[soni_sth]

Hi,
I'm running a simulation for kaplan turbine.I used the experimental data for the BCs.
I ran two set of simulation with different BCs.
1.Total Pressure inlet and static pressure outlet
2.Total Pressure inlet and mass flow outlet.

The torque produced by these two simulation should be same as I am using the data of one single experiment but it comes to be different by large number!!!!!

Also,I want to plot the performance curve of the turbine.Can anyone help me?I would really appreciate.

Thank you.

[ghorrocks]

Obviously the detail of the boundary condition is making a difference. For the pressure outlet, what pressure averaging did you use? If none then it assume a constant pressure across the boundary. For the mass flow outlet it assumes a constant velocity across the outlet.

These differences should only make a difference if your outlet is close to the area of interest, so it suggests your outlet boundary is too close.

Alternately it suggests your experimental results are not very accurate.

[lindner]

For the performance curve you should run the simulation changing the opening of the blades and guide vanes. If you use Total Pressure at inlet and static pressure at outlet what you can do is prepare 5 geometries and run all the cases with the same boundary to get the curve.
Witch boundary gives you the closest results comparing to experimental data?
Also in the end check the net head of the turbine, using (areaAve(Total Pressure)@inlet-areaAve(Total Pressure)@outlet)/("density"*"gravity")

lindner

[soni_sth]

Thanks for the replies.

@ghorrocks:If I have to compare with the experimental data,Then I have to use data from the experiments,Right?We have pressure data at four locations..at the inlet,at draft tube,just before stator and right after rotor.I used Total pressure at the inlet and static pressure at the outlet of the draft tube.The power developed does not match with the experimental value.There is large deviation!!!!
If I do not use the experimental data,how do we impose the BC values in CFX. For certain head and discharge how do we know the RPM it develops which is an input in CFX?

@Lindner:In my case,the turbine is Straflow.rotor and stator blades are fixed.How do we develop the performance curve for this?

Would really appreciate your help
Thanks a lot

[Far]

check these

http://www.cfd-online.com/Forums/cfx...-pressure.html
http://www.cfd-online.com/Forums/cfx...essor-map.html

If I have to compare with the experimental data,Then I have to use data from the experiments,Right??

Not really. Use necessary boundary conditions from experimental data like geometry, inlet total pressure, RPM and then vary the static pressure at outlet to get the performance curve. In my experience best boundary condition is average static pressure, even if boundary is close to blade.

I used Total pressure at the inlet and static pressure at the outlet of the draft tube.The power developed does not match with the experimental value.There is large deviation!!!!

are you comparing the data at same discharge and RPM?

If I do not use the experimental data,how do we impose the BC values in CFX. For certain head and discharge how do we know the RPM it develops which is an input in CFX

see above post and as far as RPM is concerned, at the moment CFX/Fluent are not able to handle the time varying RPM condition and therefore you have to specify the fixed RPM and see if the results are close to experimental data. But to be certain you must ensure first that problem setup is independent of mesh and other parameters for some known results.
You can use velocity triangle method to estimate the RPM. However as you already said that you have experimental data then use RPM from that data set.

[soni_sth]

hi Glenn,

My simulations for two experimental data comes to be similar to that of experiment.In a attached file I have plotted POwer Vs RPM. As we increase RPM and Discharge ,power increases and then it decreases.But In teh simulation power keeps increasing.I am being able to figure out what parameter that I have to change in CFX. I would really appreciate your help.

Thank you so much.