| As a complex distributed parameter system,greenhouse is a type of semi-closed and highly coupled environment,and its spatial and temporal distribution characteristics have a significant impact on crop production and greenhouse energy consumption.The environmental parameters in the greenhouse include temperature,humidity,and light,which interact and influence each other,ultimately determining the growth of crops and energy consumption in the greenhouse.The rational control of greenhouse environment to improve crop production efficiency and reduce energy consumption has always been an important issue in greenhouse technology research.In this study,orthogonal decomposition technique is used to extract the main environmental features from the computational fluid dynamics model of the greenhouse environment.By transforming the high-dimensional data of the greenhouse environment into low-dimensional data through orthogonal decomposition technique,the spatiotemporal changes of the greenhouse climate can be reconstructed based on the low-dimensional model using third-order spline interpolation,while reducing the computational complexity.Based on this,combined with the crop growth model,this study proposes a rolling time-domain optimal control strategy for the greenhouse environment.In each step of the limited time domain,the crop dry weight and greenhouse energy consumption are taken as the optimization objectives J,and the lowdimensional model is used to quickly solve the greenhouse environment response based on actual external meteorological data.The optimal control variables,such as the greenhouse shading rate and fan speed,are sought through swarm intelligence algorithm.This process of variable optimization is completed throughout the entire growing season.The main research of this project includes:(1)Developing a greenhouse environment simulation model using computational fluid dynamics methods,extracting data on the distribution of climate characteristics in the crop area,and using proper orthogonal decomposition to project the highdimensional climate feature array onto a low-dimensional feature subspace that can be reversely mapped and reconstructed in the high-dimensional space.The simulation results are validated by installing temperature sensors in the actual greenhouse to verify the accuracy of the simulation and the effectiveness of the dimensionality reduction reconstruction.(2)To improve the accuracy of greenhouse observation,a lettuce growth model considering grid distribution is established.Referring to the lettuce growth model based on energy balance method,the model is extended based on the greenhouse environment model to expand the homogenized indoor air temperature to the temperature of each grid and the average dry weight of the entire greenhouse to the dry weight of each grid,thereby obtaining a crop growth model with high spatial resolution.(3)In order to further realize the optimization control strategy of greenhouse environment considering crop growth spatial distribution,this study proposes an optimization strategy with crop dry weight and greenhouse energy consumption as optimization indicators.The method applies swarm intelligence algorithm to search for the optimal control variables within each time step of one hour,and the greenhouse environment response of each candidate solution is obtained by multi-dimensional interpolation of the POD mode coefficients.This optimization strategy is rolled out throughout the entire growing season to obtain the predicted crop growth situation and the optimal sequence of greenhouse facility control states.Compared with traditional methods,this optimization control strategy realizes the impact of finely observed environmental variables on crop growth in spatial distribution,solves the problem of high computational time cost and complexity in CFD modeling methods,and applies to optimization control,thereby achieving a high-resolution optimization scheme for greenhouse environmental parameters. |
PDF Full Text Request