Abstract : The evaluation of the thermal performance of a greenhouse requires an understanding of the energy exchanges between its components. In recent years, Computational Fluid Dynamics (CFD) proved to be efficient to predict the microclimate inside greenhouses, but few CFD works analyzed hot and arid climatic conditions. The aim of the present study is to investigate the impact of such specific conditions prevailing in the southern Mediterranean basin considering three different unheated greenhouses equipped with rows of canopy (tunnel, Venlo and plastic vertical wall greenhouse). 2D simulations were conducted in order to estimate the influence of the greenhouse geometry and cover material on the energy balance, focusing in particular on the heat losses. To establish the boundary conditions and validate the model, experiments were launched inside the greenhouses. Simulations were conducted for two contrasted conditions: the nocturnal period when the energy performance of each type of greenhouse was investigated, and the diurnal period for which a solar day was simulated taking account of natural ventilation. A good agreement was found with experiments concerning the temperature profiles and the energy losses. The predicted stable flow patterns slightly differed from one greenhouse to another for each condition: during the night they all disclosed two convective loops sweeping the floor, rising along the middle of the greenhouse, cooling down along the roof and moving down along the sidewalls of the greenhouse, while during the day flow patterns were strongly impacted by the vent openings. The obtained values (temperature and velocity) were coherent with results found in the literature for different greenhouse types and sizes. The developed tool offers potential applications in research on greenhouse climate to assess different cover material performances. It could also be used to investigate the energy consumption and partition for different greenhouse shapes, different crop densities different growth stages.
