Research On Predictive Control Based On Subspace In Gas-collector Pressure Of The Coke Oven

Coke is the main raw material in the metallurgical industry for smelling iron. Large amount of by-product gas will be generated from coke ovens in the coking process. It is a very important link in the steel and iron production, which saves energy and reduces environmental pollution. The stability of gas-collectors' pressure directly affects the quality of gas, the life-time of ovens as well as environment. The gas collecting process of the coke ovens is a complex system with multivariable, non-linear. The traditional control method is difficult to achieve the ideal control effect. Currently, the iron and steel enterprise still use the method combining PID and manual regulation, and this method has low intelligence. At the same time, there exist unstable product quality, waste of resources, environmental pollution and other issues. Therefore, it is significant for the iron and steel enterprise to study on the control of gas collecting process of the coke ovens.This paper first analyses the process of coke oven collecting gas, Then based on the factors of influence on the stability of the coke oven pressure, uses the data-driven subspace identification method for modeling of gas gathering process, and establishes the subspace predictive model of coke oven collector pressure system. This paper designs the subspace predictive controller(SPC) in the framework of generalize predictive controller(GPC). In order to make the system without static error, the paper designs the increment subspace predictive controller. Because the collector pressure system of coke oven is a typical constrained process, so in the premise of improving the control accuracy, this paper designs the incremental subspace constrained processing prediction method. Lastly, because the collector pressure system of coke oven in real industrial processes are nonlinear and time-varying, so use the online subspace based on receding window to reduce the model mismatch, design the online subspace predictive controller based on receding window.The results of simulation show that the gas-collector pressure can be controlled in a certain region and the effect of the control is good.