General Mechanical

General Mechanical

Export prestressed stiffness matrix of a sector model with cyclic symmetry

    • biao.zhou

      Hi everyone,

      I've got a problem with exporting the prestressed stiffness matrix of a sector model with cyclic symmetry. I'm looking for the way of recovering the condensed cyclic interface DOFs in the global stiffness matrix.


      For a non-prestress sector model, I managed to export its full stiffness/mass matrix by running a modal analysis for the FREE model without constrained DOFs and cyclic interfaces and then by reading the matrix from the '.full' file.




      For the sector model in the spinning condition, I have to firstly run a cyclic static analysis with the cyclic status on and constrained DOFs in the hub.



      OMEGA,0, 0, Vel_Rot_RAD
      ! BOUNDARY CONSTRAINTS: selected nodes at the disk hub 
      D, ALL, ALL, 0 

      In the subsequent modal analysis, I tried to set the cyclic status off but failed. The constraints in the hub are deleted. Then with the cyclic status on, the exported stiffness/mass matrix only covers a part of the nodes in the sector model. It seems that the DOFs attached on one of the cyclic interfaces are condensed out in the global matrix. Moreover, it seems the node mapping with the exported matrice do not match the original sector model.

      So I'm looking for the way of exporting the full prestressed stiffness matrix with all the DOFs in the original sector model (free body) included.
      Any suggestions and comments will be highly appreciated.
    • Bill Bulat
      Ansys Employee

      Are you using MAPDL directly of Mechanical with command objects?

      I'm not perfectly clear on what you want to do or if it can be accomplished. When you SOLVE after the CYCLIC command has been issued, the first things that happen (before the solution is calculated) is a duplicate sector of the original sector is created and I think too that constraint equations are internally and transparently created that relate the nodal DOFs on the periodic sectors. Looking at the documentation of the CE command,

      "The first unique degree of freedom in the equation is eliminated in terms of all other degrees of freedom in the equation."

      So I would imagine that DOFs of at least some of the nodes on the periodic sectors will not be included in the system stiffness matrix.

      Do you want a stiffness matrix of only the original sector or do you instead want one including both the original and duplicate sectors? I think the latter is possible, though my guess is that terms for some of the nodal DOFs on the periodic sectors will be missing.

      Kind regards,


      • biao.zhou

        Hi Bill,

        Thanks for your response. Yes, I am running the MAPDL codes directly. I want to obtain a stiffness matrix with all the dofs in the original base sector included.

        Let me explain clearly how I attemped to do that.

        % The reference sector model in Ansys firstly undergoes a cyclic static 
        % analysis in order to include the rotation-induced prestress effect.

        % Then a linear perturbation modal analysis for the sector model with
        % cyclic symmetry and without any constrained dof is performed in order to
        % export the sector-level mass/stiffness matrix (of only the original base sector) for % the following model reduction research. 

        After the second step, Ansys gives the exported matrices of the original base sector. However, a part of the DOFs attached onto the cyclic symmetry interfaces are condensed in the exported K and M matrix. More specifically, both the high and low cyclic interface dofs are partially retained in the global matrix. 

        Now I'm trying to write some codes in order to read the imcomplete sector matrix and further to restore the terms related to the condensed cyclic interface dofs.

        But this is not a easy task. Hope I can make it.


Viewing 1 reply thread
  • You must be logged in to reply to this topic.