Author(s): D. B. Laucelli
Linked Author(s):
Keywords: No Keywords
Abstract: Preventing degradation of river bed is strictly related to the estimation of scour depth downstream of grade-control structures. Empirical formulations for such scope can be retrieved from literature, being basically obtained by physical and regression analyses on measured data on laboratory scale observations. Some of these expressions are currently used for practical applications in real design problems, even if they shown troubles in large-scale extension. This paper try to apply a datadriven approach to the estimation of maximum scour depth at different scales, by using a new evolutionary-based technique, named Evolutionary Polynomial Regression, which combines numerical regression and evolutionary computing. Data from literature are used in order to build scour depth models, including different bed grain-size distributions and scales of the erosive phenomenon. The results are discussed considering some well-known equations available in literature. The returned equation proved efficient and physically sound.
Year: 2009