An explicit boundary condition-enforced immersed boundary-reconstructed thermal lattice Boltzmann flux solver for thermal–fluid–structure interaction problems with heat flux boundary conditions

Abstract

For the thermal–fluid–structure interaction (TFSI) problems with moving boundaries, the original implicit boundary condition-enforced immersed boundary method (IBM) by Wang et al. [1] has to generate a large correlation matrix and compute its inversion at each iteration, implying that the virtual memory requirement and computational cost would grow exponentially with the number of Lagrangian points. In this work, we proposed an efficient explicit boundary condition-enforced immersed boundary method for Neumann boundary condition (NBC) to achieve high computational efficiency and maintain similar accuracy as the original implicit IBM [1], which circumvents the needs to assemble a large correlation matrix and inverse it in the original implicit IBM [1] through second-order approximation based on the error analysis using Taylor series expansion. Most importantly, the proposed explicit IBM can efficiently solve practical physics problems with a tremendous amount of Lagrangian points involving Neumann boundary condition. The comparisons of the virtual memory and computational cost between the explicit and implicit IBMs demonstrate that the explicit boundary condition-enforced IBM is not only computational efficient, but also has memory saving performance. The proposed explicit IBM integrated with the reconstructed thermal lattice Boltzmann flux solver (RTLBFS) is validated with some classical benchmarks, and the results show that the proposed explicit IBM can successfully resolve TFSI problems with Neumann boundary condition.