Chapters
- 1 Getting started
- 2 The interface
- 3 Coordinate systems and conventions
- 4 Modelling
- 5 Materials
- 6 Sections
- 7 Supports, releases and modifiers
- 8 Loads
- 9 Load combinations
- 10 Meshing
- 11 Static analysis
- 12 Dynamic and seismic analysis
- 13 Results
- 14 Design
- 15 Scripting and the AI model builder
- 16 Import and export
- 17 Verification
- 18 Scope and limits
- 19 Troubleshooting
- 20 Glossary and notation
On this page
This chapter collects in one place the size limits of the FEA Tool, the assumptions of its analyses, and what it does not model. Knowing them is part of using the tool well: most of the limits below are stated by the tool itself when they apply, but a model is easier to plan when they are known in advance. The scope of each design module is listed on the Scope and limits page of the site.
18.1 Size limits
| Quantity | Limit |
|---|---|
| nodes (including the nodes of the area mesh) | 3,000 |
| frame members (after the automatic frame mesh) | 6,000 |
| degrees of freedom | 18,000 |
| modes in a modal, Ritz or spectrum analysis | 60 |
| time steps in a time history | 20,000 |
| node history kept for drawing a time history | 200,000 node-steps |
A larger model or record is refused with a plain message before the run. Your plan may set lower member and node limits. When the automatic area mesh would pass 85 % of the node limit, its element size is enlarged and the run says by how much (Section 10.6).
As a guide to speed, a linear static analysis of 3,000 nodes, 5,500 members and 17,400 degrees of freedom with two load patterns and a combination takes about one second in a current desktop browser.
18.2 Assumptions of the analyses
- Materials are linear elastic and isotropic in every analysis except the nonlinear static, pushover and nonlinear time history, where plastic hinges or fibre sections act. Areas are isotropic; orthotropic and layered shells are not available.
- Small displacements are assumed in the linear, P-Delta, buckling and dynamic analyses. Large displacements are available only in the nonlinear static analysis (co-rotational geometry).
- Frame members are straight and prismatic, with shear deformation where the section has shear areas. Warping torsion is not part of the frame element: the warping constant is used by the design modules only, so open sections in significant torsion should be checked separately.
- P-Delta captures the sway effect (P-Δ) at the member ends; the bowing effect between the ends (P-δ) needs the member divided into several elements.
- Damping in the time history is Rayleigh, stiffness- or mass-proportional, or modal; there are no discrete dampers.
- Ground motion in a time history is applied in one direction at a time from the window (scripts can give several records). Static loads are not included unless an initial state is chosen.
- Mass comes from the self weight, from load patterns and from node masses (Section 8.10); there is no added mass of fluids.
18.3 What is not modelled
- Solid elements, plane-strain and axisymmetric elements.
- Link elements: gaps, hooks, friction pendulums, isolators, viscous dampers and multi-linear links. Use springs for linear elastic supports.
- Cable sag: a cable is a straight member, axial only, and tension-only in the nonlinear analyses.
- Staged construction and time-dependent effects (creep, shrinkage, relaxation). The prestress panel takes losses as entered percentages.
- Moving loads and influence lines, frequency-domain and steady-state analyses.
- Soil-structure interaction beyond node and area springs.
- Strength loss of plastic hinges after the end of the plastic plateau, and P-M interaction of reinforced concrete hinges beyond a simplified surface (use fibre hinges for a section-based interaction).
18.4 Limits of particular analyses
| Analysis | Limit |
|---|---|
| P-Delta | one combination per run; superposition of patterns does not apply |
| Buckling | elastic buckling only; one element per member finds frame modes only (members are divided into four parts when no mesh is set) |
| Response spectrum and time history | area results are not given; frame and node results are |
| Nonlinear static, pushover, nonlinear time history | models with areas are refused; rigid end offsets are not used; members are not cut at intermediate joints (divide them first); member loads keep their direction |
| Ritz vectors | no Sturm check of missed modes |
| Linear acceleration and Wilson-θ below 1.37 | conditionally stable; refused for models with massless degrees of freedom or diaphragm links |
| Storey checks | the weak storey check needs storey strengths and is not covered |
| Response spectrum scaling | applied for ASCE/SEI 7-22 and NBC 2020; none for EN 1998-1 (not required) or AS 1170.4 |
18.5 Limits of combinations and design
- The combination generator leaves some actions to you (Chapter 9): fluid, earth and notional patterns in EN 1990, temperature without a factor you give in AS/NZS 1170.0, the creep case of NBC 2020 Table 4.1.3.4, and service combinations with snow and wind for ASCE/SEI 7-22.
- Design actions are envelopes: each action's extreme is taken independently, so simultaneous actions are not paired. This is conservative.
- Torsion and minor-axis shear from the model are reported but not checked by the member modules.
- The slab and wall design gives flexural and membrane reinforcement and checks one-way shear without shear reinforcement. Punching shear, crack widths and deflections of slabs are checked in their own modules or by the engineer.
- Aluminium, prestressing tendon and user-defined materials are analysed but not designed.
18.6 Responsibility
The FEA Tool supports the work of a qualified engineer, who remains responsible for the model, the interpretation of the results and the design. Results should be checked against simple hand calculations and engineering judgement before they are relied on.