Data-driven inverse modelling of ionic polymer conductive composite plates

Abstract

Analytical solutions of the partial differential equations (PDEs) governing the behavior of ionic polymer plates have not been yet obtained and therefore only time consuming discrete numerical methods can be used instead. To avoid the computational cost, of numerical solutions this paper introduces a solution construction method that exploits analytical approximation basis functions borrowed from solutions of single physics formulations associated with rectangular ionic polymer plates for artificial muscle applications. This is achieved by utilizing an inverse approach that exploits global optimization. An objective function is constructed to express the error between the experimental and analytical values of the selected state variables. Minimization of this objective function yields an efficient determination of the unknown free coefficients. Comparisons between the determined approximations and the experimental data along with computational efficiency improvements conclude this paper. © Springer-Verlag Berlin Heidelberg 2006.