Version 11

Victor

Version 11.0 is now available here

https://mecway.com/download/mecway110rc.msi
https://mecway.com/download/mecway110rc2.msi
https://mecway.com/download (this is the same file as mecway110rc2.msi above)

This is a minor update with the following changes:

• Faster graphics for big models
  
• CCX updated to version 2.15
  
• Elastic support has optional tangential stiffness.
  
• "Correct collapsed elements" is now automatic.
  
• "Don’t generate keyword" for CCX allows multiple selection.
  
• "Don’t generate STEP" for CCX.
  
• Contact without faces is ignored instead of causing the solver to fail.
  
• Clearer names for Contact parameters.
  
• Mesh tools -> Node coordinates works with faces and elements too.
  
• Diagrams in elastic support, bonded contact, and loft.
  
• "Reset to default" button in meshing parameters
  
• You can delete the default component.
  
• Multi-delete for components, named selections and loads & constraints.
  
• File -> Import no longer imports solutions but now does import formulas and stress linearization.
  
• Enabled C, T, L section shapes for modal vibration 3D and dynamic response 3D.
  
• Factor of safety color range is reversed and limited to a maximum of 15 by default.
  
• Sample files with formulas for all components of stress and displacement in cylindrical coordinates.
  
• Deprecated IGES file support.
  
• Deprecated internal nonlinear solver.
  
• Removed Tochnog Professional support.
  
• Bug fixes
  

prop_design

congrats victor,

looks like some awesome new features. can't wait to try it out.

anthony

prop_design

I ran a model from v10 using the v11rc. there is a new entry in the solution area that I can't find any info about in the mecway or ccx manuals. it is called ERROR_STR_%_. this was a nonlinear stress using composite shell elements. can you explain what this is? it's max value is around 50% whatever it is.

Victor

That's an estimate of the error in the stress that newer CCX versions (2.12-2.15) generate. It's described in the "Gradient error estimator" section of the CCX manual.

prop_design

thanks. I read the help on it. rather confusing. not sure what this really is, even after reading the help. but thanks for pointing me to the right place in the documentation.

tk1537

Looks good. Is this a final or beta v.11? I wasn't sure what the "rc" meant.

Victor

It's effectively the same as a beta but if no bugs are discovered, I'll make it the final one. One already was (regression to v9 behavior of element selection on wireframe) so I've now updated it to rc2.

Once the download page is updated to show version 11, it won't change without the version number changing too.

Sergio

Congrats Victor, those are really usefull features! Is there a way that the mehs of the first part (from imported geometry) goes directly on the default component?

Victor

That probably makes sense now that default component doesn't have to be there. I'll need to check for possible side effects but I'll add it to my list for next time.

Sergio

How does work the multiple delete in the model tree? Is only the new "Delete all" entry in the menu or is possible to select using ctrl and shift?

Victor

No multi-select, sorry. Only the new menu items. Also delete empty/delete unused/etc. for named selections.

disla

Thank you for the new update.
I can see that many of the community comments have been attended.
Great job Victor.

tk1537

Victor,
I got Mecway v.10 to use all 8 cores for stiffness, spooles, and stress using the bconvergence v2.15 CCX. However, when I run the same problem in your newly posted Mecway v.11 using the imbedded CCX and the same environmental variables, I use 8 cores for stiffness and stress, but only 1 core for spooles. Is something different between the two CCX's, or do I have something set wrong?

Victor

Yes. The CCX from dhondt.de is compiled for multicore but the one with Mecway isn't. I'm surprised it still does that with stiffness and stress though. The multicore versions usually fail more test cases than single core and sometimes produce bad results with eigenvalue problems though I've yet to test if this 2.15 does too. I've included the single core version with Mecway for safety.

tk1537

Victor, Here's the text from the CCX output file for the problem I ran on your v.11 with the embedded CCX. Do you intend to test the multicore version against the problem areas you mentioned? I'd be interested in the results.

BTW, the results from both runs (v.10 w/multicore 2.15 and v.11 with imbedded 2.15) are identical for Von Mises and Displacement Magnitude.

