Selection Of Empirical Model For Cucumber Temperature Management In Solar Greenhouse Under Different Light Conditions

With the development of science and technology,the intelligent management of solar greenhouses has been well known by farmers.The empirical model BP neural network uses the greenhouse microclimate environmental factors that can be obtained directly in the greenhouse to simulate the cucumber production environment of high-yield solar greenhouses.It has provided technical guidance for the production of ordinary farmers and has reached a high level of practice.At present,Lingyuan from Liaoning,and Yinan from Shandong are the main cucumber producing areas in solar greenhouse.The optimal temperature model for high yield of solar greenhouse cucumbers established by BP neural network model can guide the high yield production of solar greenhouse cucumbers.However,the current model established by BP neural network is established in a specific environment and cannot be directly used by farmers in various regions in production.How to let more farmers get the guidance of the model and how to choose the two models are still problems.In order to solve the above problems,in this experiment,the differences between the internal microclimates of the two modeling greenhouses were analyzed to determine the differences in the quantitative relationship between the internal microclimate factors in the greenhouses,and the differences between the models were determined by the differences in the calculation results of the models.Using the stepwise regression equation to determine the dominant internal environmental factors in the greenhouse are total solar radiation and sunshine hours.According to the calculation of multi-period and multi-factor similarity coefficients and similarity distances,firstly,the clusters of picking periods are divided according to the light intensity inside the greenhouse to construct a multi-dimensional spatial sequence.Based on the difference of the weights,a multi-period and multi-factor similarity distance and similarity coefficient are used to establish a greenhouse microclimate similarity determination system.Using Java editing program,the two models are combined with the greenhouse microclimate similarity determination system to establish a greenhouse automatic selection model system,which provides a theoretical basis for the selection of the model.It provides a reference for the practical and effective application of more models built in future experiments,and also solves the problem that most farmers in the solar greenhouse are not high-level in cucumber cultivation technology,get rid of experience,move towards digitization,gradually professionalize,and achieve and maintain high yields.Purpose.The microclimate similarity determination system established in this experiment is of great significance for guiding the cucumber production in facilities and improving the management level.The main findings are as follows:1.The empirical model of greenhouse cucumber temperature management based on BP neural network has strong regional characteristics,natural climate conditions and greenhouse lighting performance are important factors affecting the universality of the model.The same model can not be used for greenhouse management in areas where the climatic conditions are quite different.The optimal temperature model of cucumber in solar greenhouse established in Yinan,Shandong province and Lingyuan,Liaoning Province is not universal2.Use stepwise regression equation analysis to find out the total solar radiation and sunshine hours that dominate the greenhouse yield,and determine the relationship between the internal microclimate factors and yield of the greenhouses in Lingyuan,Liaoning and Yinan,Shandong.Lingyuan stepwise regression equation:Y=14.7 fz+0.983 rs-438.67(R~2≥0.863)Yinan stepwise regression equation:Y=5.307 fz+3.338 rs-1015.627(R~2≥0.776)In the formula:Y is the yield of cucumbers in the greenhouse(kg),fz is the total solar radiation(MJ/m~2),and rs is the sunshine hours(min).3.Determining the total amount of solar radiation and sunshine duration is the most important factor affecting the yield in greenhouse,and the application of the model is to achieve high yield by adjusting the temperature.Under the guidance of the suitable model,farmers can achieve high yield,so the total amount of solar radiation and sunshine duration are the main factors affecting the application effect of the optimal temperature model.4.In view of the long growth cycle of cucumber from winter to summer,there is a large error in analyzing the similarity distance and similarity coefficient based on the climate environment inside the greenhouse during the whole growth cycle,and the model cannot be directly applied to production.The light intensity in yinan and Lingyuan greenhouses during picking period was clustered and divided to establish multi-dimensional spatial sequence,and the weight ratio of total solar radiation and sunshine duration to yield in eight different spatial sequences was calculated,so as to establish microclimate similarity judgment system in solar greenhouses.Java programming will be inside the sunlight greenhouse microclimate similarity judgment system with the shandong and liaoning Ling Yuan incorporated the sunlight greenhouse cucumber optimum temperature model combining the model of automatic selection system,in the test ground temperature indoor through to the greenhouse environmental data collection for self judgment to guide of greenhouse,the instrument can uninterrupted for greenhouse production,automatic adjustment,Provide the most reliable data at all times.