Development of a physics-informed neural network for analyzing the stress state of soft biological tissues

Title	Development of a physics-informed neural network for analyzing the stress state of soft biological tissues
Authors	K. M. Urazova¹ ¹Ulyanovsk State Technical University
Annotation	This work explores the application of physics-informed neural networks (PINNs) for modeling the mechanical behavior of soft biological tissues. A comparative analysis of the accuracy of PINNs with traditional hyperelastic models, including Neo-Hooke, Yo, Mooney-Rivlin, and Halzapfel-Gasser-Ogden models, is presented. Special attention is given to the combined description of the elastic and viscoelastic properties of tissues through a modified Kelvin-Voigt model. The results demonstrate that the proposed PINN approach achieves an accuracy of 99\% with a mean squared error of $3.199 cdot 10^{-7 kPa2, which is 4 orders of magnitude better than traditional methods. This work contributes to the development of personalized modeling methods for biomechanical systems.
Keywords	PINN, neural networks, mechanics of biological tissues, hyperelastic models, viscoelastic models.
Citation	Urazova K. M. ''Development of a physics-informed neural network for analyzing the stress state of soft biological tissues'' [Electronic resource]. Proceedings of the International Scientific Youth School-Seminar "Mathematical Modeling, Numerical Methods and Software complexes" named after E.V. Voskresensky (Saransk, July 29-31, 2025). Saransk: SVMO Publ, 2025. - pp. 254-258. Available at: https://conf.svmo.ru/files/2025/papers/paper51.pdf. - Date of access: 30.08.2025.

PROCEEDINGS OF THE SCIENTIFIC CONFERENCES AND SCHOOL-SEMINARS

Development of a physics-informed neural network for analyzing the stress state of soft biological tissues

Contacts