In short. A concurrently maintainable system lets any planned maintenance, repair or replacement happen without shutting down the load. It needs a second path to the load, isolation points that actually isolate, spare capacity while a path is out, and a written procedure. It is a design property to prove, not a label to assume.
Concurrent maintainability versus fault tolerance
Concurrent maintainability covers planned work. Fault tolerance covers unplanned failure. A system can be one without the other. For example, a single UPS with a maintenance bypass is maintainable with a risk (the load is on raw mains during the work) but not fault tolerant. Many sites target concurrent maintainability for all components and fault tolerance for the most critical. See fault-tolerant architecture.
What it requires
- Alternate path: a source of power for the load while a component is out.
- Isolation: breakers, isolators or links that isolate a component with safe, verified separation.
- Capacity: the remaining path carries the full load with margin.
- Control and protection: settings and logic that work in the reduced state.
- Procedure: a method of procedure with steps, checks and rollback.
Maintenance cases to check
| Item | Question |
|---|---|
| Generator | Can one be serviced while the others carry the load? |
| UPS module or system | Can it be isolated without dropping load? |
| Switchboard or breaker | Can one be isolated and made safe with others live? |
| Transformer | Can the other carry the load? |
| Busbar section | Can it be de-energised safely? |
| Cable or feeder | Is there an alternate feed? |
| Controls and DC supplies | Can they be maintained without a trip? |
Hidden obstacles
- A bus section that cannot be isolated without taking out adjacent sections.
- A control supply common to both paths.
- No safe means to test a breaker in place.
- Limited capacity so the remaining path is overloaded during work.
- Interlocks that prevent the switching sequence needed for the work.
The risk during maintenance
While one path is out, the system is in a reduced state. A failure in the remaining path then causes an outage. Reduce the exposure by scheduling work in low-risk periods, limiting its duration, and having the remaining path verified before starting. N+2 or 2N+1 designs keep redundancy even during maintenance.
Worked illustration
A 2N system has two UPS paths each loaded to 45 per cent. To service UPS A, the load moves to path B, which carries 90 per cent for several hours. The system is concurrently maintainable but not fault tolerant during the work. To reduce the risk, the team checks path B’s batteries and breakers first, restricts other work in the building, keeps a technician at the board and has a rollback plan. That procedure is part of the design.
Proving it
- Walk through each maintenance task on paper using the single-line diagram.
- Perform a trial isolation of each major component during commissioning.
- Maintain a library of methods of procedure.
- Update the study after each modification.
The definition
A system is concurrently maintainable if any component or path can be taken out of service for planned maintenance without interrupting the critical load. Uptime-style tier frameworks use this idea at a defined level; apply it on its merits to your own site.
What it requires
| Element | Requirement |
|---|---|
| Generators | At least N+1 so one can be serviced |
| UPS | N+1 modules or two systems |
| Switchgear | Bypass or alternate paths for each breaker |
| Distribution | Alternate path to every load |
| Cooling and fuel | Redundant, with isolation valves |
| Controls | Can be maintained without stopping the plant |
The trap
While one path is in maintenance, the system is at lower redundancy. A failure on the remaining path then causes an outage. Concurrent maintainability does not mean fault tolerance during maintenance; for that, N+2 or 2N paths are needed.
Method of review
- List every component.
- For each, write the procedure to take it out of service.
- Check whether the load stays supplied throughout, including the switching steps.
- Identify any component that fails the test and decide to redesign or accept the risk.
Operational discipline
Method of procedure documents, a second person check and rehearsed sequences prevent human error during maintenance.
See N to 2N redundancy and maintenance and lifecycle.
Common mistakes
- Declaring concurrent maintainability without testing isolation.
- Ignoring the added risk during the maintenance window.
- Letting load growth remove the capacity that made it possible.
What to send us
Send the single-line diagram and your maintenance policy. Send your application details. See also maintenance and lifecycle resilience.
Questions engineers ask
Is concurrent maintainability the same as redundancy?
Redundancy is a feature that helps achieve it, but you also need isolation points, procedures and spare capacity during the work.
Does it require 2N?
Not always. Some N+1 arrangements are concurrently maintainable if isolation and paths allow it.
How do I test it without risk?
Rehearse on paper, test in commissioning before the load is live, and run live trials under controlled conditions with rollback ready.
Why does maintenance cause outages in redundant systems?
Because the system is in a reduced state and procedure errors or a second fault take down the remaining path.
This guide gives general principles. Safe isolation procedures must follow your site rules and applicable standards.