Close

Forum

Help with GEBT curve beam definition

  1. Alberto Fernandez

    I'm trying to get GEBT to calculate a curved beam and I'm finding some issues.

    I based my example on catilever1.dat. I modified to include an arbitrary position in space and a curvature. I'm able to solve the problem only wiht a high number of iterations. But, my main issue is with the position of elements. The beam is a circular arc. Using 5 elements for a member a get what can be seen in the picture attached. It seems that curvature is applied to the position of each element starting from the direction of the member. This makes the curvature of the beam correct but its position is wrong (the plot includes the end points of a member).

    I may be making a mistake somewhere, but I can not see where. Is this the expected behaviour?

     

    curved-beam.png

    Reply Report abuse

    Replying to Alberto Fernandez

    1. Qi Wang

      Hi Alberto,

      In GEBT, the first step is defining members by 1) keypoints 2)initial curvatures and 3) member frame at x1=0 (direction cosine matrix (DCM) at starting keypoint, the default value is given as I3, identity 3 by 3 matrix, i.e., defining in global frame)). Then the GEBT preprocessor subroutine will mesh the member and generate finite element (FE) nodes accrodingly. 

      I would suggest check your input file for the keypoints, initial curvatures, and member frame first. If this does not fix the problem, please provide more details and accurate description about your case, including your input files, to the forum so that we can help with debugging.

      Qi

      Reply Report abuse

      Replying to Qi Wang

    2. Alberto Fernandez

      I don't see any problem in my file. It passes all the checks. My question is more related to how intermediate nodes are generated on each member. It seems to me that the process involves taking the first point of the member and position them by using curvature by modifying their local z position without regard to where the end point of the member is.

      To make my point clear, I’m attaching the file used to generate the plot. As you can see there is no load applied. That way, we should get the geometry assumed in the calculation. That geometry is supposed to be an arc. In this case created nodes form an arc, but their position deviates a lot from what it should be (all y should be positive). The last point appears at the position it should.

      I hope it is clearer this time.

      Reply Report abuse

      Replying to Alberto Fernandez