• Registered: 2014-06-09
  • Posts: 2

Topic: Floating Point Overflow

Hi,

I am modeling a prestressed concrete I girder, and get the error "internal analysis error: floating point overflow" halfway through the analysis.  I have altered the displacement increment that is applied to the girder and the error always occurs at the same analysis step. The construction case and load cases before and after this load step are identical.

I have read through 2.4.2 in the trouble shooting manual, and have not found any obvious errors.

What does floating point overflow mean and how could I resolve this issue?

Thanks in advance,
Alex

Quote Post 1

  • dpryl
  • CC software developer
  • Offline
  • Registered: 2007-07-13
  • Posts: 1,107

Re: Floating Point Overflow

Hello, have you already checked the convergence and results for the calculated steps? Is the deformed shape as expected? Are there any surprising cracks or plastic strains?

dpryl's Website

Quote Post 2

  • Registered: 2014-06-09
  • Posts: 2

Re: Floating Point Overflow

Ok, I noticed that the girder failed at the one the supports, and that two of the 4 convergence criteria didn't converge.  I'm assuming the error occurred due to the failure at the support.  However, this type of failure didn't occur during the physical test. Would you suggest enlarging the support?  Currently I'm using a shell element for the steel plate supports.

Alex

Quote Post 3

  • dpryl
  • CC software developer
  • Offline
  • Registered: 2007-07-13
  • Posts: 1,107

Re: Floating Point Overflow

Hello, using shells for the elastic loading/support plates usually does NOT make much sense. Normal volume brick or tetrahedra are typically the best choice there.

dpryl's Website

Quote Post 4