Development Of Adaptive Control System Based On Fuzzy Neural Network In Facilities Agricultural Greenhouse

The content of this thesis is one of important part of the major project "The development and demonstration of adaptive control system in facilities agricultural greenhouse"(No:120157), which is sponsored by "Study on the system of the construction project of Taiyuan to support the new rural". This work is proposed in accordance with the low degree of automation, weak adaptive ability and serious waste of labor. According to the distribution of the management requirement of Green Planting Base of Xiaodian in Taiyuan, a adaptive control system has been developed based on computer control technology, fuzzy neural network, configuration monitoring technology and network technology. The system can improve the greenhouse management and labor productivity effectively also.Since the1980s, agricultural infrastructure construction has made remarkable results in china, and we have the ability to built a large multi-span greenhouse. However, due to the control accuracy is low, the coupling relationship of the environmental factors is neglected as well, automation levels are lower in greenhouse system. Then the life of equipments are shorter because of using frequently too big in greenhouse. And lack of the capacity of adaptive, self-learning and self-organizing in the system finally. By using high-precision sensor, this paper developed a set of PLC as control core, the lower machine is responsible for data collection, processing, and information transmission. PC is responsible for the secondary monitoring system of the PC monitor. The management software is developed in PC based on fuzzy neural network, and the lower machine monitoring software is proposed based on the data processing and the integrated control. The main contents of this thesis are as follows:After literature reviewing and site investigation, the development trend of the system about greenhouse monitor and manager is conscientiously analyzed. And the function indicators are proposed as well. The whole project is designed based on the software and hardware frame structure combined with fuzzy neural network technology, microprocessor technology.According to the site investigation, we choose and test the type and feature of sensor. The sub-station based on PLC is designed with the requirements of high accuracy and strong anti-jamming capability to achieve the purpose of automatic control about temperature, humidity, light intensity, the concentration of carbon dioxide and other auxiliary facilities. Network program, data acquisition program, human machine interface program, alarming program with high portability are developed by using modular programming methods.Using star network architecture and Kingview6.55WEB as software flat in monitoring center, a set of analysis decision system is developed, which can realize the function of real-time monitoring, alarming and information collecting, adaptive controlling. The adaptive control strategy is proposed based on fuzzy neural network, and the genetic algorithm has been used to optimize the center value and the width of the membership functions. Configuration man-machine interface has the merits such as strong interaction, simple and beautiful screen and perfect functions.Visualization is adopted in the information communication based on the scene of the construction of network environment. Two level monitoring networks are established based on RS-485protocol and Ethernet communication protocol. Seamless information integration between management layer and field control layer is completed.The system tested by the hardware and software has been finished in the laboratory, and the result indicated that the performance can meet the expectation requirement. The system has completed the installation and debugging by the lab experiments. It has verified that it can successfully running in Green Planting Base of Xiaodian in Taiyuan since December2012. And the debugging and operation results show that the system hardware has a perfect function with a compact structure and stable performance; the software programs also operate precision. And the analysis and decision-making software interface is fully functional and operable. Communication network has the characteristics of strong real-time, the whole system could meet the design requirements.