Experimental Study of Greenhouse Climate by Auto Regressive Model with External Input

Abstract

This paper presents an experimental study of modeling of greenhouse climate. Auto regressive with external input (ARX) model is used for system identification of greenhouse climate without heated and ventilated systems. In system identification of greenhouse climate regarded as a kind of MIMO system, outside temperature, wind speed, solar radiation and outside humidity are utilized as the input variables, and inside temperature and humidity are the output variables. An ARX model is made from data with 60 seconds sample time. In order to evaluate the performance of the ARX model, we compared the measured data of another day with the simulated outputs of the ARX model. Finally, the ARX model is validated by the goodness of fit.