In short. A single point of failure (SPOF) is any one element whose failure stops the supply to a critical load. Finding them needs a deliberate walk through the system, from each load back to each source, asking what one failure would do. Remove the ones that matter, and write down the ones you accept.
Method
- Pick a critical load group.
- Trace every path from the load to every source on the single-line diagram.
- List each element that all paths share.
- Repeat for control, communications, DC supply, cooling and fuel that the path depends on.
- Rank each SPOF by likelihood and consequence.
- Decide: remove, mitigate, or accept with a recorded reason.
Where SPOFs hide
| Area | Typical hiding places |
|---|---|
| Power path | Common busbar, single transformer, one cable route, shared bypass |
| Controls | One PLC, one network switch, one 24 V supply |
| Protection | Single trip battery, single relay on a critical breaker |
| Generators | Shared fuel pump, day tank, cooling circuit, starting battery |
| Support | Single chiller or ventilation fan for a critical room |
| People and process | Only one person knows the procedure or holds the key |
| Supply chain | One fuel supplier, one spare part source |
Options for each SPOF
- Duplicate the element, with independent supplies.
- Bypass it with a manual or automatic route.
- Monitor it so failures are found before demand.
- Harden it, using higher quality components and better maintenance.
- Accept it where the cost is out of proportion to the risk, and document the decision.
Prioritising
Score each SPOF on consequence (which loads, how long), likelihood (age, history, environment) and detectability (do we know when it has failed?). Fix high-consequence, high-likelihood and hard-to-detect items first.
Worked illustration
A review of a data room finds the following SPOFs. A single static bypass feeding both UPS is shared, so a fault there drops everything. Both generators take fuel through one transfer pump. One BMS server controls the load-shedding logic. The first and second are fixed by redesign: separate bypass for each UPS and a second fuel pump with a crossover. The third is mitigated: the shedding logic is moved into local relays so it continues if the server fails. Each fix targets a different hiding place.
Not all SPOFs are equal
A SPOF on a passive element such as a cable in a protected duct has a low probability of failure. A SPOF on an active element such as a controller or a fan has a higher one. Weigh accordingly, and document the reasoning so later engineers can revisit it.
Keeping it eliminated
- Review SPOFs when the system changes.
- Include SPOF review in design and acceptance checklists.
- Test the alternate path regularly.
Finding single points of failure
A single point of failure (SPOF) is any component whose failure alone interrupts the critical load. They are often not obvious on a single-line diagram.
| Typical hidden SPOF | Remedy |
|---|---|
| Shared DC control supply | Two independent DC systems |
| Single fuel transfer pump or day tank | Duplicate pumps, dual supply |
| One ATS for all essential loads | Multiple ATS or bypass |
| Common cable route | Separate routes |
| One controller or PLC | Redundant or distributed control |
| Single cooling system for electrical room | Redundant cooling |
| One key technician or one vendor | Cross-training, support contracts, spares |
Method
- Walk the power path from the load back to each source.
- At each element ask: if this fails, does the load stay up?
- List every element where the answer is no.
- Rank by likelihood and consequence.
- Eliminate, mitigate or accept with sign-off.
Dependencies beyond electrical
Water for cooling, fuel delivery, access roads, communications and staff are all dependencies. A generator without fuel supply is not a source after the tank runs out. A fuel contract with delivery times matters.
See electrical FMEA and common-mode failures.
Common mistakes
- Looking only at the power path and not at controls and support systems.
- Assuming a redundant element is independent without checking its supply.
- Failing to record accepted risks.
What to send us
Send the single-line diagram and control architecture. Send your application details. See also common-cause failure.
Questions engineers ask
Is it possible to remove every SPOF?
Rarely. The practical goal is to remove or mitigate those that matter and to record the ones accepted.
Are people a SPOF?
They can be. A procedure only one person can perform is a SPOF; training and documentation reduce it.
How often should I review?
At design, at commissioning, and after any significant change to the system or loads.
What is the difference between a SPOF and a common-cause failure?
A SPOF is one element whose failure stops supply. A common-cause failure is one event that disables several elements at once.
This guide gives general principles. The review should be done with your actual drawings and operating experience.