Development of Design‐Phase Digital Twin for Structural Optimization of a Biomedical Device

ABSTRACT

Digital twins have transformed design and verification workflows in engineering, yet their adoption within medical device development remains limited. Conventional build‐and‐test approaches restrict design space exploration and often delay identification of mechanically critical risks. This paper presents a case study on the development of a calibrated, design‐phase behavioral digital twin for a capsule‐based biomedical delivery system used in dentistry, with the annular snap‐fit nozzle‐to‐body interface selected as a representative mechanically sensitive feature. Empirical testing, analytical snap‐fit theory and nonlinear Finite Element Analysis (FEA) were integrated to characterize mating force, strain and structural safety during assembly. Strong pre‐calibration agreement was observed between the analytical and numerical models and back‐calculation of an effective polymer‐to‐polymer coefficient of friction ( μ  = 0.12) reduced prediction error relative to empirical measurements to within experimental repeatability, establishing a validated behavioral digital twin. Design optimization employed a Taguchi orthogonal array to efficiently explore strain‐constrained geometric variation. Results identified undercut depth, governed by nozzle inner diameter, as the dominant contributor to retention strength. An optimized configuration achieved a 65%–75% increase in mating force relative to the baseline design while remaining within permissible strain limits. Numerical verification predicted peak mating force of 190 N and identified localized stress concentrations in the nozzle neck region, highlighting reinforcement requirements before production tooling. Physical verification and validation of the optimized system confirmed the predicted increase in structural performance. Consistent with the digital twin maturity continuum, the framework is positioned as a design‐phase, behavioral digital twin. It operates with offline one‐way physical‐to‐digital data flow during the product realization stage, with automated bidirectional synchronization identified as the next stage of development. These findings demonstrate how digital twins can shift biomedical device development from reactive validation towards predictive, manufacturability‐aware design, strengthening patient safety, clinician confidence and regulatory assurance in Class II medical devices.