Sig Turbomachinery Feature MixingPlane Interface OpenFOAM-1.6-ext

From OpenFOAMWiki
Revision as of 19:41, 15 April 2012 by MBeaudoin (Talk | contribs)

Valid versions: OF Version 16ext.png

1 The mixingPlane interface for OpenFOAM-1.6-ext

1.1 Authors

  • Martin Beaudoin, Hydro-Québec
  • Hrv Jasak, Wikki Ltd.

1.2 Description

The mixingPlane interface is implementing the interpolation of face data between two primitivePatches using mixing plane averaging.

The mixingPlane interface is a new addition to the OpenFOAM simulation toolbox. For turbomachinery simulation, the mixingPlane is essentially a circumferential averaging interface. For steady-state simulations in hydroturbines, it is common to use a mixingPlane interface between a guide vane passage and a runner passage, or between a runner passage and a full draft tube geometry.

The main component of the mixingPlane interface is a circumferential averaging interpolator that computes and transfers average field values across the upstream and downstream patches that make up the mixingPlane. The interpolator is constructed using an intermediary cylindrical patch surface made of a stack of 360 degree ribbons shared between two specially crafted GGI interfaces.

Special care has been taken for the discretization of the cylindrical ribbon patch in order to appropriately interpolate both the coarse and fine regions of the upstream and downstream patches from the mixingPlane interface. A discretization profile is constructed based on the 2D upstream and downstream patch point distribution in cylindrical space. This 2D profile will then serve as the discretization control curve for generating the intermediary 360 degree ribbons patch. More information about the actual implementation details of the mixingPlane interface for OpenFOAM will be provided in upcoming publications.

1.3 References

   OpenFOAM TURBO TOOLS: FROM GENERAL PURPOSE CFD TO TURBOMACHINERY SIMULATIONS 
   H. Jasak, M. Beaudoin, Proceedings of ASME-JSME-KSME Joint Fluids
   Engineering Conference 2011, AJK2011-FED, July 24-29, 2011, Hamamatsu,
   Shizuoka, JAPAN 
   Steady-state capabilities for hydroturbines with OpenFOAM,
   M. Page, M. Beaudoin, A.-M. Giroux, Internationnal Journal of Fluid
   Machinery and Systems, 4(1):160–170, Jan-Mar 2011.
   Development of a General Grid Interface for Turbomachinery simulations with
   OpenFOAM, M. Beaudoin, H. Jasak, Open Source CFD International Conference,
   December 2008 


1.4 Availability

1.4.1 Availability of the mixingPlane_RC1 version

1.4.2 How to download

1.4.3 How to compile

1.5 Current limitations

The development of the mixingPlane interface for OpenFOAM is still a work in progress at Hydro-Québec. Still, it was decided to publish a first "release candidate version" in 2012 in order to share the source code with collaborators and the community at large, and to stimulate some feedbacks and comments.

We are still working on improving this interface in order to use it with "industrial" grade meshes and large hydroturbine simulation problems. We will publish our improvements on a regular basis, so you have to expect some changes to the code (certainly), to the dictionaries syntax (possibly), and to the documentation (obviously) pertaining to the mixingPlane interface. So this source code should still be considered as experimental.

As of January 2012, the limitations of the current mixingPlane implementation (mixingPlane_RC1) are:

  • The circumferential average of the various fields are computed from the cell-center values of the cells located on each side of the mixingPlane interface. We are missing a proper interpolation algorithm in order to evaluate the fields correctly when in presence of meshes where the size of the cells on either side of the mixingPlane varies significantly. The solution is known, but not implemented yet. We are actively working in order to fix this. In the meantime, this is a limitation that can be overcome by imposing constraints on your mesh generator. The size of the first layer of cells on the upstream and downstream side of the interface should be mostly homogeneous. This does not mean the size of the cells should be the same for BOTH the upstream and downstream side. They have to be homogenous on the upstream side, and homogenous on the upstream side.
  • The interface is not fully parallelized yet. In order to run mixingPlane cases in parallel, just make sure the interface and the cells on each side of the interface are located on the same processor.
  • First make it run, then make it run fast. We are at the "first make it run" stage of the development. So performance has not been addressed yet. We are using the GGI interface as an internal component for this new development, so the performance of the mixingPlane will always be closely linked to the performance of the GGI interface.

1.6 Examples of dictionaries configuration

  • constant/polyMesh/boundary
Todo
  • field value dictionaries in the time directories
Todo

1.7 Specification of mixingPlane reference frame

Todo

1.8 Specification of the mixingPlane discretization profile

Todo

1.9 Running the mixingPlane in parallel

Todo

1.10 The mixingPlaneCheck functionObject

Todo

1.11 The checkMixingPlane utility program

Todo

1.12 Tutorials

  • Simplistic examples
  • The ERCOFTAC conical diffuser
  • The ERCOFTAC centrifugal pump
  • Simple rotor-stator examples


1.13 Test harness

Todo