Some issues generating inp file

AndreaAndrea
Hi to everyone! I'm a structural engineer and I'm a CCX user from some years. I tried MecWay for generating .inp file to perform nonlinear analysis. I saw that Mecway doesn't write all the parameters:

First increment, step duration, minimum increment and maximum increment. Follow an example

*Static, SOLVER=SPOOLES
     0.10000,     1.00000, 1e-05, 1.0

Other problems is that Is not possible to manage the maximum number of increments in the step:

*STEP,NLGEOM=YES,INC=100 (Written by Mecway)

For large plastic strain is important to register more than 100 increments

Other aspect is the application of *AMPLITUDE card to *CLOAD. Mecway builds ad *AMPLITUDE for every nodal load but is possible to apply the same *AMPLITUDE card defining only one *CLOAD for all nodal loads applied to a surface:

*Cload,OP=NEW, AMPLITUDE=AMP1
      5,1, -0.56395427E-13
     59,1,   24.000000
     60,1, -0.50931698E-13
     61,1,   24.000000
     62,1, -0.52086157E-13
....................................

• Node number or node set label.
• Degree of freedom.
• Magnitude of the load

Other problem for me are the units. I prefer a unitless inp file. So one unit must be 1 mm. Mecway writes it in meters and al results are in Pa (Using CGX). Obviously I prefer MPa.

I have others advices: for example is important to compute the overall forces applied to the model before performing the analysis. It is important if one works on model with minimal supports and balanced loads.

I have other advices about how to build material models to solve in CCX....next time

Comments

  • SergioSergio
    edited November 2016
    Hi Andrea, welcome to the forum!

    About the step and static card, you can define any quantity of (fixed) increments by means of "Number of time steps" and "Time Step" on the Analysis settings window, then you must define your loads/displacements with a time function or a table taking in count this two values. For example I normally choose 100 for number of time steps and 0.01 for time step, that's gives me a total step time of 1. Then I use a table for ramping the loads from 0,0 to 1,1 (100% of load at time 1). Once you catch the idea is very simple and straightfor.

    Now if this is not enough, you can add your own STEP or STATIC card by means of "Custom model definition" and removing the Mecway owns cards (look at the CCX branch on the three). All within the program interface. Those custom CCX options are very powerfull and allow you to make any change you want to the input file, even adding multiple steps to the analysis. Abaqus CAE have a similar tool for editing the input files, as the solvers normally gets new options before they implement in the interface.

    About AMPLITUDE... yea is not gracefull the way it writed, as you say you can use the same amplitude for several loadings (the same as for contact definitions). But it works. If you would like to use the amplitud as you did with handmade input files, again you can just define groups of nodes/faces on the interface, and apply all your boundary conditions by means of custom CCX content. I did it that way in the past until I feelt comfortable with the way the input files are writted.

    About units, I'm in the same position as you, years working with mm and Newtons... but was years living in a mistake, the unit for displacement in the SI is meters, not mm! the mm and Newtons are an exception (don't know how to say it) for mechanical guys, but true is that the displacement unit is meters. Anyway, would be great to have an option to write input files in mm/N/s unit, I recognize that had a lot of problem with units for hyperelastic or hyperfoam coeficient materials where they are not so clear or direct as a simple stress unit.

    About overall forces applied on the model, in NX there is an option to sumarize all the applied loads, for some complex models with several distribuited loads on different places, is good to make this check to avoid mistakes before launching the solver. There are also some customers that ask you for a table comparing the input loads against the reaction forces on the fixation points, so yes, an option to compute this total loads would be great.

    About material models... well the only thing that I miss is the possibility of entering the stress/strain data points (hardnening rule) for plasticity analysis, but again can be done easily with custom editions.

    Regards!



  • AndreaAndrea
    Thanks for your answer
  • AndreaAndrea
    edited November 2016

  • SergioSergio
    edited November 2016
    About units on input files... if we ask for mm why not somebody could ask for inchs? Guess that this could lead to a nightmare for the developer.

    The new version just arrived, let´s give some breack to the developer. Next week we can start to ask new features :-)
  • AndreaAndrea

    Thanks for your answer. I looked at some video tutorial
    about Mecway on youtube and I saw how to manage the inp file direct from the
    pre-processor. Obviously is a good way but I'd like a pre-processor fully
    integrated with CCX...obviously is my desire but I think that many CCX user
    could desire this macros and others. About units I have a curvature on
    structural steel and so I used always N/mm/kg/s for all the parameters. I know,
    for example, that steel density on this set of units is uncomfortable: 7,85 x 10e-6
    kg/mm3 instead 7850 kg/m3 but is only a practice aspect.

