论文部分内容阅读
在炼油、化工装置中,由于有些被控过程存在随机干扰,动态特性随时间变化,纯滞后较大等原因,致使用PID控制算法难以满足生产的要求。因此我们在常减压装置的PROVOX集散控制系统的上位机上用FORTRAN77高级语言开发了LQG自校正控制软件,该软件成为集散控制系统的一个通用控制模块,对PROVOX系统上的多个控制回路实施LQG自校正控制。在控制算法上,我们从新的角度提出一种设计LQG控制器的方法,并与递推最小二乘法结合,构成了LQG自校正控制器。在算法中,被控过程用其输入输出模型描述,在模型中考虑了过程的纯滞后特性,建立一个二次性能指标,在性能指标中包含了输出与给定的偏差的多步预估,使控制器具有较好的控制性能。由于该算法无需求解Riccati方程和进行谱因子分解,计算量小,易于实现。
In refinery and chemical plant, PID control algorithm is difficult to meet the production requirements due to random disturbance of some controlled process, dynamic characteristics change with time, large pure delay and so on. Therefore, we developed the LQG self-tuning control software on the upper computer of the PROVOX distributed control system of the atmospheric and vacuum distillation unit using the FORTRAN77 high-level language. The software becomes a universal control module of the distributed control system and implements LQG on multiple control loops on the PROVOX system Self-tuning control. In the control algorithm, we propose a new method to design LQG controller and combine it with recursive least square method to form LQG self-tuning controller. In the algorithm, the controlled process is described by its input-output model, the pure hysteresis of the process is considered in the model and a secondary performance index is established. The performance index includes the multi-step prediction of output and given deviation, The controller has better control performance. Since the algorithm does not need to solve the Riccati equation and perform spectral factorization, the calculation is small and easy to implement.