How to Review Structural Analysis Output Before Preparing Drawings

Analysis output is not a drawing, and it is not automatically a design. It is a pile of displacements, forces, reactions, maybe utilisation ratios, and a warning log. Review is the act of deciding which of those numbers describe the building you intend to freeze. Starting CAD because the solver finished is how pretty elevations document a model nobody believed.

This article sits after STAAD.Pro model checking and before model to CAD drawing. The same discipline applies if the engine was MBS or another program: read deflections for sense, read reactions as a document, read governing members against the load cases you meant to run, and refuse to freeze if the story is inconsistent.

You will not find span limits, allowable ratios or combination formulas here. Those live in the project design basis and the standards it names. Review asks whether the output is coherent and ready to become approval drawings or steel sizes on a schedule — not whether a blog thinks a number is “typical.”

Output is a story; read the plot before the climax

Start with warnings, then global behaviour, then local numbers. Jumping to the highest utilisation hides a unit error that made every ratio meaningless. If the building sways the wrong way under a documented wind case, member design is entertainment.

Keep the drawings of the intended system open while you review: portal frames, bracing, base types. Output that contradicts those drawings means either the model or the drawing is already wrong. Reconcile before CAD multiplies the lie.

Displacements, forces, reactions, utilisation — four different questions

Displacements ask whether stiffness and supports are in the right universe. Member forces ask whether the load path you drew exists. Reactions ask what the foundations will be told. Utilisation asks a code-check question that is only meaningful after the first three.

Displacements

Look at a gravity case and a lateral case separately. Gravity should go down. Lateral should go the way the load points. Wild rotation at a joint is a release or connectivity problem, not a “flexible” building.

Member forces and utilisation

Governing members should be the ones you would expect — knees, braced-bay chords, crane columns — not a dummy beam you forgot to isolate. If ratios exist, confirm they used the combinations from the load-case discipline, not a trial envelope.

Reactions

Treat them as an issued document: STAAD reactions or PEB column reactions. Review is not complete if civil would still have to guess axes.

What must be frozen before CAD starts

CAD needs geometry and the sizes you are willing to show. It does not need every diagram. Freeze: envelope, member layout, section list, support story, brace bays, and the analysis file name. Do not freeze a size that only exists in an unreviewed ratio plot.

Red flags that mean stop plotting

Some output patterns are stop signs. They are not always model bugs — they can be real behaviour you have not designed for — but they are never “plot anyway.”

• Instability or zero-stiffness warnings left unexplained.
• Deflections that reverse the applied load direction.
• Pinned bases reporting significant moments.
• Tension in every “dead only” column (units or sign convention).
• A single leftover TEST combination governing all steel design.
• Reactions that do not exist for a column that CAD already showed.
• Tapered PEB frames analysed as prismatic without a documented approximation.
• Bracing members carrying bending they cannot physically take.

A review sequence you can time-box

On a simple PEB, this is a sitting, not a week. On a messy industrial model, it is still the same order. Time-boxing without order is just rushing the ratios.

1. Warnings and completion messages.
2. Displacement snapshots for gravity and one lateral combination.
3. Support reactions: sample typical, corner, braced, special.
4. Governing members versus expected load path.
5. Size list versus properties versus what you plan to draw.
6. Notes: filename, basis, unresolved items.
7. Permission to convert geometry — e.g. via Staad2CAD — only after that permission is explicit.

Output review to drawing permission

The last step is a yes/no: may CAD begin?

1. Archive the analysis file you are reviewing; do not review a moving target.
2. Read warnings to zero unexplained items.
3. Walk displacements and a sample of forces.
4. Produce or update the reaction table as a document.
5. Produce the member-size list that drawings must use.
6. Write the three-line summary (units, combinations, warnings).
7. If any red flag remains, return to the model; do not start sheets.
8. If clear, issue drawing permission with the file name CAD must cite.

Engineering tips

• Review on combinations you will issue, not only on a pretty single case.
• If two engines exist (MBS and STAAD), compare a few reactions and a few sizes; do not ignore a mismatch.
• Keep screenshots only as working notes; the issued artefacts are tables and PDFs with revisions.
• A second reviewer should be able to repeat the gravity-deflection glance without coaching.
• If utilisation is not used (sizes from another engine), still review forces and reactions in STAAD if it is a check model.
• Do not let a colour contour replace a table of maxima with combination names.

Output-review failures that CAD then cements

Starting elevations because “ratios are under 1”

Ratios can be under 1 for the wrong combinations, the wrong unbraced lengths, or a model missing self-weight. Read the story first.

Reviewing only the GUI, never a printed warning list

The GUI hides what you did not click. Print or export the log.

Freezing CAD geometry while sizes are still “being optimised”

Then the xref is stale every afternoon. Freeze sizes for the issue, or freeze only grids and wait on sizes.

Ignoring a reaction mismatch with MBS because STAAD “is only a check”

A check that disagrees is information. Resolve or document; do not shrug.

Analysis-output review checklist

All boxes before drawing permission.

• Analysis file name and revision frozen.
• Unexplained solver warnings: none.
• Displacement directions make sense for sampled cases.
• Load combinations reviewed are those from the design basis.
• Reaction table annotated (marks, axes, units).
• Member-size list matches properties.
• Dummy members did not govern.
• Bracing behaviour matches intended tension/compression story.
• Red-flag list walked.
• Summary note written.
• CAD permission explicit yes or no.
• Checker named.

Frequently asked questions

Do I need to review every member force in a large model?

You need a strategy: warnings, global displacements, supports, then governing members and a sample of typicals. Reviewing every bar without a strategy is how people skip the important ones. Specials (cranes, stepped columns) are not optional samples.

The deflection looks large — is the model wrong or the building flexible?

First check units, releases and missing cladding stiffness you may have assumed. Then compare to the project’s deflection criteria. This article will not give limits; the design basis will. If criteria are unmet, that is a design issue, not a CAD issue.

Should utilisation plots go on the approval drawings?

Generally no. Approval drawings show the building definition. Ratios belong in calculations. Sizes that result from reviewed ratios belong on the drawing schedule.

When is output “good enough” to convert to CAD?

When geometry and the sizes you will plot are frozen, warnings are explained, and reactions for that revision are issuable. Conversion tools do not make that decision. See STAAD to AutoCAD.

Do not draw a result you have not read

Reviewing structural analysis output is how a solver run becomes an engineering freeze: coherent displacements, honest reactions, intended combinations, a size list, and no unexplained warnings. CAD should wait for that freeze. Pretty contours are not a review.

Continue with STAAD.Pro model to CAD drawing once permission is yes. Output review does not replace the design basis, deflection criteria or a qualified engineer’s acceptance of member adequacy. StruTools publishes the process as education, not as a certification of results.