Ensemble Timestepping Algorithms for the Heat Equation with Uncertain Conductivity

TL;DR

This paper introduces two efficient ensemble timestepping algorithms for heat conduction with uncertain conductivity, enabling reliable temperature predictions with reduced computational costs, validated through theoretical analysis and numerical tests.

Contribution

The paper proposes novel ensemble algorithms that efficiently compute multiple heat conduction solutions with shared computations, improving stability and convergence analysis.

Findings

Algorithms reduce computational costs for ensemble simulations.

Theoretical stability and convergence are established.

Numerical tests confirm the effectiveness of the methods.

Abstract

Motivated by applications to 3D printing, this paper presents two algorithms for calculating an ensemble of solutions to heat conduction problems. The ensemble average is the most likely temperature distribution and its variance gives an estimate of prediction reliability. Solutions are calculated by solving a linear system, involving a shared coefficient matrix, for multiple right-hand sides at each timestep. Storage requirements and computational costs to solve the system are thereby reduced. Stability and convergence of the method are proven under a condition involving the ratio between fluctuations of the thermal conductivity and the mean. A series of numerical tests are provided which confirm the theoretical analyses and illustrate uses of ensemble simulations.