Scour in regions of flow separation with free-surface effects

John F Newton, Purdue University

Abstract

Downstream of essentially all overflow and drop structures two distinct flow regimes are possible, one associated with a plunging flow, the other dominated by a flow that rides along the surface. Predicting the flow regime is important because diving-jet scouring rates are substantially faster than surface flow scouring rates, and diving-jet scour holes form nearer to the structure. A bi-stable region exists in which either flow is possible with the same upstream and downstream flow conditions. Bi-stable regime boundaries were delineated for a wide variety of fixed bed structures, based on both new experiments and reanalysis of others' experiments with new dimensionless parameters, in an effort to form a generalized picture of regime phenomena. Scour downstream of what was initially a backward-facing step was examined in greater detail for both bi-stable and non-bi-stable flows to better understand basic mechanisms. Experiments with both erodible and fixed downstream beds were conducted (the upstream bed was fixed but sand-roughened). Detailed flow characteristics were measured using laser Doppler velocimetry for a fixed bed bi-stable flow and for erodible bed surface flows. Final scour-hole geometry was obtained using photogrammetry. When critical flow occurs at the brink of a backward-facing step, waves are prevalent in the surface regime. Maximum wave height occurs at the transition fromsurface to diving-jet flow; when the flow dives waves can no longer be sustained. Fixed bed surface flow experiments showed that reattachment length, XR, increases with decreasing wave height, attaining a peak value of XR/hbrink of approximately 12 as the tailwater elevation approaches the upstream critical depth water surface elevation. For submerged flows XR normalized by the step height, hbrink, decreases with decreasing expansion ratios to a minimum value of XR/h brink of approximately 5, but increasing the step height is not equivalent to decreasing the tailwater depth; additional velocity dependence was also noted. Erodible bed surface regime experiments were examined in the context of shear layer growth rates and reattachment lengths. Experiments demonstrated that scour initiated within the reattachment zone for the full range of X R/hbrink. A shear layer region of influence was defined, which approximated the equilibrium upstream scour hole slope of a 26 day erodible bed experiment. Shorter duration experiments also showed scour hole slope dependence on shear layer growth rates. The upstream scour hole slope is not stable for flows within the bi-stable regime. Bed deformation caused by scour forces spontaneous cycling between the two bi-stable states, with cycling periods initially roughly proportional to the scouring rates. The upstream scour hole slope is primarily responsible for the cycling. Scouring volumetric rates were more than an order of magnitude larger for the diving-jet phase. This underscores the importance of predicting the flow regime, which is not always clearly defined for common scour equations.

Degree

Ph.D.

Advisors

Lyn, Purdue University.

Subject Area

Civil engineering

COinS