Exploring the implementation feasibility of the sol-char sanitation system using machine learning and life cycle assessment

Abstract

Globally, 1.5 billion people still lacked access to safe sanitation facilities in 2022, which exacerbated health risks and environmental degradation. To address this, we created the Sol-Char sanitation system, a potential solution for expanding secure sanitation alternatives. This study aimed to develop a machine learning model that could evaluate the viability of implementing the Sol-Char system in 76 countries with high rates of open defecation in 2022. Using the Random Forest model, we identified suitable locations considering factors such as solar energy availability and economic feasibility. The model successfully identified 42 countries (55 %), mainly in Sub-Saharan Africa and South Asia, as appropriate candidates for implementing the system. In addition, a framework was developed to guide solar technology suitability prediction using our machine learning model. Furthermore, we conducted an ex-ante life cycle assessment (LCA) study to evaluate the environmental impacts across different implementation scenarios. The baseline scenario (Scenario 1) produced the least emissions, with 299 kg CO2-eq. In contrast, the scenario (Scenario 2) involving international transportation had the highest emissions at 395 kg CO2-eq (32 % higher), while the localized scenario (Scenario 3) landed in between with 337 kg CO2-eq emissions. The LCA and contribution analysis highlighted that optimizing materials and design was essential to reduce emissions across these scenarios. Local manufacturing, particularly in high-transportation scenarios like Scenario 2, could reduce emissions from logistics but required careful consideration of local resources and energy structures, as demonstrated in Scenario 3.