Skip to content
1 (438) 800-0382FR FrançaisGet a quote

Guides

What is an integrated systems test (IST)?

An integrated systems test (IST) is the last level of data-centre commissioning. It runs power, cooling and IT infrastructure together at full load, with load banks standing in for the servers, and stages utility problems and equipment failures to prove every critical system responds appropriately, including the UPS carrying the load while the generators start.

On this page
  1. Where the IST sits in commissioning
  2. What the IST covers
  3. UPS scenarios and what to record
  4. Prerequisites
  5. Pass criteria and the record
  6. Common errors
  7. Why the power path gets the most attention
  8. After the IST

Key facts

Where it sits:
Level 5, the last of five commissioning levels, after factory, site, pre-functional and individual system testing Aggreko
What it proves:
That power, cooling and IT infrastructure perform at full load as an integrated whole Aggreko
Test load:
Free-standing load banks for the halls, or rack-mounted loads at the point each server will sit Aggreko
Generator gap:
About 10 seconds without power while the generator starts, unless a UPS bridges it Cummins
Each test:
A sequenced failure, then a restart and a return to stable operation Schneider Electric
The record:
Expected response, actual response and pass or fail for every script line Schneider Electric

Where the IST sits in commissioning

Short answer

Commissioning runs in five levels: factory acceptance testing, site acceptance testing, pre-functional testing, individual system testing and the integrated system test (Aggreko). At level 4, UPS modules, generators, PDUs and busbar systems each get full functionality testing, which can include a load burn-in, generally on load banks, or thermal imaging. Level 5 tests them together, as an integrated whole at full load.

Schneider Electric's commissioning paper (White Paper 148) gives the reason a separate integrated level exists. Vendor start-ups were each driven by a contract that operated in a vacuum, independent of the other components, and no contractual requirement existed for the products to perform in a fully integrated manner. Whole-site commissioning developed because standard start-up procedures kept failing to find system-wide weaknesses.

Fig. 1 · The utility failure sequence an IST stages

The utility failure sequence an IST stagesUtility fails. The transfer switch senses the loss and signals the generator to start. The UPS carries the load on battery for about 10 seconds. At proper voltage and frequency, the transfer switch connects the generator. The UPS returns to its input and the batteries recharge. When the utility returns, the transfer switch returns the load to the normal source.Utility failsTransfer switch senses the loss,starts the generatorUPS carries the load on battery, about10 sGenerator at voltage and frequency:transferUPS back on its input, batteriesrechargeUtility returns: load back to thenormal source
Eaton warns the UPS may stay on battery if the generator's frequency range is too wide for its input window. Source: Cummins (transfer switch, generator papers); Eaton, generator-UPS harmony

What the IST covers

The power part starts at the high-voltage service entrance and works forward through medium-voltage distribution, parallel switchgear, transfer switches, the emergency generator, the UPS system, the monitoring system and the distribution down to the racks. Lighting and life safety systems, including emergency power off (EPO), are tested too, and WP 148 adds a short-circuit study that checks every circuit breaker and ground fault trip setting. Cooling, fire suppression, physical security and the building and energy management systems are tested alongside, and each alarm is verified.

The method is the sequenced failure. WP 148 says every piece of equipment should be tested by failing it, then restarting it and returning it to stable operation. A failure in one component, such as a generator, is communicated to a related one, such as the air conditioning, so the second can act to limit downtime or be ready when power returns. The cycle runs on each component and on the whole integrated system, which means a complete power down and an automatic restart.

UPS scenarios and what to record

The table pairs the power events in the sources with the response each one describes. Aggreko's level 5 stages machinery failures and utility problems; WP 148 adds sequenced failures and a complete power down.

Event stagedExpected response (source)What to record
Utility failure: the normal source is openedThe transfer switch detects the loss and sends a start signal to the generator; the UPS bridges the gap of about 10 seconds on battery (Cummins transfer switch paper)Time stamps for each step; the UPS output and battery parameters during the gap
Generator on lineThe transfer switch connects the generator once it reaches proper voltage and frequency (Cummins transfer switch paper). Cummins' generator paper describes the UPS sensing the voltage and frequency excursion, holding the load on battery while the generator stabilises, then ramping the load onto it. Eaton warns that if the generator's frequency range is too wide for the UPS, the UPS may treat it as unstable and stay on battery until the battery fails and drops the load.Whether the UPS returned to its input or stayed on battery, and the ramp from battery onto the generator
Utility restoredWhen utility power returns, the transfer switch returns the load to the normal source. An open transition breaks before it makes; a closed transition parallels both sources for under 100 ms when both are available (Cummins transfer switch paper)The transition type that ran and the UPS output through it
UPS to static bypass and back to normalThe UPS supports the load after switching from bypass back to normal operation; WP 148's sample error log records this step failing because the battery leads at the head of the string were disconnected (WP 148)Each transfer, and the load held through it
UPS on batteryEach test that puts the UPS on battery reduces the charge left for the next, so each segment of the integrated test takes the available runtime into account (WP 149)The UPS output and battery during the discharge, and the recharge allowed before the next major power drain test (WP 149)
A UPS module failure staged in an N+1 systemBy design, a parallel redundant system allows the failure of a single module without transferring the critical load to the utility source (WP 75)That the load stayed on the remaining modules and did not go to bypass
Generator failure stagedThe failure is communicated to related systems, such as the air conditioning, so they act to limit downtime or are ready when power returns (WP 148)Each linked system's response, then the restart and the return to stable operation
Complete power downAn automatic restart of the integrated system (WP 148)Every system's restart and the point it was stable again

The UPS-level version of these steps, in Eaton's load bank method (a battery discharge of up to 4 minutes by opening the rectifier input, and three transfers to bypass and back), is covered under UPS load bank testing.

