Study On Change Regulation Of Key Environmental Factors And Monitoring And Control System In Solar Greenhouse

With the popularization of solar greenhouses,greenhouse control system has achieved rapid development in China.It is the premise of intelligent control to establishing accurate greenhouse prediction models,which can provide theoretical guidance for greenhouse control systems and greatly improve their intelligence level.In this paper,the combination model of wavelet neural network was used to forecast the time series of key environmental factors in greenhouse.On this basis,in order to improve the prediction accuracy,the wolf pack algorithm is introduced to optimize the initial parameters of the wavelet neural network.In addition,by improving the initialization method and adaptive ability of wolf pack algorithm,the optimization effect and efficiency of wolf pack algorithm are improved,and the prediction model is further optimized.According to the actual demand of solar greenhouse,the environmental monitoring and control system of solar greenhouse based on Internet of Things was designed,and the control method of the equipment is optimized on the basis of the prediction model.Specific research contents and steps are as follows:Firstly,the prediction model based on wavelet neural network is constr ucted,with five input layers,one output layer,and six hidden layers.The model uses five data values before the current time as input and the data values at the current time as output to predict greenhouse environmental factors.The experimental results show that the wavelet neural network prediction model can accurately predict the trend of greenhouse factors,and the overall error is small.Then,in order to improve the accuracy and stability of the prediction,the initial parameters of the wavelet neural network are optimized by the Wolf Pack Algorithm with the excellent optimization ability.To improves the Wolf Pack Algorithm,the idea of reverse learning was introduced,which improves the efficiency and probability of searching for the optimal solution,and combines the running and siege process of wolf pack algorithm into an adaptive hunting process,which can adjust the step size adaptively according to the distribution of prey resources,thus greatly reducing the huge amount of computation caused by optimization in high-dimensional space.The experimental results show that the model prediction accuracy is higher after the improved Wolf Pack Algorithm optimizes the parameters.On the basis of improving the prediction accuracy,the air temperature is predicted by multi-step.The experimental results show that the improved model can accurately predict the air temperature within three steps(i.e.30 minutes),which provides a theoretical basis for intelligent control of greenhouse equipment.Finally,according to the actual demand of solar greenhouse in China,the three-tier structure of the Internet of Things is improved by referring to the domain structure of the Internet of Things,and the greenhouse demand and system functions are allocated to the corresponding levels.From three aspects of hardware circuit,communication protocol and data processing,the intelligent monitoring and control system of Internet of Things in solar greenhouse is designed in detail,which realizes real-time monitoring and reliable transmission of greenhouse environmental information and remote control of greenhouse equipment.In addition,combined with the prediction model for short-term accurate prediction of air temperature,by improving the intelligent control method of greenhouse curtain rolling machine,the intelligence of the control system is improved.