基于Elman网络的过热汽温隐式广义预测控制

Superheated Steam Temperature Implicit Generalized Predictive Control Based On Elman Network

过热蒸汽温度具有时变、大迟延、大惯性的特点,使用传统的PID控制器很难取得满意的效果。针对这一问题,提出了基于Elman网络的过热汽温隐式广义预测控制系统。采用Elman网络作为被控对象的预测模型,在求解最优控制增量时直接辩识控制器的参数,使得对象的建模方便而且控制系统的在线计算量小。仿真结果表明,此控制系统具有优良的控制性能。

The superheated steam temperature has the character of timevarying, large delay, large inertia. It is very difficult to acquire the satisfied effect by traditional PID controller. Aiming at the problem the system of superheated steam temperature implicit generalized predictive control based on Elman network is presented. The system adopts Elman network as the predictive model and identifies the parameters of controller directly when solving the control increment. So the system can model the object easily and has the little quantity of calculation. The result of simulation indicates that the control system has the perfect capability.