XANSYS Message: 174 [Go back to message list] [bookmark on del.icio.us]
No rating yet Subject: Re: Stress concentrations Author: Patrick Tibbits Date: 1999-01-11 12:38:00We have a new ASCII graphics champion!
Suggestions for the high local contact stresses:
1) If the corner of the circumferential ring has a fillet, model the fillet radius with a fine mesh.
2) If the ring has a 90 degree corner, model the elastic-plastic problem, localized plastic flow will spread the small stress concentration. I have not myself tried a problem with both contact and material nonlinearities.
3) Implement the lamina method in spring elements, whose stiffness comes from the Hertz load-displacement relation for line contact. Each spring element would have length equal to the appropriate shaft radius, and would begin at the centerline of the shaft. I used LINK10s with a gap to include the contact boundary conditions.
Patrick Tibbits, Ph.D. // Sr. CAE Engineer McGill Mfg. Co. Inc. // 909 N. Lafayette St. // Valparaiso, IN 46383 tel=219-465-2268 // fax=219-465-2263 // ptibbits@e...
> From: Brower Brent > (DTUS/CENGDE)[SMTP:Brent.Brower@p...] > Reply To: xansys@o... > Sent: Monday, January 11, 1999 10:12 AM > To: 'xansys@o...' > Subject: [xansys] Stress concentrations
> This is a problem that has bugged me in the past and I am now having > trouble with it in a contact problem that I am working on:
> I am analyzing a press fit of a "rod" into a hole. The rod that is > pressed into the hole has circumferancial rings turned on it. The > problem is that one of the elements on the edge of the rod where the > press fit is occurring has a huge stress differential across it. > Increasing the mesh density just increases the maximum stress at the > node "touching" the hole wall. The stress is so high at the one > element that when VonMise's stress is plotted it is red and the rest > of the model is blue.
> I have had this problem before, usually when constraining > displacement. Using pressure loads instead of fixing displacement > would work in the past, but since the stresses are caused by contact > this is not an option.
> Any help here would be appreciated.
> The model is 2D axysemmetric, with plane 82 elements.
> This is ASCII attempt to illustrate my problem.
> 000000 | > 000000 | > 000000 | Centerline is at the left edge. > 000000000| > "ROD" 000| > 000000000| > 000000 | > 000000 | > 000000 | > 000000000| > 000000000| > 00000000X| High Stress at the X > 000000 | > 000000 |_______