• Funding : FEDER, Indelab
  • Start year :
  • 2025

Thanks to the development of 3D metal printing techniques, medical devices are constantly evolving. Today, it is possible to produce a customised titanium implant, but the lead times involved make large-scale customisation complex to implement. The aim of this thesis is to gain a better understanding of the mechanisms involved in the design and optimisation of medical implants for bone repair, both in terms of implant geometry and internal architecture. Indeed, a new generation of medicinal implants will incorporate an internal macroporous network capable of containing an active principle, dedicated to the release of antibiotics for therapeutic or prophylactic purposes. This work will be based on mathematical modelling approaches, such as linear programming (in integers) or constraint programming, numerical simulation tools and machine learning techniques. Initially, the work will be based on the design of a database of medical images, which will be processed using convolutional neural networks to identify design rules adapted to clinical cases. These rules will then be validated using numerical simulations and mechanical tests. This work will be based on mathematical modelling approaches, such as linear programming (in integers) or constraint programming, numerical simulation tools and machine learning techniques. Secondly, using mathematical modelling approaches, an internal porous network will be developed and optimised inside the implant to accommodate the drugs. Once the key parameters have been identified, a digital tool in the form of a vector-based web interface for communication between designers and practitioners will be created to test the deployment of the overall digital solution.