NEURONS PROCESS CONTROL SYSTEM CLINKER BURNING IN ROTARY KILNS

Rotary kiln clinkers are widely used in many industrial plants for the production of cement. Clinker manufacturing pro-cess requires the large consumption of energy. Getting the quality of final product is possible by controlling and regulat-ing the set of process parameters. It largely depends on the professional level of the operator and possible unmistakable solutions. Many industrial plants in which operation are the rotary kilns are still using outdated equipment that does not match to modern requirements of process control clinker. The question of improving the quality indicators is still ex-tremely important. To improve the process of clinker burning is possible through the using of control systems with fuzzy logic controllers or modern neyrosystem control. Analysis of the literature. Currently, there is a lot of literature and scientific works in the field of automated control systems of the clinker burning process. Significant contribution to the development of control systems and intensify the process of firing have made by such scholars: V. K. Сlass [4], P. V. Besedin [1], M. A. Verdiyan [8]. In the works of these scholars are presented methods for creating control systems clinker, improving quality indicators in relation to one of the options or providing use of fuzzy logic control and neural network with one control circuit for the burning process. The analysis of the literature demonstrates that the neural con-trol systems are the promising way to improving the clinker burning process and require further development. The pur-pose and objectives. The aim of the study is to find opportunities to improve the quality of the firing of the finished material (clinker) while reducing power consumption, improving the work conditions of the operator at the plant, reduc-ing the total time of the clinker burning through the creation of automated control system with neural controller, based on the reference model and its immediate setting. Conclusions. 1. In Matlab Simulink environment management sys-tem was developed clinker burning process in the rotary kiln with neyrocontroller based reference model Model Refer-ence Controller and simulated control system. 2. Two additional process models are developed: general and reference that are used during the training phase neyrocontroller. 3. On the basis of the constructed models were trained neural network identifier and the neyrocontroller. Experimentally were determined the necessary parameters of the neyrocon-troller and the neural network. 4. During training neural network were obtained graphs showing the different stages of neural network learning and neyrocontroller based on the standard model, including graphic errors during training, the graph corresponds to the input parameters to the source, several stages of training schedules neyrocontroller and others. 5. Analysis of the graphs of the learning process and control neural network system demonstrates that the developed system matches all the quality indicators modeling is able to perform automated control of the clinker burning process in the rotary kiln with high precision and to be recommended for use in the design of such systems.