Version 17.0

Victor

Mecway 17.0 is now available at the link below. As usual, it's a free update for all existing license holders.

https://mecway.com/download/mecway170.msi

Spanish language manual thanks as always to Germán Bresciano.

Changes:

Internal solver
All the same stress and strain field variables are available for solids and shells, including laminate layers, except for in-plane principal stress.
In-plane principal stress deprecated. Use Principal stress instead which has all 3 values.
In-plane principal stress is no longer displayed on both shell surfaces together.
Solver responds to the cancel button faster.
Von Mises failure criterion for shells includes transverse shear stress.
Bending moment (legacy) is no longer generated when solving.

CCX
Suppresed loads and constraints applied directly to nodes/faces/elements no longer have their named selections written to .inp files.
ERROR_STR_% excluded from CCX solution using the NOE variable in *EL FILE.
Most errors generating the .inp file for CCX are now fatal and cause CCX not to be started.

GUI
Contour plot max/min has an option to make the range symmetric about zero.
Displacement and Node rotation show “Constrained direction” and “Constrained axis” respectively for clarity.
Relabelled Maximum element size to Maximum element size at surfaces in STEP file meshing parameters.

Automation
New API functions: named_selections() and components().
Command line accepts filenames outside the current directory unless they contain wildcards (* or ?).

Bug fixes
Project onto surface caused a crash on 2nd use.
Swap master and slave in bonded contact caused a crash when the slave is beam end faces.
Opening a file from the command line failed when it has a STEP file included.
Unrefining truss elements retain their truss flag.
Crash with abrupt exit when solving modal vibration with small models in certain cases.
“Error: Index was outside the bounds of the array.” with modal vibration, centrifugal force, and unused nodes.
The automatic scale on the contour plot sometimes didn’t include extrema on element faces not at the nodes.

FEA_Stress

Thanks Victor. I see there is a Spanish v17 manual, but is there an English version? It seems the links still go to v16.

Victor

@FEA_Stress, the English one is included in the installer. I haven't updated the web pages for v17 yet.

dculp

v.17 is slower than v.16 for modal analysis?

The following table shows the run times for the attached model (alternating between v.16 and v.17 while the other version was closed) --

Version --- Time "Solver exited" --- Total time (results displayed)
v.16 -------------- Not recorded ----------------------- 23:47
v.17 -------------- Not recorded ----------------------- 56:14
v.16 ----------------- 21:45 ------------------------------ 30:03
v.17 ----------------- 17:02 ------------------------------ 54:55
v.16 ----------------- 18:47 ------------------------------ 27:03
v.17 ----------------- 19:01 ------------------------------ 61:00

Thus, the time at which the green panel displays "Solver exited" seems roughly comparable between v.16 and v.17. However, after "Solver exited" is displayed, v.17 takes approximately 30 minutes longer than v.16 to actually display the results. (I'm not sure what the solver is doing in this interval.)

If you can't replicate the above differences, could there be a difference in my setup between the two versions?

CCX solver, Windows 10

Victor

There seems to be a problem with v17 importing CCX solutions slowly when you have formulas. It might be a bug re-evaluating the formulas multiple times, I'm not sure yet. Thanks for confirming it and I'll sort it out for v18. Sorry about that regression.