A Study On Modeling And Control Of Greenhouse

It's not only important but also difficult to control the greenhouse climate. On the one hand, the factors included in the greenhouse interact: the climate factors are important to the crop, while the growth of crop affects the climate factors. The processes include both physics processes and physiology processes. So the greenhouse is a multi-variable coupled, time-varying and non-linear system. On the other hand, the means to control greenhouse environment are limited. The main methods include ventilation, shading, heat preservation, spray and heating etc. Especially, the lower control losses are required. In this paper, the works on greenhouse modeling and control are surveyed, and the further research of greenhouse modeling and control is made. The modeling of greenhouse climate is the key to the control of greenhouse climate. At first, the main thermodynamics processes related to greenhouse climate are analyzed systemically, the physical models are set up. Then, the method of experiment modeling, which is based on the theory of system identification, is deliberated according to the input-output data. In order to overcome the deficiency existed in the previous models, this paper proposes the adaptive fuzzy logical networks model of the greenhouse climate. At first, the initial model is built according to the qualitative description about greenhouse climate. Then, utilizing experimental data or practical data, the parameters in model are adjusted by back-propagation algorithms. The quantitative model of greenhouse climate can be built. Due to its capability to handle both numerical data and linguistic information, the modeling based on adaptive fuzzy logical networks is superior to the method based on system identification and neural networks. The simulation results show that the adaptive fuzzy logical networks model of the greenhouse climate can quantificationally describe the system well and make more veracious prediction. In this paper, the two types of configurations for greenhouse climate optimal control are studied. They are hierarchical control and integrated control. In hierarchical control, set points are transferred from some decision units to a suitable low-level controller. Integrated control where the control actions are obtained by optimizing an explicit objective function is based on the optimum control theory and is solved according to state equations. Whether hierarchical control is adopted or integrated control is adopted, it's difficult to control greenhouse climate optimally. Applying expert system to greenhouse climate control is appropriate. In this paper, the expert control system combined with fuzzy logical models is constructed. According to the expert rules related to the prediction value, data measured and empirical data, the expert control system makes control decisions. The form of knowledge expression is ule frame + rule style which is suitable for agriculture knowledge. The simulation results show that the frame of expert control system is feasible.