Reading a tank dip chart without fooling yourself
A dip stick gives you a height. Turning that into a volume you'd put on a delivery note involves more traps than most people expect — and nearly all of them live in the middle of the tank.
A dip chart — a strapping table, if you're in fuel — maps liquid depth to volume for one specific tank. Drop the stick, read the wet mark, look up the number. Simple, until you notice two operators reading the same tank an hour apart and disagreeing by 400 litres.
Innage or ullage — make sure you're reading the number the chart expects
One question causes more disagreements than tank geometry ever will: is the number on the tape the depth of liquid, or the depth of empty space above it? Both are legitimate readings and both end up on paperwork, but they are not the same number, and a chart built for one will silently hand you nonsense if you feed it the other.
Innage is the liquid depth, measured up from the tank's datum plate or bottom. This is what the segment formula above expects, and what most horizontal-tank strapping charts are keyed to. Ullage — "outage" in most US petroleum-measurement usage, though site conventions vary and sometimes invert these two names — is the empty space, measured down from a fixed reference point near the top of the tank (the gauge hatch, or a marker on the gauge well) to the liquid surface. Large vertical storage tanks are very often gauged this way, because a reference point at the top is easier to seat a tape on precisely and repeatably than a bottom that may have sediment on it.
Converting between them needs one more figure: the reference gauge height (RGH) — the fixed distance from the reference point down to the tank's strapping datum, stamped on the tank's calibration certificate. From there:
Innage = RGH − ullage
A worked case. A vertical tank has an RGH of 4.000 m and an internal diameter of 3.0 m. An operator gauges an ullage of 1.150 m, so innage is 4.000 − 1.150 = 2.850 m. Treating the shell as a plain cylinder for a moment — a real dished- or coned-bottom tank needs the head correction discussed below — volume is π × 1.5² × 2.850 ≈ 20,145 litres. You can check that arithmetic directly with a cylinder volume calculator, or run the whole tank, heads included, through the tank volume calculator.
Two traps hide in that one conversion. First, "ullage" and "innage" get used loosely — some companies' gauge tables call the empty-space reading "ullage," others reserve that word for the liquid depth and call the empty space "outage," the opposite of the definitions above. Confirm which convention a specific chart uses, especially on a site that's changed operators or gauging software. Second, the reference point for ullage isn't always the top of the shell — a gauge hatch that sits proud of the roof, or a stilling-well flange, adds tens of millimetres that the certified RGH already accounts for but a quick tape-measure-up-the-outside estimate won't. Use the certified RGH, not a field guess.
Why the chart isn't linear
In an upright cylinder, it is: every centimetre of depth is the same slice of volume, and the chart is a straight line. Nobody needs a chart for that.
A horizontal cylinder is the interesting case, and it's the common one for fuel, chemicals and process storage. The liquid surface is a chord across a circle, and the wetted cross-section is a circular segment. Near the bottom, the tank is narrow, so a centimetre of depth is barely any volume. At the centreline, it's at its widest, and that same centimetre is worth the most it will ever be. Then it narrows again toward the crown.
The segment area for a tank of radius r at liquid depth h is:
A = r² · cos⁻¹((r − h)/r) − (r − h) · √(2rh − h²)
with the inverse cosine in radians. Multiply by the cylindrical length for volume. That's the whole of the "shell" calculation — every dip chart for a plain horizontal cylinder is this formula tabulated.
Tank volume calculator → Horizontal, vertical and dished-end tanks from a dip reading, in litres or gallons.What that shape means on the floor
Take a 2.5 m diameter, 6 m long horizontal tank — roughly 29,000 litres. One centimetre of dip error near the bottom (say at 10 cm depth) is worth about 59 litres. The same centimetre at the centreline is worth about 150 litres. Two and a half times the consequence for the same sloppy reading, and it happens exactly where tanks spend most of their working life.
The practical rule: a dip near half-full is the least forgiving reading you will take. If a number matters — a transfer, a reconciliation, a delivery — take it when the tank is well off the centreline, or take it three times and average.
The corrections people skip
Dished ends
Almost no real tank is a plain cylinder. The ends are dished — torispherical, ellipsoidal, or hemispherical — and each end adds volume that the simple segment formula knows nothing about. On a short, fat tank the two ends can be a tenth of total capacity. Treating a 2:1 ellipsoidal-head tank as a plain cylinder will under-read it by roughly that much, all day, consistently. That's not noise you can average away; it's a bias.
Get the head type off the nameplate or the fabrication drawing. If you only have the tank in front of you: hemispherical ends bulge out by a full radius, 2:1 ellipsoidal by half a radius, and torispherical (the most common in cheap tanks) by rather less.
Tilt
Most tanks are deliberately laid a degree or two off level so they drain to the outlet. That tilt means the depth at the dip point isn't the average depth in the tank — and the error switches sign as the tank fills. A 6 m tank sloped 1° has one end sitting about 100 mm lower than the other. Dip at one end and you're reading a value that can be several hundred litres away from the true contents.
This is one of the main reasons a calibrated chart for that specific tank, prepared by a strapping survey, beats a chart calculated from nominal dimensions. The survey folds in the tilt, the real diameter, and any deformation, because it measures what the tank actually does rather than what the drawing says.
