When Should Tank Roundness, Verticality, and Levelness Be Measured?

Direct answer: measure the named characteristic in its declared state

There is no single best measurement state for tank roundness, verticality, and named-feature geometric levelness. Each result needs its own project-defined characteristic, physical state, reference, method, purpose, and decision authority.

• Roundness has no universal best state. A project-defined roundness result becomes interpretable only when the evaluated shell region and elevations, assembly, support, restraint, project-defined fill or load, temperature, reference or fitting basis, coverage, and method are named.
• Verticality has no universal best state. A project-defined verticality result depends on the evaluated axis, generatrix, centerline, or other feature and height interval, together with the actual support, foundation, levelling, restraint, assembly, fill or load, temperature, vertical reference, observation network, and method.
• Named-feature geometric levelness has no universal best state and does not mean liquid fill level. The project must name the points, plane, rim, support features, or elevation relationship being evaluated, along with the support, foundation and levelling condition, assembly, fill or load, temperature, datum or elevation reference, and method. Liquid fill is a separate state input.

If a field that could change the result's meaning is unknown, hold comparison or acceptance. Do not insert a customary tolerance, fill condition, datum, instrument, or code rule. This editorial decision framework is not prescribed by any one cited source. A result from a different state may remain a process observation or baseline, but must not be silently promoted into delivery acceptance.

Separate the three geometry questions before selecting a state

Roundness asks about a defined shell-form characteristic

Begin by defining what the project means by roundness after the rolled shell has been released into tank assembly. Name the shell region, course, and elevations covered. State whether the requested result concerns a cross-section, radial deviations, a diameter distribution, an ovality expression, or another controlled project definition. Record the reference or fitting basis, units, sectors or points, and scan or observation coverage. This article does not define rolled-shell end condition, longitudinal-seam fit-up, or pre-seam measurement.

A check during course assembly, support installation, joining, correction, restraint, or partial assembly describes that state. It may guide work or trigger review, but it does not automatically describe the later released, supported, delivered, or installed tank. The project decides which state, definition, and criterion matter for the intended decision.

Verticality asks about a feature and vertical reference

Name the evaluated feature before asking whether it is vertical: the tank axis, a generatrix, a centerline, or another controlled feature. Also name the height interval, vertical reference, observation network, and actual support, foundation, and levelling condition. Whole-tank verticality is different from a local nozzle or equipment-interface orientation. A local nozzle coordinate record can control an interface without becoming the default reference for the tank as a whole.

A shop observation made on temporary supports or while a restraint remains active may be a legitimate process result. It cannot establish installed verticality unless the controlled requirement accepts that same characteristic and state, or an authorized project basis explicitly relates the two states.

Named-feature geometric levelness asks which elevations are related

Define the exact subject: named support points, a base or top reference, a rim, an interface plane, or another elevation set. Record the elevation or gravity reference and which contacts carry load. Named-feature geometric levelness is not liquid fill level, does not by itself establish foundation adequacy, and does not prove drainage or cleanability. The adjacent guide on installed support orientation and the named low-point relationship addresses drainage inputs; its low-point reasoning does not turn a geometric-levelness result into a drainability decision.

In one terrestrial-laser-scanning and digital-twin study of a vertical storage tank, Wu and co-authors treated radial deformation, tank inclination, and foundation settlement as separate assessment indicators.^[1] That abstract-level evidence supports keeping different geometry questions visible. It does not define this article's characteristics, require terrestrial laser scanning, transfer an accuracy, set a tolerance, establish acceptance, or prove a StelTank result.

Build one measurement-state tuple around each result

Give every planned or completed result a stable Result ID. Bind it to this tuple:

tank and revision + exact characteristic and coverage + assembly + support/foundation/levelling + restraint/release + project-defined fill/load + temperature + datum/reference realization + method/instrument/data state + purpose + decision authority

The physical state and measurement state are both necessary. A detailed tank description without a realized reference and method leaves the result ambiguous. A detailed instrument record without the tank state answers a different question. Record which contacts carry load; what temporary or permanent items are present; which clamps, braces, fixtures, connections, or other restraints are active; and what state changes could invalidate a comparison.

For the measurement side, record the datum, vertical or elevation reference, how it was physically realized, and whether it was accessible and undisturbed. Add the method and instrument or data identity, setup or observation network, coverage, evaluation or fitting rule, software or data revision, and the uncertainty, calibration, traceability, environment, and status fields required by the project. These are fields to obtain from responsible owners, not values selected by this article.

The public abstract of Peggs and co-authors' review describes multiple large-scale dimensional-metrology techniques, specialized or hybrid applications, and measurement-data traceability.^[2] That bounded metrology context supports recording the selected method and its traceability basis. It does not rank instruments, select a tank method, establish uncertainty or tolerance, or prove that shop and site methods or results are equivalent.

ISO's public record for ISO 5459:2024 identifies the standard's subject as datums and datum systems.^[3] This article uses only that public subject to reinforce the need to identify a reference basis. It reproduces no licensed clause and does not choose a tank axis, support plane, vertical or elevation reference, datum realization, tolerance, method, applicability, or conformity result.

