Open Collections

Application of Monte Carlo Simulation to the Probability Assessment in FMEA

Tailings and Mine Waste Conference (2023 : Vancouver, B.C.); University of British Columbia. Norman B. Keevil Institute of Mining Engineering

The mining industry requires risk assessment of major facilities such as tailings dams (TD) at each stage of project development. Various methodologies have been used depending on the needs, with one of the most popular applications being the failure mode and effects analysis (FMEA). The FMEA assesses the likelihood of occurrence (L) and the consequence (C) of each identified failure mode to determine the risk level (R) associated with each failure mode (FM). Subsequently, the critical FMs are selected based on specific criteria for risk classification in order to guide risk reduction efforts. The criteria evaluation form is the qualitative rating standard (e.g., low, medium, high). Two primary issues emerge. First, judgment errors may arise during the workshop-based assessment of Ls and Cs due to potential overconfidence exhibited by the experts. Second, the averaging of the scores performed by each expert to derive a final value is a deterministic analysis that overlooks the variability in expert judgments. The variability includes individual biases, such as risk aversion or risk-taking tendencies, as well as uncertainty among the scores provided by the team. It does not consider a weighted average of individual assessments. It is important to emphasize that some relevant information may be lost by considering only the averages of expert analysis. Given the limitations of deterministic analysis, several researchers have proposed tools to calibrate the behaviour of individual experts within a team. Calibration tests are utilized to measure and address overconfidence exhibited during the estimation of L or C factors. This can be fitted to obtain a stochastic representation of each of them, and then the Monte Carlo simulation can be applied to gain more information about the uncertainty in the outcomes (R). The decision maker then has a range of plausible outcomes and the probabilities of occurrence of them.

Text

2023-12-08

Attribution-NoDerivatives 4.0 International

http://hdl.handle.net/2429/86799

Unreviewed

http://creativecommons.org/licenses/by-nd/4.0/