I need to do scaled down model studies of an existing river bridge. Froude number dynamic similarity would be maintained.I need to know the criteria for scaling down the size and properties of the river bed material to be used in the model.
If you intend to maintain Froude similarity, I think you need a distorted model to use sediment scale (diferent horizontal and vertical scales). Some times a light weight bed material in the model is apropriate for that scale (grain size and density).
You can find out some theory on this in:
1 - a book by Prof Yalin - Yalin M.S.: Mechanics of sediment transport. Pergamon 1977.
2 -Recent Advances in Hydraulic Physical Modelling, Volume 165 of the series NATO ASI Series pp 39-63
and I should say that in our laboratory we sitll use and old publication to help in choosing scaling options:
Chauvin J.L.: Similitude des modèles de cours d’eau à fond mobile. EDF. Bulletin CREC N° 1-1962.
In scour studies around bridge piers, dimensional analysis based on effective variables (ds = f(flow parameters : V,h; sediment parameters: d50, σg, Vc; fluid : ρ, µ and g, Pier: D, Ks, etc.) leads to non-dimensional function as: ds/D = f1 (V/Vc, Fr, D/d50, y/D, Re, Ks=pier shape factor, Kθ=angle of attack for flow, KG=river geometry, ). For similarity all of this parameters should be same for model and prototype system. As it is difficult to satisfy this similarity criteria, you should use a distorted model. Based on above dimensional analysis, for scaling down the river bed materials the similarity for D/d50 should be satisfied, in other words, the parameter D/d50 should be same for model and prototype. For non-uniform bed materials, this similarity should be satisfies by σg too.
The problem is when your scaled down bed material size be so small that could not be cohesionless, or when sediment size be smaller than 0.6 mm that ripple forms may be formed.