An experience with surrogate functions on the design of a transcutaneous energy transfer system

Abstract

A transcutaneous Energy Transfer (TET) is a system that uses electromagnetic fields to transfer power from outside the body to an artificial organ (AO) inside the body. In this work a systematic approach to obtain an optimal TET is proposed by using a multi-objective optimization approach that minimizes the volume and the thermal effects of the TET and takes into account the electrical constraints of the device. In order to solve the multi-objective optimization problem a Compromise Programming based criterion was adopted followed by a local constrained optimization. In order to reduce computational time, the device attributes were modeled by an alternative proposed Kriging tree model, which is a class of surrogate functions based on Kriging models from a set of data simulated by Finite Element Methods applications.