Dear onelab people,
for electric machine analysis, should I use getdp or can I use gmesh FEM? Which one is recommended?
Thanks in advance,
Alexander
In the model "ElectricMachines/pmsm_cbmag", when chosing current excitation, the input value for IQ doesn't seem to have any effect. I always get a small cogging torque.
In the simplified pmsm model this works as expected.
After some thought I have thought it best to attach my efforts,so maybe the problem is clearer.
to run using gmsh:
gmsh motor3D_003.pro
the click run.
to run using command line:
gmsh -3 -format msh1 motor3D_003.geo; getdp motor3D_003.pro -solve Elasticity_u -slepc -pos post_modal_3
Please don't laugh too hard...
I think I have now understood the format of the .pre file. But I'm still having problems with 3D modal analysis (eigenvalue problem K - omega^2*M).
Is there someone who can shed some light on this or point me to an appropriate example?
PS: I'm sorry, if I sound confused. It is because I really am.
I am attaching the mentioned file: Lib_Elasticity_u.pro
Hello,
I am using getdp for a modal analysis of a motor stator and housing. Currently I have a model with first order tetrahedra and 24385 nodes. I need a mapping from Node Number (1 .. 24385) and DOF (one of {UX, UY, UZ, ROTX, ROTY, ROTZ}) to DOF number (1 .. 142278). I should be able to generate that mapping from the .pre file, but don't understand the format (Appendix A.2.2 is no big help either). At least the number of all DOF (constrained and non-constrained) is
146310 = 6×24385 = 6×nnodes. Thank you in advance for your help.
Hello,
I am using getdp for a modal analysis of a motor stator and housing. Currently I have a model with first order tetrahedra and 24385 nodes. I need a mapping from Node Number (1 .. 24385) and DOF (one of {UX, UY, UZ, ROTX, ROTY, ROTZ}) to DOF number (1 .. 142278). I should be able to generate that mapping from the .pre file, but don't understand the format (Appendix A.2.2 is no big help either). At least the number of all DOF (constrained and non-constrained) is 146310 = 624385 = 6nnodes.
Thank you in advance for your help.
Hello,
I am using getdp for a modal analysis of a motor stator and housing. Currently I have a model with first order tetrahedra and 24385 nodes. I need a mapping from Node Number (1 .. 24385) and DOF (one of {UX, UY, UZ, ROTX, ROTY, ROTZ}) to DOF number (1 .. 142278). I should be able to generate that mapping from the .pre file, but don't understand the format (Appendix A.2.2 is no big help either). At least the number of all DOF (constrained and non-constrained) is
146310 = 6×24385 = 6×nnodes. Thank you in advance for your help.
Hello,
I am using getdp for a modal analysis of a motor stator and housing. Currently I have a model with first order tetrahedra and 24385 nodes. I need a mapping from Node Number (1 .. 24385) and DOF (one of {UX, UY, UZ, ROTX, ROTY, ROTZ}) to DOF number (1 .. 142278). I should be able to generate that mapping from the .pre file, but don't understand the format (Appendix A.2.2 is no big help either). At least the number of all DOF (constrained and non-constrained) is 146310 = 624385 = 6nnodes.
Thank you in advance for your help.
Dear getdp folks,
I need some guidance on how to obtain the stiffness and mass matrices for a linear elasticity eigenvalue problem (presumably as a postprocessing step).
I need to transform them into modal coordinates.
Thanks in advance!
Thank you! This fully answers my questions. I will now close the issue.
Upon closer inspection of the file_mat_sysmat.pos{1,3}.m I am assuming that the file_mat_sysmat.pos1.m file is the stiffness matrix and the file_mat_sysmat.pos3.m is the mass matrix. Could you please confirm? Many thanks in advance for your help.
Thank you for your comment. I still have some questions:
Thank you so much for your help and sorry for being so dense.
Sorry to create a new issue.
I would still like an answer to #82. What is exactly in the *.m and *.bin files? Whom could I please contact by email? Many thanks in advance. Have a great day.
I would still like an answer to this one. Whom could I please contact by email? Many thanks in advance. Have a great day.
Observing the getdp's output it seems most time in spent in "Postprocessing(Compute)".
I noticed the following files in the simulation directory:
-rw-r--r-- 1 alexandershendi0 alexandershendi0 169M May 11 04:33 file_mat_sysmat.pos1.m
-rw-r--r-- 1 alexandershendi0 alexandershendi0 56M May 11 04:33 file_mat_sysmat.pos1.m.bin
-rw-r--r-- 1 alexandershendi0 alexandershendi0 57M May 11 04:33 file_mat_sysmat.pos3.m
-rw-r--r-- 1 alexandershendi0 alexandershendi0 19M May 11 04:33 file_mat_sysmat.pos3.m.bin
-rw-r--r-- 1 alexandershendi0 alexandershendi0 3.2M May 11 04:33 file_sol_sysmat.pos.m
-rw-r--r-- 1 alexandershendi0 alexandershendi0 3.2M May 11 04:33 file_vec_sysmat.pos1.m
-rw-r--r-- 1 alexandershendi0 alexandershendi0 3.2M May 11 04:33 file_vec_sysmat.pos3.m
The .m-files seem to be MATLAB code for generating a sparse matrix. But it isn't clear to me what the .bin files are for and what is the difference between *.pos.*, *.pos1.* and *.pos3.*.
Many thanks in advance.
I have added the model files below. This is also relevant for issue #82 (closed).
Meshing: gmsh motor3D_001.geo -3 -order 1 -format msh1
Solution: getdp motor3D_001.pro -solve Elasticity_u -pos post_modal_002 -v2
Lib_Elasticity_u.promotor3D_001.promotor3D_001.geomotor_common.pro
I have checked the Linux machine. It is a Lenovo Ideapad Duet Chromebook with ONELAB running in the ChromeOS Linux container/VM/whatever. But it seems to have libatlas installed (I installed PETSc and SLEPC as binary packages, but compiled ONELAB myself). I think this is an optimised BLAS/LAPACK.
Dear getdp folks,
I need some guidance on how to obtain the stiffness and mass matrices for a linear elasticity eigenvalue problem (presumably as a postprocessing step).
I need to transform them into modal coordinates.
Thanks in advance!
Sorry, I have to ask another question.
I have added:
Print[Sys_Mec, File "sysmat. pos"];
after SaveSolution[Sys_Mec];
but it generates only an empty file 'sysmat. pos'.
I also tried to do so in the PostOperation
section, but getdp complained about Sys_Mec
being not known.
Sorry for being so dense.