Experimental analysis and prediction of velocity profiles of turbidity current in a channel with abrupt slope using artificial neural network

• Published in: Journal of the Brazilian Society of Mechanical Sciences and Engineering 2017-11, Vol.39 (11), p.4503-4517
• Main Authors: Baghalian, Sara, Ghodsian, Masoud
• Format: Article
• Language: English
• Subjects: Artificial neural networks; Computational fluid dynamics; Engineering; Mechanical Engineering; Neural networks; Sedimentation; Storage capacity; Technical Paper; Turbidity; Velocity
• ISSN: 1678-5878; 1806-3691
• DOI: 10.1007/s40430-017-0867-9

Turbidity currents are one of the most important factors in sedimentation process in dam reservoirs. Increasing the sediment deposition in front of a dam declines its storage capacity and poses significant operational challenges. Therefore, understanding of turbidity currents fluid dynamics and associated depositional patterns is crucial for efficient operations and management of dam reservoir. In this study, turbidity currents velocity profiles in channel with abrupt slope were investigated experimentally and numerically using artificial intelligence. Experiments were carried out and velocity profiles were measured in a rectangular channel. Then using obtained non-dimensional velocity profiles, new equations were suggested for velocity profiles of turbidity currents. Finally, an artificial neural network approach was proposed and applied to predict the velocity components at some sections of the channel where experimental results were not available. The results showed that the designed artificial neural network predicts the velocity profiles with acceptable accuracy.