A hybrid CFD – Deep Learning methodology to improve the accuracy of cut-off diameter prediction in coarse-grid simulations for cyclone separators

Abstract

In many industries, cyclone separators are frequently employed to remove solid particles from the fluid flow. Cut-off diameter is recognized as a significant parameter to evaluate the performance of cyclone separators in addition to pressure drop. Computational Fluid Dynamics (CFD), a powerful computer-based method, can precisely estimate the cut-off diameter of cyclone separators. There is no arguing, however, that the CFD technique is computationally expensive and practically difficult. This research has suggested a more precise, computationally proficient hybrid CFD–DL method to improve the accuracy of cut-off diameter prediction in coarse-grid simulations for cyclone separators. It has been demonstrated that the proposed method not only requires less computational cost than typical CFD, but also delivers more accuracy results (with mean error less than 5.1% compared to experimental data). In other words, it takes advantage of the promise of a novel approach to decrease computational time while enhancing accuracy for CFD simulations.