CAE · FINITE ELEMENT ANALYSIS · OPTIMISATION
Simulation-led lightweighting of a gearbox shift fork through non-linear FEA, design iteration, convergence and analytical validation.
How can a gearbox shift fork be redesigned to significantly reduce mass while maintaining sufficient strength, stiffness and realistic load transfer?
The study uses finite element analysis to understand the structural behaviour of the component, progressively refine its geometry and identify a lightweight design that remains within the defined operating constraints.
Define geometry, material and functional constraints.
Apply constraints, contact, load and mesh strategy.
Compare linear and non-linear structural response.
Remove material based on stress and deflection behaviour.
Mesh convergence, static checks and dynamic comparison.
The initial 1.28 kg design was first assessed using linear analysis before introducing non-linear contact. The non-linear model was considered more representative because the applied load was transferred through contact with a massless rigid ring rather than through a uniformly distributed pressure.
| Analysis | Max Stress | Max Deflection |
|---|---|---|
| Linear | 29.51 MPa | 0.0155 mm |
| Non-Linear | 22.26 MPa | 0.00794 mm |
The rail mounting was fixed while a massless rigid ring represented the gearbox interface. This enabled the force to be introduced through contact and allowed the influence of contact position and resulting turning moment to be studied.
Structured partitioning reduced the original swept-mesh node count while improving regularity and numerical stability. Final convergence was then assessed using quadratic tetrahedral elements with progressively smaller seed sizes.
Eight reported iterations progressively removed material from low-stress regions while monitoring stress, displacement and contact behaviour. The methodology itself also evolved when the original contact model produced an unrealistic turning moment.
Maximum stress began to stabilise at approximately 1.5 × 105 nodes. Deflection remained effectively constant throughout refinement, while further refinement was constrained by the educational licence node limit.
The final fork was simplified to a cantilever-beam representation to independently check displacement and bending stress.
A linear dynamic analysis using the Lanczos algorithm extracted the ten lowest eigenvalues. Mode 3 was identified as the dominant mode from its effective mass, with a natural frequency of 1657.3 Hz.
SIMULATION-LED LIGHTWEIGHTING An 84.4% mass reduction was achieved through iterative geometry development while remaining within the deflection limit.
METHODOLOGY MATTERS Revising the contact representation was necessary when the original load application produced unrealistic behaviour.
VALIDATION DEFINES CONFIDENCE Analytical checks exposed the limitations of simplified static and dynamic representations rather than being treated as exact matches.
“The most valuable part of the project was not simply reducing mass, but understanding how modelling assumptions, contact behaviour, convergence and validation influence the confidence placed in an optimised design.”