CFD PREDICTION OF THE NATURAL VENTILATION IN A TUNNEL-TYPE GREENHOUSE: INFLUENCE OF WIND DIRECTION AND SENSIBILITY TO TURBULENCE MODELS

A contribution to the development of a new model for the characterisation of the climatic conditions in a tunnel-type greenhouse is presented, which takes into account the crop (tomato) like an active 3D region where both the momentum transfer (porous media) and the heat and humidity transfers between the crop and the inside airflow are considered. The effects of the wind direction on the climatic parameters inside the greenhouse are simulated together with the use of different turbulence models available in the CFD code. The model consists in the determination for each node of the crop grid of the energy balance between the transpiration flux and the radiation flux. The heat and humidity transfer coefficients are deduced from the leaf laminar boundary layer characteristics which are calculated with the local velocity of air in the crop. This model is included in the Fluent CFD package in the form of a User Defined File (UDF) added to the main processing unit. The 3D geometry includes a tunnel-type greenhouse with five vents on each side and a five rows mature tomato crop and its external direct environment. The wind boundary conditions for the velocity distribution are deduced from experimental data and the direction of the wind, relative to the longitudinal axis of the greenhouse, can vary from 0 to 90 degrees. Three turbulences model are tested: the standard k-e model, the renormalisation group (RNG) model and the realizable ke model. The results of the simulations performed for 0, 45 and 90 degrees show clearly the influence of the wind direction on the velocity, temperature and humidity distributions inside the greenhouse. The computations of the velocity field using the different turbulence models also show noticeable difference in the velocity, temperature and humidity patterns inside the greenhouse and confirm the importance of the choice of the closure model for the modelling of turbulence. INTRODUCTION Computational fluid dynamics (CFD) packages are nowadays powerful tools for modelling of airflows and climate patterns in agricultural structures like greenhouses. Simulations with CFD allow determining the influence of the greenhouse design parameters like orientation relative to wind direction and location of the ventilators, on different climatic parameters like temperature and humidity distributions, ventilation flow rate, etc...The use of CFD is to date the only way of predicting the distributed climate in a greenhouse and in the inner crop by resolving the equations of heat and mass transfer for an accurate mesh of discrete locations (Boulard et al., 2002-1). Furthermore, the perfecting of these predictions, especially for relative humidity, constitutes a decisive contribution to the integrated pest management (IPM) (Boulard et al., 2002-2). Since about twenty years, some authors have performed experimental studies relative to the effects of the natural ventilation on the crop transpiration (Bot, 1983; De Jong, 1990; Fernandez and Bailey, 1992; Boulard et al., 1996). At the same time several authors have developed global air exchange models based on the “big leaf” assumption (Stanghellini, 1987; Yang et al., 1990; Boulard and Wang, 2000) and energy balance Proc. IC on Greensys Eds.: G. van Straten et al. Acta Hort. 691, ISHS 2005 458 methods (Kindelan, 1980; De Halleux et al., 1991; Wang and Deltour, 1996).All these studies about global models are summarized in a recent review paper (Roy et al., 2002-1). With regard to numerical simulations, some authors have been used commercial or self-developed CFD codes to perform simulations of different greenhouse configurations. From these studies, summarized in a review paper by Reichrath and Davies (2002), it can be deduced that the use of a 3D model and the taking into account of the inner vegetation are two essential points for the consistency of the numerical results. In a previous paper (Roy et al, 2002-2) we presented a model that combines different scales of a 3D model (the tunnel and its direct environment; the inner domain of the tunnel) and that takes into account an active crop by modelling the crop as a porous medium and determining the heat and water vapour exchanges at the crop level. The objective of this study consists of the determination of the influence of physical conditions (wind direction) and numerical conditions (sensibility to the turbulence model) on the prediction of the inner climate.