基于史密斯预估补偿及自整定PID的过热汽温控制方案

The Control Scheme of Overheating Steam Temperature Based on Smith Predictive Compensation and Self-adjusting PID

介绍了史密斯预估器的基本原理,利用预估器对传统的控制方案进行改进,将串级控制改为带补偿环节的单PID控制。论述了基于梯度下降法的RBF网络训练方法,采用RBF网络对PID参数进行自整定。将两种方法结合起来,设计了基于史密斯预估补偿及自整定PID的过热汽温控制方案,仿真结果表明,这种改进控制方案相比于原方案调节稳定,抗扰动能力强,并缩短了调试时间。

The fundamental principles of Smith predictor are presented, and it is utilized to improve the traditional control scheme,changing the cascade control into single PID control with compensation link. The RBF network training method based on gradient descent method is expounded and the RBF network is applied to the self-adjusting of PID parameters. Then based on a combination of the two methods, a control scheme of overheating steam temperature based on Smith predictive compensation and self-adjusting PID is designed. The simulation results show that, compared with the original scheme, the improved control scheme offers a more stable adjustment and has a more powerful disturbance-rejection performance. Moreover, it can also shorten the debug time.