The power goes out at twenty to four on a Sunday morning. The transfer switch does its job, the standby generator picks up the load, and for about four minutes everything looks like the continuity plan working exactly as written. Then the engine loses power and stops.
What happens next is predictable enough that maintenance engineers can describe it before they arrive. Someone pulls the primary fuel filter and finds it packed with a dark, soft mass. The generator was fine. The switchgear was fine. The problem was the fuel, which had been sitting in the tank for two years and was no longer doing the job it was bought to do.
Most organisations spend real money on business continuity. They buy the generator, they sign the maintenance contract, they run the annual test and file the certificate. Very few of them ever ask what condition the fuel is in, and that is the part of the plan that quietly degrades on its own.
What is stored fuel degradation?
Stored fuel degradation is the loss of diesel quality that happens inside the tank while nobody is using it. Water collects, microbes grow at the boundary between the water and the fuel, and the fuel oxidises. The result is sludge and sediment that block filters and foul injectors when the engine is finally asked to run.
Diesel is bought to wait. Standby fuel might sit for three years before anything asks it to burn, and it is a refined product with a working life measured in months rather than years.
What goes wrong inside a tank of standby diesel?
Three things, and they compound. Water condenses out of the air the tank breathes in, microbes colonise the boundary between that water and the fuel, and the fuel itself oxidises. All three end up as material that blocks a filter.
Water. Tanks breathe as temperatures rise and fall, drawing in humid air. That moisture condenses on the tank walls and runs down under the fuel. Nobody adds it and nobody sees it arrive, but over a few winters a measurable water layer collects at the bottom of the tank.
Microbial growth. Bacteria and fungi are present in diesel as delivered, dormant and harmless in dry fuel. Once there is free water beneath the fuel, the boundary between the two becomes a habitat: water on one side, hydrocarbon food on the other. Colonies grow there and produce a sludge, and their waste products acidify the water layer and attack the tank floor from the inside.
Oxidation. Fuel exposed to air over long periods forms gums and insoluble sediment that either settle out or stay suspended in the fuel.
None of this shows up when you look into the tank hatch, but all of it ends up in the same two places: the filters, which block far ahead of schedule, and the injectors, which foul and stop atomising fuel properly. That is why a generator can start cleanly and then die under load.
How long does stored diesel actually last?
Less time than most continuity plans assume, and less than it did fifteen years ago. The product itself changed. Diesel sold in the UK and Europe today is low in sulphur and carries a biodiesel component, and both of those changes shortened the practical storage life of a full tank.
BS EN 590, the standard that defines automotive diesel here, allows up to 7 per cent by volume of fatty acid methyl esters. FAME holds more dissolved water than mineral diesel and is more readily digested by microorganisms, so a tank running to current specification reaches trouble sooner than one filled in 2005 would have. Reducing sulphur removed a component that had incidentally suppressed microbial growth.
Both changes were environmentally sound. Neither was made with three-year storage in mind, and nobody sent a memo to the people writing continuity plans when the specification moved.
Which businesses are most exposed?
Any business with a generator has some exposure, but it concentrates in facilities where backup power is not optional: data centres, hospitals, airports, telecoms exchanges and bank operations centres. All of them hold standby fuel because regulation, insurers or the service levels they have sold to customers require it.
These are also the sites where an outage is most expensive. An hour of unplanned downtime in a colocation facility is measured against contractual penalties and lost customer confidence; in a hospital it is measured against patient safety. NFPA 110, the standard covering emergency and standby power systems, treats installation, maintenance, operation and testing of the whole supply system as one subject rather than as a set of separate machines. The fuel is part of that system. In practice, it is usually the part with no owner.
Why does the annual generator test not catch this?
Because the test measures the wrong thing. Running the generator proves the engine starts and the alternator produces power. It says nothing about the water content, microbial population or oxidation state of the fuel still sitting in the tank, and those are the conditions that cause the failure.
There is a second problem with test runs. A routine monthly or annual exercise is usually short and often at partial load, so it draws a fraction of the fuel that a real outage demands. A partially restricted fuel path passes that test comfortably and fails a full-load emergency start. The test regime that was meant to give confidence is frequently the reason the problem stays hidden.
How do you test the fuel itself?
By sampling the bottom of the tank rather than the middle. ASTM D6469 covers microbial contamination in fuels and fuel systems, and it is the water and the growth sitting at the tank floor that a laboratory needs to see. A sample drawn from mid-tank in a contaminated tank will often come back clean.
Anyone who has drawn a bottom sample from a tank that has stood for a few years knows what it looks like next to fresh fuel: dark, cloudy, solids held in suspension and a separate layer at the base of the jar. Two jars side by side usually settle the argument faster than a laboratory certificate, because they show the condition of the fuel the generator is actually being asked to burn.
Cleaning a tank is not the same as protecting one
When a site does discover contamination, the usual response is a one-off tank clean. Someone drains the water, removes the sludge, polishes or replaces the fuel, and the tank goes back into service. That fixes the condition on the day. It does nothing about the mechanism, because the tank will keep breathing, water will keep condensing, and the cycle starts again.
Continuous conditioning takes the other approach. The tank contents are circulated on a schedule through filtration and water separation, so water is removed before it can support growth and particulate is taken out before it reaches an engine.
| One-off tank clean | Continuous conditioning | |
|---|---|---|
| What it addresses | The contamination present today | The mechanism that creates it |
| When it happens | After a problem is found | On a schedule, between tests |
| Effect on water ingress | None, the tank keeps breathing | Water removed before microbes can use it |
| Condition between visits | Degrades again from day one | Held within limits |
Firms working in this field, Evergee among them, frame the difference plainly: a tank clean is treatment, and fuel conditioning systems are protection. The first is something you do after a failure. The second is what stops the tank reaching that state between tests.
What to check this quarter
None of the following needs a contractor, and all of it can be answered from existing paperwork or a short walk to the tank.
- Find the date the fuel was last replaced or tested.
- Ask whether anyone has ever drawn a bottom sample.
- Check the tank has a working water drain point.
- Confirm the annual test run reaches full load.
- Read the maintenance contract to see who owns the fuel.
- Add fuel condition to the continuity plan review.
If the first question has no answer, that is the finding. A tank nobody can date is a tank nobody is managing.
The weakest link
A business continuity plan is only as strong as its least maintained component. Most plans are audited on the assets that are easy to see. The generator has a service record, the UPS has a battery report, the transfer switch has a test date.
The fuel usually has nothing. It was delivered once, signed for, and then left alone in a steel tank outside the building to change slowly for the next several years, on the assumption that fuel is fuel. For most of a plan’s life that assumption costs nothing. It gets tested once, on the night the power fails.
Leave a Comment