Prediction of sand erosion in CFD with dynamically deforming pipe geometry and implementing proper treatment of turbulence dispersion in particle tracking

Abstract

Predicting more accurate sand erosion rate in oil and gas production is important for maintaining safe and reliable operations while maximizing output efficiency. Computational fluid dynamic (CFD) is a powerful tool for erosion prediction and provides detailed erosion pattern in complex geometry. To improve accuracy of erosion prediction, first part of this paper proposes an algorithm to include the effect of dynamically changing pipe geometry in CFD based erosion modelling and second part of this paper provides guidelines for proper treatment of turbulence dispersion in CFD based erosion modelling.Sand erosion is a time-dependent phenomenon where most severe erosion scars are localized and occurs under disturbed flow conditions. These localized high erosion rates cause substantial change in pipe geometry, creating cavities which modify the flow field, the particle trajectories and eventually the erosion pattern. This work has illustrated that the erosion prediction can be influenced by including the effect of dynamically changing pipe geometry in CFD modelling. An algorithm was developed to couple erosion rate calculations with a Moving-Deforming-Mesh, which dynamically deforms the solid wall surface geometry and computational mesh based on the local erosion rate at regular time intervals. This effect will be significant for geometries involving sharp corners and/or regions of recirculation zones.Sand erosion largely depends on particle impact location, impact angle and impact velocity - which is the commutative effect of particle trajectory leading to the bounding surface. The particle trajectory is determined based on mean and the turbulence of the flow field. In steady state CFD simulations with Reynolds Averaged Naiver Stokes (RANS) turbulence models, turbulent dispersion is obtained from mean turbulence quantities using a stochastic random walk model, which estimates the instantaneous characteristics of a turbulent eddy. However, stochastic random walk model parameters will have significant effects on turbulence dispersion of the particles and hence the erosion prediction. Specific guidelines for these model parameters can improve the accuracy and consistency in CFD based erosion prediction.