Cracks almost always start at a shape, not in the middle of a flat wall
Real-world part failures concentrate at specific geometric features far more often than they start in an unremarkable flat section, because sharp transitions in geometry concentrate stress into a small area. A sharp internal corner, a hole with no relief at its edge, a sudden jump in wall thickness, or a notch cut into a loaded section, each of these acts as a stress riser, a place where load funnels into a small region well beyond what the average stress across the part would suggest. This checklist looks for the most common of these shapes in your design.
How the checklist works
Each of the four flagged shapes gets its own specific fix rather than a generic "add fillets everywhere" answer: a sharp corner needs a fillet radius, an unrelieved hole needs a chamfer or relocation, a sudden section change needs a taper, and a notch needs a rounded base and possibly relocation. If more than one shape is flagged, the result recommends prioritizing whichever sits in the highest-stress region of the part, since not every flagged shape carries equal real-world risk.
Worked example
A design with a sharp internal corner and a sudden section thickness change, no unrelieved holes or notches: the tool flags both shapes, recommends adding a fillet radius at the corner and tapering the thickness transition instead of stepping it abruptly, and suggests fixing whichever one sits closer to the part's main load path first if both cannot be addressed immediately.
A common mistake
A frequent error is fixing every flagged stress riser equally instead of prioritizing the one nearest the actual load path. A sharp corner in a low-stress cosmetic region matters far less than the same shape sitting right where load concentrates; spend the limited redesign effort where it actually reduces failure risk.
Limitations
This is a shape checklist for common stress-riser patterns, not a stress analysis. It cannot calculate actual stress concentration factors or predict a specific failure load; validate a genuinely load-bearing or repeated-cycle design with a real test part.