| Greenhouse crop production system is a complex system, which include crops, facilities, environment, management technology, and the interaction of these factors. A simulation model with better precision and the quantitative relationship of each factor in the system is the premise for the optimization control of the greenhouse environment. The inside airflow, temperature and humidity distribution can be predicted by CFD technology under different external weather conditions. The simulation results not only provide the detailed data information, but also as the theory support for the greenhouse control strategy. The dynamic change of temperature and humidity under different time scale in a single span venlo type greenhouse was simulated by the CFD technology and verified by the experiments. The most advantages of CFD simulation is that the detailed flow field data can be produced, but it has a long running time due to the complex model structure. To make full use of the advantages and in the same time effectively overcome the defects, system identification tenology convert the complex CFD model into a simple model which can be easily applied in the control system design.The specific job and conclusion as follows:(1) The physical model of radiation, convection and thermal conduction between outside and inside air, cover,soil, crop was analyzed based on the mass and heat transfer theoretical model. It can be provided the theoretical support for the CFD modeling.Based on the air flow character inside the Venlo type greenhouse, the standard k-ε turbulence model was used to CFD numerical computation. The fluid motion close to the wall was computed by the stardard wall function. DO radiation model was adopted to simulate the effect from the solar radiation.The natural ventilation caused by buoyancy was added to the momentum conversation equation as a source item by Boussinesq hypothesis. The porous model was used to simulate the crop in the greenhouse.(2) The 3-D unsteady mathematical model of natural ventilation condition in Venlo type greenhouse was built up by CFD method. The boundary conditions were based on solar radiation, external air temperature, wind velocity and direction changed over time.The result showed that the inside temperature has the same variation tendency with outside temperature during the whole simulation period. The RMSE was 0.688 ℃ between simulated and measured value, maximum relative error and average relative error was 8.9% and 2.8%, respectively. The established CFD model is effective.(3) The dynamic cooling and dehumidification process in Summer plum rain season under natural ventilation were simulated by CFD unsteady model. Duo to the indoor airflow movement and outdoor climate conditions, different position in the greenhouse had its different respond time and amplitude.The area close to the west entrance has the maximum cooling amplitude with 10.3℃ within 30 minutes, while the canopy area in the central greenhouse was 2.2℃ .The cooling and dehumidification amplitude was 6.8℃ and 13.6% at the southeast coner above the canopy level,respectively. When the outside wind speed changed from 0.81 m·s-1 to 1.2 m·s-1 and the wind direction changed from south-southwest to west, the local airspeed near the west windward entrance appeared firstly enhanced and then decreased. The airflow speed in the east area was markedly increased. The average temperature of x=1.5m cross section at the entrance of ventilation and crop canopy dropped 1.87℃ and 0.92℃ in 180s. Cooling effect of natural ventilation was remarkably affected by wind speed and direction.(4) Based on the mass and energy exchange mechanism and energy conversion between greenhouse air and spry system, divide the greenhouse into 3 computational domain by the interface where the spray nozzle located, the spray system simulation model was built up by CFD technology. The sensitive heat and latent heat exchange between spray system and inside air was add into the control equations as the source item. The result showed that the RMSB of temperature and humidity was 0.525℃ and 2.9%, respectively. The established CFD mode is verified effectively.(5) The input/output time series data of the system was provided by the unsteady CFD numerical simulation was inserted into greenhouse environment control system. Then the CFD model was converted into system control model based on data by system identification.It was showed by simulated experiment that the RMSE was 0.5888℃ between CFD-ARMAX model output and measured value. So the model can describe the actual temperature change in the greenhouse. Based on the CFD-ARMAX model, the predictive controller was designed, then the optimal control result was obtained at the temperature set point by simulation. |
PDF Full Text Request