New flow tunnel
In FILOSE project, a huge flow tunnel is used for the experiments. The experiment set-up is in this large flow tunnel. The tunnel has a main disadvantage, which is that the researchers have to climb into the water and prepare the experiment in the water, which is really inconvenient, and really slow. The idea comes, that we should construct a new flow tunnel, where we can do the set-up easier. My task was to investigate the water tunnels all over the world and do some Computational Fluid Dynamics simulations for different suitable solutions, what we can achieve with our available aquarium.
Applied flow tunnels in other institutes:
This flow tunnel is used in a laboratory at the University of Linköping. The basic idea of our new flow tunnel comes from this layout, but our available aquarium determines some differences, for example the lengths of the test section.
I divided the construction to two main parts. The first will be the upper part, where we have the inlet pipe, the flow conditioners, the contraction section, test section (available aquarium), and the drain section. The other, lower part will be the propulsion and the pipeline system.
The aquarium modelled with the help of Solidworks:
The main criterion for the flow tunnel is the steady flow in the test section. The steadiness of the flow depends on the layout and the applied flow conditioners. In the first picture we can see that, after the test section two drain pipes were used, because the generated vortices decrease each other’s effect to the test section. If we would like to use the whole test section we should use this two drained layout.
I made basic simulations for this two-drained section, and it gives what I expected, that the drain section doesn’t have main effect on the test section. (The u velocities for different displacement from the beginning of the test section, the orange line means the section before the drains.)
After this result, I neglected the drain parts in my simulation, because it had many curved surfaces, where I can not make a really good meshing.
The next step was to make different layouts for the contraction elements, and determine the flow conditioners.
In the simulation I had two really hard part, because I had to determine the velocity inlet boundary conditions, without knowing the lower part of the flow tunnel, and the properties of the flow conditioner elements.
For solving the first problem I wrote a velocity profile file, where I used a really disturbed velocity distribution, with 3 velocity components, and I used this as a velocity inlet boundary condition. After this, the turbulence parameters were determined, with the help of ANSYS manual. For the correct coefficients the correct velocity inlet is needed, so I used only some guessed data, according to my former simulations.
To determine the coefficients of the flow conditioner, we made an experiment in the old flow tunnel, with the suitable flow conditioner, and without it. I compared the pressure data, and I determined the pressure loss in the conditioner in case of different flow velocities. From these data, I can easily determine the porous media coefficients for the simulation.
We will use an inlet from the back surface, because this inlet will have mainly u velocity components, and we can reach better uniformity than when the water comes from the under surface.
The final simulation results for five different layouts. The layouts can be seen in the left picture, the x velocity distribution in the beginning of the test section can be seen in the right column. (The y and z velocity components can be neglected, because of the thick flow conditioner, which was modelled with porous zone media.)
The first layout is without contraction element, the inlet of the water will be on the back side of the aquarium:
The second layout is when a longer inlet and a wider contraction element were used:
The velocity distribution is better, but a really big vortex can be seen before the porous zone, so a better layout is needed.
The third configuration, when two porous zone were used in the contraction element. The aim of the first (thinner) porous zone is to avoid the vortex. The wide of the contraction element was also reduced, and the next results were given:
The velocity distribution is worse than the former cases, but the streamlines are better.
The fourth layout was investigated with a solid plane inside the inlet element.
This gave a really good result, so it should be used in the future, but I made another simulation for two conditioners which was the fifth layout:
This also really good, and probably this case will be better than the former one, but this simulation was made with small mesh element number.
- Inlet from the back side of the tunnel is better than from the lower part.
- We can use simplified construction for the contraction element. (Straight walls instead of spline shaped walls.)
- The used flow conditioner is really good to create one velocity component.
- For the uniformity we have to use some larger planes in the contraction element, the size and the arrangement of these planes depends on the propulsion and the final shape of the contraction element.