FEAP User Forum

FEAP => Programming => Topic started by: fethio on August 01, 2017, 11:18:11 PM

Title: change/correction in fiber structure tensor implementation?
Post by: fethio on August 01, 2017, 11:18:11 PM

I noticed a difference in the definition of the fiber structure tensor between versions 8.4 and 8.5 exists. In 8.4 it was defined as:

Code: [Select]

      tau(1) = a(1)**2
      tau(2) = a(2)**2
      tau(3) = a(3)**2
      tau(4) = a(1)*a(2)*2.d0
      tau(5) = a(2)*a(3)*2.d0
      tau(6) = a(3)*a(1)*2.d0
while in version 8.5 it reads (apart from the name change tau -> mm)

Code: [Select]

      mm(1) = a(1)**2
      mm(2) = a(2)**2
      mm(3) = a(3)**2
      mm(4) = a(1)*a(2)
      mm(5) = a(2)*a(3)
      mm(6) = a(3)*a(1)
(Note the missing *2 in the last three items.)

Is this correction of a previous error, or is it due to the way things are handled differently in 8.5?

Title: Re: change/correction in fiber structure tensor implementation?
Post by: K.Li on August 02, 2017, 12:35:01 AM

I think mm(6) is the Voigt notation of a(3) \otimes a(3),  there is no 2 here. Check  the formulation of this model in Prof Holzapfel's paper in 2000

Title: Re: change/correction in fiber structure tensor implementation?
Post by: Prof. S. Govindjee on August 02, 2017, 10:42:18 AM

I believe that the factors of two in the expression for tau(4:6) are errors in version 8.4.  The Cauchy stresses are stored in Voigt notation so, for example, sig(4) = sigma_{12}  NOT = 2*sigma_{12} [note, the Berkeley ordering for entries 4:6].  The expressions in version 8.4 are giving incorrect factors of two in the shear stresses. 

If you are using 8.4, you should delete the *2.d0 on tau(4:6).

Title: Re: change/correction in fiber structure tensor implementation?
Post by: fethio on August 03, 2017, 01:06:22 AM

Thank you for the clarification Prof Govindjee.
My user element was based on the fiber in ver8.4, and had to be corrected based on the discussion in this post.