Application of neuro-fuzzy model to estimate the characteristics of local scour downstream of stilling basins

Abstract

The local scour phenomenon in the vicinity of bridge piers and stilling basins has received considerable attraction from designers due to its consequences which may endanger these structures. Various factors govern the pattern of scour evolution which results In the complexity of this phenomenon. Many researchers indicated that the use of fuzzy logic in modeling this phenomenon could be a promising alternative to reflect the vagueness and ambiguity of effective parameters. The aim of this study is to investigate the performance of a neuro-fuzzy model based on Takagi and Sugeno's theory in estimating the maximum depths, pattern and time evolution of scour hole downstream of a stilling basin of U.S.B.R. type I. The investigation was conducted under various discharges, tail-water depths (low, balanced and high), different bed materials and model sizes. The characteristics of the equilibrium state of the scour phenomenon as well as the time to reach the maximum scour depth were considered. The results showed a significant conformity between estimated and experimental data which recommends an acceptable outcome using a neuro-fuzzy model to forecast the properties of scour hole downstream of stilling basins. This would contribute to predicting the design geometry of stilling pools and taking the appropriate precautions to protect the downstream channel bed. © IWA Publishing 2010 Journal of Hydroinformatics.