Optimization of manipulated cement mill variables using AI models

Reducing energy consumption without sacrificing the quality of the products is a modern research topic. The cement industry is considered as an industrial sector of high pollution. The operators of a cement plant can only influence a few parameters of the system named manipulated variables. In this work Artificial Intelligence models which predict the non-manipulated variables are incorporated into an optimization process using the Differential Evolution method in order to extract optimal values for the manipulated variables in the TITAN cement plant in Greece. The optimization depends on the mill feed, the cement particles, the mill and the separator energy consumption and the mill vibrations. The results have shown that the system can be switched to an optimized state of low energy consumption while respecting the parameters set on the cement particles.