Module 02 / 6
From sketch to FEA
Geometry, materials, boundary conditions — and what not to simplify
Finite-element analysis does not make a plan scientific. It makes your assumptions visible. That is the only honest use of it in a teaching clinic.
Clinicians sketch. Engineers mesh. The gap between those two habits is where most 'digital periodontology' currently stops.
A sketch encodes judgement: which wall of the defect is the problem, where the papilla is still biologically possible, which root is a pier and which is a passenger. Finite-element analysis cannot replace that judgement. What it can do is refuse to let the judgement remain implicit.
The chain is short and unforgiving.
Geometry. From CBCT and intraoral scan you can build a virtual patient. That is necessary and, in the hands of groups such as the Semmelweis / Dent.AI line of work, increasingly fast. It is still a surface (and a segmentation). Surfaces do not carry load until you give them thickness, interfaces, and a ligament.
Materials. Assigning a single elastic modulus to 'bone' is how you get a colourful plot that cannot change a flap design. Cortical and cancellous bone, PDL as a biphasic or at least hyperelastic medium, dentin, enamel, and — if you are honest — a cementum layer: each is a claim. If you do not know the claim, you should not plot the stress.
Boundary conditions. Where is the model held? Where does the occlusal load enter, at what angle, over what contact area? Is the antagonist a point, a facet, a group function? Most published dental FEA papers are more sensitive to these choices than to the brand of implant they claim to compare.
What not to simplify. Do not delete the PDL. Do not freeze the mandible to a rigid block and then interpret millimetre-scale crestal peaks as clinical truth. Do not treat a 2D mid-root section as a 3D furcation. You may start in 2D — this laboratory does, on purpose — but you must label the start as a start.
The teaching simulator on this site is a labelled start. It is not a solver. If a picture cannot survive a caption that says so, it should not be in a surgical conversation.
In the chair
If you cannot say which surface is fixed, which surface is loaded, and which material you assigned to the crestal millimetre, you are not looking at a simulation. You are looking at a picture.
Educational and illustrative teaching models. Not a medical device. Not for diagnosis, treatment planning, or patient-specific simulation.