Simulation Of Single-pan Plastic Greenhouse Microclimate For South China

Single-span plastic greenhouse which has original structural characteristics is extensively used in southern China, where internal temperature, humidity and other environmental conditions are the most important factors that affect crop growth and development. This article is based on single-span plastic greenhouse in south, by using three methods to simulate and predict the daily average value, amplitude and spatial distribution of internal environment under different conditions of Season, weather and ventilation measures. The main content and results of this research are as follows:1. Based on the meteorological data both inside and outside the plastic houses in Cixi, Zhejiang and Agricultural Meteorological Experimental Station of Nanjing University of Information Science and Technology, three BP neural network models were established. Outdoor solar radiation, air temperature, relative humidity and wind speed were used as input variables, and indoor temperature and relative humidity were applied as output variables. Results showed that the root mean square error (RMSE) and the relative error (RE) between trained air temperature and measured value for three models were less than2℃and about4%respectively. RMSE and RE between trained relative humidity and measured value were less than a percentage of7and7%respectively. RMSE between predicted air temperature and measured value for three models were about2℃. RE between predicted air temperature and measured value were less than6%in spring, while larger in winter. RMSE and RE between predicted relative humidity and measured value were less than a percentage of7and9%respectively. Indoor temperature and relative humidity can be simulated quite precisely with BP neural network, both of which satisfied the forecast requirements for plastic greenhouse microclimate. The precision of temperature for three models was higher than that of relative humidity obviously.2. Based on the meteorological data of different season and weather condition in plastic greenhouse from2006to2009in Cixi, the relationship between a variety of air temperature amplitude inside and outside the plastic greenhouse under three kinds of weather conditions as cloudless day, broken sky and overcast sky in winter and spring was acquired by using Harmonic analysis to analyze the second-order harmonic of the daily air temperature inside and outside the plastic greenhouse. According to the harmonic model parameters, the simulation model inside plastic greenhouse was established by applying stepwise regression whose independent variables was meteorological elements outside plastic greenhouse. The results showed that the simulation accuracy was comparative in winter and spring, and the model accuracy under three weather conditions had the same trend. The coefficient of determination (R2) between simulated air temperature and measured value was above0.93both in cloudless day and broken sky while the RMSE and the AE was no more than3.0℃and2.4℃. R2between simulated air temperature and measured value in overcast sky was approximately0.79, with a RMSE less than3.0℃and a RE about2℃, which had a poor accuracy compared with that in cloudless day and broken sky. Temperature phase inside plastic greenhouse was slightly ahead of outside, which was significantly obvious in cloudless day and broken sky than that in overcast sky and so it was between Spring and winter. Daily minimum temperature inside greenhouse was always lower than that outside greenhouse, and was significantly obvious in spring than winter. It had the biggest difference in cloudless day, followed with broken sky and the overcast days at last.3. Based on the indoor and outdoor observed data, the simulation and verification of air temperature were carried out under ventilation by the use of computational fluid dynamics (CFD) technology. Temperature simulation and actual measurement determin the coefficient of R2at1.2meters height to be0.7578,and1.5meters height to be0.6141,with the actual root mean square error (RMSE) to be1.52℃and1.9℃, meanwhile the low temperature difference value accounted for86.1%and80.6%respectively. Maximum and average relative error of temperature simulation and measured values are8.1%and3.6%at1.2m height plane, and8.2%and4.6%at1.5m height plane espectively. The simulation results coincide the measured results well.4. Simulation of greenhouse ventilation under different methods and different wind speeds were studied by using CFD technology. Compared with unilateral window ventilation, windows and doors joint ventilation, made average air temperature of the canopy height cooling of0.25°C and of1.5m height dropped by an average of0.32℃. Cooling rate of door and window joint ventilation of greenhouse’s upper plane is slightly larger than the lower plane. When the wind speed increased by0.5m/s to1m/s, it made average air temperature of the canopy height cooling of0.4℃, average air temperature of1.5m height dropped by an average of0.58℃. Greenhouse temperature field showed the trend from low to high from the south side of the window air inlet to the north side window vent caused by wind speed. Although the cooling range and cooling distribution are different, but all are at widening. As long as accuracy is sufficient, CFD technology can assess the greenhouse three-dimensional flow field characteristics and optimize greenhouse design, which provides a reference for the production design management of plastic greenhouses.