Scholarlink

Polylactic acid (PLA) is a bio-based alternative to petro-plastics, yet its environmental performance is strongly influenced by the energy sources used during production. Conventional life cycle assessment (LCA) often relies on static assumptions, failing to capture future changes in energy systems. This study applies a machine learning (ML), enhanced dynamic LCA (dLCA) to evaluate the impacts of PLA production from cane sugar in Australia under projected electricity grid decarbonization. An LCA model was developed in OpenLCA using the ReCiPe 2016 v1.03 Midpoint (Hierarchist) method, with inventory data sourced from literature, databases, and process simulations. ML models, linear regression, XGBoost, and Random Forest, were trained on historical data from the Australian Energy Market Operator and International Energy Agency Net Zero scenarios to forecast grid carbon intensity from 2025 to 2050. Results identify lactic acid fermentation and polymerization as the dominant global warming potential (GWP) contributors, together responsible for over 70% of emissions due to high electricity and heat demands. Under a high-renewable scenario, GWP decreases from 406 kg CO2 eq/t PLA in 2025 to −1357 kg CO2 eq/t by 2050, representing a reduction of ~51%. These findings demonstrate that aligning PLA production with renewable electricity and low-carbon heat sources can deliver substantial climate benefits. The integration of ML with dLCA enhances the ability to model future energy transitions, enabling more realistic environmental assessments. Findings highlight the importance of coupling PLA production with renewable electricity and low-carbon heat to achieve long-term emission reduction targets and support the transition toward net-zero manufacturing systems.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Dynamic Life Cycle Assessment of Polylactic Acid in the Context of Energy Transition: A Machine Learning–Enhanced Sustainability Approach

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Md. Monjurul Islam,
  • Nahid Tanjum,
  • Nawshad Haque,
  • Muhammed Bhuiyan,
  • Biplob Kumar Pramanik

摘要

Polylactic acid (PLA) is a bio-based alternative to petro-plastics, yet its environmental performance is strongly influenced by the energy sources used during production. Conventional life cycle assessment (LCA) often relies on static assumptions, failing to capture future changes in energy systems. This study applies a machine learning (ML), enhanced dynamic LCA (dLCA) to evaluate the impacts of PLA production from cane sugar in Australia under projected electricity grid decarbonization. An LCA model was developed in OpenLCA using the ReCiPe 2016 v1.03 Midpoint (Hierarchist) method, with inventory data sourced from literature, databases, and process simulations. ML models, linear regression, XGBoost, and Random Forest, were trained on historical data from the Australian Energy Market Operator and International Energy Agency Net Zero scenarios to forecast grid carbon intensity from 2025 to 2050. Results identify lactic acid fermentation and polymerization as the dominant global warming potential (GWP) contributors, together responsible for over 70% of emissions due to high electricity and heat demands. Under a high-renewable scenario, GWP decreases from 406 kg CO2 eq/t PLA in 2025 to −1357 kg CO2 eq/t by 2050, representing a reduction of ~51%. These findings demonstrate that aligning PLA production with renewable electricity and low-carbon heat sources can deliver substantial climate benefits. The integration of ML with dLCA enhances the ability to model future energy transitions, enabling more realistic environmental assessments. Findings highlight the importance of coupling PLA production with renewable electricity and low-carbon heat to achieve long-term emission reduction targets and support the transition toward net-zero manufacturing systems.