Topological Optimization of Brake Caliper Adapter

A steel brake caliper adapter, responsible for transmitting braking torque from caliper to wheel hub, taken through five successive material-removal iterations to cut mass while holding every design safely under a fixed stress ceiling. Each iteration was validated by FEM under the same worst-case combined braking load before further material was removed.

To make sure the optimized geometry's performance wasn't an artefact of how the load was modelled, the final design was independently re-checked using a second, physically distinct load-application method — giving high confidence in the result before committing to prototyping.

The optimisation proceeded as five successive versions, each removing material identified as non-load-critical from the previous step and re-validated by FEM before moving on. Starting from a solid reference geometry with lateral stiffeners, the first iteration removed those stiffeners; the second opened up the material bridge between the two mounting bolt holes; the third enlarged that central cavity into a more organic shape optimised for material distribution; and the fourth and final version added two lateral cavities and reformed the base geometry into a fully organic, mass-minimised shape. At every step, the design constraint held firm: maximum stress had to stay below 200 MPa.

All five versions were tested under the same worst-case load: a 17 kN resultant force representing the combined braking torque reaction transmitted from caliper to adapter, decomposed via a defined unit direction vector into its two orthogonal components and applied through rigid multi-point-constraint elements connecting a virtual reference node to the mounting bolt holes. This load-application method let the same loading scenario be applied consistently across every geometry variant in the sweep.

Despite the mass dropping by more than a quarter across the five iterations, peak stress moved only slightly — from 192.0 MPa on the original design to 197.2 MPa on the final optimised version — staying comfortably within the 200 MPa ceiling at every step. Displacement rose modestly in step, from 0.272 mm on the original to 0.314 mm on the final version, an acceptable trade against the mass saved. The final version, with its two added lateral cavities and reformed base, delivered the best balance of weight and structural performance of the five candidates tested.

To rule out any dependency on the load-application technique itself, the first and final design iterations were re-analysed using a second, independent virtual-node concentrated-load method and compared directly against the original results. The two methods agreed closely: the first iteration returned 189.8 MPa against 190.2 MPa from the primary method (a 0.2% difference), and the final optimised design returned 196.5 MPa against 197.2 MPa (a 0.4% difference). That level of agreement across two physically distinct load-application approaches gave strong confidence that the optimised geometry's structural performance was robust and not an artefact of modelling choices, clearing the design for prototyping.