聚合釜温度自适应预测控制设计与仿真

Design and simulation of adaptive predictive temperature control in polymerization

在聚合反应升温到设定温度的过渡阶段,由于聚合颗粒的不断产生影响聚合釜传热系数,其温度被控对象特性也会发生变化,如容量滞后变化、滞后时间的变化等,温度控制具有的非线性、滞后、参数慢时变的特性,本文提出一种基于CARIMA模型的自适应广义预测控制,以工艺常见的一阶惯性滞后对象为研究对象,利用递推最小二乘法不断在线辨识修正其CARIMA模型参数,并通过改进的JGPC广义预测控制计算最优化的控制量,从而能够达到自适应被控对象变化,实现无超调最优平滑控制。通过仿真实验,得出该方法能够在对象特性发生改变时仍能保持良好的曲线特性,与传统PID相比,具有优良的自适应控制效果。

In the transition stage of the polymerization reaction heating to the set temperature, the characteristics of the controlled object will change because of the constant influence of the polymer on the heat transfer coefficient of the polymerizer, such as the change of the capacity lag, the change of the lag time and so on. The temperature control has the characteristics of nonlinear, lag, and slow time-varying parameters.In this paper, an adaptive generalized predictive control based on CARIMA model is proposed, which takes the common first order inertia time lag object as the research object. The method uses recursive least square method to identify and amend its CARIMA model parameters continuously and on-line, and can achieve self-optimal control by improving the JGPC generalized predictive control. The adaptive control object is changed to achieve the optimal control without overshoot. The simulation experiments show that this method can still maintain good curve characteristics when the object features change. Compared with the traditional PID, the proposed scheme has excellent adaptive control effect.