You should be able to add the coding from Stanum as a post-processing script in WB. If you still have trouble getting WB to do this, you can have it write out an input file that you could submit as a batch job in APDL. This would have the added advantage of allowing you to see inside the ANSYS...
You won't be able to use the finite element method to determine the stress concentration factor for this geometry because the sharp corner is a singularity. If you were to keep refining the mesh and look for the highest stress each time, you would see it keep rising without bound. This...
You also need to understand what loads the part will experience because it is probably the stress or strain that is one of the design drivers (besides weight as IRstuff mentioned).
Mustaine3 is correct that only linear elastic behavior is used in a frequency analysis. If you feel it is important to account for nonlinear behavior, you could do a full (direct) transient solution to find the response to a particular loading condition.
The default for MODOPT is mass normalization of mode shapes if you are using APDL. (I don't know the default for WorkBench, but it would be bad if it wasn't the same.)
Flash,
Be sure to let us know the outcome. We had a user who found a big bug with Workbench Random Vibration and it took a year before a Class3 Error Report was distributed by ANSYS (WB2011-24: Don't exclude any modes when calculating responses).
Take a look at the SHPP command to control element shape checking. You can get Warning messages instead of Errors. You can also change the shape criteria.
1. One option is use NUMMRG. If you have coincident keypoints, you can merge the keypoints and lines. Or you can mesh each part individually, then merge the resultant coincident nodes.
2. One approach is to use NSYM and ESYM.
In WB you can have it create a file of line mode commands, so you would then have an APDL batch file. I don't know how to go the other way and I'm not sure there would be any benefit. The advantage of WB is a slicker connection to your CAD files. But once you make changes to the model in...
If you want to apply a 1g static load to the structure, use the ACEL command (details in Help).
If you are doing something like a 1g sine sweep, you could do a spectrum analysis. An example is in 6.4 of the Structural Analysis Guide for ANSYS 13.0.
Then you can look at your log file to see what commands were generated by the GUI and see if they are different from the ones you used in your unsuccessful batch submittal.
Your link is blocked by my IT group, but I'll try to help.
1. It is the usual practice to do a steady-state solution at near-zero time to establish the initial conditions for a transient solution, so you should be OK.
2. I'm more familiar with APDL than WB, but the basics are the same. Your...
One approach that might be helpful is to write a macro for each type of load you will impose. You could have the safety factor as one of the arguments. If you have uniform increments in the safety factor or other parameters, you could use a loop (or nested loops) to invoke the macros. If the...
If you want to count local peaks you loop through each node one at a time. For each loop you have a parameter set to the magnitude of the node's displacement. Then select attached elements and the nodes of those elements (omitting the key node). Find the maximum magnitude of displacement...
Just to amplify on flash3780's response, the magnitudes of the deflections listed for the mode shapes (eigenvectors) is arbitrary (usually mass normalized). You can't tell anything by comparing deflections from one mode to another. The values only have relative meaning within the mode. You...
189 has a midside node while 188 only has nodes at the ends (not counting the orientation node). You can make the problem 2D by applying appropriate boundary conditions (for example, UZ, ROTX, ROTY). I don't know about WorkBench, but you can still use Legacy Elements in APDL (they just aren't...
If you have nonlinear behavior, then modal and harmonic solutions are eliminated. You can do a direct transient solution to find the response of interest. Whether the "frequency change" you are studying is an input load or a function of geometry, you can still see how your response varies as...
Can you pick something about the mode of interest that is unique? For example, maybe your mode is characterized by having the displacement at one pair of nodes being equal while another pair is equal magnitude but opposite sign. Then you can loop through your modes until you find the first...
I don't think you necessarily have a conflict between /Post1 and /Post26. Your contour plots are of 3 individual mode shapes. All 3 (and any others that are included) will contribute to the measured deflection that results from the applied load. So it looks like the maximum is occurring...