The UPS input window and walk-in are explained in UPS and generator. For the bypass paths themselves, see static and maintenance bypass.

Prerequisites

  • Levels 1 to 4 complete (Aggreko); every subsystem tested to its manufacturer's start-up process, with the start-up results in the commissioning team's hands before commissioning starts (WP 148).
  • A written test script: line by line procedures, communicated by the command center and executed by the testing team. The commissioning agent writes it, and the team rehearses it beforehand (WP 149).
  • A command center to give permission for each next step, handle all communication and hold every contact and emergency number; on a small site, one person can fill the role (WP 148).
  • Cooling, electrical and monitoring systems ready to run when the load banks arrive (WP 148).
  • Heat loads in place: free-standing load banks sized for the halls, or rack-mounted loads at the server positions (Aggreko).
  • A checklist of the expected inputs and outputs for each critical component in the test schedule (WP 149).
  • Emergency operating procedures posted on each piece of equipment and at each EPO station (WP 149).
  • A shift plan, so every tester has a backup who can step in (WP 149).

Pass criteria and the record

WP 148's script report sets the expected response beside the actual response for each line and designates each line pass or fail.

WP 148 lists three documents that commissioning should produce. The "as built" script report records what was tested and how, and gives a line by line account of each pass or fail, with the test procedure, expected response and actual response. The component error log, also called a failure mode effects analysis (FMEA), records each component that failed and how the failure affected equipment upstream and downstream. The trending report closes the job with an action plan and a validation statement from the commissioning agent.

WP 148's sample error log shows the kind of fault an IST catches. A UPS failed to support the load after switching from bypass back to normal operation, because the battery leads at the head of the string were disconnected. The corrective action was to have every battery lead checked and the test rerun.

Common errors

WP 149 collects ten errors that undermine commissioning. Six bear on an IST, three of them on the UPS directly:

  • Outdated test procedures, such as testing a delta conversion UPS with procedures written for a double conversion UPS.
  • A script that is unclear or never rehearsed, so it is misinterpreted, and steps are left incomplete.
  • Only the UPS gets tested: large exterior resistive load banks wired to the UPS output leave the cooling system out. Resistive heating units installed in the racks let the UPS, emergency power and cooling be tested together.
  • Battery charge run down. Each test that puts the UPS on battery reduces the runtime left for the next; WP 149's best practice is to allow sufficient recharge after a major power drain test.
  • Fatigue. WP 149 calls it the principal cause of human error during commissioning, with staff logging 12 to 18 hour days; rotate shifts.
  • A commissioning agent engaged late, and budget cuts that compress the test scope.

Why the power path gets the most attention

Power remains the leading cause of impactful outages in Uptime Institute's 2026 outage analysis, and failures involving UPS systems, transfer switches and generators dominate. The same report notes that grid failures should not cause a data center outage, yet transfers to generators sometimes fail. Uptime's Data Center Resiliency Survey 2026, cited in the report, finds that power-related issues caused more than one in five IT service outages reported over the past three years. Among power causes, UPS failures are cited most often, followed by transfer switch and generator issues.

Those are exactly the transfers an IST stages. For redundancy designs and the failures each one tolerates, see N+1 and 2N; for UPS work on a site preparing for commissioning, see UPS service for data centres.

After the IST

The yearly load bank simulation Aggreko recommends is covered under load bank procedure. WP 148 treats the commissioning record as the baseline for spotting changes and trends, and recommends bringing the commissioning agent back periodically or at major business events to review the integrated system. WP 149 adds a disciplined process to keep the documentation current. For the annual UPS test, see UPS load bank testing; for a new UPS, see installation and commissioning.

Questions

What is the difference between individual system testing and an IST?

Individual system testing (level 4) proves each UPS module, generator, PDU and busbar system on its own, which can include a load burn-in on load banks. The IST (level 5) runs power, cooling and IT infrastructure together at full load and stages failures across them.

Does an IST need load banks?

Yes, a heat load stands in for the IT load: free-standing load banks sized to the halls, or rack-mounted loads at the server positions (Aggreko). WP 149 warns that exterior load banks wired to the UPS output test only the UPS.

Who writes the IST script?

The commissioning agent, according to WP 149; WP 148 adds that without one, the design engineer or the construction company can usually perform the commissioning.

Why recharge the UPS batteries between tests?

Each test that puts the UPS on battery reduces the charge left for the next one. WP 149 warns that after several such tests the available runtime is severely reduced, and calls sufficient recharge after a major power drain test a best practice.

Sources, checked 2026-10-04:

Need a UPS serviced, tested or replaced?

Tell us the site, the UPS and the timing. We reply and quote within 4 business hours, or call 1 (438) 800-0382

Timing

We reply and quote within 4 business hours.

More details
Call 1 (438) 800-0382Get a quote