Implementation of the Arclength Method in FEAP

[Flex]

Hi guys,

I'm working with FEAP for a while now and wanted to implement an arclength method with an own constraint function. In order to do so,
I wanted to understand the code arclen.f as it is implemented in FEAP. I do understand how the routine works, but I stumbled across this block:

c Arc Length (unscaled displacement vector)
 alfa   = dot(du,u1,neq)
 alfa1 = dot(u1,u2,neq)
 if (abs(alfa1).gt.tol*abs(alfa)) then
  alfa = -alfa/alfa1
 else
  alfa = -alfa/(alfa1 + sign(tol, alfa1))
 endif

My main question would be:
Is this the arclength implementation following Riks (1972) as it is described for example in Wrigger's " Nonlinear Finite Element Methods"? Is there any other literature, where this implementation is mathematically formulated?

Thanks in advance!

[Prof. R.L. Taylor]

This is a standard arclength part, the only modification is the second option where the tol is violated.  This is to prevent wild changes in the alph value.  I do not currently have access to the references but assume it is as described by Wriggers.

[ilinca]

I do not have the book with me right now, but there is also the paper
Schweizerhof, K H, and P Wriggers. “Consistent Linearization for Path Following Methods in Nonlinear FE Analysis.” Computer Methods in Applied Mechanics and Engineering 59, no. 3 (1986): 261–79.

[Flex]

Thanks for the fast replies! I will look it up right away.

[A.Harandi]

I wanna use arcl for displacement control situation However when I put arcl,,5 or arcl,,4 in my input file the simulation stops and it appeared that no load has applied.
Is there anyone here who uses arcl with a displacement control problem with step by step load application?

[Prof. R.L. Taylor]

The use of displacement control requires additional data.  If you first do the check in interactive mode it will tell you what data is needed.  It appears the data should be this:

BATCh
   ARCL,,<4,5>
END
 node dof disp_inc