Uncertainty analysis based on probability bounds (p-box) approach in probabilistic safety assessment

Area of Science:

• Engineering Safety Assessment
• Risk Analysis
• Uncertainty Quantification

Background:

• Safety assessments of engineering systems inevitably involve uncertainties.
• Existing methods like Monte Carlo simulation have limitations when precise probability distributions or parameter dependencies are unknown.
• Addressing these limitations is crucial for reliable decision-making in safety assessments.

Purpose of the Study:

• To present an uncertainty analysis method for level-1 probabilistic safety assessment (PSA) that overcomes limitations of traditional probabilistic approaches.
• To introduce and validate the probability bounds (PB) approach for propagating uncertainties in safety assessments.
• To compare the PB approach with existing methods like two-phase Monte Carlo simulation.

Main Methods:

• The study employs a probability bounds (PB) approach, integrating probability theory and interval arithmetic.
• This approach generates probability boxes (p-boxes) to represent and propagate uncertainties.
• A practical case study was conducted using developed code based on the PB approach.

Main Results:

• The PB approach provides a rigorous framework for comprehensive uncertainty propagation.
• Probability boxes (p-boxes) effectively capture and manage uncertainties when precise probabilistic information is unavailable.
• The developed PB approach demonstrated its utility in a practical safety assessment case study.

Conclusions:

• The probability bounds (PB) approach offers a valuable alternative for uncertainty analysis in safety assessments, particularly when dealing with imprecise or incomplete data.
• PB analysis enhances the reliability of safety assessments by providing a more comprehensive understanding of uncertainty impacts.
• This method supports improved decision-making in engineering safety by offering a robust way to handle complex uncertainties.