Approximation of extracted features enabling 3D design tuning for reproducing the mechanical behaviour of biological soft tissues

Abstract

This article describes a new method, inspired by machine learning, to mimic the mechanical behaviour of target biological soft tissues with 3D printed materials. The principle is to optimise the structure of a 3D printed composite consisting of a geometrically tunable fibre embedded in a soft matrix. Physiological features are extracted from experimental stress–strain curves of several biological soft tissues. Then, using a cubic Bézier curve as the composite inner fibre, we optimised its geometric parameters, amplitude and height, to generate a specimen that exhibits a stress–strain curve in accordance with the extracted features of tensile tests. From this first phase, we created a database of specimen geometries that can be used to reproduce a wide variety of biological soft tissues. We applied this process to two soft tissues with very different behaviours: the mandibular periosteum and the calvarial periosteum. The results show that our method can successfully reproduce the mechanical behaviour of these tissues. This highlights the versatility of this approach and demonstrates that it can be extended to mimic various biological soft tissues.