Thermal convective turbulent flows are composed of coherent structures such as thermal plumes, whose behaviour is erratic and not yet fully understood. The most intense plumes can be linked to significant phenomena such as mantle hotspots or thermal fatigue in nuclear reactors. Recent research shows that turbulent convection organises hierarchically from the wall, with a continuous process of plume formation and aggregation forming large-scale patterns. This raises once again the issue of global heat transport modelling, linked to the local dynamics of thermal plumes in a highly fluctuating environment. This internship aims to apply numerically the hierarchical analysis of plumes to 3D DNS data of Rayleigh–Bénard convection in water across a large range of Rayleigh numbers. Statistical analysis of temperature fields will help understand how the turbulent regime affects plume networks. The approach may later be extended to DNS with rough walls or experimental data from Lyon.