Data-driven variational multiscale reduced order modeling of vaginal tissue inflation

Area of Science:

• Biomechanics
• Computational modeling
• Biomaterials

Background:

• Vaginal tissue exhibits complex geometry and material properties, leading to significant challenges in finite element (FE) simulations.
• Nonlinearities in FE simulations of vaginal tissue increase computational time substantially, limiting their practical application.

Purpose of the Study:

• To develop and evaluate reduced order modeling (ROM) techniques for accurate and efficient simulation of rat vaginal tissue deformations.
• To compare the performance of data-driven variational multiscale ROM (DD-VMS-ROM) with classical Galerkin ROM (G-ROM).

Main Methods:

• Utilized proper orthogonal decomposition (POD) to extract key information from FE simulations.
• Extended a data-driven variational multiscale (DD-VMS) framework for ROM.
• Compared DD-VMS-ROM with a classical Galerkin ROM (G-ROM) for simulating vaginal tissue inflation.

Main Results:

• Both G-ROM and DD-VMS-ROM reduced FE computational costs by orders of magnitude with minimal loss of accuracy.
• DD-VMS-ROM demonstrated improved accuracy over G-ROM at a slightly increased computational cost.
• The study successfully captured experimentally measured displacement fields of rat vaginal tissue.

Conclusions:

• ROM techniques offer a promising approach for efficient and accurate computational modeling of vaginal deformations.
• These methods have the potential to significantly advance research in areas related to maternal health.
• The developed DD-VMS-ROM framework provides a robust tool for complex biomechanical simulations.