Research On Simulation And Prediction Of Greenhouse Microclimate In Summer By Computer

With the deeply development of research on greenhouse, intelligent controltheory has played a more and more prominent role in the greenhouse study. This paperset up summer microclimate model in greenhouse in Shanghai fOr simulation andprediction with this theory, then research the variation law of microclimate in thegreenhouse imported from abroad in the term of summer climate in china and analyzethe sole effect of every parameter on microclimate in greenhouse, so as to provide thetheoretical foundation for the learning and management of advanced greenhouse fromforeign countries.China has introduced many modern greenhouses from fOreign countries in thelast years, and the most typical greenhouse is the Holland Venlo type greenhouse. Butthe running result of these greenhouses is a little satisfying in china, especially in thebroiling summer, the energy-consuming for temperature-lowing in greenhouse is soenormous that the high cost and the low benefit cannot maintain the normal produceof these greenhouses. Therefore, a greenhouse microclimate model in summer wasbuilt fOr simulation and prediction, and then the variation discipline of microclimatein greenhouse was researched to provide the theoretical reference on the optimizationdesign of greenhouse imported in structure; the effect result of every parameter onmicroclimate in greenhouse was analyzed to determine the best control device andmake fOr the descending of energy-consuming in greenhouse, the microclimate stateof greenhouse in the typiFal climate days was prediction ed to avoid the occurrence ofcorPs ca1amity. -The microclimate in greenhouse is characterized by complicated, nonlinear anduncertain environment, and is extremely difficult to set up the precise mathematicsmodel fOr it, so this paper atilizes inielligent control theory, especially the artificialneural network and genetic algorithm to build the empirical model for themicroclimate in greenhouse. Firstly, on the basic of normal error back propagationalgorithm (BP a1gorithm), the model was added the inertia impulse item in theupdating formula for weight, and then let learning rate and inertia parameter adjustllself-adaptively, so the improved BP algorithm (Improved BP algorithm, IBP algorithm) formed. Secondly, the weight and threshold of BP algorithm was preliminarily optimized by genetic algorithm with powerful capability in global-optimization. Finally, the whole microclimate model in greenhouse was set up with the powerful local-optimization of IBP algorithm. During the course of model-building, in view of the influence of the corps in greenhouse on microclimate, the input parameters included the leaf area index (LAI) of crops. In addition, Visual C++ advanced language is characterized by the object-oriented programming, the friendly development environment, the convenient interface of man-machine conversation, and the high efficiency in development, which is just the reason that this language was used.By the experiment of simulation and prediction, the precision of simulation and prediction reach to 1.5% and 3%, respectively. This result demonstrated the model reasonable and reliable. Every parameter of microclimate in greenhouse was analyzed respectively by using the model, the results are that the temperature and humidity outside are the most important factors respectively, and solar radiation and LAI strongly influence the temperature and humidity inside. By analyzing and predicting the microclimate in greenhouse in typical weathers, the conclusions are that high temperature with low humidity weather and high temperature with high humidity appear one after the other. The measures are proposed as the followings: to the former, it should be taken to decrease the temperature with soppy shade-fan or spray inside or outside; to the latter, it should be taken to enlarge the ventilation area, to fit spray apparatus at the top of roof outside, to adopt havelock outside instead inside, to coat the roof white, et...