Do our design guidelines represent the art of nature? Simulating cross-bar block ramps with FLOW-3D
Mario Oertel, winner of Flow Science's
4th 30th Anniversary Simulation Contest
This application note was contributed by Mario Oertel, Senior Researcher in the Hydraulic Engineering Section,
Civil Engineering Department of the University of Wuppertal in Wuppertal, Germany, the winner of the 4th Flow Science 30th Anniversary Simulation Contest.
Block ramps are modern constructions to conquer large river steps with fish climb possibilities.They are classified into block carpets and block clusters, while clusters are separated into structured, unstructured and self-structured blocks (Tamagni et al. 2010). Limiters of block ramps are due to water depths and flow velocities as well as stability criteria.
Cross-bar block ramps are part of the structured ramp family and separate a river bottom slide into several basins with adequate velocities and acceptable water levels. Low discharges are flowing through lower openings in the cross-bars (Fig. 1), larger discharges lead to stone overtopping flows (Fig. 2 and 3). German guidelines (e.g., DVWK 232) give analytic solutions to calculate water depths within the basins.
Physical and Numerical Models
This research investigation deals with physical and numerical simulations of a cross-bar block ramp to identify the art of nature flow and to compare this with available analytical guideline solutions. The physical model is scaled 1:15. Hence, the full ramp dimensions of L = 2.0 m (length) and 0.8 m (width) represent a prototype of 30 m in length and 12 m in width. The level difference between the upstream and downstream end of the ramp is 6.7 cm (1 m in prototype). The cross-bars on the ramp are arranged by recommendations of DVWK 232. The rows have a distance of approx. 35 cm (5 m in prototype); the large boulders are approx. 6 cm in height (0.9 m in prototype). Discharges vary between 1 l/s and 50 l/s in the physical model. FLOW-3D is used for numerical calculations.
Figure 1: Physical model run with Q = 3 l/s
Figure 2: Physical model run with Q = 35 l/s
Figure 3: Physical model run with Q = 35 l/s, side view
The numerical model is calibrated using previous physical model runs with and without any boulder arrangements (see Oertel et al. 2010). Model scaled and prototype scaled simulations confirm negligible scaling effects due to viscosity and surface tension influences. The numerical model is made of one mesh block with 1 cm cell size (Fig. 4). 1.82 million cells, while 1.56 million cells are actively used. The bottom geometry and the cross-bars are included by using an STL file. Based on further investigation (Bung et al. 2008) the RNG turbulence model is selected. At the inflow and outflow area the velocity boundary condition is chosen. Therefore, water depths and velocities at the inlet and outlet are those measured in the physical model. The finish time is 60 seconds, the output interval 1 second.
Figure 4: FLOW-3D model
Simulations are running on a Windows 7 system with an Intel Core Quad 3.00 GHz processor, 4 GB RAM and a parallel solver. The simulations stop after steady state conditions are reached. Simulations run for 3-12 hours, depending on the flow conditions.
Results and Conclusion
Three results are analyzed within this investigation: (1) physical model, (2) numerical model, (3) guideline approaches. For low discharges (Q < 15 l/s) the guideline's solution fit pretty well, but with increasing discharges this analytical approach leads to unrealistic results (Fig. 5).
Figure 5: Comparisons between physical model, FLOW-3D, and national guideline approach. In the last two frames,
there are no experimental results, because these flow rates could not be obtained in the laboratory conditions.
Comparing FLOW-3D results (Fig. 5 and 6) with the physical model, a good conformity can be found for every discharge regime. The stepped as well as the waved flow and, for very high discharges, a quasi uniform flow are reproduced in an accurate way. Especially for discharges over 50 l/s the main advantage of using FLOW-3D can be found. Caused by pump capacities only 50 l/s are available in the University of Wuppertal's Hydraulic laboratory.
Figure 6: Graphical results of FLOW-3D simulations, velocities in [m/s]
After calibrating and generating trustable results with FLOW-3D it is possible to run the model with changing discharges and geometries in a cost-effective way. Here, simulations with discharges up to 200 l/s are arranged to identify water surface elevations and flow conditions for flood events. In summary, FLOW-3D is used as a strong instrument for hybrid flow simulations. It is helpful to reduce laboratory costs, to increase model run possibilities, and to find new solutions for design guidelines.
Bung, D.; Hildebrandt, A.; Oertel, M.; Schlenkhoff, A.; Schlurmann, T. (2008) Bore Propagation Over a Submerged Horizontal Plate by Physical and Numerical Simulation, ICCE 2008, Hamburg.
DVWK 232 (1996) Fish climb constructions, orig. Fischaufstiegsanlagen, German Association for Water Management and land improvement.
Oertel, M.; Heinz, G.; Schlenkhoff, A. (2010) Physical and numerical modeling of rough ramps and slides, Proc. First European IAHR Congress, May 04.-06., Edinburgh.
Tamagni, S.; Weitbrecht. V.; Boes, R. (2010) Design of unstructured block ramps: A state-of-the-art review, River Flow 2010, Bundesanstalt für Wasserbau.