Jump to content

Mr. Konstantin Kreutzer

Members
  • Content Count

    5
  • Joined

  • Last visited

Posts posted by Mr. Konstantin Kreutzer


  1. Hi Arne,

     

    thanks for answering that quickly. If there is nothing like an instant-solution, I think it would be an overkill to add an extra-option. In this case I just have to get better in predicting the calculation time. But anyhow, thanks for the offer!

    Concerning my workflow: Provided that your computation power is limited as well as your amount of time, wouldn´t it be the best to run a Sobol with a number of variants that exceeds (in the best case just a bit) the expected calculations which can be computed in the given time?

     

    Best regards,

    Konstantin


  2. Hi everyone,

     

    since we are planning to run various optimization processes with a respectable number of variations it would be great to use two or more processor-cores.

    The way we are doing it right now is to enable asynchronous update, which works fine. There is just one tiny inconvenience left.

    Activating asynchronous update leads inevitably to CAESES preparing all the variations before starting the actual calculation. It means that one has to wait for CAESES to stop the remaining "stil-to-evaluate"-designs after canceling the overnight-calculations in the next morning (using a conservative number of expected designs because the calculation time is not always very predictable). This can consume a bit of time.

     

    If there is an easy (maybe one-click?) solution to start a Design-Engine using multiple cores without creating all the designs in advance?

     

     

    Thanks!

    Konstantin


  3. Hi all,

     

    I am trying to program an automatic framewise parametrization of the ship hull. The overall structure of my solution is divided in several sub-problems:

     

    -LOOP (a sufficent number of frames in x-direction for a smooth parameter-curve)

          1.)Automatic generation of frames from the IGES-import via intersection curves with single surfaces, collected in one polyline

          2.)Generation of the parameters from the frame

          3.)Add the actual value of parameter to the parameter-curve which represents the parameter-distribution over the x-direction

    END LOOP

     

    -Parametrize the parameter-curves (which will be the design parameters)

     

    -Create meta-surfaces from the parametrized parameter-curves.

     

    For the first step in the loop, I need a way to get acces to the IGES-surfaces inside a feature-definition. Is this somehow possible by setting a scope as input for the feature or is there a way to acces the project-objects directly in the feature definition?

     

    Thanks!

    Konstantin

×
×
  • Create New...