Development Of Control System And Control Strategy For Climate Environment In Greenhouse

Computer control is one of the key technologies in modern greenhouse, which is one of the most important facilities in high efficient agricultural production. It goes without saying that a large amount of research work has been done in this field. However, climate control in greenhouses, viewed as "an unending challenge", is an extraordinarily complicated problem, in which there is still a lot of work to do. For meeting the objective of "Research and Development of Intelligent Horticultural Plastic Greenhouses (No.021102542) ", which is one of the key research projects in agriculture in Zhejiang province. A computer control system was designed based on the actual demand of agricultural production and the climate characteristics of Zhejiang province. And the strategy of climate control in greenhouse and the collection of climate information were also studied in this project. Main research constents and results are listed as follows:(1) Neural Network Model (NNM) of greenhouses and Proportional-Integral-Differential (PID) control. It is difficult to establish a precise mathmatic model of greenhouse because greenhouses are easy to be affected by outdoor climate. The Radial Basis Function Neural Network (RBFNN) is simple in structure and has an excellent performance in function approximation and training speed. In this research work, RBFNN-based indoor temperature prediction models were established and trained with the data collected from the greenhouse in the Experiment Center of Plant at Zhejiang University. A satisfied result was achieved by useing this model, the root mean square error between the predicted temperature and the actual temperature was only 0.0073. A method based on proportional integral plus feedforward (PI&FF) control was investigated in order to save power and precisely control the temperature in greenhouses. This method, based on the difference between measured indoor temperature and the set temperature, and the other parameters, such as outdoor temperature, illumination, wind speed and so on, was applied to calculate the temperature of hot-water pipe. Genetic arithmetic (GA) with NNM of greenhouses were combined together to optimize the parameters in the PI&FF control. The results indicated that identifying the parameters in PI&FF control based RBFNN could improve the efficiency. With root average square error between the desired temperature and indoor temperature as the evaluation function in GA , self-adapted adjusting of PI control parameters could be achieved.(2) Fuzzy control (FC) in greenhouse. To overcome the defection of PID control, some methods were presented to control the temperature of heating pipe in greenhouses and the...