A fuzzy-random fields approach for assessing the natural frequencies of corroded pipes conveying fluids

Abstract

Evaluating the dynamic behavior of corroded pipelines conveying fluids is essential to guard against resonance and fatigue failures. However, existing solutions for obtaining the dynamic characteristics, such as natural frequencies, lack an updating methodology based on new corrosion inspection data, hindering effective planning for preventing corrosion damages. Thus, this article introduces a novel framework for characterizing the natural frequency of corroded pipelines conveying fluids. The framework simulates fluid-structure parameters using Gaussian random fields, modified using fuzzy membership functions to account for corrosion effects as identified from field inspection. Based on these fuzzy-random fields, piecewise fuzzy-probabilistic formulations are developed for the natural frequency of pipeline systems function of (1) dimensionless fluid-structure parameters to facilitate adaptation for various pipeline systems; (2) a pipeline condition index (PCI) to enable updates with new corrosion inspection data; and (3) a user-defined safety margin to accommodate different safety requirements. Numerical examples are presented, showcasing the ability of the proposed piecewise formulations to estimate the natural frequencies with R2 ranging from 90.8% to 99.8%.