基于PSO的预测控制及在聚丙烯中的应用

PSO Based Predicted Control and Its Application to Temperature Control of Polypropylene Reactor

输入输出受限非线性系统的预测控制问题,可以看作是一个难以直接求解的约束非线性优化问题。针对预测控制在解决此类优化问题时,存在易收敛到局部极小或者非可行解,对初始值敏感等缺点,提出了一种基于微粒群优化方法的非线性预测控制算法。采用微粒群优化算法(PSO)作为模型预测控制的滚动优化方法,在线实时求解最优控制律。将PSO与序贯二次规划(SQP)算法进行对比仿真实验,求解两个标准函数优化问题,结果表明PSO能够快速有效地求得全局最小点,而SQP则很容易陷入局部极小点。将该算法应用于丙烯聚合反应过程的温度控制中,仿真结果显示了该方法的有效性。

Predictive control algorithms for a multivariable nonlinear system with limited inputs and outputs can be formulated as a constrained nonlinear optimization problem which is hard to solve. To the problem that traditional optimization methods are found often be trapped in illegal solutions or local minima, and sensitive to initial values in solving the above problems, a new nonlinear model predictive control algorithm based on particle swarm optimization （PSO） is proposed, in which PSO is employed as the optimization art in determining the series of optimal manipu- lated variables online. PSO is applied to benchmark functions comparing with sequential quadratic programming SQP. And PSO can get the global minimum quickly while SQP is going to be trapped in local minimum point. The proposed method is applied to the temperature control of the polypropylene reactor and simulation result shows the effectiveness.