Non-linear FEA for parts that yield, flex or seal
When a part deflects a long way, yields, buckles or relies on contact that changes under load, linear analysis gives the wrong answer. Non-linear FEA follows the real behaviour through the whole load path.
Linear analysis rests on three assumptions: deflections are small, the material is elastic and contact does not change. Many important parts break all three. A snap fit bends far enough to change its own stiffness. An O-ring is squeezed by a quarter of its section. A thin panel buckles. A steel part yields locally and sheds load into the material around it. In these cases a linear result can be wrong by a wide margin, in either direction.
More about non-linear analysis
Non-linear FEA applies the load progressively and updates the stiffness as the geometry, the material and the contact change. The result is the full load-deflection curve, the permanent set left after unloading, the force needed to push a fit home, the contact pressure a seal achieves and the load at which a structure collapses. Material behaviour ranges from elastic-plastic metals with strain hardening to hyperelastic rubbers and elastomers.
The engineers running these studies are F-CAD®’s design team, and the aim is a better geometry: a clip with the right insertion force, a seal groove that holds pressure, a press fit with the grip you need, a thin section that stays stable.
Typical parts: O-rings and gaskets, moulded snap fits and clips, springs and flexures, overload and energy-absorbing members, thin-walled tubes and panels, press-fitted bushes and bearings.
What a non-linear study answers
Will it yield, and will it spring back?
The extent of plastic strain at peak load, and the permanent deformation left once the load is removed.
What force does the fit need?
Insertion and removal force for snap fits, clips and press fits, with the strain in the flexing feature.
Will the seal seal?
Contact pressure and contact width for O-rings, gaskets and lip seals at the squeeze, pressure and tolerance stack you specify.
At what load does it buckle or collapse?
The load at which a thin wall, strut or panel loses stability, with the effect of imperfections included.
How much reserve is there beyond first yield?
The margin between first yield and collapse, which can be considerable for ductile metals and decides whether an overload is survivable.
How does contact change as load builds?
Gaps that close, faces that lift and parts that slide, and how each of those redistributes the load.
What we model
Large deformation
Geometric non-linearity for parts that deflect or rotate far enough to change their own stiffness, such as springs, clips, flexures and thin membranes.
Plasticity and yield
Elastic-plastic metals with strain hardening from the material’s stress-strain curve, giving plastic strain, permanent set and residual stress after unloading.
Changing contact
Contact that opens, closes, slides and sticks with friction, for fits, seals, clamps and interlocking parts.
Hyperelastic materials
Rubbers and elastomers represented with hyperelastic material models that capture their nearly incompressible, strongly non-linear response.
Buckling and collapse
Loss of stability in thin-walled tubes, panels and struts, with imperfections included so that the collapse load is realistic.
Press and interference fits
Contact pressure and hoop stress from interference, including local yielding in the hub or on the shaft.
Overload and misuse cases
Loads beyond normal service, such as a jammed mechanism or a user leaning on a part, to see whether it bends safely or fails.
Before you start
What to send us
- CAD of the parts, including mating parts for fits and seals
- Stress-strain curves, or grade and temper for metals
- For rubbers: Shore hardness at minimum, ideally tension or compression test data
- Loads or displacements, and the order in which they are applied
- Tolerances that matter, such as seal squeeze or interference range
- The limit you are designing to: force, strain, deflection or pressure
How do we know we need non-linear analysis?
If deflections are large compared with the part’s thickness, if stresses pass yield, or if contact changes significantly under load, a linear study is likely to mislead. We will tell you which applies to your part.
What material data is needed for a rubber seal?
Shore hardness gives a starting point. Uniaxial tension or compression test data on the actual compound gives a much better fit, especially at the high strains seals see in service.
Can you predict exactly when a part breaks?
Non-linear analysis shows yield, strain and collapse. Fracture also depends on defects and toughness, so results are reported as strain against the material’s elongation at break. That shows how close the part is to fracture; the exact moment of failure also depends on defects the model does not contain.
Why is an idealised buckling load not enough?
Real structures have imperfections, and thin shells and tubes are sensitive to them. A collapse analysis that includes imperfections can give a much lower load than an idealised linear prediction.