• State Verified Answer
  • +1 person also asked this people also asked this
  • Replies 3 replies
  • Answers 1 answer
  • Subscribers 29 subscribers
  • Views 1933 views
  • Users 0 members are here

RMBRIDGE Calculation Time.

Pablo Bernal Sahun
Offline

 

We are working in a large model wich requieres about 15 minutes to run without CS and 2.5 hours with CS (internal formulas). It is a cable stayed bridge linear construction stage analysis with a total of 1003 stages.

We have reduced the CS load cases to a minimum because we have seen CS are demanding most of the calculation time.

Also we have tried to reduce the number of elements that have CS using TStop for the elements where this action can be neglected, but we are not sure if using TStop reduces calculation time in an effective manner or there is another way to avoid CS calculations in certain elements.

Is there an alternative method to avoid CS calculations in some elements?

Any other idea to reduce calculation time?

Parents
  • +1
    Hi Pablo,
    I think you have already had some good points like using "internal formulas" which can have a big influence on the calculation time. In general creep can be a crucial factor for the calculation time in RM Bridge as it uses a very accurate calculational model for the creep calculation: So for example each element has its own time axis depending on its age and activation time. Further as creep itself causes additional load effects, these calculated loads are again considered in addition for the consecutive creep calculations (i.e. “creep of the creep of the creep…”). Moreover creep is interacting with the shrinkage process etc. So it is clear that the amount of stages and the number of elements have a strong influence on the calculation time of creep.
    I don’t know which version you are using, but there have definitely been some improvements exactly in regard to this in the recent time. Also RM Bridge can use more cores/threads by now.
    And by the way 2.5 hours doesn’t actually sound like such a long calculation time as there can be projects that have to be run over night as well…
    I hope I could at least illustrate a bit the background of the issue.

    Answer Verified By: MarcosBeier 

  • 0 Offline in reply to Diether
    Thanks Diether, I agree 2.5h is not too much for this complex model. However there is a big number of elements in my model and creep of some of these elements is negligible. Maybe calculation time could be improved by skiping creep and srinkage calculations on these elements. Is there other way to avoid creep ans srinkage calcs on some elements than TStop command?
Reply Children
  • 0 in reply to Pablo Bernal Sahun
    Hi Pablo,
    I am afraid there is no effective way of reducing the calculation time in regard to what you explain. There would only be a workaround of excluding certain elements from the creep calculation when using the time dependent functions by assigning a different material group to these elements where the functions are set to zero. However, in general the approach using the time dependent functions would take considerably more time as when using the approach of the internal formulas, so that there wouldn't be a win of time in the end, but probably the other way around.
Communities
Support and Services
Training and Learning
Social Media