A generator’s autonomy is set by the usable fuel it can reach, not by the size written on the tank. Diesel is drawn from 25 to 50 mm above the tank floor, so a 1,000 litre tank gives about 900 litres of running. This page covers tank arrangements, the autonomy arithmetic, fuel quality and the storage routine.
What decides autonomy, and what the tank does not decide
Autonomy is a division: usable litres divided by litres per hour at the load the set will actually carry. Every other figure on a tank drawing – the nominal volume, the fill point, the gauge – only feeds that division. Two sites can have the same 1,000 litre tank and a difference of a working day in autonomy, because it is the load and the consumption that move the answer, not the tank.
Four numbers settle it. The first is the running load in kW, and it is almost never the rating of the set: a 500 kVA set carrying 300 kW is at 75 % of its prime rating, and that is a different consumption from the full load figure. The second is the fuel consumption at that load, in litres per hour or litres per kWh, which comes from the engine fuel curve; the figures for common sizes are in diesel generator fuel consumption. The third is how much of the tank can actually be used, which is less than the volume printed on the plate. The fourth is the hours wanted, and on most sites that is set by the length of the outages and by how often a fuel truck can reach the site, not by preference.
The tank decides none of those four. It does not decide consumption, which belongs to the engine and the load; it does not decide whether the fuel will still be within specification in a year, which belongs to how the fuel is stored and tested; and it does not decide the refill interval, which is a question for the site and the fuel supplier. What the tank decides is the usable volume – and that is the figure a drawing most often gets wrong, because a nominal 1,000 litres is not 1,000 litres of running time.
| Number | Where it comes from | How it moves the answer |
|---|---|---|
| Running load | The load list, in kW, not the set rating in kVA | A set at 75 % of its rating burns about 20 % less per hour than at full load |
| Fuel consumption | The engine fuel curve, quoted at the same load and the same site conditions | Sets the litres per hour the tank has to deliver |
| Usable volume | The nominal volume, less the heel below the pick-up and the ullage above the fill point | Commonly about 90 % of nominal, and less on a shallow or long tank |
| Hours wanted | The outages the site actually sees, and the refill interval it can rely on | Multiplies litres per hour into the litres the tank must hold |
Tank arrangements: base tank, bulk tank and day tank
There are three arrangements in normal use, and they answer different questions. A base tank is built into the set, usually in the baseframe under the engine and alternator; it travels with the machine, needs no separate foundation, and is the default on an open or canopied set. A bulk tank is a separate vessel, from about 1,000 litres to tens of thousands, sited near the set or at a distance with a transfer line to it; it is what turns a few hours into a few days. A day tank is a small tank close to the engine, fed from the bulk tank, whose job is to keep a short head of clean fuel at the engine and to limit how much fuel sits inside a building or a container.
Volume alone does not tell you which one to use. A base tank is limited by the footprint of the baseframe, so on a 500 kVA set it is measured in hundreds of litres, not thousands, and it is normally sized for a single shift of running rather than for a multi-day outage. A bulk tank has no such limit, but it brings the fill point, the containment, the vent, the gauge and the delivery access with it, and those are siting decisions that have to be made before the concrete is poured. On a containerised set the tank may sit inside the container base, which keeps the footprint down and removes a separate vessel, at the cost of a larger enclosure and a lower ceiling on volume.
The arrangement also decides how the fuel reaches the engine. A base tank under the set is gravity-fed through a short line, which is the simplest case. A bulk tank at a distance is drawn through a longer suction line, which has to be sized so the engine’s lift pump is not working against a restriction, and which is where most priming and air-lock problems start. A day tank removes that problem by putting a small volume close to the engine, but it adds a transfer pump, a level control and a second place for fuel to leak. The choice is therefore about where you want the fuel to physically be, and how much complexity you want between the tank and the injectors.
| Arrangement | Where the fuel sits | What it is good for | What it does not solve |
|---|---|---|---|
| Base tank | In the baseframe, under the set | Transport, a single shift of running, no separate foundation or pipework | Long autonomy; volume is capped by the baseframe footprint |
| Bulk tank | A separate vessel, near the set or at a distance | Multi-day autonomy, deliveries in volume, one tank for several sets | Containment, siting and fire separation; a long suction line has to be designed |
| Day tank | Small tank close to the engine, filled from the bulk tank | A short clean supply, fuel kept outside a building, volume inside the plant room kept small | Total autonomy; it depends on the bulk tank and on the transfer pump |
The autonomy arithmetic, and sizing a tank for the hours you want
Work the calculation from the load outwards, never from the tank inwards. Take the running load in kW, find the litres per hour at that load, multiply by the hours wanted, and divide by the usable fraction to get the nominal volume – then round up to a standard size and check the result against the refill interval.
