换热网络模型预测控制中的滚动优化策略

Steady-state optimization research of model predictive control for the heat-exchanger network

针对换热网络具有非线性、滞后性、约束性等特性,提出将模型预测控制算法应用到换热网络的控制中,并对其滚动优化策略进行深入研究。首先,建立以综合费用最少的换热网络模型及目标函数,然后引入最速下降法对换热网络进行优化,以获得下一步的最优旁路开度,实时计算,实现换热网络模型预测控制中的滚动优化策略。通过算例证明:采用滚动优化策略能够快速、准确地计算出每一步的最优旁路开度,在保证出口温度达到目标值的同时使系统经济效益有所提升。

The heat-exchanger network has the characteristics of nonlinearity,delay and binding,etc.In terms of the characteristics above,it was proposed applying the model predictive control to the control of heat-exchanger network,and a steady-state optimization research of model predictive control for the heat-exchanger network was carried out in this paper.The heat-exchanger network model was firstly established,and the minimum total cost was taken as objective function.The heat-exchanger network was then optimized by the steepest descent method to obtain the optimum bypass adjustment fractions for next step.An example showed that the optimum bypass adjustment fraction for every step can be calculated quickly and accurately using the optimization algorithm.Using the model predictive control not only ensures the outlet temperature reaching the objective value,but also increases the economic efficiency of the system to some extent.