Real-Time Hybrid Simulation Framework for Dynamic Industrial Process Optimization

Abstract

Industrial process optimization has become increasingly important in Industry 4.0 due to the growing demand for higher productivity, improved product quality, reduced energy consumption, and adaptive decision-making under dynamic operating conditions. However, conventional optimization methods often fail to respond effectively to process disturbances and rapidly changing environments. This study proposes a Real-Time Hybrid Simulation Framework for Dynamic Industrial Process Optimization that integrates physics-based modeling, data-driven prediction, real-time Industrial Internet of Things (IIoT) data synchronization, and intelligent optimization within a unified closed-loop architecture. The proposed framework was developed through five stages: system requirement analysis, hybrid model development, real-time data integration, optimization module implementation, and performance validation under dynamic disturbance scenarios. Experimental evaluation demonstrated that the proposed framework consistently outperformed conventional PID control, physics-based optimization, and standalone data-driven approaches. The framework achieved the lowest tracking error (ITAE = 5.47), the highest prediction accuracy (R² = 0.97), and the lowest prediction errors (MAE = 1.76 and RMSE = 2.71). Furthermore, it reduced energy consumption by 50.4%, increased production throughput by 70.7%, improved product quality by 18.3%, enhanced response time by 64.2%, and increased operational stability by 45.6% compared with conventional PID control. These results demonstrate that the proposed framework effectively improves prediction accuracy, process stability, and real-time optimization performance in dynamic industrial environments. The novelty of this research lies in the seamless integration of physics-based and data-driven models with continuous real-time synchronization and multi-objective optimization, providing a scalable and adaptive solution for intelligent digital twins and next-generation smart manufacturing systems.