Take a 500 kVA set at 75 % load, so 300 kW, which burns about 80 litres per hour. Eight hours of running needs 640 litres of usable fuel; at 90 % usable that is a nominal 710 litres, so a 750 litre tank is the smallest that will do and a 1,000 litre tank is the usual answer once the gauge accuracy, the heel and the site’s own margin are allowed for. The same set at full load, 400 kW, burns about 101 litres per hour, so the same eight hours now needs 808 litres usable and a tank of about 900 litres nominal. That difference between 75 % and 100 % load is one of the reasons two quotations for the same outage can come back with different tanks.
Two habits keep the arithmetic honest. First, use the load the set will carry and not the rating on the nameplate; a set sized with headroom runs at part load, and part load uses more litres per delivered kWh, so the autonomy per litre is worse, not better. Second, treat the nominal figure as an upper bound that is never reached: the pick-up pipe stands above the tank floor to keep water and sludge out of the fuel system, and the space above the normal fill level is ullage for expansion rather than running volume.
| Set and load | Fuel | 4 hours | 8 hours | 12 hours | 24 hours |
|---|---|---|---|---|---|
| 100 kVA at 50 % | 13 L/h | 52 L | 104 L | 156 L | 312 L |
| 500 kVA at 75 % | 80 L/h | 320 L | 640 L | 960 L | 1,920 L |
| 500 kVA at 100 % | 101 L/h | 404 L | 808 L | 1,212 L | 2,424 L |
| 1000 kVA at 75 % | 155 L/h | 620 L | 1,240 L | 1,860 L | 3,720 L |
The day tank: when a second tank earns its place
A day tank is not a bigger base tank; it is a different arrangement. It holds a small volume close to the engine and is refilled from the bulk tank, usually by a transfer pump controlled by a level switch, sometimes by gravity where the bulk tank can be set at a higher level. Its value is in three things: the suction line at the engine becomes short, so priming and air locks stop being a commissioning problem; the volume of fuel inside a building or a container can be kept small while the bulk of it stays outside; and on a site with more than one set, a single bulk tank can feed each machine through its own day tank.
Sizing follows the same arithmetic with a different target. A day tank is commonly sized for one to four hours at full load, which is enough for the engine to run while the transfer pump keeps up and for the set to ride through a short interruption in the transfer. It is not there to provide the site’s autonomy, and a day tank without a working transfer pump is a set with one to four hours of fuel. The level control has to be arranged so that the pump starts before the tank is low, stops before it overfills, and sends the overflow back to the bulk tank rather than onto the floor.
Two details are settled at the design stage. The first is elevation: a day tank set above the engine feeds by gravity, but the inlet pressure at the engine has a limit, so a high head has to be checked against the engine maker’s figure rather than assumed. The second is the return line, because a diesel engine returns fuel it has warmed to the tank, and a day tank without a proper return path, or with a return that airlocks, is a set that will not hold load reliably.
| Site condition | Day tank | Why |
|---|---|---|
| Bulk tank remote from the set, or below it | Yes | A short suction line at the engine removes priming and air-lock problems |
| Two or more sets on one bulk tank | Yes | Each engine gets its own controlled supply instead of competing on one line |
| Fuel stored inside a building or a container | Yes | The volume inside the structure stays small while the bulk stays outside |
| Base tank under the set, single set, outage shorter than a shift | No | Gravity feed over a short line already answers the question |
Fuel quality: water, microbes and what polishing does
Diesel does not mix with water, but it carries water in suspension and takes more in from two directions: condensation on the tank walls as the air above the fuel expands and contracts, and contaminated deliveries. Because diesel is lighter than water, the water settles to the bottom and collects at the lowest point in the tank, which is why the drain and the water separator exist and why the pick-up is set above the floor.
