ENHANCING REAL-TIME OPTIMIZATION WITH MACHINE LEARNING PREDICTIONS OF LOWER CALORIFIC VALUES

This paper presents an enhancement to a Real-Time Optimization (RTO) strategy for industrial furnaces by integrating machine learning tech-niques. The developed methodology utilizes historical process data to build pre-dictive models that estimate the LCV, enabling more effective optimization of the fuel gas mixture, composed of different fuel gases. Implementing this strategy resulted in a significant reduction in the production cost without compromising the quality of the heating process. The results demonstrate the potential of integrating machine learning models into Real-Time Optimization systems, contributing to cost reduction and increased sustainability in the industry.