The Process Control System Design Of Rotary Kiln Based On Hierarchical Control Strategy

Rotary kiln occupies an important position in cement, limestone and some non-ferrous metal productions, and it is used to do heat treatments, for example, roast, chemical raw materials, ores and intermediate products. China uses the most rotaries in the world. Calcination process temperature is one of the most important parameters in rotary kiln process control. Improving the accuracy of rotary kiln calcination temperature control is very important for product quality improving, energy saving and costs reducing.The calcination process of rotary kiln is a complicated course and has characteristics as a lot of inertia, pure time-delay and nonlinearity. The accurate temperature model of rotary kiln is not easy to set up. The process of rotary is featured by multi-scales, such as flame temperature and gas flow with faster and more frequent changes, while the input quantity is relatively stable at different time scales. In this article, a multi-scaled method of factor analysis, combined with the practical production data, is used to divide the rotary kiln system into a few independent loops of multi-level. Meantime, in this article, the method of system identification is applied to identify a number of transfer functions' models, according to a large number of historical data, and the model is applied to the simulation study.The process control of rotary kiln is complicated and changeable, and the control effect is not good with traditional PID controller. The fuzzy controller is designed based on integrated expertise knowledge, not requiring precise mathematical models of control objects. In the fuzzy control system, when the fuzzy approach and fuzzy decision-making method were selected, the appropriate selected fuzzy control rules can achieve good control effect well. In this paper, the idea of "variable domain" is drawn. Under the premises of the same fuzzy rules, the domain expands or shrinks along with the error changes, and the domain migrates. In this paper, a multi-level complex control system is designed using the method of fuzzy control. A large number of simulations has shown that the controller made the multi variables controlled at different time-scale. When major changes of parameters happen to the control objects, control accuracy will be more improved to some extent than the conventional controllers, by adjusting the scope of domain to achieve stable control.