Concurrent maintainability changes what an Authorised Person appointment has to cover. On an industrial site, you can usually plan an outage window, isolate the circuit, and get the work done without much argument about downstream consequences. On a data centre running N+1 or 2N electrical topology, taking a distribution path out of service for testing or maintenance is meant to be invisible to the load. That's the whole point of the redundancy. It also means the HV and LV Authorised Person structure, and the testing regime that sits under it, has to account for switching sequences and capacity headroom in a way that generic electrical compliance guidance doesn't really address.
Most of what's published on Authorised Person appointment and electrical testing schedules is written for a broad industrial or commercial audience. It covers the legal minimum and the general shape of a competent persons scheme well enough, but it says almost nothing about what changes when the asset can't tolerate an unplanned drop to N or below, or when a single line diagram has six parallel paths that all need to stay in a known, provable state during a maintenance window.
Why the industrial competent persons model doesn't map cleanly onto a data hall
The Electricity at Work Regulations 1989 set the legal baseline: anyone working on or near electrical systems must be competent for the work they undertake, and the duty holder has to manage that risk. The regulations don't define "Authorised Person" as a term, and they don't separate HV and LV appointment into distinct categories. That structure, a Senior Authorised Person overseeing HV switching, Authorised Persons carrying out defined HV operations, and Competent Persons handling LV work, comes from the safety rules tradition that much of UK industry inherited from electricity supply industry practice, formalised in guidance such as HSG85 and adopted widely by facilities and estates teams.
What that framework assumes, in its original form, is a fairly static electrical topology: a substation, a switchboard, a known set of feeders. A data centre has a live electrical hierarchy that changes hour to hour as load shifts between UPS modules, generators come on and off test, and maintenance takes individual PDUs or breaker positions temporarily out of the redundant set. An Authorised Person appointment that doesn't reflect the current state of that hierarchy, only the design intent on a drawing from commissioning, is not really telling you what's safe to switch today.
HV Authorised Person appointment when the switching sequence matters more than the switchgear
Appointing a Senior Authorised Person for HV work at a data centre is not simply a matter of confirming HV competency and signing an authorisation certificate. The person also needs site specific authorisation tied to the current switching schedule, because the sequence in which HV switchgear is operated is what determines whether concurrent maintainability actually holds during the work. Getting the sequence wrong doesn't necessarily cause an immediate safety incident. It can quietly remove the redundancy the facility is contracted to guarantee, without anyone noticing until the next fault.
That's part of why HV authorisation at a data centre tends to be reviewed more frequently than the underlying competency certificate would strictly require, and why some operators tie HV authorisation to a specific building or electrical zone rather than treating it as a blanket site authorisation.
LV Authorised Person duties and where the overlap with HV actually sits
LV work happens far more often than HV switching on most data centre sites, simply because breaker maintenance, PDU testing, and UPS bypass operations are more frequent events than HV network reconfiguration. The LV Authorised Person or Competent Person carrying out that work needs authorisation that's specific to the distribution board or PDU in question, not a general LV competency sign off, because the consequence of an error, tripping a board that's carrying load for a rack that has no redundant path available at that moment, is a direct service impact rather than an abstract risk.
The overlap with HV appointment shows up at points like generator paralleling gear and main LV switchboards fed directly from HV transformers, where LV work can have HV implications if it's not sequenced correctly. Testing regimes that treat HV and LV as entirely separate authorisation streams tend to miss that overlap, which is one reason we've seen operators build a single switching authority record that covers both, rather than maintaining them in parallel systems that don't talk to each other.
Scheduling electrical testing around uptime constraints rather than a fixed calendar
BS 7671 and manufacturer guidance give a starting point for LV inspection and testing intervals, and HV switchgear typically has its own manufacturer-defined maintenance schedule. Neither of those was written with an uptime SLA in mind. In practice, data centre electrical testing gets sequenced around available redundancy capacity: you test a UPS module when the others can genuinely absorb the full load, not when the calendar says the test is due.
That means the testing schedule is only as good as the visibility of what else is happening on the electrical system at the same time. A cooling system permit that's also drawing down redundancy on the mechanical side, or a second electrical permit on an adjacent PDU, can turn a routine test into a real exposure if nobody cross-references them. This is close to the same problem the wider industry manages through SIMOPS visibility, applied to a facility where the concurrent operations are electrical and mechanical rather than process and maintenance.
What's worth checking in your current compliance record
A useful exercise, whatever system currently holds this information, is to pull the record for a single HV switching operation and see whether it shows the redundancy state before and after the switch, not just the fact that it happened. The same test applies to LV work: can you retrieve, for any completed PDU maintenance task, which redundant path was carrying load at the time and who authorised the switch. If that information sits in a switching log, a separate permit system, and a BMS export that don't reference each other, reconstructing it after the fact is slow and occasionally impossible.
We've written before about the practical detail of managing electrical permit to work in data centre and contractor environments, and about the narrower question of when live working on electrical systems is actually permitted, both of which sit underneath the Authorised Person structure described here.
Where this fits alongside wider critical equipment assurance
HV and LV authorisation and testing are one part of a broader critical equipment integrity picture at a data centre, alongside cooling, fire suppression, and structural systems that all have their own performance standards and verification regimes. We've looked at how that assurance picture holds together in critical equipment integrity for data centres, which covers where electrical testing sits against the rest of that framework.
If you're reviewing how HV and LV Authorised Person appointments are currently recorded and linked to your switching schedule, we can look at the existing workflow with your electrical and facilities teams and identify where the gaps in cross-referencing actually sit.