Abstract
This paper presents a comparative study of Conjugate Heat Transfer (CHT) problems using a partitioned approach with separate solvers. The two solvers, one for the Reynolds Averaged Navier-Stokes (RANS) equations and the other for the heat conduction equation, are integrated into a coupling framework, allowing multiple solid-fluid models to run concurrently and perform several internal iterations before exchanging data. The aim of the work is to examine how the choice of the solid solver affects the stability and convergence of the coupled problem. Two solvers are considered: an explicit Finite-Volume (FV) solver equipped with multigrid acceleration, and a standard Finite-Element (FE) solver. The FV solver is used with Neumann interface conditions, eventually applying a relaxation parameter to the heat flux, while the FE solver requires Robin-type interface conditions to stabilise the computation. This complicates the comparison of computational efficiency between the two solvers, since each can perform better depending on the specific conditions of the coupled problem. We discuss these mechanisms for various test-cases, including a Low Pressure Turbine stage with annexed stator well.