That layer of water is also where the problems start. Micro-organisms grow at the fuel and water interface and produce the dark sludge that blocks filters and reaches the injectors; the same water corrodes the tank from the inside and shortens the life of a steel vessel. The first symptom is usually a filter that blocks long before its interval, or a set that starts to hunt or lose power under load. On a coastal site this is not an occasional event: the day and night temperature swing drives condensation all year, and the mechanisms are set out in diesel generators for hot and humid climates.
Fuel polishing is the answer to contamination that has already happened. It circulates the stored fuel out of the tank, through a coalescer and filters, and back in, on a timer or on a schedule, removing free water and particulate. It is worth fitting where fuel sits for months, where the tank is large enough that a full replacement is expensive, or where a site has already had a contamination event. What polishing does not do is repair fuel that has gone out of specification chemically; it removes water and solids, and the fuel’s own quality still has to be established by a test.
| What is seen | What it usually is | What to check | What clears it |
|---|---|---|---|
| Filters blocking before their interval | Water and microbial sludge carried from the tank | The tank drain, the water separator bowl and a fuel sample | Drain, clean, treat, and polish if the tank is large |
| Dark sludge in the separator or on the filter | Microbial growth at the fuel and water interface | Free water at the tank bottom and a laboratory test | Remove the water first, then treat; sludge does not clear on its own |
| Hard starting, hunting or lost power under load | Restricted flow at the filters or at the pick-up | Filter differential, pick-up condition, suction line for air | Restore flow, then find where the contamination came from |
| Rust or pitting inside a steel tank | Water standing at the bottom over a long period | The tank bottom, the drain point and the tank age | Remove the water, keep the tank fuller, and inspect the vessel |
Storage life, testing and the top-up habit
Diesel is not a product that keeps indefinitely. Kept clean, dry and out of the sun, it commonly stays within specification for something between six months and a year, and the interval shortens with heat, with water in the tank and with repeated small top-ups that never turn the stock over. The grade matters too: the fuel the engine is rated on is named in the datasheet, and in most markets that is a low-sulphur automotive diesel to EN 590 or ASTM D975 rather than a heavier gas oil.
Testing is how the site knows rather than hopes. A laboratory fuel test normally reports water and sediment, particle count and microbial growth, using the standard methods for each – water and sediment by ASTM D2709, particle count by ISO 4406 – and it is worth having one annually, and again after any event that puts water into the tank. The result is what tells you whether the stock is still fuel or has become a disposal problem, and it is also the evidence a supplier needs if a delivery turns out to be contaminated.
The habit that keeps a tank usable is simple and unglamorous. Keep it as full as practical, because a full tank has less air above the fuel and therefore less condensation. Draw the water off the separator and the tank drain on the interval in the maintenance schedule rather than waiting for a symptom. Run the set under load so the fuel is turned over instead of standing, and record the level each time so a slow loss becomes visible as a trend rather than as a surprise. Additives have a place where growth has already been found, but a biocide treats the symptom; the water it grows in still has to come out.
| Check | Interval | What it tells you |
|---|---|---|
| Water drawn from the separator and the tank drain | Monthly, and after every delivery | Whether water is arriving, and how fast |
| Level reading against the log | Monthly | Consumption, and any loss that is not consumption |
| Run on load and record the readings | Monthly | That the fuel reaches the engine and the set carries load |
| Laboratory fuel test | Annually, and after a water event | Whether the stock is still within specification |
| Bund, vent and fill point inspected | Annually, and after any spill | That containment is intact and the fill point is not a leak path |
| Filter and separator element change | The engine maker’s interval, shortened if contamination is found | That flow is not being restricted by loading the element early |
Containment, level and leak: what the tank has to have
A tank is also a piece of containment, and on most sites the containment is what the local authority and the insurer look at rather than the volume. The common arrangement for a single bulk tank is a secondary containment – a bund, a concrete curb or a double-skinned vessel – sized to hold the contents of the largest tank plus a margin, with 110 % a widely used figure. On a double-skinned tank the same duty is done by the interstitial space and a leak detector rather than by a bund around the outside.
