Continue to Site

Welcome to MCAD Central

Join our MCAD Central community forums, the largest resource for MCAD (Mechanical Computer-Aided Design) professionals, including files, forums, jobs, articles, calendar, and more.

Register Log in

Advanced dynamic random w/ mechanica

E

ebear1

New member
Has anyone found the advanced dynamic random studies in mechanica to provide useful results? I have read PTC's website and other sites related to advanced random studies, the information is limited and not in a scope with real world applications with respect to mechanica discussing solutions, I have talked with PTC customer support, they have not been helpful as usual. I need to solve a problem that requires advanced random approach to ensure that a mechanical assembly is structurally sound. Background: I have an input ASD (g^2/Hz) curve that I have used in conjunction with a base excitation loading scheme and accounted in xyz for gravity (386.4) scalar to excite the assembly. I have looked at the modal analysis to ensure convergence (16 natural freq's). I have created measures on a point where I believe one will see the most displacement (xyz) during the random qualification. Since mechanica is working in the RMS realm, it seems that the results may be subjected to error that is all negative numbers result in positive ones? Integrated mode using WF 2.0 I can not get von Mises stress results over the model (in the past, I was able to use pro2001 and get a fringe plot of von Mises over the models)? Results related to the displacements are in the x, y and z directions, do these max displacements need to be combined into a resultant form. I gather that mechanica is dealing in a probability solution? Is the approach similar to Miles eqns? Anyone that has used mechanica to solve a real world random analysis on a project, please discuss the results scheme of mechanica in some detail with me, your knowledge and guidance would be greatly appreciated.
 
ebear1


I'm in the same situation, somehow... I have to simulate a random vibration qualification, and my problems are the same or wider.. The online help is very unuseful giving only brief description of options..


The only thing I can tell you is that Mechanica in indipendent mode gives back Von Mises results. At least, it seams so..Moreover you can combine acceleration or displacement results in indipendent mode.. even if it is not clear why you can obtain back the magnitude.. THen I run mechanica in integrated mode and then, in order to verify results, open the file in indipendent mode.


one of my problems is:what is the difference between base excitation in the support and ground? aren't the support constrained somehow to the ground? or supports are something differentfrom constrained area?


greets


Edo
 
I have recently used Pro-Mechanica for analyzing the response of a complex assemby to simultaneous pseudo-random excitations. I was only interested in dipalcements at certain points; therefore, the measures I selected were located at those points. The results came out very close to actual experimetalmeasurents, I'd say within 10%.


I didn't look for stresses because I knew they were not an issue for this application. But I don't see why I wouldnt have gotten vm stresses instead of displacements had I asked for them.


Did you select the option of 'at each step' when youdefined the measures?


Then you can get a Stress vs Frequency/Time plot. Then you run a Dynamic analysis at that frequency that gave the highest value during the Random Analysis.


I hope this helps, if not, disregard.
 
Edo777,


It is good to know I am not the only one in the boat when it comes to Mechanica in the advanced dynamic application mode. I have learned much since I first posted the original message.


I can tell you that even though mechanica lets you display von Mises stress results in a random simulation, these results areincorrect. Stress results in a random situation are in one sigma rms values. PTC has posted an announcment that even though the functionality may exist in old versions of mechanica,the results are not accurate and the functionality is incorrect.


In selecting relative to ground or supports during a random vibration or shock analysis, you have to decide on the frame of reference you want the results to reside. To see the difference, set up a simple simulation creating 2 cases one with the results relative to supports and one with ground. Play the animation for both and note the behavior and displacement results of both. I typically select relative to supports for my analysis and the displacements match experimental data.


In order to help you with more details on the last question, I need to know howyou typically set up yourmodal analysis. Do youuse base excitation or are you loading the modelto excite it?


Another important note about random if you do not already know is that you should define measures in locations of interest as mechanica does not create any predefined ones unfortunatly. For example, you can place datum points on curves or surfaces that you want to look at in detail with respect to accelerations, velocities and displacements. There are a number of measures you can create from rms values to maximums, take a look and experiment with simple models to learn all the different ones you can create. Also to havemechanica calculate these defined valuesat every iteration ensure that you select ensure that you select thecorrectbox.


Hope this helps...
 
You must log in or register to reply here.

Sponsor

Articles From 3DCAD World

Back
Top