Evaluation of Computational Fluid Dynamics Modeling for Erosion in Elbow With a Large-Scale Erosion Database

Abstract Computational Fluid Dynamics has been widely used in various engineering applications. Solid particle erosion of material is a multi-physics problem that requires a great deal of expertise to simulate and predict. Evaluation of the CFD models available is critical to formulate the correct CFD settings to perform the task of adequately and accurately predicting erosion and uncertainties from the CFD models. Collecting high quality erosion data is also very challenging, which has been the objective of Erosion/Corrosion Research Center at The University of Tulsa over the past decades. Continuous efforts have resulted in a large number of reliable data collected which can be used for validating different CFD models. The data is collected with high accuracy fixed mounted and temperature compensated ultrasonic transducers that will not interfere with flow and particle regimes inside the elbow. Air and water is used as the carrier fluids. The sand concentration is monitored to guarantee consistent diluted particle-laden flow condition. Experiments are repeated for validity. With data available, CFD models are established to predict erosion for all the collected cases which is the first trial ever to evaluate the robustness of applying CFD to predict erosion under various flow and particle conditions. The significance of the work underlies that obtaining an overall impression of how accuracy is CFD modeling for predicting erosion and a calibrated CFD framework for CFD-based erosion prediction can be derived from numerous practices in the process. The study uses RANS models for turbulence modeling and one-way particle tracking under engineering application context. The effect of global mesh size is examined to reach the best practice for representing erosion results. All the results from CFD are compared to data and have shown that CFD models with delicate calibration can predict erosion accurately, which provides a valuable tool for engineering design and prediction under complex conditions.