The fill point is the other place a spill happens, and it is worth arranging so that a hose left open or a tanker that overfills discharges into the containment rather than across the yard: a drip tray or a small local bund at the fill connection, an overfill alarm or a high-level cut-off, and a vent arranged so that liquid cannot be pushed out of it. Level indication belongs on the tank as well as in the office – a local gauge for the person standing at it, and a low-level alarm that reaches the set’s controller or the site’s monitoring so that an emptying tank is noticed before the engine stops.
Two items are frequently left out. The first is bonding and earthing: filling a tank from a road tanker generates static, the tank and the fill point are bonded to each other and to earth, and the bond is made before the hose is connected rather than after. The second is the marking and the paperwork – the fuel grade, the capacity, the fill point and the drain point labelled, and a spill kit within reach of the fill connection rather than in a store. Where the tank is inside a plant room, the same items apply plus whatever fire separation and ventilation the local code requires for that room.
| Item | What it is for | Typical form |
|---|---|---|
| Secondary containment | Holds the contents if the tank or a connection fails | A bund or curb sized to about 110 % of the largest tank, or a double-skinned vessel |
| Fill point and drip tray | Catches what spills when a tanker is connected or overfills | Camlock or dry-break coupling inside a local tray |
| Vent | Lets the tank breathe while filling and emptying | Flame-arresting vent, arranged away from openings |
| Level indication | Shows how much fuel is there before it runs out | A local gauge, plus a low-level alarm wired to the controller |
| Overfill protection | Stops a delivery overfilling the tank | A high-level alarm, or a cut-off that stops the transfer or the delivery |
| Drain valve | Lets water and sediment out of the lowest point | A valve at the tank floor, accessible and lockable |
| Bonding and earthing | Discharges static before and during filling | A bond from the fill point to earth, made before the hose is connected |
| Leak detection | Turns a slow loss into an alarm rather than a stain | An interstitial sensor on a double-skinned tank, or a gauge trend |
Keeping the fuel usable on a standby site, month by month
A standby set is the hardest place to keep fuel in good order, because the fuel sits rather than circulates. The routine that works is the same on a hospital, a data hall and a telecom site, and it is built on three ideas: water comes out early, fuel is turned over rather than stored, and the level is recorded so a trend can be read.
Each month, draw the water off the separator and the tank drain into a clear container and look at it; a small amount is normal, a steady amount is a delivery or a condensation problem, and a growing amount is a tank problem. Record the level against the last reading, so that consumption and loss can be told apart. Run the set on load for long enough to reach normal temperature, and record the readings rather than simply noting that it started. Each year, have the fuel tested, inspect the bund, the vent and the fill point, and change the filters on the engine maker’s interval – and sooner where the test says the fuel is carrying water or solids.
Where a site cannot follow that routine, the honest answer is a different arrangement rather than a longer interval. A tank owned and replenished by a fuel supplier, a site with no staff to take a reading, a plant room whose local code sets its own inspection regime, and a tank whose fuel is never consumed all sit outside the pattern above, and each of them is settled on the site’s own rules and the fuel supplier’s schedule. This page describes general practice; the tank, its containment, its venting and its inspection belong to the local authority, the fire code and the insurer for the particular site.
The mistakes that cost the most
The failures we see around fuel storage are rarely exotic, and nearly all of them are decisions taken at the quotation or the layout stage rather than at the tank.
| Mistake | What it costs | What prevents it |
|---|---|---|
| Reading the nominal tank volume as autonomy | A set that stops hours before the site expects it to | The usable-volume calculation, with the heel and the ullage taken off first |
| Sizing the tank on the set rating rather than the running load | A tank that is too small at part load, or money spent on volume that is never used | The load list in kW, and the consumption at that load |
| No refill plan for a long outage | Autonomy that ends at the first day of a multi-day event | Agreeing the refill interval with the fuel supplier before the tank is ordered |
| A bulk tank at a distance with an undersized suction line | Priming, air locks and a set that will not hold load | Pipe sizing and a day tank arrangement where the run is long or the tank is low |
| Never draining water from the tank | Microbial sludge, blocked filters and injector damage | The monthly drain, recorded, and an annual fuel test |
| Fuel left standing for years without a test | A tank of out-of-specification fuel at the moment it is needed | Turning the stock over, testing annually, and polishing where the volume is large |
| Containment and bonding treated as optional | A spill, a failed inspection, or a static discharge at the fill point | Bunding sized to the local rule, and a bond made before the hose is connected |
Questions about fuel tanks and autonomy
How do I work out how many hours a tank will run?
