Reference Comparisons

Purpose

This page collects the cases where Armatura is compared against stored reference values rather than only closed-form textbook formulas.

That matters because not every useful benchmark is a single-member hand-calculation case. Some are small frame problems where a trustworthy target value is still worth checking even if the derivation is not reproduced in the docs yet.

Current candidate in the repository

The best current candidate is in SamLabs.Armatura.Test/ComparisonTests.cs:

• UpsideDownU_CompareToExpectedValues
• UpsideDownU_CompareToExpectedValues_2
• UpsideDownU_CompareToExpectedValues_150x80_rotated90

These tests model a small upside-down-U portal frame and compare selected Armatura results to stored expected values.

Benchmark model

The first comparison test uses:

• frame height: 2.5 m
• frame span: 2.5 m
• fixed supports at both base nodes
• rectangular section
• midspan point load of 10 kN downward on the top beam

The test then compares:

• horizontal reaction at the left support
• equal and opposite horizontal reaction at the right support
• total vertical reaction sum
• vertical displacement magnitude at beam midspan

Stored reference values in the test

For UpsideDownU_CompareToExpectedValues, the code uses:

• expected left support horizontal reaction: 1249.68 N
• expected midspan vertical displacement magnitude: 0.000337 m

For the other variants, the expected displacement changes with section orientation/rotation while the same general comparison structure remains.

Why this is a good validation page

This is the first strong candidate for a page that says:

here is a frame problem with a stored reference target, and here is how close Armatura is to it.

That is a different kind of confidence than the beam closed-form pages. It shows that the solver behaves plausibly on a small indeterminate frame, not just on textbook single-member cases.

Armatura result

Populate this after running the test and capturing the actual output.

Quantity	Reference	Armatura	Abs. error	Rel. error
Left support Fx	1249.68 N	TODO	TODO	TODO
Right support Fx	-1249.68 N	TODO	TODO	TODO
Sum of vertical reactions	10000 N	TODO	TODO	TODO
Midspan vertical displacement magnitude	0.000337 m	TODO	TODO	TODO

Editorial note

This page should clearly state where the reference values came from once that is documented. Right now the tests store them as accepted targets, but the docs should eventually answer the next question a careful reader will ask:

• were these values taken from a hand calculation,
• from CALFEM,
• from another solver,
• or from a previously trusted Armatura run after a bug fix?

Until that provenance is written down, this page should remain explicit that the values are stored reference targets from the test suite.