Research On Modeling And Control Strategy In Sunlight Greenhouse Environment

Sunlight greenhouse with good light and heat function, used for low season thermophilic crops cultivation. Sunlight greenhouse can cope with the cold, frost and other extreme weather conditions on crops, increase crop yield and quality. Modern sunlight greenhouse, can not only sealed insulation, ventilation and cooling, at the same time has the function of regulating the indoor temperature and humidity, illumination and so on environment, creating the best environment needed by the crops by computer automatic control conditions.In unsuitable crop season for crops to provide optimum growth conditions.The key to achieve sunlight greenhouse control is to establish a small climate models. Due to the indoor temperature and humidity environment not only affected by indoor environment factors and the structure of the greenhouse, also influenced by outdoor weather conditions, and has a very complex relationship between these factors, therefore, the greenhouse microclimate environment is a highly nonlinear, strong coupling, large time delay system, establish a mechanism of sunlight greenhouse environment model is relatively difficult.Based on the complex properties of the sunlight greenhouse, this paper use artificial neural network, establish sunlight greenhouse temperature and humidity forecasting model based on Bayesian regularization algorithm.In this model, the principal component analysis using input simplified network model, decoupling the effects of coupling relationship factors. Then the network model was optimized by using the Bayesian regularization algorithm, improve the network generalization ability. In order to verify the effect of the established prediction model, this paper has chosen common neural network algorithm(traingd,traingdm) to establish a predictive model of temperature and humidity. By comparison, the use of Bayesian regularization algorithm for high prediction accuracy of the model established strong network generalization ability.Prediction model based on the established sunlight greenhouse environment, this paper puts forward the control strategy of the sunlight greenhouse temperature environment. The proposed self-learning adaptive ability of fuzzy logic and neural network knowledge skills combined, design of fuzzy neural network controller. Using fuzzy neural network control method for the simulation verification of greenhouse environment temperature, the simulation results show that the fuzzy neural network for sunlight greenhouse etemperatur controlled environment is feasible,and by memory neural network fuzzy inference rules, can improve the control precision at the same time, enhance the system’s adaptive ability.