Multi-objective Bayesian optimization accelerated design of TPMS structures

Triply periodic minimal surface (TPMS) is an effective filling architecture in porous ceramic artificial bone for its great bionic characteristics and self-supporting properties. However, design optimization of TPMS architecture remains a huge challenge due to the convoluted multi parameter-property relationship. In this study, optimization of TPMS structure for composite titania ceramic is accelerated by using the multi-objective optimization algorithm guided finite element method (FEM) simulation. According to the FEM analysis on thickness (P-t), array number (P-a) and constant number (P-c) as verified by compression testing and fluid experiment, P-t and P-a in TPMS structures are the key factors to force reaction, while P-a mainly determined the pressure drop. With P-c increasing, the pressure drops initially increased and decreasing after P-c=0. Bayesian optimization (BO) method is used to optimize both strength and permeability iteratively, in which the optimal Pareto frontier converges in less than 10 iterations. The parameter combination (P-t=0.28, P-c=-0.49, P-a=3.5) in best performing TPMS structure yields suitable modulus and permeability required for practical bone filling application, with 1.21 GPa in strength and 4.03 x 10(-9) m(-2) in permeability.