Artificial Intelligent Systems and Machine Learning

Nirjhar Bar ¹, Asit Baran Biswas ¹, Manindra Nath Biswas ², Sudip Kumar Das ¹

Affiliations
1 Department of Chemical Engineering, University of Calcutta, 92, A. P. C. Road, Kolkata-700009, IN
2 Government College of Engineering & Leather Technology, LB Block, Sector III, Salt Lake City, Kolkata-700098, IN

Abstract

Helical coils are extensively used in different process industries. The two-phase flow through the coils is more complex than that of straight pipe due to presence of centrifugal forces, the flow through coils is always developing in nature. The knowledge of frictional pressure drop is an important hydrodynamic parameter used for the designing the coil. The applicability of the Artificial Neural Networks (ANN) methodology was investigated using experimental data obtained from our earlier studies on the frictional pressure drop for gas-non-Newtonian liquid flow through helical coils in horizontal orientation. Multilayer Perceptron (MLP) trained with back propagation algorithm using four different transfer functions in a hidden layer is used in ANN. Statistical analysis is used for the comparison to identify the best network. The ANN’s capability to predict the two-phase frictional pressure drop across the coils is one of the best estimation methods with high accuracy.

Keywords

Artificial Neural Network, Multilayer Perceptron, Frictional Pressure Drop, Gas-Non-Newtonian Liquid Flow.

Full Text