| dc.description.abstract | The physical basis of natural ventilation is reviewed and the ventilation performance examined of six different greenhouse and tunnel types equipped with either roof or side, or both roof and side vents. Altogether, sixteen different ventilation configurations are analysed. Neural networks of ventilation rate were first developed to select relevant parameters influencing performance and these confirmed that the surface area of the vent-opening and wind speed together explain the largest part of the variance of air exchange rate measurements. Ventilation performance is then analysed using physical models derived from Bernoulli's theorem. The effects of vent-opening, their position and type, and the influence of a crop are all considered, together with different models combining the two major driving forces for natural ventilation, namely, wind and stack forces. Combining both wind and stack effects did not improve the predictive power of the model for the greenhouses equipped with only roof or side vents, but when both roof and side vents are open together, model performance is improved by combination of the two driving forces. It is shown too that the presence of a crop decreases the ventilation efficiency and that the linear dependence of the air exchange rate on wind velocity, when the chimney effect can be neglected, is only partially supported by the experimental data. (C) 1997 Silsoe Research Institute. | en |
