Energy Planning Toward Absolute Environmental Sustainability: Key Decisions and Actionable Insights Through Interpretable Machine Learning

Ghuys, Nicolas;Coppitters, Diederik;Van den Oever, Anne;Messagie, Maarten;Jeanmart, Hervé;et.al.
(2025) ESCAPE 35 - European Symposium on Computer Aided Process Engineering — Location: Ghent, Belgium (6.July.2025)

Files

Coppitters_paper_escape35.pdf
  • Open Access
  • Adobe PDF
  • 111.95 KB
Download

Details

Authors
Show more
Abstract
Energy planning models traditionally support the energy transition by focusing on cost-optimized solutions that limit greenhouse gas emissions. However, this narrow focus risks burden-shifting, where reducing emissions increases other environmental pressures, such as freshwater use, solv-ing one problem while creating others. Therefore, we integrated Planetary Boundary-based Life Cycle Assessment (PB-LCA) into energy planning to identify solutions that respect absolute envi-ronmental sustainability limits. However, integrating PB-LCA into energy planning introduces chal-lenges, such as adopting distributive justice principles, interpreting trade-offs across PB indicator impacts, and managing subjective weighting in the objective function. To address these, we em-ployed weight screening and interpretable machine learning to extract key decisions and action-able insights from the numerous quantitative solutions generated. Preliminary results for a single weighting scenario show that the transition scenario exceeds several PB thresholds, particularly for ecosystem quality and mineral resource depletion, underscoring the need for a balanced weighting scheme. Next, we will apply screening and machine learning to pinpoint key decisions and provide actionable insights for achieving absolute environmental sustainability.
Affiliations

Citations

  • APA
  • Chicago
  • FWB

Ghuys, N., Coppitters, D., Van den Oever, A., Messagie, M., Contino, F., & Jeanmart, H. (2025). Energy Planning Toward Absolute Environmental Sustainability: Key Decisions and Actionable Insights Through Interpretable Machine Learning. ESCAPE 35 - BOOK OF SHORT PAPERS, 35(1), 294. (Original work published 2025)