An adaptive neuro fuzzy controller for cement kiln

In this paper a neural- fuzzy controller is used to control cement kiln. The fuzzy controller is in the TSK form. The controller is trained during the control action due to cope with the plant changes. The most important aspects of this controller are first using couple of smaller controllers instead of a complete centralized one and second using the same framework that kiln operators use .i.e. the input variables that the controller use are the same input variables that the kiln operators use to control the same controlled variables. Joint together, decentralized fuzzy controller instead of a centralized fuzzy one has fewer parameters which need less memory and processing power of the controller. The proposed controller is tested on a simulator model which made on the real data of Saveh cement factory. The simulation results show the efficiency of the proposed controller.