Simulation of Heat Conduction in Complex Domains of Multi-material Composites using a Meshless Method

TL;DR

This paper introduces a high-accuracy meshless method using domain decomposition and cloud-based interpolation to simulate heat conduction in complex multi-material domains with discontinuous properties, avoiding interface smearing.

Contribution

The paper presents a novel meshless algorithm employing polyharmonic splines with appended polynomials and cloud-based interpolation for accurate heat conduction simulation in complex, discontinuous domains.

Findings

Achieves high accuracy in 2D and 3D problems with sharp conductivity discontinuities.

Numerically better conditioned than previous meshless algorithms.

Effectively handles complex geometries with practical boundary conditions.

Abstract

Several engineering applications involve complex materials with significant and discontinuous variations in thermophysical properties. These include materials for thermal storage, biological tissues with blood capillaries, etc. For such applications, numerical simulations must exercise care in not smearing the interfaces by interpolating variables across the interfaces of the subdomains. In this paper, we describe a high accuracy meshless method that uses domain decomposition and cloud-based interpolation of scattered data to solve the heat conduction equation in such situations. The polyharmonic spline (PHS) function with appended polynomial of prescribed degree is used for discretization. A flux balance condition is satisfied at the interface points and the clouds of interpolation points are restricted to be within respective domains. Compared with previously proposed meshless algorithms with domain decomposition, the cloud-based interpolations are numerically better conditioned, and achieve high accuracy through the appended polynomial. The accuracy of the algorithm is demonstrated in several two and three dimensional problems using manufactured solutions to the heat conduction equation with sharp discontinuity in thermal conductivity. Subsequently, we demonstrate the applicability of the algorithm to solve heat conduction in complex domains with practical boundary conditions and internal heat generation. Systematic computations with varying conductivity ratios, interpoint spacing and degree of appended polynomial are performed to investigate the accuracy of the algorithm.