基于模型预测的日光温室温度优化控制  

Optimal temperature control of solar greenhouse based on MPC method

作　　者：杜太行 刘德 孙曙光 钱春阳 梁倩伟[1] DU Taihang;LIU De;SUN Shuguang;QIAN Chunyang;LIANG Qianwei(School of Artificial Intelligence,Hebei University of Technology,Tianjin 300130,China;Information Institute,Tianjin Academy of Agricultural Sciences,Tianjin 300192,China)

机构地区：[1]河北工业大学人工智能与数据科学学院,天津300130 [2]天津市农业科学院信息研究所,天津300192

出　　处：《中国测试》2022年第1期85-91,167,共8页China Measurement & Test

基　　金：天津市农业科技成果转化与推广项目(201901060);天津市科技计划项目(17YFZCNC00280)。

摘　　要：针对日光温室室内温度的优化控制,在切换控制的基础上提出基于模型预测(model predictive control,MPC)的温度控制方法。首先以室外环境因素、设备开关量以及室内温度为输入,建立NARX神经网络室内温度预测模型;进而将其输出作为目标优化函数的输入,该函数以温度的控制误差、设备切换次数、系统能耗为约束,从而将控制问题转化为目标函数的求解问题,在此基础之上得到局部最优的开关设备控制序列;最终通过滚动优化机制进入最佳温度切换控制状态。实验结果表明:基于模型预测的切换控制算法可以有效调控温室内的温度,降低温室系统能耗。Aiming at the optimal control of indoor temperature in solar greenhouse,a temperature control method based on model prediction is proposed on the basis of switching control.First,take outdoor environmental factors,equipment switching values,and indoor temperature as inputs to establish a NARX neural network indoor temperature prediction model.And then use its output as the input of the objective optimization function,which takes the temperature control error,the number of equipment switching times,and the system energy consumption as constraints,so as to transform the control problem into the objective function solution problem.On this basis,obtain the local optimal switchgear control sequence.Finally enter the optimal temperature switching control state through the rolling optimization mechanism.The experimental results show that the switching control algorithm based on model prediction can effectively regulate the temperature in the greenhouse and reduce the energy consumption of the greenhouse system.

关 键 词：日光温室 切换控制 NARX神经网络 模型预测控制 节能 

分 类 号：TP273.1[自动化与计算机技术—检测技术与自动化装置]