Comments

• [Deleted User]

  April 2016

  Nice set of improvements Victor, Congrats
• Victor

  April 2016

  Cheers prop_design. And thanks for your help on the gyroscopic forces.
• Sergio

  April 2016

  Great work Victor!
• apadzak

  April 2016

  Hi Victor, just had my few first ettempts done and CCX was my stop point. This time for a change it says I have access denied to the path where the CCX.exe is. Avast scanned MY and all went ok so I am wondering why?
  Hope I am not the problem creator.
  Thanks,
  apadzak
• apadzak

  April 2016 edited April 2016

  I figured it out. The problematic bit was tick in the "Use the same directory as LIML file" in the options, where CCX is being set up. Solving samples from Mecway instalation directory with CCX still creates error as above. Solving another model in other than MY instalation directory was successful.
  I suspect it is something to do with administrator priviliges, which I normally need to confirm if I want to save something in the c:/Program Files.
  Thanks,
  apadzak
• Sergio

  April 2016 edited April 2016

  Can you allow to modify the angle for paint selection surfaces, and expand the selection automaticly without needed of dragging the mouse (just picking one face)? Maybe for next v6 release?
  
  About the arrow keys for moving between increments of nolineal analysis, as they work also for pan the model in the view, maybe could be better to add an specific shortcut (shift + arrows keys?)
  
  For the lab option for reduced integration element, can be moved to something in the model file and not in the program? Is very easy to forget activated between sessions and finishing making standard analisis with reduced elements. Maybe a warning in case of launching a lineal analysis having that activated.
  
  On CCX nolineal imported results, can be limited the scale of the timeline to the real computed increments and not the desired number? Guess that V4 work like that, now is a little confusing to have the cursor over noncomputed increments
  
  I have found that if I select Components in the results tree (CCX imported), I can have a visualization of the deformated model without results (every component whit his own mesh color). Can be removed the scale of colors then in this visualization mode?
  
  For all the rest... amazing job :-)
  
  Regards
• Victor

  April 2016

  Thanks for the feedback Sergio and apadzak.
  
  For "Use the same directory as LIML file", i'll try to clarify that and the error message. Most people should leave it turned off.
  
  For angles, I hope to avoid having to enter numbers. You can zoom in and out to change the size of the region it includes.
  
  The arrow keys follow the common Windows convention of acting on the control that has focus. This is to help usability because it's familiar.
  
  Labs options are just quick hacks so you have to take care with them. When/if I find out a suitable place to put it, it'll become a proper persistent setting saved with the model.
  
  For the time line, I agree it's misleading. Hopefully I'll get that improved for a future version but not this one.
  
  Color scale - good idea.
  
  
  
• kwip

  April 2016

  Victor, thanks a lot! Great work again.
• apadzak

  April 2016

  Hi Victor, I am not sure how new that idea is but I am wondering how easy would that be to make Mecway to use a spreadsheet to read material values? It is quite convinient as you can download material data from material websites as Excel and it is much easier to manipulate data stored in this way.
  Thanks,
  apadzak
• Victor

  April 2016

  Good suggestion. Importing material data is something I'm already hoping to do for a future version.
• apadzak

  April 2016

  I am sorry I am doing it now and not few months ago but life was too fast for me to remember... I have spoted that Electrical Energy Density in 3D Static Analysis is shown in [Pa], SI units for it is [J/m^3] if I am not wrong. I saw it in v4 but completely forgotten to say - sorry. Z88 Aurora stores material data base in txt files which one can manipulate via Z88 interface. It frees Linux users from MS Office but maybe using Open Office format could be an answer, since Mecway is becoming Bi Operational on systems. BTW, Z88 Aurora authors have made Topology Optimization tool, which will be GNU licence I think, 3days to premiere of Z88 Akron. I am not suggesting anything but it could be a powerfull interlink to Mecway to have such heavy guns CCX and Topology Optimization.
  Thanks,
  apadzak
• Victor

  April 2016

  Pa and J/m^3 are the same thing but I agree it would be clearer if the more intuitive units appeared. This can lead to complications where people who see it from different points of view have a different idea about what's the obvious way to write the unit. For now I've just gone with a bare minimum and tried to avoid redundancy. There's a similar issue with moment per unit length which has units of force (eg N) but feels like it should be eg N.m/m. Those changes will have to wait for a future version though, as will any non-trivial new features that aren't vital.
  
