Performance evaluation of tunnel type sediment excluder efficiency by machine learning

ABSTRACT

A tunnel type sediment excluder is a preventive fluidic device constructed on the bed of the river in front of head regulator to mitigate the sediment in the off-taking canal. The water approaching the head regulator is separated into two parts by the roof slab of the tunnels. The top part having relatively clearer water enters into the offtake canal, while the bottom layer which carries most of sediment flows through the tunnels and is discharged from the undersluices to downstream side. Very crude methods for design of sediment excluder are in vogue to date which are based on past experiences and physical model study. So, a rational approach of machine learning methods; multivariate adaptive regression splines (MARS), group method of data handling (GMDH) and M5P has been utilized for estimating the removal efficiency of sediment excluder (E). These machine learning methods have also been compared with the conventional models given by earlier researchers. Sensitivity study indicated that extraction ratio (R) was the most influencing parameter, which is followed by size of silt (S). Parametric study further suggests that with increase of R, E is getting enhanced.

KEYWORDS:

• Tunnel type sediment excluder
• multivariate adaptive regression splines
• M5P
• sensitivity
• parametric study

Disclosure statement

No potential conflict of interest was reported by the authors.