Erosion in oil and gas pipelines occurs when sand particles impact pipeline walls due to a change in flow direction or any restrictions in the line of flow. In uninhibited terrains like deep sea deserts and mountains through which pipelines pass it’s difficult to monitor the pipe fittings and it is important to reliably predict the erosion rates and make suitable design changes to alleviate this problem. The current solid particle erosion prediction models are either empirical or semi empirical in nature and the suitability over wide ranges of operations is questionable. Erosion models have been implemented in a computational fluid dynamics (CFD) framework. The ratio of actual to predicted erosion rates using this framework can range from 0.01 to 100. The treatment of particle shapes and the particle – wall particle – particle and particle – flow interaction are simplified in a CFD framework and an improvement to modeling these interactions could help reduce the errors in prediction. In a coupled CFD – Discrete Element Method (DEM) framework the fluid is modeled using the CFD approach the solid particles are modeled using DEM in the Lagrangian frame and the interactions are also modeled. The DEM framework accounts for particle shapes and the particles rebound from the wall when the particle surface impacts the wall. Since erosion is a direct result of energy lost by the particle by impact (normal and tangential) the cumulative energy loss of particles on the pipe wall can be quantified and the erosion on the pipe wall can be predicted. In this work the CFD – DEM approach is used to model erosion in a pipe elbow and the results are compared with the experimental results and predictions using the single particle erosion models in a CFD framework.