  In my view, adaptive refinement has a higher priority than topology optimization. So that's way down the track.
• apadzak

  April 2016

  Thanks for clarification. Adaptive refinement sounds very good! I will get quite then, not to disturb you
  
• apadzak

  April 2016

  Me again Victor, I just realised that all of the manipulation geometry tools have copy optino but scale doesnt. I just hoped to do scale and copy in one step and not will have to work arround. From user point of view sounds like easy and usefull change... is it?
  Thanks
• Sergio

  April 2016

  Hi Víctor, with adaptative refinement you meant remesh in stressed areas automatically?
• apadzak

  April 2016

  Oh! I forgotten, there is workarround this (move+copy) but it it requires a bit of typing and thinking (time spending) to choose cylindrical coordinate system when it comes to such geometries. Personally I think it could enhance user interaction.
  
• Victor

  April 2016

  Good point about consistency with scale/copy. I'm reluctant to change old features after the beta though. Please put new change requests in another thread so this one can be about the beta version.
  
  Yes, automatically remeshing areas with more error/etc. Not to say I'll get there any time soon but I think that's more useful than, and perhaps a necessary part of, topology optimization.
  
• apadzak

  April 2016 edited April 2016

  OK, my v5beta version generates quite strange output -not really realistic. I must admit I haven't spend a lot of time on it and I run out of 1000 nodes which doesn't probably help as well with regards to elements quality. It is Hyperelastic non linear analysis with contact. I have used Money-Ravlin values, CCX contacts and rigid body motion (it says it isn't allowed but does the job).I have played with "Slope of pressure-overclosure curve K" values (1 then 10 then 100), I changed convergence tolerance to more conservative from 0.1 to 0.001 (it didn't affect solution time that much). I am not sure about D1 value in M-R material model as I have it in [MPa] and Mecway has it in [m2/N]? If that was the other way around then units would be consistent [N/m2]=[Pa] but in this case I am a bit lost. My D value from a paper is "1/(332MPa)" and if I was to stick it in Mecway I would do "1/(332MPa)"=3,012048e-3 and to get it to [Pa] I would multiply it by 1x10e6. But I am not sure if this is what MY wants . I have two models as in one case I just wanted to have exact geometry of the simulated hyperelastic body so I replaced contacting rigid bodies with pressure and other files has got clamping bodies but the volume was reduced by half to satisfy 1000 nodes limit (I could have done it in other way but deleting was the quickest option). As I said, not much time to consider what could cause the problem but both of the models don't behave as expected. Analysis set up though was much quicker than in ANSYS, I am still trying to overcome few silly issues there
  Any suggestions welcome!
  Thanks,
  apadzak

  O_ring_1a.liml 65K

  O_ring_1.liml 96.5K
• Victor

  April 2016

  Thanks for trying that.
  
  The main problem is D1 is too big. In this case it should be 3e-9 m^2/N (=1/(332MPa) ). I'll try to clarify what that means in the program and perhaps include 1/MPa because different people use different definitions for their (in)compressibility constants. Thanks for pointing out this confusing aspect.
  
  When parts pass through each other like that, it can mean the pressure-overclosure slope is too small. Unfortunately this is a tricky parameter that you need a bit of trial and error to tune. I found 100 000 Pa/mm works better but it somehow explodes near the end of the time steps.

  O_ring_1a_fixed.liml 65K
• apadzak

  April 2016

  Thanks a lot I will have a play later to get this better. I have decreased the timestep so maybe this will lead to better solution... I will see. Thanks for help.
• apadzak

  May 2016

  Hi Victor, I have recently came back to that O-ring example and have added selfcontacting element faces. RIght now I have 3 CCX contact features in the tree and there is no way to assign their names. Is it possible to include such option? More contacts in more complicated models may lead to chaos if one gets back to analysis after some time and will try to figure out what is what.
  Thanks,
  apadzak
• Victor

  May 2016

  Thanks for the suggestion. I'll add it to the list for future versions.
• Victor

  May 2016

  Updated to version 5 beta 2:
  
  http://mecway.com/download/app/mecway50beta2.msi
  
  - Improved scaling on high resolution (high DPI) displays.
  - Components display in solution view doesn't show a color key
  - Automatic component colors aren't chosen to be so close to the blue selection color
  - CalculiX working directory option is clearer
  - Removed Ogden hyperelastic material
  - Some display bugs fixed
  
• Sergio

  May 2016 edited May 2016

  Hi, I just tried the new beta, very nice the new visualization mode for results (deformed / no results), I found usefull for analysis with very big deformations (soft materials), and for modal analysis as well.
  
• DaveStupple

  May 2016

  Victor, scaling now much better, buttons are now a sensible size. I was just about to feed back on some other issues, but I think the Beta 2 has them sorted, the main one being an annoying flicker when selecting.