Related Experiment Video
Updated: Jul 26, 2025
Fracture Apparatus Design and Protocol Optimization for Closed-stabilized Fractures in Rodents
Published on: August 14, 2018
Parameter optimization in a finite element mandibular fracture fixation model using the design of experiments
Michaela Maintz1, Bilal Msallem2, Michael de Wild3
1Medical Additive Manufacturing Research Group (Swiss MAM), Department of Biomedical Engineering, University of Basel, Allschwil, Switzerland; Department of Oral and Cranio-Maxillofacial Surgery, University Hospital Basel, Basel, Switzerland; Institute for Medical Engineering and Medical Informatics IM(2), University of Applied Sciences and Arts Northwestern Switzerland FHNW, Muttenz, Switzerland.
Abstract:
Only a few mandibular bone finite element (FE) models have been validated in literature, making it difficult to assess the credibility of the models. In a comparative study between FE models and biomechanical experiments using a synthetic polyamide 12 (PA12) mandible model, we investigate how material properties and boundary conditions affect the FE model's accuracy using the design of experiments approach. Multiple FE parameters, such as contact definitions and the materials' elastic and plastic deformation characteristics, were systematically analyzed for an intact mandibular model and transferred to the fracture fixation model. In a second step, the contact definitions for the titanium screw and implant (S-I), implant and PA12 mandible (I-M), and interfragmentary (IF) PA12 segments were optimized. Comparing simulated deformations (from 0 to -5 mm) and reaction forces (from 10 to 1'415 N) with experimental results showed a strong sensitivity to FE mechanical properties and contact definitions. The results suggest that using the bonded definition for the screw-implant contact of the fracture plate is ineffective. The contact friction parameter set with the highest agreement was identified: titanium screw and implant μ = 0.2, implant and PA12 mandible μ = 0.2, interfragmentary PA12 mandible μ = 0.1. The simulated reaction force (RMSE = 26.60 N) and surface displacement data (RMSE = 0.19 mm) of the FE analysis showed a strong agreement with the experimental biomechanical data. The results were generated through parameter optimization which means that our findings need to be validated in the event of a new dataset with deviating anatomy. Conclusively, the predictive capability of the FE model can be improved by FE model calibration through experimental testing. Validated preoperative quasi-static FE analysis could allow engineers and surgeons to accurately estimate how the implant's choice and placement suit the patient's biomechanical needs.
