Influence of PIG Mass, Launching Time and Turbulence Model on 3-D CFD Transient Simulation of PIG Motion

Abstract

Pigging procedures are common maintenance operations used to perform cleaning, draining and pipeline inspection in order to improve flow efficiency and operation cost. Despite these procedures are commonly used, questions still remain regarding the flow and the PIG motion features due to the complex interaction among pig, wall and flow, and the changes in internal fluid pressure and local fluid density. Currently, the PIG dynamic predictions are based on experimental data from short scale laboratory experiments and numerical models founded on physical simplification. So far, the transient of PIG motion calculated by methods that combine CFD and fluid-structure interaction in a 3D model and the influence of the physic and numerical features over the pig dynamics has not been analyzed yet. To provide a better understanding of pigging runs, this paper proposes a CFD methodology to obtain a 3D transient simulation of PIG motion. A moving control volume attached to the PIG let to solve the governing equation in a stationary mesh. This methodology is used to obtain the transient simulation of a PIG launched in a straight water pipeline for different PIG mass, launching time and turbulence models in order to study its influence over the PIG dynamics. The numerical results show a linear relation between the mass and the pressure drop in the transient state, but with no influence over the final stationary state. Also, an asymptotic relation between the transient pressure drop and the launching time was observed with no influence over the PIG terminal velocity. Besides, it is exposed the influence of the turbulence models (κ-ε, SST and BSL Reynolds Stress) in the results of pig motion; appreciable difference between the drop pressure of Omega-Based Stress Models (SST and BSL) and κ-ε turbulent model at steady state is shown, and, finally, a comparison of the velocity profiles at the interstice for each model was developed, this one shows an inaccuracy of the κ-ε model to describe the velocity profile in the walls proximities.