The flow and heat transport phenomena developing in a real-scale, underground hot-water storage tank intended for central solar systems and made of concrete walls are studied numerically and experimentally. The cubic tank with a volume of 8 m(3) has been equipped with two linear diffusers extending over its entire width. For the numerical computations, charging of the tank at a constant flow rate and three different inlet-temperature histories was considered. One of these corresponded to a simple constant value, a second one to solar-collector heating and the third one to electric heating. In the last case experimental data were also obtained. The charging process was simulated by dynamic models based on the multinode and plug-flow approaches, as well as two-dimensional (213) computational fluid dynamics (CFD), for which both low-Re k-epsilon and two-layer turbulence models were used. The distinct features of the flow and temperature fields for each charging mode as obtained from the models have been analyzed and compared to each other. For the electric heating case, preliminary comparisons between models and experiments were made, showing good qualitative agreement, while quantitative agreement was achieved only for parts of the entire transient process. The effects of turbulence-model choice and water-surface heat losses were also demonstrated and found to be important factors in the modeling procedure. (C) 2009 Elsevier Ltd. All rights reserved.