Session: 07-03-01 Risk-Informed Decision-Making

Paper Number: 132344

132344 - Dealing With Epistemic Uncertainties in Reliability-Based Pipeline Integrity Management 

Abstract: 

One of the key characteristics of reliability-based integrity management and risk-informed decision making is that they provide a consistent framework for explicit consideration of various uncertainties in the modelling and decision making process. 

From a practical application point view, it is useful to classify uncertainties into two categories, the aleatory uncertainties and the epistemic uncertainties. The aleatory uncertainties are uncertainties that are considered "intrinsic" to system, and thus cannot be further reduced without major advances in the state-of-the-art understanding or in the time horizon that is relevant. Epistemic uncertainties, sometimes called lack of knowledge uncertainties, on the other hand are uncertainties that can potentially be further reduced by ways such as data refinement and model improvements that can be done in near term.

The classification of aleatory and epistemic uncertainty has two key implications: (1) the predicted reliability performance and risk level from reliability-based methods are typically the results from a combination of both aleatory and epistemic uncertainties, and can be different from the actual reliability performance of a system due to additional uncertainties from the lack of complete and perfect knowledge; (2) the modelling and decision-making process thus naturally involves both understanding and characterizing uncertainties, as well as evaluating options for further reduction of epistemic uncertainties so more refined risk and reliability results can be used. 

This paper discusses two commonly used approaches for dealing with epistemic uncertainties in reliability-based integrity management and risk-informed decision making. The first approach involves performing reliabiltiy-based analyses first without concerning the types of uncertainties. The differentiation of aleatory and epistemic uncertainties is subsequently accounted for in a sensitivity analysis that evaluates the cost and benefits associated with various uncertainty reduction options. This approach provides a comprehensive decision making process for dealing with epistemic uncertainties but it could require significant more modelling and analysis effort. The second approach is to have an explicit separation of epistemic and aleatory uncertainties from the onset to produce a set of reliability and risk results corresponding to different level of epistemic uncertainties. This approach provides an quick overview of the sensitivity of the results to epistemic uncertainties. However, in some cases, the separation of epistemic and aleatory uncertainties can be challenging due to the subjectivity that is involved and potential correlations between epistemic and aleatory uncertainties. Examples of pipeline integrity and risk management are presented in the paper to illustrate the two approaches and their advantages and disadvantages are demonstrated and discussed.

Presenting Author: Dongliang Lu TC Energy

Presenting Author Biography: Dongliang Lu is a senior pipeline integrity and risk engineer with Asset Management, Integrity & Reliability team at TC Energy. His area of expertise is risk and reliability analysis of engineering systems, structural integrity analysis, in-line inspection tool performance evaluation, mathematical modelling and optimization, statistical data analysis and data visualization.

Authors:

Dongliang Lu TC Energy
Peng Gong tc energy
Aleksandar Tomic TC Energy