Prediction Model Based on Neural Networks to Optimize Thermal Efficiency and Emission Control in Firetube Boilers

This study addresses the critical challenge of enhancing thermal efficiency in industrial firetube boilers within the fishing industry, a sector burdened by significant fuel consumption and associated costs amidst rising fuel prices; achieving even marginal improvements in boiler efficiency can result in substantial economic savings and environmental benefits. Utilizing the Peruvian technical standard for efficiency determination, alongside recommendations from boiler manufacturers and operational conditions, this research employs artificial neural networks (ANNs) to model and predict efficiency outcomes based on various operational parameters, including fuel type and combustion conditions specifically, the study explores the impact of excess air and fuel regulation on thermal efficiency and pollutant emissions, employing applied research methods and a comprehensive analysis of boiler operation at 80% and 100% load conditions. Results demonstrate the capability of neural network models to accurately predict thermal efficiency, with optimized configurations achieving significant reductions in CO2 and CO emissions by 43% and 55%, respectively. The findings underscore the potential for neural network applications in optimizing boiler operations, offering a pathway to economic and environmental improvements in industrial processes. The study concludes with optimal operational parameters that balance efficiency gains with emission reductions, highlighting the practical implications for the fishing industry and beyond.