From geometry to a number someone trusts
A tank calibration chart looks simple. A column of levels, a column of volumes, and a relationship between them that inventory, custody transfer and process control all depend on. The simplicity is the point — and it hides a fair amount of decision-making about how those numbers were produced.
If you are specifying a calibration and comparing quotes, the questions worth asking are about method and datum rather than equipment. This explains how a chart is produced from 3D scan data, where the published standards apply and where they do not, and what a competent deliverable contains.
Where the standards sit
Tank calibration is a well-standardised field, but the standards are narrower than people assume.
The ISO 7507 series covers the calibration of vertical cylindrical tanks, across several methods: manual strapping, optical reference line, optical triangulation, and internal and external electro-optical distance ranging. The API Manual of Petroleum Measurement Standards covers the same ground from a different direction, with Chapter 2.2A addressing upright cylindrical tanks by manual strapping, and Chapter 2.2E covering horizontal cylindrical tanks by manual methods.
The shape of that landscape tells you something useful. The standards are strongest where the geometry is regular and an assumed shape holds. They are most stretched where it does not — horizontal tanks with dished ends, vessels that have deformed in service, and anything with internal structure displacing volume. Those are exactly the cases where measuring the real geometry, rather than assuming an idealised one, changes the answer.
If your application is subject to a specific standard or a regulator’s requirement, confirm which one applies before scope is agreed. It determines the method, not the other way round.
What the scan actually gives you
Terrestrial laser scanning measures the surface it can see. For internal calibration that means the vessel needs to be empty, clean enough to scan, and safely accessible — a real constraint, and one worth establishing early because it usually drives the programme more than the survey does.
What you get in return is the vessel’s actual shape rather than its nominal one. Ovality, shell distortion, the true profile of the dished ends, the seating of internal structure. For a vessel that has been in service for twenty years, the difference between actual and nominal is not noise. It is the reason the old chart no longer agrees with the level instrumentation.
How the chart is built
The sequence matters, and it is where methods differ. Ours runs like this.
The registered point cloud is cleaned first, then a best-fit cylinder is fitted to the shell data as a geometric reference — not as the model, but as an aid for assessing the real surface against an ideal one. The shell and dished ends are then meshed from the measured data and merged into a single watertight mesh, repaired, with normals unified. That watertight mesh is the volumetric model. Everything after this is a calculation against it.
The datum is set next, and this is the decision that most affects the result. The dip reference is placed at the top of the striker plate as zero, so the chart reads against the point the operator actually dips to. Volume is then calculated from elevation in fine increments — typically 2 mm steps from the datum — producing the level-to-volume relationship across the full working range.
The last step is conversion into whatever format the site actually uses, whether that is a traditional dip chart or a level-sensor calibration table, and a QA pass comparing the summed chart against the enclosed volume of the watertight mesh. Any material mismatch gets investigated before issue rather than explained away.
The slope decision
One point deserves singling out, because it is counter-intuitive and it is where charts go wrong.
An installed horizontal tank is rarely level. It sits on its saddles with whatever slope it was set with, and that slope is usually deliberate — for drainage, or for the dip point to work as intended. The temptation when processing is to level the model to its own centreline, because a levelled cylinder is tidier to calculate against.
That is a mistake. Levelling the model removes the installed slope, and the volume-at-level relationship it produces is no longer the relationship the operator sees when they dip the tank. We keep the vessel in its surveyed orientation for the final calculation for exactly this reason, and we have tested the alternative: a centreline-levelled trial did not reconcile against the mesh volume, and preserving the surveyed slope did.
If you are comparing methods, this is a good question to ask. How is the installed orientation handled, and does the issued chart reconcile against the model volume?
What a complete deliverable contains
At minimum: the calibration chart in the format your systems consume, the increment it was produced at, the datum definition, and a statement of how the chart was verified.
Beyond that, a tank and vessel calibration scope would normally include the registered point cloud, the watertight mesh the volumes were derived from, and a note on any region where the chart was extrapolated rather than measured — the volume below a striker-plate sensor offset, for example, which is commonly populated from the slope-corrected model where the client asks for it.
That last point is worth stating plainly, because it is the kind of thing that quietly differs between suppliers. Values derived rather than measured should be identified as such in the deliverable.
The secondary benefit
A calibration scan captures the whole vessel, not only what the volume calculation needs. That dataset supports damage and deformation assessment against a future survey, reverse engineering of components where drawings are missing, and as-built modelling of the surrounding pipework if the capture is extended.
If a vessel is being opened and made safe for internal access, that access is the expensive part. It is usually worth deciding at scoping whether the capture should cover more than the calibration alone requires.
Scoping a calibration
The practical starting point is the vessel type, its installed orientation, the format your systems need the chart in, and whether a standard is mandated for your application. With those four, we can define a method and tell you what access the work requires.