PROJECT APPLICATION EXAMPLE

chetn-iv-55
Page Navigation
5

system. These were taken to be the "prototype cases." Then, distorted models of these same

configurations were tested, and the measured velocities of the distorted models were scaled and

compared to the prototype results. Turbulence generated by flow past vertical edges was in

similitude in distorted models because the turbulence was mainly in the horizontal plane with small

vertical velocity fluctuations. Turbulence generated at sloping edges exhibited a scale effect closer

to the bottom, but near the surface the flow was in near similitude. There were no scale effects

within the main region of nonturbulent flow. Finally, turbulence created by flow past a horizontal

step was also shown to be in similitude because the turbulence was mostly manifested in the vertical

plane with only small horizontal turbulent fluctuations.

Another potential cause for dissimilar flow patterns is nonturbulent flow around a bend which is

known to generate secondary or helical flows. Hughes and Pizzo (in preparation) performed a

theoretical analysis of potential scale effects and concluded that scale effects would exist in a

distorted model, but this may not be critical provided the model has adequate bottom roughness to

help balance the effect of the cross-channel centrifugal acceleration. It was noted that most

numerical modeling neglect the convective accelerations that will not be in similitude which implies

these terms have minor influence. No experimental data were given to support this hypothesis.

In summary, there will be scale effects present in geometrically distorted models where large-scale

turbulence features such as gyres are generated by solid boundaries. The magnitude of the scale

effect is difficult to ascertain, but differences between model and prototype decrease as the

magnitude of the vertical turbulent fluctuations decreases. Because distorted models have steeper

slopes that decrease the magnitude of the vertical turbulence components generated by the slope, it

should be expected that the prototype might experience stronger vertical turbulence than

demonstrated in the model. Once again, whether or not these scale effects degrade the model results

will depend on the goals of the modeling and the relevance of the turbulent flow processes to the

specific regions of interest within the study area.

PROJECT APPLICATION EXAMPLE: U.S. Army Engineer District, Alaska, sponsored flow

table studies to examine the hydrodynamic flow regime in upper Cook Inlet in the vicinity of the

Port of Anchorage. Shoaling at Anchorage Harbor during the summer months has required annual

dredging that averages between 200,000 and 400,000 yd3 per year with occasional larger deposition

quantities between 800,000 and 1,000,000 yd3. The flow table models helped to clarify the flow

regime, and flow visualization techniques indicated that shoaling was likely caused by ebb tide flow

separation occurring at a headland (Cairn Point) located just upstream of the port. Figure 3

illustrates the approximate line of flow separation and the reduced flow region in the lee of Cairn

Point adjacent to the port. This particular finding had not been hypothesized prior to the flow table

tests. Details of the study are in Hughes and Pizzo (in preparation).

Idealized Flow Table Models. Two types of flow table models were designed and constructed

for this study: (a) Idealized models at two different scales with bathymetry represented as two or

three horizontal terraces, and (b) a 3-D model that reproduced actual bathymetry. Figure 4 shows