Temperature
Liquids expand. Hydrocarbons expand a lot — roughly 0.1% per °C for gasoline, less for diesel. A 30,000 litre tank of gasoline that warms 10 °C over a day gains about 300 litres of volume without a single molecule being added. The dip goes up, the mass is unchanged.
Where volume is money, this is handled by correcting observed volume to a standard temperature (15 °C or 60 °F, depending on the standard). If your reconciliation shows tanks mysteriously gaining in the afternoon and losing overnight, you're seeing thermal expansion, not theft.
The standard tool for that correction is a volume correction factor (VCF) — a multiplier looked up from tables such as API MPMS Chapter 11.1 (the API/ASTM/IP tables) by product density and observed temperature, that converts an observed volume at whatever temperature it was gauged to a volume at the standard reference temperature. A VCF of 0.9930, for instance, means "this observed volume reads 0.70% high" — multiply by it to correct down. The tables aren't a flat percentage-per-degree because the expansion coefficient itself varies with density: light products like gasoline expand noticeably faster per degree than a heavy fuel oil, so the same 10 °C swing corrects by a different amount on two different tanks.
It matters most on custody transfer, high-value product, or any large tank with a wide day-to-night swing — a few tenths of a percent on a multi-million-litre terminal tank is real money, and skipping the correction is exactly how two honest gaugings a day apart end up looking like a discrepancy worth investigating. It matters far less on a small process tank you're topping up by eye, or over a reconciliation window short enough that morning and evening temperatures roughly cancel. And it's a different mechanism entirely from the volume change you get when two miscible liquids are blended — see why alcohol plus water is less than the sum for a case where volume shrinks from mixing rather than temperature, and where both effects can be acting on the same tank at once.
The bit at the bottom
Sludge, water bottoms and sediment sit below the liquid you care about, and the chart counts their depth as product. Water-finding paste on the stick tells you where the water line is; subtract it. In a tank that's been in service for years, the unusable bottom can be a surprising fraction of what the chart calls "contents".
Troubleshooting: when the reading doesn't add up
Most "the tank must be leaking" calls turn out to be one of a short list of mundane causes. Work through these before you trust an alarming number.
- Stuck float. Mechanical float-and-tape or servo gauges ride a guide wire or stilling well, and product buildup — wax, crystallised salts, polymerising resin — can gum the float or its guides so it stops tracking the surface. The tell is a reading that barely moves across a fill or draw that should have changed it a lot, or one suspiciously identical to yesterday's. Pull the gauge and check it moves freely through its whole travel; a float that's absorbed product and gone heavier than its calibration assumes will ride low and read high.
- Wrong chart revision. Any internal change — a new heating coil, an agitator, a repaired and re-profiled bottom, even a replaced ladder or column — shifts the strapped volume at every depth below the change, and the tank should get a new chart. Old revisions have a habit of staying taped up nearby, or saved under the same filename as the current one. If a tank's numbers have been quietly off by a consistent amount since a turnaround or repair, check the chart's revision date and issue number against the tank's modification history before assuming the tank itself has moved.
- Datum has shifted. A tank on a soft foundation settles over years, unevenly — really the tilt problem above, but slow enough that nobody notices until it's several hundred litres worth. A resurvey resets the datum; a chart from ten years and one ground-settlement cycle ago is no longer describing the tank in front of you.
- Units or scale mismatch. A chart printed in inches read against a tape marked in millimetres, or a chart tabulated in whole-litre steps being interpolated to a tenth by eye — sounds too obvious to matter, but it's a common cause of a "discrepancy" that vanishes the moment someone checks the chart's header row instead of just the numbers in it.
- Reading during movement. Worth repeating as a troubleshooting step in its own right: if a suspicious number was taken during or right after a fill, transfer or agitation cycle, re-gauge once the surface has settled before treating it as real.
Building your own chart
You can calculate one from dimensions, and for a tank holding water in a workshop that's fine. Measure the internal diameter, not the external one — the wall thickness matters. Grab the outside diameter of a 3 mm-wall, 1.5 m tank by mistake and you're off by 0.4% on diameter, which is already pushing 0.8% on volume. Then measure the cylindrical length between the tangent lines where the heads start, get the head type, and tabulate the segment formula every centimetre.
For custody transfer, product accounting, or anything an auditor will look at, use a strapped chart from a proper survey. The calculated version tells you what an ideal tank of those dimensions would hold. The strapped one tells you what your tank holds, which is the number that has to balance.
Strapping chart generator → Produce a depth-to-volume table at your chosen increment, ready to print for the tank.A short checklist
- Dip against a fixed reference point, the same one every time — a datum plate, not "the top of the flange, roughly".
- Confirm whether the chart is keyed to innage or ullage, and get the reference gauge height from the calibration certificate, not a tape measure up the shell.
- Let the surface settle. Reading during a fill or right after agitation is guesswork.
- If it's a mechanical float gauge, confirm it moves freely through its full travel before trusting a static-looking reading.
- Check for water bottoms before you trust a low reading.
- Know whether your chart is calculated or strapped, whether it includes the heads, and check its revision date against the tank's repair history.
- Note the temperature alongside the dip if the product is worth a VCF correction.
- Treat near-half-full readings with extra care — that's where a centimetre costs the most.