A combination of computational fluid dynamics (CFD) and adaptive neuro-fuzzy system (ANFIS) for prediction of the bubble column hydrodynamics

Abstract

This paper shows a combination of computational fluid dynamics (CFD) and adaptive neuro-fuzzy inference system (ANFIS) to propose a new viewpoint for multiphase flow modeling, including the accuracy of soft computing techniques in the prediction of a three dimensional (3D) bubble column reactor. Since there are some difficulties (i.e., high computational time in numerical methods and expensive equipment in experimental techniques) in predicting bubble column reactors, particularly at different column locations and various operation conditions, soft computing methods can be developed as a favorable replacement for conventional measurement and prediction techniques. This study employs CFD beside the ANFIS method to simulate the bubble column hydrodynamics for homogeneous regime. Existing experimental, numerical and correlation results in the previous studies have been used to validate the implementation of the current CFD investigation. The liquid velocity, turbulent kinetic energy and gas hold-up (air volume fraction) have been used as input training data in the ANFIS model. The ANFIS results have been also compared with the CFD results, using root-mean-square error (RMSE), coefficient of determination (R2) and Pearson's coefficient (r). Both CFD and ANFIS prediction methods illustrate that, towards the bubble column center, the gas hold-up is higher than wall regions. The results show that ANFIS is a robust method to predict bubble column hydrodynamics properties.