Examination of A Standard Waterways Experiment Station (WES) Spillway Under Various Flood Conditions
A 2D examination of a standard Waterways Experiment Station (WES) spillway profile under various flood conditions was analyzed using FLOW-3D and compared with actual experiments.
FLOW-3D predicts an error of less than 10%. FLOW-3D error was less than a 1% for the Upper Nappe profiles and greater than 10% but less than 20% for the gauge pressure distribution along the crest for the beginning and end of the H/Hd=1.33 results and for the beginning of H/Hd=1.00.
Design Challenge and the Validation Solution
Many existing spillway structures are under-sized to cope with the amended probable maximum flood levels. Formation of lifting pressure and cavitations over spillway crest under flood condition can be encountered. Early dam structures were designed with limited hydrological information. The US Army Corps of Engineering scaled physical models which were expensive and time consuming.
From Figure 1 the gauge pressure distribution among the various heads compared to the experimental show correct behavior. The upper Nappe profiles compared in Figure 2 show an almost exact overlay between the experimental and FLOW-3D.
An experimental validation shows a close match between simulation and reality
Figure 1: Normalized gauge pressure heads compared with experimental results.
Figure 2: Normalized upper Nappe profiles for three different heads compared
with experimental and the crest surface.
Negative gauge pressure (Pa) for H/Hd=13.3 above spillway.
The geometry of the spillway profile came from Hydraulic Design Chart 111-2/1 (Prepared for Office of the Chief of Engineers, 1952; revised in subsequent years). This represents as close to a true 2D problem because the measured results are not influenced by 3D edges since the areas of interest are far from it. The design head was 10m and three different heads, H/Hd=1.00, H/Hd=1.33, and H/Hd=0.5, were examined. The mesh consists of .25 size mesh for the x and z-direction. The aspect ratio was kept close to unity, 1.00 and 1.38 in the x and z directions. Constant density of 1000kg/m3. Gravity was 9.8 m/s2. A turbulent model using the two-equation (k-ε) model with the turbulent length of .65, .91, and .24 respectively for the given heads. The initial height of the fluid head started at the top of the crest. The simulation ran for 15 seconds where it reached a steady state.
- Platform: AMD Athlon 64 x2 Dual Core Processor 4200+
- Processor Speed: 2.2GHz
- RAM: 4.0GB
- Simulation Time: 28 min