Publicar

27 - PRIORIZAÇÃO DE RISCOS EM PILHAS DE MINERAÇÃO: UMA ABORDAGEM INTEGRADA DE ESTATÍSTICA E INTELIGÊNCIA COMPUTACIONAL SOBRE DADOS FMEA

Doi: 10.47094/COBRAMSEG2026.2/27

PALAVRAS CHAVE: Análise de Riscos, FMEA, Mineração; Pilhas de Rejeito; Análise Estatística; Inteligência Computacional.

KEYWORDS: Risk Analysis, FMEA, Mining, Waste Rock Pile, Statistical Analysis, Computational Intelligence.

ABSTRACT:
The safety of mining waste disposal structures is critical to prevent environmental and human disasters. This study applies multivariate statistical analysis to assess risks in 21 waste rock piles in Brazil, using FMEA data (2023-2024), aiming to identify patterns among variables such as structural components, risk priority, and analysis reliability. Results show that 93% of failure modes were classified as "Very High" or "High" priority, with surface drainage failures standing out (58.82% in critical cases). Cluster analysis (K-means) classified the structures into four risk levels, while ordinal logistic regression revealed that higher analysis confidence reduces risk priority (coefficient: ?1.7465; p < 0.001). However, 52.18% of assessments lacked reliability, reaching 76% for the most critical risks. The study concludes that risk management should prioritize data quality and components such as drainage and stacking systems.

Autores

  • Claudia Martins Bhering Dominoni

  • Iara Cota Pontes Soares

  • Julia Gonçalves Ribeiro

Download