protection design, not to a section number you are guessing at. Confirm gas detection, smoke and heat detection, the suppression system, deflagration venting or explosion-prevention ventilation, and the spacing and barriers, and confirm the interlocks between them all operate. Deflagration venting and explosion prevention are often handled through the venting and explosion-control standards the design references, and the local fire code and the AHJ govern what is actually enforced. When the standard, the listing, and the local code disagree, the stricter controlling document and the AHJ win.

The punch list workflow: cold punch, hot punch, closeout

The punch list is how the deficiencies get tracked, sorted, and closed, and a BESS punch list runs in two clear stages tied to the cold-to-hot arc. The cold punch is every deficiency found in the de-energized phase, and the rule on cold punch items is hard: the ones tagged as blocking energization have to be closed, verified, and signed before the system is energized. That is the gate that keeps a static defect from becoming a hot incident.

The hot punch is the deficiencies found during energized functional and performance testing: a protection function that did not operate at its setting, a capacity result short of spec, a control point that does not read right, a thermal channel that runs hot. Each item gets logged with enough detail that someone who was not there can understand it, a severity that says whether it blocks turnover or is a minor closeout item, and an owner responsible for the fix.

Closeout is where the open items get resolved, re-tested, and re-witnessed, and the witnessed acceptance is where the owner, the commissioning agent, and usually the engineer of record sign against the closed list and the test records. Track every item to closure with its before and after, because a punch list with items quietly marked complete and never re-verified is worse than no list. It looks like proof and is not. The signature at the end is only worth what the tracking behind it is worth.

What to document

The commissioning record is what a future operator and the warranty both rely on, so it has to let someone who was not on site reconstruct the test and check the result against the spec. Capture the static checks, the energization, the performance numbers, the protection results, the fire and life-safety verification, and the punch list with its closures, all tied to the specific equipment by serial number and rating.

Record enough that the result is defensible years later: the torque and insulation values, the polarity confirmation, the BMS addressing and trip verification, the capacity and round-trip-efficiency measurements with the conditions they were taken under, the protection and interconnection test results against their settings, the fire and gas system verification, and who witnessed each acceptance. If anything was adjusted and re-tested, record the before and after so the next person can see what changed and that it was proven after the change.

Common mistakes

• Energizing the system with open cold punch items that were tagged as blocking energization.
• Accepting nameplate capacity on faith because the system charged and discharged, instead of running the capacity test.
• Acknowledging BMS faults away to keep testing moving, masking a real cell or sensor problem.
• Leaving the fire, gas, or suppression system in bypass or test and never proving it operates.
• Reversing DC polarity into the PCS, or skipping the polarity check on a string you cannot turn off.
• Treating lockout as if it makes the battery safe to touch, when the racks hold full voltage regardless.
• Insulation-testing the battery the way you would a dead cable instead of following the manufacturer's procedure.
• Running the thermal loop with air trapped in it, starving a module and chasing the trip as an electrical fault.
• Testing the BESS alone and never running the integrated systems test that proves the transfer.
• Marking punch items complete without re-testing and re-witnessing the fix.

Field checklist

Standards and references

Several bodies govern different parts of BESS commissioning, and naming the right one for the point is what separates a credible record from a guess. NFPA 855, the standard for the installation of stationary energy storage systems, is the installation and fire-safety framework, covering detection, suppression, spacing, ventilation, explosion control, and hazard mitigation analysis. It runs on a three-year cycle and the 2026 edition expanded its scope, so confirm the adopted edition with the AHJ. UL 9540 is the system safety standard the BESS is generally listed to, and UL 9540A is the separate thermal-runaway fire-propagation test method whose data drives spacing, separation, and fire protection.

On the electrical side, the NEC, NFPA 70, covers the installation, with Article 706 specific to energy storage systems and its disconnecting-means and shutdown requirements, alongside the usual articles for the AC and DC wiring and overcurrent protection. NFPA 70E covers electrical safety and the arc-flash work practices, which on a BESS apply to a DC source with serious available fault current. For acceptance testing of the electrical equipment, ANSI/NETA ATS gives the field test and inspection requirements, recent editions added coverage for battery energy storage systems, and insulation-resistance methods reference IEEE 43.

Where the BESS interconnects, IEEE 1547 and its test companion IEEE 1547.1 govern the interconnection and the grid-support and ride-through functions, with the utility and the jurisdiction setting the applied settings and category. IEEE also publishes battery and energy-storage characterization and test practices, including the 1679 series for evaluating storage technologies and the lead-acid maintenance and test practices where a lead-acid system is used. Above all of these sit the manufacturer's instructions and the project specification, which set the actual numbers, and the authority having jurisdiction, which has the final say on what is enforceable. When a standard and the spec disagree, the stricter controlling document wins.

Units and terms

BESS work runs on a handful of units and acronyms, and reading the wrong one is how a result gets accepted that should not be. Energy is in kilowatt-hours and megawatt-hours, the capacity the battery stores. Power is in kilowatts and megawatts, the rate it charges or discharges, and the ratio of energy to power gives the duration. The DC side is rated in volts DC, often hundreds of volts up past 1000 V DC, and that voltage is present whenever the cells are charged.

State of charge and state of health describe the battery, round-trip efficiency describes the losses, and the PCS and BMS are the two pieces of hardware whose acronyms show up on every drawing. Keep them straight, because a spec that buys a round-trip efficiency at the AC terminals and a system that reports it at the DC bus are talking about two different numbers.

BESS

Battery energy storage system, the cells, BMS, PCS, controls, and enclosure as a unit

kWh / MWh

Energy stored, kilowatt-hours and megawatt-hours, compared against nameplate capacity

SOC

State of charge, how full the battery is now, as a percentage of usable capacity

SOH

State of health, the battery's capacity now versus when it was new

RTE

Round-trip efficiency, energy out on discharge divided by energy in on charge

PCS

Power conversion system, the bidirectional inverter between the DC battery and the AC grid

BMS

Battery management system, which monitors and protects the cells and trips the battery

Related tools

Calculators and readiness checks for this work

Compare your options

FAQ

People also ask

• How long does data center electrical commissioning take?
• How long is a generator load bank test?
• What kind of batteries do data centers use?
• What is a rack BBU?
• What is a BESS?