Fibre Reinforced Facade Encyclopedia: Design and detailing
Finite element analysis is a numerical method that divides a panel into a mesh of small elements, solves the equations of equilibrium across them, and returns the stress and deflection field under a set of applied loads. On a facade it is the tool used to size thin shell panels whose geometry is too complex for hand calculation, to locate anchors, and to understand where a panel is actually working. It is an analysis of a model, not of a panel, and its answer is only as good as the material properties, boundary conditions and loads put into it.
A flat rectangular panel spanning between four anchors can be checked by hand in an afternoon. The same panel folded into a double curved shell with six anchors and a return on three sides cannot, and that is the point at which FEA earns its cost.
At a glance
| Property | Typical value | Note |
|---|---|---|
| Typical element type | Thin shell elements | Solid elements only at local details such as anchors |
| Mesh size, general field | Panel thickness to a few times thickness | Refined at anchors, corners and re-entrants |
| Convergence check | Refine until peak stress stabilises | Typically within about 5 percent between refinements |
| Loads applied | Self weight, wind pressure and suction, handling | Plus thermal and imposed structural movement |
| Wind actions source | AS/NZS 1170.2, with local pressure factors | Corner and parapet zones govern many panels |
| GRC design basis | Long-term aged LOP for serviceability | Not the 28-day MOR |
| FRP design basis | Deflection usually governs, not strength | Low elastic modulus; orthotropic laminate |
| Deflection limit, typical | Span over 180 to span over 360 | Project and system specific |
| Validation | Physical test of the panel or connection | FEA supplements testing, it does not replace it |
When FEA is warranted
Most facade panels do not need it. A flat, rectangular, uniformly supported panel is a plate problem with published solutions, and a competent facade engineer will size it by hand faster and more reliably than by building a model. FEA earns its place in specific circumstances.
- Geometry with no closed-form solution: double curvature, folded plates, deep returns, perforations, and panels whose stiffness comes from shape rather than thickness
- Anchor layouts that are irregular, sparse, or dictated by the structure behind rather than by the panel
- Handling, demoulding and lifting cases, where the panel is green, unsupported at its service anchors and loaded in a direction it was never designed for, which is frequently the governing case on a thin shell
- Local details: the stress field around a cast-in anchor, a rib to skin junction, a corner return or an opening
- Optimising rib layout, where the question is where to put stiffening rather than whether the panel works
Modelling a thin shell panel
A facade panel is thin relative to its other dimensions, so shell elements are right for the general field: they represent bending and membrane action in a surface at a fraction of the cost of solids. Solids are reserved for local sub-models where through-thickness behaviour matters, typically at a cast-in anchor or bonded bracket pad, driven by forces from the global shell model.
Mesh density is decided by convergence, not appearance. The general field can take elements of the order of the panel thickness to a few times the thickness, refined at anchors, corners, re-entrant geometry and openings where stress gradients are steep. The test is to refine and re-run: if peak stress moves significantly between densities the mesh is too coarse, and if it keeps rising without settling the model contains a singularity, usually a sharp re-entrant corner or a point support. Singularities are a modelling artefact, so the response is to model the actual radius or the anchor’s real bearing area rather than to report the number the mesh produced.
Material properties must be design values, not headline ones. For glass fibre reinforced concrete, serviceability is checked against the long-term aged limit of proportionality rather than the 28-day modulus of rupture, because aged properties are what the panel has for most of its life. For fibre reinforced polymer the elastic modulus is low, so deflection almost always governs before strength does, and the laminate is orthotropic, so a single isotropic modulus gives the wrong answer on a directional lay-up.
Boundary conditions and loads
Boundary conditions are where most facade FEA goes wrong, because anchors are not the idealisations the software offers by default. A flexible anchor is not a pin; it is a component with finite stiffness in several directions and it exists precisely so the panel is not restrained. A gravity anchor carries load in one direction and should not be modelled as restraining others. Modelling all six degrees of freedom as fully fixed produces artificially high corner stresses and an artificially stiff panel; modelling everything as a pin produces the opposite.
Represent each anchor by springs or connector elements with stiffnesses from the anchor manufacturer’s data or from a test, and run the sensitivity: analyse with the anchors stiff and soft and confirm the panel works across the range. Where a panel is anchored to a structure that itself deflects, that imposed movement is a load case in its own right, and it is the case that cracks panels in service.
Why FEA supplements testing rather than replacing it
An analysis predicts the behaviour of an idealised model of an idealised material, and a thin cementitious or composite shell is neither, so professional practice is to use analysis and testing together. Material properties come from batch testing: glass fibre reinforced concrete is batch tested for modulus of rupture and limit of proportionality to the GRCA specification, and those results feed the model. Connections are proven by physical pull-out and shear tests on the actual panel construction, and whole systems by facade testing to AS/NZS 4284 where the project requires it, and by prototypes and mock-ups that are loaded and inspected.
The same distinction applies absolutely to fire. Finite element analysis says nothing about reaction to fire or non-combustibility, and no fire performance may be claimed from it. Fire classification comes from tests to standards such as AS 1530.1, AS 1530.3, AS 5637.1 and AS 5113, each carried out by a testing body named in the report. Used properly the sequence is: analyse to understand and to size, test to prove, then re-run the analysis with measured properties so the model becomes a calibrated tool for the rest of the panel types.
Standards and references
- AS/NZS 1170.2 Structural design actions, wind actions, including local pressure factors
- AS 1170.0 Structural design actions, general principles, for load combinations and serviceability
- GRCA Specification for the Manufacture, Curing and Testing of GRC Products, including batch testing and aged design values
- BS EN 1170 parts 1 to 8, test methods for GRC, for the properties used as model inputs
- AS 5216 Design of post-installed and cast-in fastenings in concrete, for anchor capacity and stiffness
- AS/NZS 4284 Testing of building facades, for system level verification of the analysed design
- AS 1530.1, AS 1530.3, AS 5637.1 and AS 5113 for fire classification, which analysis cannot provide (check the edition of the NCC adopted in your jurisdiction)
Where this matters on a ShapeShift project
Spanda: A large freeform composite structure where the geometry itself provides the stiffness, which is the class of problem that cannot be resolved by hand calculation.
Related terms
Questions specifiers ask
When is FEA actually needed on a facade panel?
When the geometry has no closed-form solution: double curvature, folded plates, deep returns, perforations or shape-derived stiffness. Also for irregular anchor layouts, local details around anchors and ribs, and handling, demoulding and lifting cases. A flat rectangular panel on four anchors does not need it.
Shell elements or solid elements?
Shell elements for the general field, because a facade panel is thin relative to its other dimensions. Solid elements only in local sub-models where through-thickness behaviour matters, such as around a cast-in anchor or bonded pad, usually driven by forces from the global shell model.
How fine should the mesh be?
Fine enough to converge. Start around the panel thickness to a few times the thickness, refine at anchors, corners, re-entrants and openings, then re-run finer. If peak stress keeps rising without settling, the model has a singularity from a sharp corner or point support, which is an artefact rather than a real stress.
Does FEA replace physical testing?
No. It predicts the behaviour of an idealised model of an idealised material, and a thin GRC or composite shell is neither. Material properties come from batch testing, connections from physical pull-out and shear tests, and systems from facade testing and mock-ups.
Have geometry that will not resolve by hand calculation? We analyse panels and their connections in house and back the analysis with physical testing. Talk to a facade engineer
Reviewed by Richard Halvorsen, General Manager, ShapeShift Technologies. Last updated September 2026.