EUBCE 2026 - Somrat KERDSUWAN - A Machine Learning Approach for Predicting the Higher Heating Value of Municipal Solid Waste Based on its Physical Composition

A Machine Learning Approach for Predicting the Higher Heating Value of Municipal Solid Waste Based on its Physical Composition

Short Introductive summary

Municipal solid waste (MSW) is considered a high-potential bioenergy source generated daily by human activity. Incineration technology is widely viewed as a practical pathway for converting MSW into electricity. The quality of MSW, in terms of heating value, plays, of course, a significant role in electricity production potential. Since MSW comprises a diversity of material from high energy-containing materials like plastic, paper, and textile, to low energy-containing materials, e.g., food waste with high moisture content, the heating value, which is a key parameter for the design of a waste-to-energy plant, differs depending on geographic location, socio-economic factors, waste separation method, etc. Modern studies have explored alternative approaches to predict HHV in MSW using Machine Learning (ML); however, most studies still rely on either ultimate or proximate analyses. This study focuses on using a Machine Learning (ML) approach to predict the HHV of MSW from its physical composition, enabling cost-effective implementation at a landfill site. The effects of data quality and quantity were studied to optimize the ML model.

S. Kerdsuwan, King Mongkut's University of Technology North Bangkok, Bangsue, THAILAND
K. Laohalidanond, King Mongkut's University of Technology North Bangkok, Bangsue, THAILAND
K. Tantiwattanakul, King Mongkut's University of Technology North Bangkok, Bangsue, THAILAND