Frictional Pressure Drop for Gas - Non-Newtonian Liquid Flow through 90 degrees and 135 degrees Circular Bend: Prediction Using Empirical Correlation and ANN

Experiments have been carried out to determine the two-phase frictional pressure drop across 90 degrees and 135 degrees bend for gas-non-Newtonian liquid flow on the horizontal plane. Empirical correlation has been developed to predict the two-phase friction factor using the physical and dynamic variables of the system. The applicability of Artificial Neural Networks (ANN) methodology have also been reported. The ANN prediction have been reported using Multilayer Perceptrons (MLP) trained with five different algorithms, namely: Backpropagation (BP), Scaled Conjugate gradient (SCG), Delta-Bar-Delta (DBD), Levenberg Marquardt (LM), Quick-Prop (QP). Four different transfer functions were used in a single hidden layer for all algorithms. The A -square test confirms that the best network for prediction of frictional pressure drop is when it is trained with Backpropagation algorithm in the hidden and output layer with the transfer function 4 in hidden layer having 13 processing elements for 90 degrees bend. The A - square test also confirms that the best network for prediction of frictional pressure drop is when it is trained with Levenberg - Marquardt algorithm in the hidden and output layer with the transfer function 1 in hidden layer having 7 processing elements for 135 degrees bend. Both the methods are equally predictive in nature but the empirical correlation is based on the physical and dynamic variables of the system, whereas the ANN prediction is not dependent on the individual relationship between the input variables.