Practical uncertainty analysis in Life Cycle Assessment

This course aims to strengthen practice of uncertainty assessment in Life Cycle Assessment. The course will provide the background for understanding the basic math of LCA-modeling and how uncertainty propagates from input data to model output; methods to account for correlations between variables are presented. It also introduces the practice of global sensitivity analysis to determine how much each input parameter contributes to the output variance.

The course

Subject

The course aims to strengthen practice of uncertainty assessment in Life Cycle Assessment. The course will provide the background for understanding the basic math of LCA-modeling and how uncertainty propagates from input data to model output. Methods of uncertainty propagation accounting for correlations between variables are presented and applied on a case study. It also introduces the practice of global sensitivity analysis to determine how much (the variance of) each input parameter contributes to the output variance.

Course outline

Lectures and practical sessions using python notebooks:

  • Introduction on concepts and types of uncertainties
  • Mismatch between research question and modeling choices: model uncertainty
  • Background basic mathematical formulation of LCA to show how the uncertainty propagates from input data to model output
  • Analytical method for uncertainty propagation
  • Sampling methods for uncertainty propagation (e.g. Monte Carlo)
  • Including correlations in uncertainty propagation
  • Overview of methods for global sensitivity analyses
  • Handling models corresponding to real-life case studies

Form and academic recognition:

Form: 12 hours lectures, 12 hours practical sessions distributed in 3 full days
Academic recognition: 2 ECTS points. This includes reading a mandatory list of literature.

Learning outcomes:

  • Knowledge on uncertainty types, concepts and methods for uncertainty analyses
  • Understanding of the basic math behind uncertainty propagation in LCA
  • Ability to run uncertainty analyses on an LCA case study through Python notebook
    • Ability to apply analytical method for uncertainty propagation
    • Ability to apply a sampling method for uncertainty propagation
    • Ability to account for correlation between variables in uncertainty analysis
  • Ability to determine how much each input parameter contributes to the output variance (global sensitivity analysis)
  • Ability to communicate on uncertainty results