Font Size: 

Reduced order model constitutes an efficient option to decrease the high computational cost of dynamical systems governed by partial differential equations (PDE). The technique based on proper orthogonal decomposition (POD) was first presented in the article[1]to generate a reduced set of basis functions for Galerkin representation of PDEs which results in approximate the simulation at any time point by solving an ODEs of time dependent coefficients. Our approach in this article targets the development of a non-intrusive reduction technique. We keep the same manner of obtaining basis functions, while approximating the time dependent coefficients using Kriging based surrogate model. The proposed method is then illustrated with an application to the simulations of heat diffusion systems on a thin rod and on a square plate. The numerical results illustrate the simulation using the proposed idea.

modeling;simulation;function approximation