焦炉集气管压力的模糊预测控制

Fuzzy predictive control of coke oven gas collector pressure

在炼焦厂中焦炉集气管压力系统是一个难以用传统控制方法控制的多变量负耦合系统。考虑到焦炉集气管压力控制过程中的各种干扰影响因素和控制对象具有时变和非线性等特点。为了使焦炉集气管的压力稳定在一个恒定值,目前常用4个PID控制器分别控制每一个集气管压力,这种方法造成集气管压力系统不稳定和能源浪费。通过对集气过程的输入输出数据进行系统模型的参数辨识得出集气管压力系统模型。在数学模型的基础上分别采用PID控制、预测控制和模糊预测控制三种控制策略,通过Matlab/Simulink仿真验证这三种算法的控制效果。采用PID控制器时,系统不仅有较大的超调,还会产生震荡造成系统不稳定;采用模型预测控制器时,系统能够快速地达到稳态,当焦炉集气管压力系统模型参数改变时,由于模型不匹配造成系统超调,甚至不稳定;使用模糊预测控制器时,系统不仅控制效果较好,而且能克服模型参数改变,在多种工业情况下均能达到控制的要求。模糊预测控制器相较于单纯的模型预测控制能够对系统模型参数改变有更好的控制效果,鲁棒性更强,而且相对于传统的PID控制器更加适用于多变量系统。

In the coking plant,the coke oven gas collector pressure system is a multi⁃variable negative coupling system which is difficult to control with the traditional control methods.The various interference factors in the process of pressure control of the coke oven gas collector and the fact that the control object has the characteristics of time⁃varying and non⁃linearity are taken into account.The purpose is to stabilize the pressure of the coke oven gas collector at a constant value.Usually,four PID controllers are used to separately control the pressure of each gas collector at present,which causes instability of the gas collector pressure system and waste of energy.By the input and output data parameter identification of the system model in the gas gathering process,the model of the gas collector pressure system is obtained.Three control schemes named PID control,predictive control and fuzzy predictive control are adopted on the mathematical model.The control effect of the three algorithms is verified by Matlab/Simulink simulation experiments.When the PID controller is used,the system not only has a large overshoot,but also generates oscillations and causes the system instability.When the model predictive controller is used,the system can reach a steady state rapidly.When the model parameters of the coke oven gas collector pressure system are changed,the system will overshoot or even be unstable due to model mismatch.When the fuzzy predictive controller is used,the system not only has a better control effect,but also can avoid the change of model parameters,so it can meet the control requirements under a variety of industrial conditions.In comparison with the single model predictive control,the fuzzy predictive controller has a better control effect on the change of system model parameters,it is more robust,and it is more suitable for multivariable system in comparison with the traditional PID controllers.