多区域暖通空调温湿度鲁棒模型预测控制策略

Temperature and Humidity Robust Model Predictive Control Strategy for Multi-zone Heating,Ventilation and Air Conditioning

多区域暖通空调(heating,ventilation and air conditioning,HVAC)系统相比基础的单区域空调系统能够更好地描述空调房间各区域内的动态热响应过程,传统模型预测控制(model predictive control,MPC)算法能较好实现温湿度控制,但室内区域之间传热系数的耦合作用,使得模型中存在不确定性扰动,这种情况下,传统MPC算法难以实现控制。针对这类问题,设计了一种基于线性矩阵不等式的改进鲁棒模型预测控制(robust model predictive control,RMPC)策略,将模型中存在的扰动用多胞不确定集来表示。最终将一个无穷时域控制量无限的最小化最坏情况的min-max问题转化为有限时域内求解状态反馈控制率的问题,同时考虑了状态与输入约束条件,通过在线实时求解即可得到每一时刻最优解。最后对多区域HVAC空调房间系统进行仿真,结果表明本文算法下3个区域温湿度的跟踪精度较高,抗干扰能力也较强。

Compared with the basic single-zone air-conditioning system,the multi-zone heating,ventilation and air conditioning(HVAC)system can better describe the dynamic thermal response process in each zone of an air-conditioned room.The traditional model predictive control(MPC)algorithm can control temperature and humidity well,but the coupling effect of heat transfer coefficients between indoor areas leads to uncertain disturbances in the Model.In this case,the traditional MPC algorithm is difficult to achieve control.To solve these problems,a robust model predictive control(RMPC)strategy based on linear matrix inequality(LMI)was designed.The disturbances in the model were represented by multicellular uncertainty sets.Lastly,a min-max problem with infinite time-domain control quantity was transformed into a problem with finite time-domain to solve the state feedback control rate.Considering the state and input constraints,the optimal solution at every moment could be obtained by online real-time solution.Finally,through the simulation of Multi-zone air conditioning room system,the results show that the algorithm has higher tracking accuracy of temperature and humidity in the three areas,and stronger anti-interference ability.