| The modeling and control of greenhouse microclimate is very complex,because it involves many factors,such as many indoor environmental factors,a variety of control equipment,the growth status of indoor crops and outdoor weather conditions,and so on.The Venlo-type glass greenhouse widely used in the south of China is taken for example,and the modeling and automatic control of greenhouse microclimate is studied,by considering indoor air temperature,relative humidity,energy consumption,and various on-off control equipment.The main contents are as follows.(1)A dynamic pseudo-physical model of greenhouse air temperature and humidity for climate control purposes is constructed.Dynamic physical models of greenhouse microclimate are not suitable for climate control,not only because their structures are very complicated,but also the influence of equipment on indoor microclimate are seldom considered.Therefore,a dynamic pseudo-physical model of greenhouse air temperature and humidity for climate control purposes is constructed.Compared with physical models,the pseudo-physical model is greatly simplified,because many factors,such as the surface temperatures of cover layer and soil,are considered as boundary conditions,instead of constructing their dynamic models.At the same time,according to the switching characteristics of the common equipment,the operating mode of the greenhouse system is divided into several sub-modes,so the influence of equipment on the greenhouse microclimate is considered adequately in the model.The experimental results show that the simulation error of indoor air temperature is 2.7°C and the root mean square error is 1.0°C,while the simulation error of indoor air humidity is 13.1% and the root mean square error is 5.2%.Therefore,the pseudo-physical model can accurately simulate the dynamic changes of indoor air temperature and relative humidity,which provides a theoretical support for studying various automatic control methods of greenhouse microclimate.(2)The IARX online prediction models of indoor air temperature and humidity are constructed.Online prediction models of greenhouse environmental factors are the basis of various predictive control methods.ARX predictive models are suitable for online identification because of simple structure and small amount of computation.However,the external input variables are very different in different ARX models.By analyzing the dynamic pseudo-physical model constructed above,incremental auto regressive models with exogenous inputs(IARX models)of indoor air temperature and humidity are constructed successively under the three cooling modes,i.e.,natural ventilation mode,mechanical ventilation mode and pad-fan cooling mode.Compared with typical ARX models,the IARX models have fewer coefficients.When the number of environmental data used for model identification is more than 10 groups,the maximum prediction errors of indoor air temperature and humidity is 1.3°C and 8.9% by the IARX models,which are significantly less than the corresponding prediction errors of indoor air temperature and humidity by the typical ARX model(2.8°C and 15.2%).(3)A multi-objective compatible switching control method based on the IARX prediction models of indoor air temperature and humidity is proposed.For the actual greenhouse control system,indoor air temperature,relative humidity and energy consumption are the key factors to be considered.However,there are some conflicts in the control of these factors.Therefore,a multi-objective compatible switching control method based on the IARX prediction models of indoor air temperature and humidity is proposed,by integrating model predictive control,switching control and conflicting multi-objective compatible control.The new control method includes three parts: model prediction,switching optimization and feedback correction.In the model prediction part,the IARX models are used to predict the indoor air temperature and relative humidity over a specified horizon.In the switching optimization part,on the premise that the predictive values of indoor air temperature and relative humidity are controlled within the set ranges,the energy consumption is taken as the optimization index to design the switching rules.In the feedback correction part,the model predictive performance and model coefficients are corrected in real time by the feedback of prediction error and online identification.At first,the compatible control method of energy saving and cooling was studied by considering the indoor air temperature and energy consumption.The simulation results show that the maximum error and the mean square error of indoor air temperature are 0.7°C and 0.4°C respectively,which are both less than the corresponding errors by using a reference switching control method(2.0°C and 0.8°C).When several cooling modes are switched,the compatible switching control method can save energy more than 13%,compared with the reference switching control method.Then the multi-objective compatible switching control method was studied by considering indoor air temperature,relative humidity and energy consumption.The simulation results show that the maximum error and the mean square error of indoor air temperature are 1.0°C and 0.5°C respectively,while the maximum error and the mean square error of indoor air humidity are 3.8% and 2.2%.The problem that operating modes are switched frequently can be effectively solved by adjusting the setting ranges of indoor air temperature and humidity.(4)The compatible switching control method was verified in a glass glasshouse by considering indoor air temperature and energy consumption.The results show that the control method is feasible.The maximum deviation and mean square error of indoor temperature exceeding the upper limit are 0.7°C and 0.2°C respectively,which are obviously smaller than the corresponding errors(1.7°C and 0.7°C)produced by the reference switching control method.In addition,the compatible switching control method also shows certain energy saving potential. |
PDF Full Text Request