Prediction of melting characteristics of sludge coal blending based on Multi-task Learning

Abstract In response to the urgent need for sustainable energy solutions, this study pioneers the use of a Multi-task Learning (MTL) framework to predict the melting characteristics of sludge-coal mixtures, a key aspect in optimizing co-firing processes. The integration of sludge, a byproduct of wastewater treatment, with coal in co-firing processes, presents a promising avenue to enhance energy efficiency while reducing greenhouse gas emissions. Addressing the complexities of predicting combustion behavior due to variable chemical compositions and temperatures, the MTL approach enables learning from diverse datasets, ensuring accurate and generalizable predictions. Outperforming conventional machine learning models, our framework signifies a breakthrough in predictive accuracy. Despite facing challenges such as dataset limitations, the research underscores the framework’s potential for broad industrial application, aiming for future improvements that include dataset expansion and real-time data integration. This study not only advances machine learning applications in energy and environmental domains but also contributes to the global pursuit of sustainability.