Saher Salama

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%.

A finite element formulation is presented for the natural vibration analysis of pipes conveying fluids. The solution of the resulting quadratic eigenvalue problem generally yields complex eigenvalues and eigenvectors. The present study then develops a robust mathematical procedure that combines the real and imaginary components of the eigenvectors to form physically attainable (i.e., real) mode shapes. The procedure yields a family of solutions that is more general than previously known solutions. The well-known classical mode shape is shown to be recoverable as a special case from the present solution. The study provides new insights on the effects of viscous damping, axial compressive force, and the flexibility of intermediate pipe supports on the response. Additionally, the study develops a novel algorithm based on Hermitian angles between eigenvectors to automate the tracing of mode evolution in the frequency-velocity plots.