Setpoint optimization based on the net photosynthetic rate and predictive control of the greenhouse model

This paper proposes a temperature prediction and control method for agricultural greenhouses based on hierarchical control. Currently, most greenhouse temperature setpoints are determined based on experience, without considering the actual needs of crop growth. This study proposes an optimal method for acquiring temperature setpoints, with the goal of maximizing the net photosynthetic rate. By analyzing the mechanisms behind greenhouse environmental changes, a general model structure for agricultural greenhouses is developed. Given that many physical parameters in the model are difficult to measure, this paper employs a dynamic model parameter correction method based on a particle swarm optimization algorithm to reduce the impact of unmodeled dynamics on model performance. The study uses a Venlotype greenhouse as an example and proposes a control system that includes four primary devices (roof windows, side windows, shading nets, insulation nets) and two secondary devices (boiler heating and fans). The energy consumption of the primary devices mainly arises from switching the device states, and a traversal method is used to optimize the control strategy. Secondary devices are precisely adjusted through a trained neural network controller. The core of this control strategy is to ensure the optimal use of energy and the best control of the greenhouse environment, while meeting the temperature conditions required for crop growth.