Waste to energy spatial suitability analysis using hybrid multi-criteria machine learning approach

Area of Science:

• Environmental Science and Engineering
• Artificial Intelligence in Geospatial Analysis

Background:

• Developing countries often rely on landfills/dumpsites for municipal solid waste (MSW).
• Waste-to-Energy (WTE) systems offer sustainable MSW management and energy diversification.
• Geospatial data and AI present opportunities for effective WTE site selection.

Purpose of the Study:

• To develop and evaluate an AI-supported framework for optimal Waste-to-Energy (WTE) facility siting.
• To identify critical geospatial and environmental parameters influencing WTE facility location.
• To assess the suitability of areas for WTE facilities using machine learning and AHP.

Main Methods:

• An Analytical Hierarchy Process (AHP)-based framework integrated with machine learning algorithms (Gradient Boosted Tree, Decision Tree, Support Vector Machines).
• Consideration of social, legal, environmental, economic, morphological, and land cover parameters across 11 geospatial raster layers.
• Incorporation of Gaussian dispersion modeling for air pollution assessment.

Main Results:

• Gradient Boosted Tree (GBT) achieved the highest accuracy (94.6%) in site suitability prediction.
• Distance from existing landfills, coastline, and elevation were the most critical siting factors.
• 16.6% of Sharjah, UAE, was identified as extremely highly suitable for WTE facilities.

Conclusions:

• The proposed AI-driven framework effectively identifies optimal locations for WTE facilities.
• Proximity to existing landfills is a primary determinant for WTE site selection.
• The research provides valuable insights for developing guidelines for sustainable WTE facility development.