Computational fluid dynamics study on three-dimensional polymeric scaffolds to predict wall shear stress using machine learning models for bone tissue engineering applications

Abstract

Geometrical patterns and dimensions of the polymeric scaffold play a major role in controlling the degradation and mechanical stimuli for osteogenic differentiation. Wall shear stress (WSS) analysis of scaffold provides a better understanding of the body fluid flow dynamics. A computational fluid dynamics (CFD) study was carried out to understand velocity profile and WSS distribution when the strands are arranged in rectangular and triangular pitch for the different strand diameters and spacing. The number of scaffold surfaces with less than 30 mPa and maximum and average WSS was estimated to check the suitability of the scaffold for loading stem cells. This situation is favorable to induce osteogenic activity and cell viability. Higher spacing/pitch between the strands increases the chances of scaffold surface having WSS less than 30 mPa. When the spacing and diameter are smaller, there is no significant variation in WSS and pressure drop between rectangular and triangular pitch arrangement is observed. Machine learning (ML) models were developed to predict WSS distribution and to reduce the computational cost involved in solving the Navier–Stokes equation. XG Boost and support vector machine (SVM) models outperform the other models in predicting the WSS with high R² and five-fold cross-validation accuracy and are helpful in predicting the optimal design parameters of a three-dimensional scaffold.