Computational Modeling of Flow over a Spillway In Vatnsfellsstífla Dam in Iceland

1 / 57

Computational Modeling of Flow over a Spillway In Vatnsfellsstífla Dam in Iceland - PowerPoint PPT Presentation

Computational Modeling of Flow over a Spillway In Vatnsfellsstífla Dam in Iceland. Master’s Thesis Presentation Chalmers University of Technology 2007 – 02 - 02. Presentation Schedule. Introduction and background Method Theory Results Conclusions and future work.

I am the owner, or an agent authorized to act on behalf of the owner, of the copyrighted work described.

PowerPoint Slideshow about 'Computational Modeling of Flow over a Spillway In Vatnsfellsstífla Dam in Iceland' - Olivia

Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author.While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server.

- - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - -
Presentation Transcript

Computational Modeling of Flow over a SpillwayIn Vatnsfellsstífla Dam in Iceland

Master’s Thesis Presentation

Chalmers University of Technology

2007 – 02 - 02

Presentation Schedule
• Introduction and background
• Method
• Theory
• Results
• Conclusions and future work
The spillway – characteristics
• Function: cope with accidental flooding
• Height above stilling basin bottom: 27.5 m
• Lenght of spillway crest: 50 m
• Equipped with a splitter wall and cover to prevent overtopping of the chute sidewalls
• The velocity of the water is above 20 m/s (=72 km/hour!) where it flows into the stilling basin
The stilling basin – characteristics
• Function: Decrease flow velocity in order to decrease risk for erosion in the river wally downstream the basin
• Equipped with 28 energy dissipating baffles (height from 1.5 to 2.0 m)
• Length ca. 33 m and the width increasing from 22 m in the upstream part to 33 m in the downstream part, depth ca. 7 m
• Downstream the stilling basin is a 35 m long rock rip-rap made of rocks with diameter of

0.4 – 1.2 m

Background and goals
• In 1999 Vattenfall in Sweden did hydraulic experiments for the spillway with a 1:30 model
• In the experiments flow was investigated over the spillway, through the bottom outlet and in the stilling basin
• Goals of the present study:
• investigate flow over the spillway and in the stilling basin with computational methods (CFD)
• compare CFD-results with experimental results
Aspects
• Spillway:
• water head in the reservoir vs. the discharge capacity of the spillway
• Water level along the chute sidewalls
• Pressure acting on the chute bottom
• Stilling basin:
• Water level
• Pressure acting on the baffles and the end sill
• Flow velocity out of the basin
Method
• Identify the computational domain to be modeled (according to the goals!)
• Draw the computational domain in 3D in Autodesk INVENTOR
• Import the geometry into the mesh making software GAMBIT and divide the computational domain into computational cells of different size in GAMBIT
• Import the mesh into the CFD-solver FLUENT, set up the numerical model, compute and monitor the solution
• Postprocessing with FLUENT and MATLAB; examine the results and consider revisions to the model
The computational domain
• Three different domains:
• One for head vs. flow discharge
• One for water level and pressure in the spillway chute
• One for water level, pressure and flow velocity in the stilling basin
• Why different domains?
• to spare computational power and get more precise results
Grids nr. 1 – 7 as seen from above- one grid for each of the seven different cases with flow discharge of 50 – 350 m3/s, ca. 653 000 cells/grid
Cut through grids nr. 1 and 7 in the downstream end of the reservoir by the spillway crest – different water levels
• Grid to the left: designed for flow discharge of 50 m3/s
• Grid to the right: designed for flow discharge of 350 m3/s
Grid nr. 8: finer in the chute than grids nr. 1 – 7, ca. 1393 000 cells
• The mesh in the spillway bottom
• To the left: mesh 7 which is NOT specifically designed to investigate pressure and water level in the spillway chute
• To the right: mesh 8 which is specifically designed to investigate pressure and water level in the spillway chute
Mesh nr. 8: finer in the chute than meshes nr. 1 - 7
• The grid perpendicular to the splitter wall
• To the left: mesh 7 which is NOT specifically designed to investigate pressure and water level in the spillway chute
• To the right: mesh 8 which is specifically designed to investigate pressure and water level in the spillway chute
Grid nr. 9: different types of mesh; consisting of both hexahedron cells and tetrahedron cellsca. 498 000 cells
Grid nr. 9 includes the stilling basinthough coarse in view of the size of the computational domain
Setting up the numerical model
• Define
• Material properties (air, water, concrete)
• Boundary conditions (inlet, outlet, walls,

air pressure,...)

• Operating conditions (air pressure, gravity, temperature...)
• Turbulence model (standard k-ε)
• Initial solution (nB: steady flow)
• Convergence criteria
Theory – equations of motion and the VOF method
• The continuity equation for incompressible flow:
• The momentum equation for incompressible flow:
• VOF method in FLUENT
• assumes that the two phases (air and water) are not interpenetrating
• denoting αq as the volume fraction of the q-th phase three possibilities for a given cell can be noted:
• i) : the cell is empty of the q-th phase,
• ii) : the cell is full of the q-th phase,
• iii) : the cell contains the interphase between the q-th phase and one or more phases.

Main results!Comparison to the experimental results

Water reservoir head vs. flow discharge; Q=CBH3/2where Q= flow discharge, C= discharge coefficient, B = length of crest, H=head
Volume fraction of water in the basin (longitudinal profile) – determines the water level
Total pressure on the basin end sill- a view under the water surface in the downstream end of the basin
Main results - summary
• Good agreement is reached between the experiments and CFD calculations for the following aspects:
• head vs. discharge capacity (Q=CBH3/2)
• pressure in the spillway chute
• flow velocity above the basin end sill
• Worse agreement is reached for:
• pressure on baffles in the upstream end of the basin
• water depth along chute sidewalls and in the left upstream corner of the basin
• pressure on the basin end sill
Future work – what might to be done better or added?
• Calculate the flow through the bottom outlet
• Better resolve the turbulent boundary layers close to walls
• finer mesh
• more computational power
• even parallel processing