Switchgear to busway, proven under load and under failure
UPS testing and power train commissioning software for the chain between utility feed and the rack that can't afford to break.
The power train — utility switchgear, generators, UPS, static transfer switches, PDUs and busway — is the subsystem where a single unproven joint or setting can take down every rack downstream. Deskely holds generator paralleling, UPS testing, STS transfer and load bank evidence against each device, with failure-mode testing gated the way it has to be before any of it carries IT load.
- Chain
- Utility to switchgear to UPS to rack
- Test
- Load bank and failure mode
- Redundancy
- N+1, 2N paralleling
- Gate
- Proven before IT load
Choose your project type
The problem
Every device in the power train passes its own test; the chain still hasn't been proven end to end.
A UPS can pass its own factory and site acceptance tests, a generator can pass its own load bank run, and a switchgear lineup can pass its own protection relay settings check — while the transfer sequence between all three has never actually been exercised together.
That gap is invisible until the first real utility outage, which is precisely the moment nobody wants to discover it.
Deskely models the power train as a chain of dependent subsystems, so a device's own commissioning is one gate and the integrated transfer and failure-mode test across the whole chain is a separate, explicit gate above it.
Setting up the register
Switchgear, generators, UPS, STS, PDUs and busway are tagged with their upstream and downstream dependencies.
Every breaker, generator, UPS module, static transfer switch, PDU and busway run is a tagged asset carrying its own protection settings, capacity rating and the specific upstream source and downstream load it connects.
Single-line diagrams and relay coordination studies are parsed into the register, so a protection setting check and a paralleling switchgear commissioning record reference the same drawing without re-entry.
That dependency model is what makes it possible to ask, and answer, exactly which racks are affected if a specific breaker or UPS module is taken out of service.
Testing the chain
Generator paralleling, UPS testing and STS transfer are proven under load bank conditions, not on paper.
Generator paralleling is tested for synchronisation, load sharing and step-load response; UPS testing covers battery discharge, bypass and rectifier/inverter transfer; static transfer switches are proven for transfer time under both planned and unplanned source loss — each against load banks simulating the design IT load.
Results, waveforms and transfer times are captured against the device tag with the acceptance criteria they were tested to, so a marginal transfer time is visible immediately rather than buried in a PDF appendix.
Failure-mode testing deliberately fails a source or a device — pulling utility power, tripping a UPS module — to confirm the redundant path actually picks up the load within the required window.
Certification
The power train is certified as a chain, not as a set of individually passing devices.
The power train certificate references every device's own test evidence plus the integrated failure-mode results, so a reviewer can see both that each device passed and that the chain behaves correctly when one link is removed.
Preservation routines for switchgear, generators and UPS that are commissioned but sitting unloaded ahead of IT load follow the same tag structure, keeping battery health, fluid levels and settings checks current.
Any change to protection settings or paralleling configuration after certification is tracked against the same asset, so the power train's tested state and its current state never quietly diverge.
The battery string is the weakest, least visible link
A UPS looks healthy right up until the string of batteries behind it fails the one time it's actually needed.
Battery discharge testing at commissioning proves the string can hold the design runtime on the day it's tested, but VRLA and lithium-ion strings degrade unevenly — a handful of weak cells can drag the whole string's effective runtime down long before any single-cell alarm fires on the monitoring system.
An arc flash study and the resulting PPE category labelling on switchgear and busway is the other piece of evidence that has to be current rather than assumed: a paralleling reconfiguration or a capacity upgrade changes available fault current, which changes the arc flash incident energy at every downstream panel, whether or not anyone updates the label.
Deskely holds battery discharge trend data and impedance test history against the UPS tag so degradation is visible over years rather than reset at each test cycle, and ties the current arc flash study revision to the switchgear asset so a label that's out of date after a configuration change is a flagged discrepancy, not a silent risk.
Evidence
The power train dossier
The power train stands or falls on load bank data and battery health readings taken at defined intervals, not a single pass/fail. These are the records an owner's engineer will pull first during any dispute over UPS performance.
Level 4 functional performance test script for the UPS module
Rectifier, inverter, bypass and transfer logic operate correctly on load and on simulated fault
After UPS installation, before integration with generators
Load bank step-load test report
The UPS and generator train accept and shed load in defined increments within voltage and frequency tolerance
During Level 4 and Level 5 testing
Battery impedance baseline record
Each battery string's internal impedance is captured as the reference point for future trending
At commissioning, before energisation
Black-building pull-the-plug test record
The power train transfers from utility through generator start to UPS ride-through without a load drop
Final Level 5 IST scenario
Thermal ride-through data log
Battery and inverter temperatures remain within design limits through the full duration of a ride-through event
Captured during the black-building test
SLA acceptance and witness pack
The power train's tested performance meets the contracted availability and capacity commitments
Compiled at the close of testing, before handover
How it runs
From individual device tests to a chain proven under failure.
The power train is the subsystem where dependency modelling matters most, because a passing device test says nothing about the transfer between it and its neighbours.
- 01
Tag the chain
Switchgear, generators, UPS, STS, PDUs and busway registered with explicit upstream and downstream dependencies.
- 02
Commission each device
Factory and site acceptance records completed per device against its own protection settings and capacity rating.
- 03
Load bank test
Generator, UPS and STS performance proven against load banks simulating the design IT load and step changes.
- 04
Run failure-mode tests
Sources and devices deliberately failed to confirm the redundant path picks up within the required transfer time.
- 05
Certify the chain
Power train certificate issued referencing individual device evidence plus integrated failure-mode results.
FAQ
Questions about UPS and power train scopes.
Keep reading
Data centre commissioning
The full sector view: commissioning levels, integrated systems testing, punch and the certification ladder to IT load.
Read moreITR records
Component verification, functional performance and IST records defined once per equipment type and instantiated across every asset.
Read morePreservation
Routines for switchgear, UPS and generators that sit energised but unloaded for months before IT load arrives.
Read moreOther data centre project types: Hyperscale new build data centres, Colocation facility commissioning and SLA acceptance, Edge and modular data centre commissioning, Retrofit and capacity upgrade commissioning, Cooling, CRAH/CRAC and liquid cooling systems, Standby generators, fuel and switchgear, White space rack, power and busway fit-out, Integrated systems testing and Level 5 commissioning, Campus substation and utility interconnection.
