A 2D Image-Based Approach for CFD Validation of Liquid Mixing in a Free-Surface Condition


1 Institute of Engineering, National Autonomous University of Mexico (II-UNAM), Mexico City,04510, Mexico

2 Department of Systematics and Aquatic Ecology, National Council of Science and Technology-The Southern Border College (CONACYT-ECOSUR), Chetumal, 77014, Mexico

3 Coastal Processes and Physical Oceanography Laboratory, Department of Marine Resources, Center for Research and Advanced Studies of the National Polytechnic Institute (CINVESTAV), Merida, 97310, Mexico


This study proposes an image-based approach to evaluate the validity of numerical results for cases where the setup can be assumed to be two-dimensional (2D) and mixing between liquids of different densities occurs under a free-surface condition. The proposed methodology is based on the estimation of the relative errors of the model through density matrices generated from images of the experimental and numerical results (i.e., post-processing snapshots of the density field). To demonstrate the use of the methodology, experimental tests and numerical simulations were performed for a double-dam-break problem with two miscible liquids. For the experiments, a high-speed camera was employed to capture details of the fluid interactions after the dam breaking. For the numerical simulations, an OpenFOAM® multiphase solver was employed to reproduce the benchmarking tests. Three turbulence approaches were tested: a zero-equation RANS model, a two-equation (k-epsilon) RANS model, and a Large-Eddy Simulation (LES) model. The experimental results compared favorably against the numerical results, with averaged relative errors of ~17 and ~19 % for the zero-equation and the two-equation turbulence models, respectively, and ~14 % for the LES model. From the results obtained, it can be inferred that the two-equation (k-epsilon) model had limitations in reproducing the mixing between the liquid phases in terms of relative errors. The LES model reproduces the mixing between phases more accurately than zero and two-equation RANS models, which were seen to be more suitable for capturing the formation of large eddies in the initial phase of the experiment. The present methodology can be improved and extended for different multiphase flow configurations.


Volume 13, Issue 5
September 2020
Pages 1487-1500
  • Received: 28 August 2019
  • Revised: 02 February 2020
  • Accepted: 08 February 2020
  • First Publish Date: 02 May 2020