Abstract The effective interfacial area was a vital important parameter for designing the rotating packed bed (RPB) and then the accurate prediction was essential. In this study, five types of artificial neural network (ANN) models were used to simulate the effective interfacial area in the RPB, including feed-forward back propagation neural network (FFBP) model, generalized regression neural network (GR) model, cascade-forward back propagation neural network (CFBP) model, radial basis neural network (RB) model and elman-forward back propagation neural network (EFBP) model. The dimensionless number groups were set as input data, including gas Reynolds number (ReG), liquid Reynolds number (ReL), liquid Froude number (FrL), liquid Weber number (WeL) and packing character parameter (ψ). And the variable of effective interfacial area was set as output data. The mean square error (E2) and coefficient of determination (R2) were used to test the accuracy of the model. The simulation results were compared with the experimental data from references, which showed that the five ANN models well predicted the effective interfacial area and GR model exhibited the best performance. Then GR model was further used to predict the effects of high gravity factor, gas flow rate and liquid flow rate on the effective interfacial area. The predicated results indicated that the effective interfacial area increased with the increase of high gravity factor and gas flow rate. And the effective interfacial area increased with the increase of liquid flow rate at low liquid flow rate, but decreased when the liquid flow rate was above a certain value.
