变风量空调系统的分布式预测控制

Distributed predictive control for VAV system

基于预测控制策略的多输入多输出(MIMO)控制器可以解决单通道变风量系统的控制问题,但在输入输出变量较多时,计算就变得非常复杂。在分析系统工作机理和系统动力学行为的基础上,建立了内部模型,在此基础上采用分布式预测控制技术将系统分解成四个子系统,通过设定参考轨迹、输入输出约束及加权矩阵,解决了系统凸二次型的全局优化和纳什问题。通过仿真研究验证了系统控制的效果。

A single-duct VAV system can be controlled successfully by a Muhi-Input-Muhi-Output（MIMO） controller with a Model Predictive Control（MPC） strategy.However,an obvious drawback of this way is that it is computationally intensive,when the moderate number of inputs is large.In this paper,a internal model has been built by analyzing the working mechanism and the dynamics of the system,and the whole system is decomposed into four subsystem based on distributed predictive control strategy. MPC solves a constrained convex quadratic plant-wide and Nash optimization by defining weighting factors and constraint limits. Simulation results demonstrate the validity of the approach proposed.