Intelligent Steam Consumption Control in Tire Vulcanization Processes Using ANFIS-GA Digital Twin Modeling and MPC

Steam consumption is a critical energy-related variable in tire vulcanization processes, directly influencing production efficiency, operating costs, and process stability. This study proposes an intelligent control framework that integrates an Adaptive Neuro-Fuzzy Inference System optimized by a Genetic Algorithm (ANFIS-GA) with Model Predictive Control (MPC) for steam consumption optimization in industrial tire vulcanization systems. A data-driven digital twin of the vulcanization process was developed using an industrial dataset consisting of 200 operating samples collected under practical production conditions. Vulcanization temperature and process coefficient (α), identified as the dominant factors affecting steam demand in the investigated production line, were selected as model inputs, while steam consumption was considered as the output variable. The ANFIS-GA model was employed as the internal prediction model within the MPC framework to represent the nonlinear dynamics of the process. The developed digital twin achieved a coefficient of determination (R²) of 0.886 on the test dataset, indicating satisfactory predictive capability for closed-loop control applications. Although this accuracy level is moderate for standalone prediction tasks, the receding horizon feedback mechanism of MPC effectively reduces the impact of model uncertainty during online operation. The proposed MPC strategy was evaluated under varying operating conditions and compared with a conventional PID controller. The results demonstrated substantial improvements in control performance, reducing RMSE from 0.1073 kg/h to 0.0387 kg/h and MAE from 0.0893 kg/h to 0.0259 kg/h. In addition, IAE and ISE were reduced by approximately 71% and 87%, respectively, while the settling time decreased from 80 sampling steps to only 2 sampling steps. These findings demonstrate that the integration of ANFIS-GA digital twin modeling with MPC provides a practical and effective solution for energy optimization and intelligent control in industrial tire vulcanization processes.