Sheet metal shell meshing for FEA preprocessing

Sheet metal structural analysis is one of the most common FEA workflows in the world — body-in-white, HVAC enclosures, electronic housings, agricultural equipment frames. And yet 'the software does not effectively support shell meshing for sheet metal applications' and 'meshing module for sheet metal needs to be improved' appear as complaints across multiple products in this dataset. This isn't a niche problem.

The reason it persists is that shell meshing for sheet metal requires mid-surface extraction from solid CAD geometry — a step that sounds simple and is actually brittle. When the CAD model has fillets, cutouts, embosses, or variable thickness regions (all normal in real sheet metal parts), the mid-surface extractor either produces gaps, overlapping surfaces, or fails entirely. The engineer then spends hours manually stitching the mid-surface before they can even start meshing. The big preprocessors treat this as a preprocessing problem the user has to solve, because fixing it properly would require tight integration with the CAD kernel — which is controlled by a different vendor.

This creates a structural gap that nobody has incentive to close. The CAD vendor doesn't care about mesh quality. The simulation vendor doesn't own the geometry kernel. The engineer is caught between them, manually cleaning up mid-surfaces in a tool that was never designed for it.

What a standalone tool can do is focus entirely on this one handoff: take a solid CAD body known to be sheet metal (recognized by thickness-to-span ratio), extract a clean mid-surface automatically with explicit handling of the common failure modes (T-junctions, hole edges, flange transitions), apply a shell mesh with user-specified element size and quad-dominant pattern, and export directly to the major solver formats. The scope is narrow enough to do well. The user doesn't want a general preprocessor — they want to stop spending their morning cleaning up geometry before they can start the actual analysis.

The business recurs because sheet metal designs change constantly — every ECO (engineering change order) means a new CAD file and a new meshing session. A team running twenty active programs might remesh a given component six times over its design life.

What to build

Build a mid-surface extraction and shell meshing tool specifically for sheet metal solid bodies — STEP input, automatic fillet suppression and T-junction handling, quad-dominant shell mesh output — that exports to Nastran BDF and Abaqus INP without requiring any manual surface repair.

Where to start

Start with a specific geometry class: single-thickness steel brackets and mounting plates with one or two flanges, which are simple enough to extract reliably and common enough across every industry vertical that you can find beta customers in a single Slack community or LinkedIn search.

The hard part

The hardest part is handling the long tail of CAD pathologies — models built by mechanical designers who don't think about simulation — without requiring the engineer to manually intervene, because the moment you add a manual repair step you've recreated the problem you were hired to solve.

How it makes money

Usage-based pricing per mesh job for small teams and freelance analysts, with a flat annual seat license for teams running more than 50 jobs per month — priced so that a single analyst can expense it without a procurement process.