<![CDATA[Background: Vertical Roller Mill (VRM) is the newest type of equipment in the cement milling process, which consists of grinding, drying and separation processes that have high energy efficiency. Objective: This research was conducted to create and develop a Model Predictive Control (MPC) Random Forest Regressor (RFR) in a process system that aims to improve the performance of the cement grinding process, where currently process control is still carried out using a conventional control system by humans/operators. Methods: Model creation is carried out by preparing input variable data, manipulated and output variables, data conditioning, statistical analysis, model development, validation, testing, and evaluation. Results: The MPC-RFR model achieved R²=0.99936, MAE=2.488, MSE=122.354, with SEC reduced from 35.47 to 29.46 kWh/ton (16.94% reduction) using MPC-RFR, and further to 27.47 kWh/ton (22.55% reduction) with SLSQP optimization, yielding potential annual savings of IDR 8.6–11.5 billion. Conclusion: The MPC-RFR-SLSQP approach achieved 22.55% SEC reduction in VRM cement grinding, demonstrating significant potential for industrial energy efficiency and production cost optimization in the cement sector.]]>
Copyrights © 2026
