Data based modeling approach for greenhouse air temperature and relative humidity

Abstract

Greenhouse air temperature and relative humidity are two important variables that can be influenced and controlled by many parameters. The present research refers to the prediction of these two variables in a greenhouse by a dynamic data-based model approach. Measurements of air temperature and relative humidity in several positions inside an experimental polyethylene covered arched roof greenhouse located in the experimental farm of the University of Thessaly in Central Greece, along with measurements of outside climate conditions were performed. Model inputs were the outside air temperature and relative humidity and model outputs were inside air temperature and relative humidity. Heating tube temperature and net radiation were disturbances to the system. The resulting transfer function models, for four different sensor positions along the greenhouse, were mostly first order models. Second order models were found for the positions between and above the plant canopy in the middle of the greenhouse, when the heating system was switched on and off, respectively. The average coefficient of determination RT 2 ranged between 0.76 and 0.99. The resulting models could be further used for greenhouse climate control purposes and energy use optimisation.