************************************************************

CalculiX Version 2.15, Copyright(C) 1998-2018 Guido Dhondt
CalculiX comes with ABSOLUTELY NO WARRANTY. This is free
software, and you are welcome to redistribute it under
certain conditions, see gpl.htm

************************************************************

You are using an executable made on Mon Feb 11 18:46:00 2019

The numbers below are estimated upper bounds

number of:

nodes: 33984
elements: 22336
one-dimensional elements: 0
two-dimensional elements: 0
integration points per element: 4
degrees of freedom per node: 3
layers per element: 1

distributed facial loads: 112
distributed volumetric loads: 0
concentrated loads: 0
single point constraints: 507
multiple point constraints: 1
terms in all multiple point constraints: 1
tie constraints: 0
dependent nodes tied by cyclic constraints: 0
dependent nodes in pre-tension constraints: 0

sets: 2
terms in all sets: 44784

materials: 1
constants per material and temperature: 2
temperature points per material: 1
plastic data points per material: 0

orientations: 0
amplitudes: 2
data points in all amplitudes: 2
print requests: 0
transformations: 0
property cards: 0

*INFO reading *STEP: nonlinear geometric
effects are turned on

*WARNING reading *STATIC:
the minimum increment 0.0000000000000000
is smaller then 1.e-6 times the
step time;
the minimum increment is changed
to 9.9999999999999995E-007
which is the minimum of the initial
increment time and 1.e-6 times the step time

STEP 1

Static analysis was selected

Newton-Raphson iterative procedure is active

Nonlinear geometric effects are taken into account

Decascading the MPC's

Determining the structure of the matrix:
number of equations
101445
number of nonzero lower triangular matrix elements
3989742

increment 1 attempt 1
increment size= 1.000000e+00
sum of previous increments=0.000000e+00
actual step time=1.000000e+00
actual total time=1.000000e+00

iteration 1

Using up to 8 cpu(s) for the stress calculation.

Using up to 8 cpu(s) for the symmetric stiffness/mass contributions.

Factoring the system of equations using the symmetric spooles solver
Using 1 cpu for spooles.

Using up to 8 cpu(s) for the stress calculation.

average force= 0.011443
time avg. forc= 0.011443
largest residual force= 0.000025 in node 32028 and dof 2
largest increment of disp= 7.460448e-07
largest correction to disp= 7.460448e-07 in node 2097 and dof 2

no convergence

iteration 2

Using up to 8 cpu(s) for the symmetric stiffness/mass contributions.

Factoring the system of equations using the symmetric spooles solver
Using 1 cpu for spooles.

Using up to 8 cpu(s) for the stress calculation.

average force= 0.011444
time avg. forc= 0.011444
largest residual force= 0.000000 in node 30133 and dof 2
largest increment of disp= 7.460448e-07
largest correction to disp= 7.058205e-11 in node 2097 and dof 2

convergence

the increment size exceeds the remainder of the step and is decreased to 0.000000e+00

Using up to 8 cpu(s) for the stress calculation.


Job finished

Victor

Yes, at some stage, I'll look into those earlier reported eigenvalue problems. Just a note that when it says "Using up to 8 cpu(s)", that range does include 1 and it used to be only 1 for the single core version.

Victor

I've reproduced the problem that the ccx.exe 2.15 on dhondt.de can have obvious random errors with modal vibration (*FREQUENCY) that the single-threaded version included with Mecway doesn't. Here is a solution from the multithreaded one on the left and the single threaded one on the right.

I had the environment variable OMP_NUM_THREADS = 8 and my CPU has 8 logical cores (4 actual cores). It worked OK with OMP_NUM_THREADS = 4, 2, and 1. So it might be OK as long as you set that variable appropriately.

Victor

@Sergio

> Is there a way that the mehs of the first part (from imported geometry) goes directly on the default component?

I'm thinking it'll have to stay the way it is because it's hard to know the user's intent for that default component. If they're already using it for something or are planning to, and then it gets clobbered by a new one, that might be unexpected. If Mecway checks for any use of it like having elements, been renamed, having material, changed color, use in Table, etc. then the behavior would be hard to predict because you might sometimes unknowingly change something that causes it to be retained.

Ideally, I'd rather have no default component and one just gets created when needed, but you can't create a material without it.

Sergio

@Victor, have been working in another complex skid this week, and I decide to don´t delete (and not use for any component) the default one. Guess that was a good move as in this way I know exactly where new elements will be created (on the visible default). As you say probably a workaround will be having a "destination component" in the element creation windows, but surely most of times will be setted on the default one.