MODEL ADAPTIV NEURO-FUZZY PENTRU SISTEMUL DE CONTROL AL PROCESULUI DE MACINARE A CLINKERULUI IN MORILE CU BILE DIN FABRICILE DE CIMENT

Abstract

The main purpose of this study consisted in the realization of the development model of a decision support system for the cement grinding process in ball mills, including the acquisition, processing and analysis subsystems of data regarding the progress of the technological process, based on hardware technologies and intelligent software. To create the decision-making models, the graphic environment for the development of adaptive neuro-fuzzy systems from the Matlab program package was used. The paper presents a model based on the techniques proposed and developed with the application of fuzzy logic and artificial neural networks. The input/output variables, the linguistic qualifiers and the membership functions specific to each shredding batch were identified. The inference rules were defined. A standard defuzzification method was applied to defuzzify the results obtained from the inference process. At the same time, the results of the simulations of the proposed models in the Matlab environment were also presented. The testing and verification of the data obtained with the proposed inference model was carried out by comparison with the experimental data.