Unveiling and interpreting the relationships among multi-pollutant emission factors in municipal solid waste incineration by machine learning.
Journal:
Waste management (New York, N.Y.)
Published Date:
Nov 25, 2025
Abstract
Effective control of key parameters is critical for regulating pollutant emissions in municipal solid waste incineration (MSWI), but existing research on these parameters remains limited and lacks comprehensiveness. This study used over 600,000 industrial data records (June 9-15, 2024) from 140 sensors-covering the emission concentrations of four pollutants (HCl, SO2, NOx, CO)-employed the AntDE-DTFS algorithm for feature selection, and constructed eight machine learning models; among these, XGBoost achieved the best predictive performance, with R2 values of 0.92 (HCl), 0.87 (SO2), 0.89 (NOx), and 0.75 (CO). Results showed that temperature is the predominant factor for HCl and CO emissions: maintaining the furnace flue gas treatment outlet temperature below 140 °C effectively reduces HCl output, while optimizing the economizer inlet temperature control valve position feedback to exceed 11 % significantly mitigates CO emissions. In contrast, SO2 emissions are primarily governed by pressure, with steam drum pressure maintained below 6.8 MPa minimizing SO2 emissions, and NOx emissions are affected by both temperature and pressure-with the temperature at the upper-left of the second flue gas duct as the key factor, and elevating this temperature above 780 °C contributing to lower NOx concentrations. These findings provide valuable insights for emission control strategies, offering a scientific basis to optimize incineration processes and enhance environmental management.
