Hi,
I'm a new user of Code_Saturne and I've got the CS 2.1.7 version installed on my workstation.
I'm interested in turbomachinery applications for the CFD simulation of axial turbines/compressors. I've tried the Genova's pump case on the CS solver and I've found it very interesting: I've tried both steady (frozen rotor) and unsteady simulations with mesh coupling. In order to determine the machine performances many flow solvers (typically adopted in the turbomachinery field) have the option to perform steady calculations using MIXING PLANES between the blade rows. Is it possible to set a mixing plane between the turbine blade rows in order to determine the machine performances (efficiency and output power)? I need to perform a steady calculation that allows to find the machine performances (so the frozen rotor approach is not useful in this case...). Thank you very much in advance for the answers.
Sir E.
Turbomachinery and mixing planes
Forum rules
Please read the forum usage recommendations before posting.
Please read the forum usage recommendations before posting.
Re: Turbomachinery and mixing planes
Hi,
The mixing plane method is not available in the current version of Code_Saturne. Maybe it will be integrated in a future version.
Of course, the machine performance can be determined thanks to a full unsteady computation integrated in time but it is a "little" more time consuming...
Best regards,
Benoît
The mixing plane method is not available in the current version of Code_Saturne. Maybe it will be integrated in a future version.
Of course, the machine performance can be determined thanks to a full unsteady computation integrated in time but it is a "little" more time consuming...
Best regards,
Benoît
-
SiR2007
Re: Turbomachinery and mixing planes
Hi,
Thank you for the response.
Best regards,
Sir E.
Thank you for the response.
Best regards,
Sir E.
-
rwthler
Re: Turbomachinery and mixing planes
Hello,
I am also highly interested in an test/example case for turbomachinery cases.
I am also highly interested in an test/example case for turbomachinery cases.
-
Yvan Fournier
- Posts: 4077
- Joined: Mon Feb 20, 2012 3:25 pm
Re: Turbomachinery and mixing planes
Hello,
There are several threads on this Forum from about 6 months ago relating to turbomachinery cases, which may help you get started. They relate mostly to version 3.2, but things have not changed too much in this regard since (except turbomachinery cases may now be handled using the GUI).
I can probably post a simple case from our verification suite, though I need to check if it can be published.
Note that version 2.1 is obsolete an not maintained anymore (I assume you installed it using a Linux distribution package manager. This is practical, but versions become obsolete (2.0 is a long-term maintenance release, maintained until release of 4.0 next spring), but 2.1, 2.2, 2.3, and 3.1 are no more, and 3.2 and 3.3 will soon be dropped also, when a first 4.0-beta is released (the code was branched today, so tarballs will be released in a few days, but 4.0-beta is now available for those who pull from the Subversion mirror).
Regards,
Yvan
There are several threads on this Forum from about 6 months ago relating to turbomachinery cases, which may help you get started. They relate mostly to version 3.2, but things have not changed too much in this regard since (except turbomachinery cases may now be handled using the GUI).
I can probably post a simple case from our verification suite, though I need to check if it can be published.
Note that version 2.1 is obsolete an not maintained anymore (I assume you installed it using a Linux distribution package manager. This is practical, but versions become obsolete (2.0 is a long-term maintenance release, maintained until release of 4.0 next spring), but 2.1, 2.2, 2.3, and 3.1 are no more, and 3.2 and 3.3 will soon be dropped also, when a first 4.0-beta is released (the code was branched today, so tarballs will be released in a few days, but 4.0-beta is now available for those who pull from the Subversion mirror).
Regards,
Yvan
Re: Turbomachinery and mixing planes
Hi
The flow case I am interested in looks like a fan within a duct.
Ultimately I 'd like to get a serie of the instantaneous flow-fields to look at the acoustics afterwards.
I would like you to let me know :
(1) if it is possible AND ALSO relevant
to select the frozen-rotor paradigm (which allows in a steady way and without mesh problem at the interface between the fixed sub-domain and the rotating sub-domain) to account for the Coriolis source term (associated to the rotor forced motion)
IN COMBINATION WITH
the selection of an unsteady integration scheme of the NS equations to get the time-varying turbulent flow
(2) a strategy based on the mixing-plane paradigm is available in the recent version of the code
Thank you in advance for your comments
Henri
The flow case I am interested in looks like a fan within a duct.
Ultimately I 'd like to get a serie of the instantaneous flow-fields to look at the acoustics afterwards.
I would like you to let me know :
(1) if it is possible AND ALSO relevant
to select the frozen-rotor paradigm (which allows in a steady way and without mesh problem at the interface between the fixed sub-domain and the rotating sub-domain) to account for the Coriolis source term (associated to the rotor forced motion)
IN COMBINATION WITH
the selection of an unsteady integration scheme of the NS equations to get the time-varying turbulent flow
(2) a strategy based on the mixing-plane paradigm is available in the recent version of the code
Thank you in advance for your comments
Henri
-
Brian Angel
Re: Turbomachinery and mixing planes
Hello,
If your model contains just a single row of fan blades, i.e. a single blade row, then you do not need to use a mixing plane, frozen rotor or any other approach developed for single or multi-stage flow. All you will need to do is run this simulation with some sort of a time varying upstream boundary condition to get an unsteady flow field that could impact the noise generated by turbulence.
If your model contains at least two blade rows, i.e. at least a single stage, then the frozen rotor approach can be used but, depending on the position of the two blade rows relative to one another, I suspect that you will get different levels of turbulence hence noise due to how the wakes from the upstream row flow through the blade passage and interact with the blades of the downstream row.
If your model does contain at least two blade rows and you want to account for the impact of how the wake from the upstream row interacts with the downstream row then you have two possible approaches that I can think of:
1) Use a frozen rotor type method and run several unsteady simulations with the upstream row in different positions relative to the downstream row.
2) Run an unsteady calculation with the rotor row moving in time. This way the wake chopping by the downstream row will be simulated and any impact on the turbulence field accounted for.
Be aware that if there is a non-integer ratio between the number of rotor and stator blade rows you will need to simulate the whole annulus in 1) and 2) above.
Regards,
Brian Angel.
If your model contains just a single row of fan blades, i.e. a single blade row, then you do not need to use a mixing plane, frozen rotor or any other approach developed for single or multi-stage flow. All you will need to do is run this simulation with some sort of a time varying upstream boundary condition to get an unsteady flow field that could impact the noise generated by turbulence.
If your model contains at least two blade rows, i.e. at least a single stage, then the frozen rotor approach can be used but, depending on the position of the two blade rows relative to one another, I suspect that you will get different levels of turbulence hence noise due to how the wakes from the upstream row flow through the blade passage and interact with the blades of the downstream row.
If your model does contain at least two blade rows and you want to account for the impact of how the wake from the upstream row interacts with the downstream row then you have two possible approaches that I can think of:
1) Use a frozen rotor type method and run several unsteady simulations with the upstream row in different positions relative to the downstream row.
2) Run an unsteady calculation with the rotor row moving in time. This way the wake chopping by the downstream row will be simulated and any impact on the turbulence field accounted for.
Be aware that if there is a non-integer ratio between the number of rotor and stator blade rows you will need to simulate the whole annulus in 1) and 2) above.
Regards,
Brian Angel.
Re: Turbomachinery and mixing planes
Brian - are you saying that in the case of two blade rows (say a rotor and stator row) it is possible to have a single moving rotor mesh passage which slides past a single stator mesh passage (provided there is an integer ratio between the number of blades) in code_saturne!? If so I would be very interested in knowing how this could be done.