Keep requirement, datum, method, evidence, and delivery state together in a controlled drawing-handoff package; that broader guide does not replace the result-level register below. For large assemblies that move between states, a StelTherm guide discusses reference realization across shipping and reassembly states, but it selects no tank datum, instrument, tolerance, or acceptance state.

Reopen a result when its state changes

Reopen means review whether the project needs a comparison, repeat measurement, or authorized disposition. It does not mean the tank moved, failed, or must be remeasured after every operation. Start the review when a change touches a field on which the earlier result depends.

Review altered support contact, skid, frame, leg, site-foundation, or levelling conditions. Review activation or release of clamps, braces, fixtures, strongbacks, temporary attachments, connected piping, or other load paths. Review assembly changes involving courses, bottom, head, roof, supports, nozzles, internals, attachments, lining, insulation, or agreed connected items. The responsible project authority decides whether a change affects the characteristic and evidence role.

Sun and Dilger's abstract reports different deformation responses associated with external restraint and welding sequence in its own thin-plate arc-welded specimens.^[4] That qualified analogy supports recording whether a result was taken in a held or released state. It is not a tank study and supplies no tank movement, fixture choice, release sequence, tolerance, or mandatory recheck.

Also review a change to the project-defined fill or load, thermal condition, correction state, transport state, installation, anchoring, or connection. A publisher abstract by Zarubin, Zimin, and Kuvyrkin describes nonuniform temperature, axial compressive load, local generatrix deviation, and radial displacement in one model of a reinforced cryogenic cylindrical tank during filling.^[5] That specific model supports treating fill, thermal, and load conditions as visible state fields. It does not set an empty or full condition, temperature, waiting period, correction, method, or acceptance rule for a chemical tank.

The project's hydrostatic test state guide may identify fill, support, instrumentation, and procedure inputs for a separately controlled test. It does not make hydrostatic testing the required measurement state here, and no test pressure, fluid, venting, safety, hold, or acceptance rule transfers. Likewise, a StelTherm discussion of restraint and release alignment comparison offers an adjacent state-recording example; it provides no tank support condition, tolerance, or result.

Reopen the relevant entry if a datum is damaged, inaccessible, or recreated, or if the network, coverage, fitting rule, method, software, data revision, or instrument state changes. Support, foundation, levelling, anchorage, shim, elevation, load-distribution, and site-correction decisions remain inputs from responsible design and installation authorities; this article neither selects nor validates them.

Assign the evidence role before assigning acceptance

Label what each record may prove before anyone compares its value with a criterion. The four roles below are editorial categories, not categories imposed by a paper or standard:

• Process observation: supports manufacturing control, review, or correction in its named state; it does not accept another state.
• Baseline: is retained for an authorized comparison after a change; it is not a tolerance result or acceptance record by itself.
• Factory handoff result: describes the named shop or delivery support and assembly state; it does not automatically establish installed verticality or named-feature geometric levelness.
• Potential acceptance record: becomes acceptance evidence only when the controlled requirement names the same characteristic, physical and measurement state, reference, method, criteria, result, retained record, and accepting authority; without that complete chain it remains potential evidence.

A contract may intentionally define a shop-supported state as the acceptance state. This framework does not override that choice. It requires the record to show the choice rather than implying that a later installed state is equivalent. If someone tries to promote a result beyond its recorded state, mark the decision acceptance pending and send it to the named authority. Do not label the result verified, within tolerance, code compliant, or accepted without the complete local requirement, result, record, and authority chain.

Use the Tank Geometry Measurement-State Register

The Tank Geometry Measurement-State Register is an editorial specification tool, not a standard, inspection plan, code checklist, tolerance table, survey procedure, instrument selector, StelTank capability statement, or evidence of measurement. Use two stacked, three-column tables joined by the same Result ID. One paired Part A and Part B row represents one result, not one field; add a pair for each additional Result ID.

The executable key is Result ID + named characteristic + purpose. Keep those facts together in the first Part A cell so an identifier cannot be detached from the geometric question it is intended to answer.

Part A - State record. One row represents one project-defined Result ID, named characteristic, and purpose. It records inputs only and is not evidence of a measurement or acceptance by itself. It is incomplete without the same Result ID in Part B.

In Result ID / characteristic, enter a unique Result ID, tank identity, controlled drawing or model revision, exact characteristic, evaluated feature or shell region, elevations, units, point, sector, or scan coverage, and purpose.

In physical state, enter assembly and relevant surface, lining, or insulation state; support, foundation, and levelling contacts; active restraint, connection, fixture, and release status; project-defined fill or load; and required temperature state. Keep an unknown open with an owner instead of presuming a shop condition.

In reference and method state, enter the datum, vertical or elevation reference, physical realization, accessibility, and condition; method, instrument or data identity; network, coverage, and evaluation or fitting rule; and project-required uncertainty, calibration, traceability, environment, and status fields.

Part B - Evidence and decision record. One row uses the same project-defined Result ID as its Part A row. It is incomplete and non-evidentiary by itself and does not establish acceptance unless the named physical and measurement state, requirement, criteria, result, retained record, authorized disposition, and reopen rule are all present.

