Unveiling Hydrogen Fluoride Emission Mechanisms in Municipal Solid Waste Incineration Using a Machine Learning...

Area of Science:

• Environmental Science
• Chemical Engineering
• Data Science

Background:

• Hydrogen fluoride (HF) emissions from municipal solid waste incineration (MSWI) present significant environmental and health concerns.
• The complex mechanisms governing HF formation during MSWI are not fully understood, hindering effective pollution control.

Purpose of the Study:

• To develop an integrated machine learning framework for predicting and understanding HF formation mechanisms in MSWI.
• To identify critical operational parameters and temperature thresholds for mitigating HF emissions.

Main Methods:

• An integrated framework combining XGBoost, SHAP, Structural Equation Modeling (SEM), Generalized Additive Models (GAMs), and SA-NSGA-II was employed.
• High-frequency sensor data from a waste-to-energy plant was analyzed using 5-fold cross-validation.
• Feature importance, mechanistic pathways, critical thresholds, and optimization strategies were investigated.

Main Results:

• XGBoost achieved high prediction accuracy (R² = 0.755), with flue gas temperatures identified as dominant factors by SHAP analysis.
• SEM confirmed the grate incineration zone as a primary HF source.
• GAM identified specific temperature thresholds for HF emission control, and SA-NSGA-II optimization reduced HF emissions by an average of 17.61%.

Conclusions:

• The study provides a novel, data-driven framework for mechanistic understanding of HF formation during MSWI.
• Identified key factors and thresholds offer practical strategies for sustainable MSWI operations and pollution mitigation.