ABSTRACT
The road pavement network is vital to national infrastructure and economic functionality. However, the paving industry faces increasing demands to reduce greenhouse gas (GHG) emissions and energy consumption, driven by climate change mitigation efforts and policy targets. To address these challenges, the industry has adopted secondary materials and low-temperature production technologies. While numerous life cycle assessment (LCA) studies have evaluated the environmental benefits of such approaches, most rely on secondary data or simplified models that fail to capture the complexities of real-world production environments. This study utilizes a year-long dataset from an instrumented asphalt plant and applies machine learning algorithms to predict energy consumption during production operations. It evaluates algorithm performance and uses SHapley Additive exPlanations (SHAP) to rank variable importance and analyze their impacts. The findings provide actionable insights for researchers, industry practitioners, and manufacturers, supporting the development of more sustainable asphalt production practices.
