Prediction of pressure drop using artificial neural network for non-Newtonian liquid flow through piping components

Numerous investigations have shown that ANN can be successful for correlating experimental data sets for macroscopic single phase flow characteristics. The approach proved its worth when rigorous fluid mechanics treatment based on the solution of first principle equations is not tractable. Evaluation and prediction of the frictional pressure drop across different piping components such as orifices, gate and globe valves and elbows in 0.0127 m piping components for non-Newtonian liquid flow are manifested in this paper. The experimental data used for the prediction is taken from our earlier work (Bandyopadhyay and Das, 2007). The proposed approach towards the prediction is done using a multilayer perceptron (MLP), which is trained with backpropagation algorithm because the function approximation is achieved with very good accuracy using MLPs.