Rectangular plate bending — thin plate, uniform & central point load
Plate bending — production ShellMITC4 slab vs Timoshenko thin-plate theory
Technical features tested
- ShellMITC4 — production flat-plate slab element
- ElasticMembranePlateSection (Mindlin plate)
- Simply-supported & clamped edges
- Uniform pressure + central point load
- 1:1 and 5:1 aspect ratios (8 cases)
- h-refinement convergence to closed form
Problem description
A rectangular plate carries an out-of-plane load and its center deflection is compared to the Timoshenko–Woinowsky thin-plate series solution. Following CSI SAP2000 Verification Example 2-005 (geometry after MacNeal–Harder), eight configurations are run — a square (1:1) and a 5:1 rectangular plate, each simply-supported and clamped, under a uniform pressure and a central point load. STRIX recovers the dimensionless deflection coefficient α = w·D/(q·a⁴) (uniform) or w·D/(P·a²) (point), which is scale-invariant and compared to the published table value. The plate is a production flat-plate slab, meshed with the exact ShellMITC4 + ElasticMembranePlateSection element a real STRIX slab uses. It is kept thin (a/t = 200) so transverse shear is negligible and this Mindlin element converges to the thin-plate coefficient (the thick-plate regime is covered separately). At structural scale the automatic rotational stabiliser is negligible and is removed.
Geometry, properties & loading
Geometry
- Width a
- 2000 mm
- Aspects
- 1:1 & 5:1
- Thickness t
- 10 mm
- a / t
- 200
Material
- Model
- Elastic · thin plate
- E
- 30000 MPa
- ν
- 0.3
- Units
- N · mm
Loading & BC
- Uniform q
- 1×10⁻⁴ N/mm²
- Point P
- 100 N
- Supports
- SS & clamped
- Probe
- center defl. w
Reference solution
- α — SS, 1:1, uniform
w = α·q·a⁴/D (Table 8)= 0.00406 - α — SS, 1:1, point
w = α·P·a²/D (Eq.147, Table 23)= 0.0116
Timoshenko & Woinowsky-Krieger, Theory of Plates and Shells, 2nd ed. (1959): Table 8 (SS, uniform), Eq.147 & Table 23 (SS, central point), Table 35 (clamped, uniform), Table 37 (clamped, central point), ν = 0.3. The eight-case layout follows CSI SAP2000 Verification Example 2-005 (geometry after MacNeal & Harder, 1985). STRIX recovers the scale-invariant deflection coefficient α (= w·D/(q·a⁴) uniform, w·D/(P·a²) point) and compares it to the published table value; a = short side.
Results comparison
| Response quantity | Probe | STRIX | Reference | Δ | Verdict |
|---|---|---|---|---|---|
| Deflection coefficient α | SS · 1:1 · uniform | 0.00406 | 0.00406 | +0.064% | PASS |
| Deflection coefficient α | SS · 5:1 · uniform | 0.01294 | 0.01297 | −0.264% | PASS |
| Deflection coefficient α | Fixed · 1:1 · uniform | 0.00126 | 0.00126 | +0.336% | PASS |
| Deflection coefficient α | Fixed · 5:1 · uniform | 0.00259 | 0.0026 | −0.506% | PASS |
| Deflection coefficient α | SS · 1:1 · central point | 0.0116 | 0.0116 | −0.0020% | PASS |
| Deflection coefficient α | SS · 5:1 · central point | 0.01694 | 0.01695 | −0.069% | PASS |
| Deflection coefficient α | Fixed · 1:1 · central point | 0.00559 | 0.0056 | −0.165% | PASS |
| Deflection coefficient α | Fixed · 5:1 · central point | 0.00721 | 0.00725 | −0.621% | PASS |
Convergence
| Mesh | alpha | Δ |
|---|---|---|
| 8×40 | 0.00689 | −4.944% |
| 16×80 | 0.00715 | −1.329% |
| 24×120 | 0.00721 | −0.621% |
For every case the first-order MITC4 plate approaches the Timoshenko coefficient monotonically from below (a coarse mesh is slightly stiff), reaching the closed form to a fraction of a percent at the finest mesh — comfortably better than the commercial verification tolerance for the same problem. §7 shows the governing (worst final) case; all eight appear at the finest mesh in §6. The plate is thin, so the residual transverse-shear of this Mindlin element is under 0.1%; the auto rotational stabiliser is stripped and was measured to shift the answer by under 0.2% even when present.
Conclusion
Across all eight CSI 2-005 configurations — two aspect ratios, simply-supported and clamped, uniform and central point load — the recovered deflection coefficient matches the Timoshenko–Woinowsky thin-plate solution to well within tolerance, converging monotonically. The production slab element ShellMITC4 + ElasticMembranePlateSection is verified for out-of-plane plate bending under both distributed and concentrated load and both support conditions. Thick-plate (transverse-shear) behaviour is verified separately against Roark.
References & analysis files
- S. Timoshenko & S. Woinowsky-Krieger, Theory of Plates and Shells, 2nd ed. (1959) — Tables 8, 23, 35, 37 & Eq.147 (ν = 0.3); see §5.
- CSI SAP2000 Software Verification, Example 2-005 — Rectangular Plate with Static Loads (geometry after MacNeal & Harder, 1985); same eight-case layout, center-deflection check.
- STRIX headless harness: production flat_plate slab → generatePyForLoadCase → opensees.pyd; ShellMITC4 + ElasticMembranePlateSection, auto rotational stabiliser removed.
- Engine
- v1.0.6 (opensees.pyd)
- Run date
- 2026-08-19
- Record
- records/SB5.json
- Evidence archive
- verif-evidence-eng1.0.6-win-x64.zip · SB5/
- sha256
- (pending publish)