Backup power

A backup never tested on load is an assumption, not a backup.

Divisions assembled: Power and backup supply Secure buildings

The issue

Contracted supply describes a contract; a load study describes a need.

The most common method is to take the supply contracted with the distributor and order a set of similar rating. It almost always produces an oversized installation that runs lightly loaded, soots up and fails sooner.

At the other end, an installation sized to the minimum collapses when a motor or a compressor starts. Between the two, only one method works: list what must be backed up, separate permanent from intermittent loads, account for starting inrush, and adopt a rating that leaves margin without installing an engine that will never work.

Then comes the link most often forgotten. Batteries must hold until the generator takes over; cooling must be backed up alongside the IT load; and stored fuel degrades. The failure occurs at start-up — the one moment the installation is relied upon.

Backup power chain, schematic

01Public grid: Normal source, whose availability cannot serve as a design assumption.
02Generator set: Replacement source, sized on the load assessment and its growth allowance.
03Automatic transfer switch: Switches over without human intervention, with a delayed return to the grid.
04Main low-voltage board: Protection, discrimination and separation of critical and ordinary feeders.
05UPS: Covers the generator start-up window, during which no source is supplying.
06Critical loads: Supervision, security, servers and links whose loss is an incident.
07Ordinary loads: Sheddable feeders, whose interruption carries no operational consequence.

The marked points are reported to supervision: grid presence, source state, fuel level, engine hours and load-test results.

Schematic. No rating, no equipment and no real installation appears on it.Typical sizing figures, not measured on an installation. The transfer time covers generator start-up and load acceptance; the UPS autonomy covers that window, plus a clean shutdown if the generator fails to start.

Outcome

What you get

A complete chain tested on load, from grid to generator, whose real autonomy is measured and recorded, whose status is visible remotely, and whose periodic tests are scheduled and kept in a register.

Scope

What the solution includes

Load study: A list of the circuits to be backed up, measured consumption where an installation exists, starting inrush, power factor, non-linear loads, and an extension margin settled with the client.
Uninterruptible supply: Autonomy calculated on the load actually connected rather than on nominal rating, maintenance bypass, battery health monitoring, and clean shutdown at a threshold.
Generator set: Prime and standby ratings kept distinct, derating for site conditions, base, ventilation, exhaust, containment, and access for maintenance as well as replacement.
Source transfer: Automatic transfer panel, timers, order in which circuits are restored, interlocking between sources, and behaviour when the grid returns.
Fuel: Day tank and bulk tank, transfer, gauging, filtration and draining, with autonomy derived from the real resupply lead time at the site.
Monitoring and maintenance: Machine status, fuel level, faults and test results reported; maintenance contract, wear parts in stock, and a register of tests and interventions.

Sequence

How it is run

01
Study
The load study precedes any order. It determines the rating, not the contracted supply and not the available budget.
02
Design
Chain architecture, adopted autonomy, redundancy decided on acceptable downtime, and civil works arrangements settled with the building fabric.
03
Delivery and testing
Installation, then a test of the complete chain on load — not off load. Most failures observed in real conditions would have been revealed by an on-load test.
04
Operation
Remote monitoring, the periodic test programme and the register put in place. Without them the installation decays in silence.

Commitments

What is verified at acceptance

These commitments can be held against us. Written into a tender, they filter out the responses that will not hold — including ours, if it does not hold them.

A written load study separating permanent loads, intermittent loads and starting inrush.
Restoration time measured by on-load transfer test, never declared.
Real battery autonomy measured by discharge test, and recorded as the initial state.
The complete chain tested from grid to generator: a UPS and a set tested separately do not prove their transfer.
Ventilation arrangements verified in continuous running, not on a short test.
The disposal route for spent batteries, its carrier and the traceability of collection.
A periodic test programme and its register, handed over with the installation.

Before committing

Questions to settle internally

An organisation that cannot answer these questions is not ready to launch the programme. Saying so is better than selling it a study.

Which circuits genuinely need backing up, and which can stop?
How long must the site hold without fuel resupply?
Is the technical room’s cooling backed up alongside the servers?
Is there a window during which the installation can be shut down for servicing?
Who runs the periodic tests, and where is the result recorded?
What downtime does the organisation accept, stated by the executive and not by the technical team?

Evidence

What we have delivered on this scope

Energy and critical infrastructure2024
Mining site — server room and backup power
Construction of a secure technical room and its power chain, on an isolated site with no reliable public grid.
Weeks of measurement before sizing
4
Load tests per year
12
Reference KP-2024-008