Research On Predictive Control For Gas Collector Pressure Of The Coke Oven Based On Subspace

Coke is the main material in the metallurgical industry for smelling iron. In the coking process, large amount of by-product gas will be generated. The recycling of the gas is very important for the iron and steel production, which not only saves energy but also reduces environmental pollution. Whether the gas collector pressure is stable or not directly affects the quality of gas, the life-time of ovens as well as the producing environment. Therefore, studying on the control of gas collecting process of coke ovens is of important significance.The gas collecting process of coke ovens is a highly multi-variable complex industrial process with disturbances and coupling, so it is hard to acquire good control results using traditional modeling methods and classical control theory. Based on the analysis of physical properties and process characteristics of the system of gas collector pressure, the closed-loop subspace identification method for modeling of gas collecting process is used, and the subspace predictive model of gas collector pressure system is established. This thesis designs the subspace predictive controller in the framework of generalize predictive controller. As the system of gas collector pressure is a typical constrained process, to improve the control accuracy, this thesis designs the incremental subspace prediction method with constrained processing. The result of simulation shows the gas collector pressure can be controlled in a certain region and the effect is good.Then the thesis presents the definition of Laguerre functions. In order to reduce the number of variables that are being solved in the optimization procedure, the Laguerre functions that involve the parameterization of the input signal is presented. The benefits of this parameterization can be seen from the reduction of computational time, without sacrificing the accuracy. At last an analysis on the effects of Laguerre parameters on the control performance is presented. The results of simulation show that the subspace model predictive control algorithm with Laguerre can effectively control the gas collector pressure, which proves the effectiveness of the algorithm.