| dc.description.abstract | Model-based greenhouse climate control combined with use of reliable sensors can achieve ideal balance of good greenhouse climate and energy saving; however, it requires development of robust models describing both the greenhouse environment and crop response to environmental variables. This paper presents an approach to the dynamic modelling of greenhouse air temperature. Also it was investigated to what extend the resulting models could be used to describe the inside air temperature of different types of greenhouse under the different climate conditions. Two greenhouse experiments were conducted under both Belgian and Greek climate conditions with two types of greenhouses. Outside air temperature, relative humidity, heating tube temperature, and global solar radiation were used as input variables for the model. A second order transfer function including the four input variables was suitable to describe the greenhouse temperature evolution with an average coefficient of determination R 2 T of 0.80 for the days of minimum ventilation. However, this model failed to describe the inside temperature of the greenhouse under the Greek conditions. On the other hand the first-order transfer function including all four input variables well described (with an average R 2 T value of 0.99) the dynamic behaviour of greenhouse temperature under the Greek experimental conditions. This study suggests the most suitable control and disturbance variables to model greenhouse temperature and yet can be used for control purposes under minimal ventilation conditions. However, model structure should be adapted at different climate conditions and greenhouse types. In order to catch the dynamic behaviour of the greenhouse temperature an online estimation of the model parameters should be employed. | en |