Myocardial infarction can cause irreversible damage to the heart muscle, which can lead to heart failure. The difficulty of the treatment mainly arises from the anisotropic behavior of the myocardium fibrous structure. Patches or cardiac restraint devices appear to be a promising approach to post-infarction treatment. In this study, we propose a new model-assisted method to design patterned membranes. The proposed approach combines computer experiments and statistical models to optimize the design parameters and to meet the requirement for the post-infarction treatment. Finite element model, global sensitivity analysis, random forest model and response surface model are the key components of the strategy implemented in this study, which is applied to design a real membrane. The metamodel-based design method is able to estimate the equivalent Young's modulus of the membrane in a few seconds and optimization results have been validated a posteriori by laboratory measurements. This solution opens up new prospects for the design of customized membranes with technical specifications tailored to each patient.
Keywords: Cardiac membrane; Computer experiment; Finite element model; Mechanical behavior; Sensitivity analysis; Statistical model.
Copyright © 2024 Elsevier Ltd. All rights reserved.