In Result ID, repeat the exact Part A identifier and its state-input revision. Never join the tables by row position alone.

In requirement and retained evidence, enter the evidence role, requirement source, characteristic definition, criteria, governing authority, result and units, raw data or report identity, deviation and relevant uncertainty statement, and any authorized comparison or equivalence basis. A document title without retrievable identity and revision does not close the field.

In authority, disposition, and reopen trigger, enter decision owner and written authority, disposition and date, open fields with owners, due gates, and holds, plus each change requiring comparison review, remeasurement review, or another authorized disposition. A disposition applies only within the authority and state shown.

Read Part A to establish characteristic, purpose, physical state, and reference or method state. Read the matching Part B row for requirement, evidence, authority, disposition, and reopen rule. Compare results only after both pairs show which fields are the same, different, or covered by an approved equivalence basis. Keep the decision open if either half is missing.

The blank entries are intentional. Do not invent a tank ID, physical state, datum, method, tolerance, value, result, role, disposition, or authority to make the template look complete. On a narrow screen, each table should scroll inside its own region while cells wrap; the stable Result ID remains the join key. The two caption guards must remain with their tables if either table is extracted or reused.

Route owners, holds, and the handoff package

Assign every open field to the authority that can actually close it. The allocations below are prompts to confirm in the controlled project record, not a universal contract split or evidence of a StelTank service:

• Customer or design authority: provides the characteristic, decision rule, governing documents, required acceptance state, functional references, and acceptance or disposition authority.
• Site, foundation, or installation authority: provides the approved foundation, support, elevation, levelling, anchorage, and installed-reference state. This article does not perform that design.
• Process or test authority: provides any required fill or load, fluid, temperature, stabilization, temporary-test, safety, and procedure state.
• StelTank manufacturing, only within written project scope: reviews measurement access and sequence, proposes process checks, performs agreed execution and recording, and escalates state conflicts assigned to that scope. This list does not prove equipment, accuracy, certification, or capability.
• Metrology or inspection owner: implements the controlled method and reference realization; records instrument and data status, coverage, evaluation, required traceability or uncertainty fields, and limitations.
• Acceptance owner: disposes the result only within recorded project authority.

Hold the affected decision when the characteristic, state, reference, method, criteria, retained record, or authority is missing; when support, restraint, assembly, fill, temperature, reference, or method changes; when results conflict without an authorized comparison basis; or when process evidence is being promoted into acceptance. A hold may apply to one Result ID without stopping unrelated records, but its boundary must be explicit.

Send one controlled handoff package: the current layout, drawing, or model; Result IDs; the paired one-row-per-result register; approved support and foundation input; required fill, load, and thermal state; datum and reference definition; method and evidence request; responsibility record; and open-item list. Mark assumptions as open rather than filling them with shop defaults. The recipient should be able to find each source document, match both table halves, identify the decision owner, and see what change reopens the result.

Evidence boundary and final stop line

No first-party StelTank tank drawing, support or foundation definition, measurement plan, instrument or calibration record, before-and-after dataset, capability study, tolerance result, disposition, or customer acceptance record supports this general article. The cited external sources support only bounded state and measurement-planning principles. They prove no StelTank capability, method, accuracy, or project result.

If the exact characteristic, paired state and evidence record, and authorized acceptance basis cannot be named, the team cannot yet defend the result as delivery acceptance. Complete the open register fields and return them through controlled project review.

References

1. Y. Wu, G. Yang, J. Sun, L. Cui, and M. Wang, "Construction of a Digital Twin Model for Vertical Storage Tank Deformation Assessment Using Terrestrial Laser Scanning Technology," Measurement Science and Technology, 35(3), 035206. DOI. Abstract-level use only; the study does not define a StelTank method, accuracy, characteristic, or acceptance. Back to citation
2. G. N. Peggs et al., "Recent Developments in Large-Scale Dimensional Metrology," Proceedings of the Institution of Mechanical Engineers, Part B, 223(6), 571-595, 2009. DOI. Public-abstract and review-context use only; no tank method ranking, tolerance, uncertainty, or equivalence follows. Back to citation
3. ISO 5459:2024, Geometrical Product Specifications (GPS) - Geometrical Tolerancing - Datums and Datum Systems. Official ISO record. Public title and subject only; no clause, project datum, tolerance, method, applicability, or conformity is supplied. Back to citation
4. J. Sun and K. Dilger, "Influence of Welding Sequence and External Restraint on Buckling Distortion in Thin-Plate Arc-Welded Joints," Journal of Advanced Joining Processes, 8, 100157, 2023. DOI. Institutional-abstract-level use; it does not establish tank behavior or a released-state rule. Back to citation
5. V. S. Zarubin, V. N. Zimin, and G. N. Kuvyrkin, "Deformation of the Shell of the Cylindrical Cryogenic Tank During Its Filling," Russian Aeronautics, 62(3), 364-372, 2019. DOI. Publisher-abstract-level use; the specific cryogenic model supplies no universal fill, temperature, method, or acceptance rule. Back to citation