02 · Inverse design

Dimension optimizer

Give the search a target output and a box to work inside. It runs a deterministic multistart solve and returns the geometry that lands closest while respecting your bounds and the size envelope.

01Objective

Minimise the total absolute deviation from one target strain, summed over the channels you select.

Target gage strainµε

Objective channels

FxIn-plane shear along the X beam pair.FzVertical load through the platform.MxBending about the X axis.MzTorsion about the vertical axis.

02Search bounds

millimetres

Lower and upper bounds per dimension
Key	Lower	Upper	Lock
aelastic beam
belastic beam
celastic beam
dboundary beam
eboundary beam
fboundary beam
gplatform

Locking a dimension sets its upper bound equal to its lower bound, holding it fixed through the search. The gage position x is an input, never a variable.

Gage position xmm

Size envelope limitmm

03Model, material and loads

Model variant

Current seven-parameter model. Elastic beam width and height are independent.

Material

Forces · N

Moments · N·m

Rated loads from the draft specification are pre-filled. Fy and My load the orthogonal beam pair and mirror Fx and Mx.

Solver settings

Multistarts

Random seed

Forward analysis

—Search space

midpoint of the bounds

Drawn at the midpoint of each bound — an indication of the scale the search is working at, not a result.

Ready

No optimisation yet

Set the objective and the bounds, then run the search. Results arrive with the achieved strain on every channel and the margin to yield.