A data-driven framework to inform sustainable management of animal manure in rural agricultural regions using emerging resource recovery technologies

Abstract

Thermochemical conversion technologies are emerging as preferred resource recovery practices for managing animal manure in agricultural regions. Although the implementation of such technologies has been previously studied, difficulties exist in maintaining balance between high rate of resource recovery and low environmental, economic, and social impacts, particularly in rural regions with limited resources. We developed a data-driven framework by integrating machine learning with life cycle thinking that can be used as an open-source tool to help overcome these barriers. The framework was applied to compare two emerging technologies: pyrolysis versus hydrothermal carbonization for managing the excess poultry litter in a rural agricultural region. Among different machine learning models, random forest regression was the most successful to predict resource recovery of both technologies. Next, sustainability analysis indicated that the environmental (global warming), economic (annual worth), and social (system intrusiveness) impacts of pyrolysis was lower than hydrothermal carbonization. Finally, the framework revealed that implementation of pyrolysis at 600 °C for 1 h with the heating rate of 20 °C/min would result in the highest rate of resource recovery that corresponded to the lowest impacts. These results can be helpful in providing operational conditions for implementing emerging resource recovery technologies in rural agricultural regions.

Keywords

• Machine learning
• Life cycle assessment
• Techno-economic analysis
• Pyrolysis
• Hydrothermal carbonization
• Carbon dioxide removal