    I bought a commercial license of Mecway this morning and I
    hope step by step I will understand how to use it in a better manner. I used
    Roshaz for some years and it was a very good project and some capabilities (all
    quad mesh - loads summation before launch the solver, Gencoz force distribution
    and so on) could be a developement for Mecway.

  • SergioSergio
    edited November 2016
    Great, I use Roshaz also for quad meshes, that after I import in Mecway to create hexas. Do you know if is possible still to buy a full licence of Roshaz? I have looked for a simple quad meshes unsuccessfuly the last months.
  • VictorVictor
    Thanks for those answers Sergio, and the suggestions Andrea. I'll just add a little rationale:

    I don't intend Mecway to be a tool for making .inp files. Rather, a self-contained FEA program that can optionally use the CCX solver. That means the .inp file is just an internal detail and not meant to be manually read nor the solution used with the unitless CGX. So ideally, nobody would care what units it used because nobody would be using the raw unitless data. Likewise, nobody would care about ugliness in the layout of the .inp file. However, I realize that people still need to work on it manually to do more advanced things that Mecway can't do, and that's a bit of a problem.

    Having a separate *AMPLITUDE block for each node is because there are incompatibilities between CCX and Mecway that still require multiple *AMPLITUDE blocks for the same load where multiple loads overlap, so it's easier to just separate all of them regardless instead of searching for special cases.


  • AndreaAndrea
    I don't know...sometimes I look at the website but all is still at 2011. I think that project is dead and so I decided to contribute to Mecway.
    I must say that after 7/8 hours in Mw I appreciate the simpicity and the quick learning curve. For the rest I have to working manually on the inp file like first because Roshaz doesn't support latest CCX capabilities: for example considers contacts in a manner that is not longer supported by CCX (set of node VS set of elements).
    I like that this communities is active and the project goes ahead.

    Greetings
  • rhokarhoka
    hi,

    I have minor suggestion regarding default units saved by MW, and I agree with Andrea due to fact most FE analysis are for micro models in dimension not for large/macro models - default small dimension units e.g mm-MPa are best suit for this case. so it can be quickly adapt for test data of material properties and results.

    another opinion, each engineer usually used one consistent units (eg. SI/MKS/Kip-in) and using one setting for all input box (coordinate, material & loads) are best approach to reduced repetitive task.

    currently MW have different unit setting for each input value in dialog box, look it's rare condition.
  • VictorVictor
    After you use a unit in a model, it should remember what you prefer for other values with the same quantity. You can also create a starting model that has a dummy material with the units you want. Then if you build a new model off that, new materials will default to your preferred units.

    The different coordinate units can be useful for things like feet for length and inches or mm for thickness of plate or sheet metal.
  • rhokarhoka
    great suggestion, ... using blank template with predefined units setting and material properties. thank you
  • SergioSergio
    The actual system is good enough, with the first unit selection almost all are then uniform, but still there are some places were this is not happening. There is also some unit (guess that is Nmm or something like that that doesn't give the option to enter in mm
  • kwipkwip

    A note regarding Roshaz, John Horspool, owner and sole developer of Roshaz passed away in 2011.

    R.I.P John ...so sadly this project can be regarded discontinued.

    regards.

  • AndreaAndrea

  • AndreaAndrea
    I knew....I'm not a programmer but could be interesting to continue that project. I found a lot of good ideas on Roshaz and a good algorithm for tetra free mesh, quad free mesh, fast way to apply loads (over line, line pressure, face pressure, cone/cylinder pressure, lug/pin force distribution over cylindrical and spherical faces), mesh quality check, load check before launch the solver and a good post-processor. It has some bugs never solved and it didn't import .step file. I think that last aspect was in mind of Horspool because the latest version appear step stp file but doesn't run.
    R.I.P.

  • rhokarhoka
    hi Sergio,

    MecWay has supported GMSH file format, it can be open/import directly so you can get full quad mesh of your plate/shell model and then extruded to create hexahedral mesh if you want to use solid element.

    attached pictures of my benchmark test for GMSH full quad algorithm, exciting me..
Sign In or Register to comment.

Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

Sign In Register