Study On Control Arithmetic And Data Management Of Intelligent Greenhouse

There are many characteristics of greenhouse environment, such as variable nonlinear and uncertainty, it's hard to build an accurate mathematics model. The effect of the conventional control method (for example, PID algorithm and optimal control) is not satisfying, in order to enhance the level and precision of greenhouse control, so intelligent control method is proposed for controlling the greenhouse environment. In the second part of this paper, low-cost software is developed; it not only can inspect and control the environment, but also manage the data. As a good assistant of the greenhouse control system, it promote the greenhouse control towards intelligent and simplification.In the first part, a fuzzy neural network controller (FNNC) is designed. It gathers the strongpoint of fuzzy logic and neural network to promote the study and control ability of the system. The FNNC not only can deal with fuzzy information, finish reasoning function, but also has some characteristics of the neural network, such as nonlinear and self-learning. Fuzzy Neural Networks can be viewed as a parallel and distributed network, in which each neuron represents parameter in fuzzy system and each output of neuron is connected to the neuron in the next Layer through the weight parameter. The fuzzy neural networks are trained continuously by inputting specimen data; then the membership function parameters and the weights of fuzzy logic rules are optimized by using back propagation algorithm. The parallel processing network makes the self-adaptation of the membership functions and the self-organization fuzzy logic rules possible.In addition, a greenhouse environment temperature model is built by using artificial neural networks method. The network model is trained by inputting specimen data. The experiments results demonstrated the model can well reflect the nonlinear character of the temperature model.At last the simulation to control the greenhouse is partly carried out. The simulation results showed that intelligent control did better than PID control in improving overshoot and steady-state error. The latter is more suitable to be used in intelligent greenhouse controlling.The second part designed a suit of software which is being to be used in greenhouse data management. Using Visual Basic 6.0 as the developing tool, graphic interface, drop-down menu and pop menu were adopted to facilitate the user. Through the database that built, we achieved the function such as data storage, data query, curve display of the history data, data report creation and print. On the base of communication realization that designed, it can show the inspected value of the environmental factor in real time, display the changing trend with dynamic curve. Expert decision and consult database storied a series of values of the effect factor that the expert recommended to some crop. Manager not only could reference to the expertise, but also could transfer it to parameter setting module, and it was used to control the environment factor. As long as being farther perfected, it could be used as an assistant in intelligent greenhouse control.