Estudo da aplicação de um soft sensor de finura de cimento para otimizar a troca de cimento no processo de moagem e separação em uma indústria cimenteira

Abstract

With the rapid advancement of artificial intelligence today, it is increasingly necessary for industries to invest in technologies that contribute to the automation of their processes. In the cement industry, based on the need for continuous performance improvement and the search for operational excellence, the implementation of techniques that are able to analyze the vast volume of data that is generated during the production process not only contributes to the final quality of the cement, but also generates significantly positive results, the biggest one being the reduction of costs with losses. The cement manufacturing process presents a great opportunity for modernization, especially the grinding process. As it is one of the steps that retains most of the expenses, the implementation of a soft sensor demonstrates essential gains such as an increase in productivity and cement quality, assertiveness in the analyzes and reduction in the specific consumption of electricity. The objective of this work is to evaluate the application of a soft cement fineness sensor to optimize the cement exchange in the grinding and separation process, comparing the application and viability of two Machine Learning models (Decision Tree and Neural Networks). Through the analysis of absolute errors, standard deviations, and how much the values were within the defined tolerance of each model, a more assertive behavior of the Neural Networks model was identified, with values 83.51% within the tolerance, standard deviation of 0.021 and error of 2.13% compared to the results of blaine obtained in the laboratory. In view of this, the model was put under further analysis, resulting in values such as R² of 74% and MAPE of 2.75%, consisting of promising results for the implementation of the sensor in a cement industry, since the main objective of the process is to guarantee the quality of the cement.