Over the years, significant efforts have been made to develop advanced modelling techniques for simulating multi-scale and multi-physics problems. One main strategy for addressing these complex systems is the code coupling technique, which aims to enhance simulation capabilities by combining the strengths of individual coupled codes. In this context, the objective of this work is to develop a framework for coupling the finite volume (FV) code OpenFOAM and the in-house finite element (FE) code FEMuS. The adopted strategy for exchanging heterogeneous fields involves using an intermediate representation that is common to all codes, thereby enabling the possibility of coupling multiple codes. To achieve this, the communication paradigm relies on the MEDCoupling library, which provides efficient in-memory data exchange and a common framework for representing different data structures. This paper tests the proposed code coupling framework through its application to the thermally driven cavity problem, where natural convection governs the motion. In particular, the volume data transfer algorithm is applied to a natural convection scenario in which the combination of the Prandtl (Pr) and Rayleigh (Ra) numbers ensures the development of a turbulent regime within the volume of a square cavity. The simulation of the turbulent natural convection problem is performed in cavities filled with low-Prandtl number fluids that reach the turbulence regime at low Rayleigh numbers. In this setup, the FV code solves for the Reynolds-Averaged Navier-Stokes (RANS) equations, while the FE code is employed to solve the turbulent energy equation using the kθ-ωθ model. The aim of this work is to address the complexity of the physical phenomenon by combining the strengths of both codes. Specifically, the more advanced thermal turbulence model available in the FEMuS code is used to enhance the simulation results obtained with OpenFOAM.