No announcement yet.

LUSAS, arc-length

  • Filter
  • Time
  • Show
Clear All
new posts

  • LUSAS, arc-length

    Hello, my name is Adrian and I'm a spanish civil engineer. I am analysing a reinforced concrete shell with the british software LUSAS and I'm trying to follow all the arc-length (Crisfield) iterations during a geometrical and material non-linear analysis, in order to understand perfectly all the process and calibrate further modifications.

    My problem arrives when tracking variables DELTL and DLMDA in the report given by the program after running the model. I run it with an automatic load incrementation, under Crisfield arc-length algorithm, and I realized that variable DELTL (radius of the arc-length for each load step) doesn't match with the next formula, but seldom it does:

    Increment Li = Increment Li-1 * (Nd/Ni-1)^0.5 [Lusas Theory Manual Volume 1, page 42]

    (is the automatic load adjustment defined by Crisfield in his paper "Solution strategies for the delamination analysis based on a combination of local-control arc-length and line searches")

    It basically means that we are changind the radius of the new arc of the new load step, as function of the precision achieved converging the solution on previous load step. The precision is respect to a preset (input) number of desired iterations to converge.

    I understand that inside a load step, the way DLMDA varies until reaching TLMDA to jump to the new step, depends on DELTL .

    Summarising, my question is, if does anybody else noticed that DELT is not always following the formula for "Automatic Load adjustment" and what is it doing in those cases. (Actually is not doing a "line-search" since coefficient ETA remains constant and equal to 1 (default) at any time).

    Thank you very very much, and apologize for this long paragraph.

  • #2
    I am afraid that I am not familiar with LUSAS. Perhaps someone else can provide feedback...

    Jorgen Bergstrom, Ph.D. PolymerFEM Administrator