Fuzzy-probabilistic evaluation for the dynamic instability of corroded buried pipes conveying fluids

Abstract

This study develops fuzzy-probabilistic models to investigate the dynamic instability (i.e., the first instability point) of corroded buried pipes conveying fluids. The models are developed via a novel hybrid of random fields and fuzzy logic methods to capture both aleatoric uncertainty, stemming from the stochastic nature of pipeline, fluid, and soil parameters, and epistemic uncertainty, arising from corrosion inspection challenges, especially in urban areas. Key features of the models include: (1) applicability to various pipe geometries, fluid types, and soil stiffness distributions; (2) Pipe Condition Index, on 0–100 scale, updatable from field inspection to reflect corrosion levels; and (3) uncertainty quantifiers including the random field significance level and fuzzy model coefficients. Results showcase the corrosion’s significant influence on dynamic instability. Furthermore, non-uniform soil stiffness distribution, with minimum stiffness at mid-span, increases the susceptibility to dynamic instability compared to the uniform soil case. The developed models achieve a high goodness-of-fit, with R² within 0.91–0.96, underscoring their accuracy in predicting the dynamic instability of corroded pipelines conveying fluids.