Use of artificial neural network to predict pressure drop in rough pipes
- Resource Type
- Conference
- Authors
- Sahai, Swati; Kulkarni, Tanmay; Tikhe, Shubhangi; Mathpati, C. S.
- Source
- 2017 International Conference on Computing Methodologies and Communication (ICCMC) Computing Methodologies and Communication (ICCMC), 2017 International Conference on. :452-455 Jul, 2017
- Subject
- Communication, Networking and Broadcast Technologies
Components, Circuits, Devices and Systems
Computing and Processing
Power, Energy and Industry Applications
Robotics and Control Systems
Signal Processing and Analysis
Artificial neural networks
Mathematical model
Friction
Fluids
Training
Data models
Predictive models
artificial intelligence
neural network
pressure drop
Moody chart
- Language
Prediction of frictional pressure drop in rough pipes is a challenging task for process engineers. All the chemical and allied industries have large and complex piping networks made up of varied materials. The pumping system design needs accurate estimation of frictional pressure drop in pipes. This is done using Moody chart which relates friction factor with two key variables, namely Reynolds number and roughness parameter. Obtaining exact mathematical relationship explicit in nature is very difficult and hence use of artificial intelligence using neural network systems can be a promising approach. In this work, the Moody chart has been digitized and an ANN model has been trained and tested. The regression coefficient of 99% was obtained for training and validation steps.