Toward a modeling and control approach based on the drying product in greenhouses

Solar drying in greenhouses has become popular as an alternative to conventional drying because it provides a barrier that reduces contaminants and allows it to dry large volumes of product. However, it is still necessary to investigate the optimal management of greenhouse components to improve drying conditions, increasing the quality of the final product. This research evaluated two on/off type controllers and a model predictive control (MPC) in a greenhouse-type solar dryer at the Universidad Autónoma Chapingo, Mexico. Tomato slices were used as the product to be dehydrated. A model, based on the system identification N4SID algorithm, capable of predicting the temperature of the slice was obtained. The contrast of on/off controllers arises from using air and slice temperatures as the basis for the control logic. Using the model, the parameters for the model predictive control were obtained using the slice temperature as the predicted variable. In all cases, the exhaust fans were the actuators. The results showed that the controller based on slice temperature consumes less energy than the one based on air temperature, and the MPC controller is the most efficient compared to the two on/off controllers, reducing the energy required to drive the fans. Furthermore, the MPC can keep the slice close to 50°C. The same behavior is achieved with the on/off control based on slice temperature but not on the air temperature-based control, where a significant variability in the slices’ temperature was observed.