Divide the usable fuel by the litres per hour at the load the set carries. A 1,000 litre tank is about 900 litres usable; a 500 kVA set at 75 % load, so 300 kW, burns about 80 litres per hour, which gives roughly 11 hours. The same tank on the same set at full load, 400 kW and about 101 litres per hour, gives about 9 hours, so the load has to be stated before the number means anything.
What size tank does a 500 kVA generator need for eight hours?
At 75 % load, 300 kW, the set uses about 80 litres per hour, so eight hours needs 640 litres usable, which is a nominal tank of about 710 litres – a 750 litre tank as the smallest that will do, and 1,000 litres as the usual practical size once gauge accuracy and a margin are allowed for. At full load, 400 kW and about 101 litres per hour, the same eight hours needs a nominal tank of about 900 litres.
What is a day tank, and when is one needed?
It is a small tank close to the engine, refilled from the bulk tank, commonly sized for one to four hours at full load. It is worth fitting where the bulk tank is remote from the set or below it, where two or more sets share one bulk tank, or where fuel is stored inside a building or a container and the volume inside the structure has to be kept small. It is not there to provide the site’s autonomy.
Does diesel go off in storage?
It does, and the practical figure is commonly six months to a year for fuel kept clean, dry and out of direct sun, and less in high ambient temperature or where water is present. The failure is not usually the fuel itself but what grows in it: microbes live at the fuel and water interface and produce the sludge that blocks filters. An annual laboratory test, with water and sediment and particle count reported, is how a site knows rather than assumes.
What is fuel polishing, and does it fix bad fuel?
Polishing circulates the stored fuel out of the tank, through a coalescer and filters, and back in, removing free water and particulate. It is useful where fuel sits for months, where the tank is large enough that replacing the contents is expensive, and after a contamination event. It does not restore fuel that has gone out of specification chemically – that is decided by a test, not by the polishing unit.
How much fuel should be kept in a standby tank?
Keep it as full as practical, because the air space above the fuel is where condensation forms, and top up after every draw so the stock turns over. A tank that is kept full and never used is the other failure mode: the fuel ages in place. The workable habit is to keep the tank near full, run the set under load so the fuel circulates, test the stock annually, and record the level so a loss is visible as a trend.
Can fuel be stored in a plastic tank?
Small portable containers are one thing; bulk storage is another. Bulk diesel is normally held in a steel vessel with secondary containment, a flame-arresting vent, a drain at the lowest point and bonding to earth at the fill point, because the risks are static during filling, fire separation and the volume that would escape if the vessel failed. The size, the material and the containment are set by the local code, and that is the document to check rather than a general answer.
Send the load profile, not only the kVA
Give us the running load in kW for each group of loads, the hours per outage and per year the site expects, the longest gap between fuel deliveries, and whether the tank will be under the set, separate from it, or inside a building. The reply carries the litres per hour at your load, the usable and the nominal tank volume, the autonomy in hours, and whether the arrangement needs a day tank.
Written and maintained by the KVA Grid sales engineering desk. The consumption figures behind every autonomy calculation on this page are the ones set out in diesel generator fuel consumption, and the effect of ambient temperature and humidity on stored fuel is in diesel generators for hot and humid climates. Tank siting, containment, venting and inspection are settled by the local authority, the fire code and the insurer for the particular site, and the fuel grade is the one the engine datasheet is rated on; this page describes general practice rather than a particular rule. A set that has been drained for shipment is handled as described in generator lead time, packing and shipping terms, and the cost of the fuel over the life of the machine sits inside total cost of ownership. Nothing on this page is a quotation for a particular project. Reviewed 11 October 2026.
