| |
In modern scientific discourse, special attention is paid to the development and improvement of technological processes using innovative information technologies. Intelligent systems, in particular artificial neural networks (ANNs), are a promising tool for analyzing and optimizing complex production mechanisms. This dissertation work is devoted to the development and testing of neural network models for improving grinding processes in drum mills, which is important for effective energy consumption management and improving the quality of final products.
Materials and methods. To achieve these goals, an integrated approach was chosen, including the collection and analysis of experimental data, the formation of a knowledge base, the development of structural parametric and simulation models, as well as the intellectualization of the management process. The research is based on data obtained during experimental work at production mills. To train the INS, data on the operating parameters of the equipment, the physical and mechanical properties of the processed materials and the characteristics of the final product were used. Validation of the model was carried out by comparing the results obtained using the INS with data from independent experiments.
Results. The results indicate a significant increase in the efficiency of grinding processes due to the use of developed neural network models. It was found that INS contribute to optimizing energy consumption and improving product quality. The INS models demonstrate high accuracy in predicting changes in grinding quality, which allows you to quickly adjust the process parameters in real time. Trained neural networks are able to adapt to changes in input parameters, which ensures their applicability in conditions of production variations. An important result is also the development of software for the integration of the INS into the automatic process control system.
Keywords:grinding, quality control, neural network technologies, granulometric composition, artificial neural networks, predictive analysis, efficiency assessment
|
LEGAL
INFORMATION:
