XANSYS Message: 161 [Go back to message list] [bookmark on del.icio.us]
Average Rating: 5 (1 votes) Subject: Re: A Very Basic Question Author: Peter Budgell Date: 1999-01-08 18:13:00The question of when to use element versus nodal averaged stresses, of course, "depends."
If you use the AVRES command (with PowerGraphics) you can optionally get discontinuity of the stresses at changes of element REAL value, and element MAT (material) value. PowerGraphics also plots with discontinuity of stress with events such as shell elements meeting at an angle. This can make it possible to get improved information about high-stress regions, while getting the "cleaner" appearance of nodal stress averaging away from intersections.
Fatigue analysis techniques (in my limited experience) will be industry dependent. An average stress might be taken from the FEA work, and factors applied that are load, industry, material, and connection method dependent. A more exact stress may be taken from a very fine element mesh and examined using other rules. Fracture analysis may be indicated. With a good very fine element mesh, the elemental and nodal stresses should be similar with acceptably shaped elements, except at the intersections of the type addressed by PowerGraphics and AVRES. If the stresses are not similar, you may want a locally finer mesh and/or better element shapes for certain fatigue work.
Nodal averaged stresses may wash out local stress variation caused by element shape variation. Nodal stress averaging becomes dangerous when it hides stresses with large elements, at intersections of shell elements, etc. In the past, I have seen nodal stress averaging abused with unfortunate results. If there are significant differences in element stresses (PLESOL) and nodal averaged stresses (PLNSOL) where fatigue is to be examined, or where stresses are high, I expect that a finer mesh may be wanted, and PowerGraphics with AVRES should be considered.
Peter C. Budgell pbudgell@o... Senior Development Engineer 519-740-0757 ext.249 Innovative Steam Technologies FAX: 519-740-2051 200 Avenue Road http://www.otsg.com Cambridge, Ontario, Canada N1R 8H5
> Under what conditions should one use elemental stress as opposed to > nodal > stress. Is there any particular reason using elemental stress in > fatigue > calculations. Thanks in advance for your response.