Related Experiment Video
Updated: Jun 20, 2026
09:39
Physical, Chemical and Biological Characterization of Six Biochars Produced for the Remediation of Contaminated Sites
Published on: November 28, 2014
35.1K
AI-based tree modeling for multi-point dioxin concentrations in municipal solid waste incineration.
Heng Xia1, Jian Tang1, Loai Aljerf2
1Faculty of Information Technology, Beijing University of Technology, Beijing 100124, China; Beijing Laboratory of Smart Environmental Protection, Beijing 100124, China.
Journal of Hazardous Materials
|September 21, 2024
Summary
This study introduces a novel whole-process machine learning model for measuring dioxin (DXN) emissions from municipal solid waste incineration (MSWI). The model accurately predicts DXN generation, adsorption, and emission, enabling better pollution control.
Area of Science:
- Environmental Science
- Chemical Engineering
- Artificial Intelligence
Background:
- Municipal solid waste incineration (MSWI) is a significant source of dioxin (DXN) emissions.
- Current DXN emission control strategies are hindered by the lack of real-time, whole-process online measurement capabilities.
- Existing models primarily focus on stack emissions, neglecting DXN generation and absorption dynamics.
Purpose of the Study:
- To develop a comprehensive, data-driven model for online measurement of DXN concentrations throughout the entire MSWI process.
- To establish a novel framework for DXN modeling that integrates generation, adsorption, and emission phases.
- To support optimal pollution reduction control through a mechanistic understanding of DXN behavior.
Main Methods:
- Application of advanced tree-based machine learning algorithms, including deep and broad learning, adaptive deep forest regression, and fuzzy forest regression.
- Analysis of diverse data characteristics (high-dimensional small samples, low-dimensional ultra-small size samples, medium-dimensional small samples) across different DXN phases.
- Utilizing a deep understanding of the DXN mechanism to inform data characteristic determination and model construction.
Main Results:
- A robust, whole-process tree-based model for DXN was developed and validated using nearly one year of authentic data from an MSWI plant in Beijing.
- The model demonstrated effectiveness in handling various data complexities inherent in DXN generation, adsorption, and emission phases.
- The proposed framework provides a novel approach to DXN modeling, moving beyond stack-focused analysis.
Conclusions:
- The developed whole-process model offers a significant advancement for online DXN monitoring and control in MSWI.
- This approach facilitates a deeper exploration of DXN mechanism characterization.
- The framework provides essential support for optimizing pollution reduction strategies in waste incineration.
