Learning Forum

[biomeriel]

Hello, nI'm running a model of a residual limb that has a silicone liner and the thigh soft tissue is model with Mooney Rivlin 2 parameter properties. nAt the moment, i get a deformation error in the liner frequently as it sort of puckers? i think it might be called hourglassing ?? but it looks like scrabbled egg.. Anyway, i saw a discussion mentioning keyopts but when i type in :KEYOPT,matid,6,1 in the command APDL for the liner and thigh i will get an 'unknow error' and the simulation doesn't run ncan someone please explain how to use the keyopts command for quadratic tetrahedron? thank you! nMeriel n

[peteroznewman]

Hello Meriel,nPlease review this tutorial. /forum/discussion/18002/convergence-during-large-deflection-of-hyperelastic-materialsnIf you need more advice, please reply with images showing the force convergence plot, and the N-R Force Residual plot and show the elements around the Maximum value.nRegards,nPetern

[biomeriel]

Hi peter,That what i used before, but my keyopt command causes the solver not to work. is it that i can only have it on the liner and not the hyperelastic thigh tissue?? nI dont have newton residuals set up i'm afraid, i never knew what number to put in and when i did put in a number i wasn't sure what it was telling me. nhere's an image of the deformed liner, i've hidden the socket and thigh to make it clearer n

[biomeriel]

sorry forgot to attach with the last post n

[peteroznewman]

Click on the Solution Information folder and type 3 in for the N-R Residual plots. That means it will save a plot for the last 3 iterations. Also type a 3 into the Element Violation line under the N-R Force Plot line.nYou need to change the Analysis Settings and make the Initial and Minimum Substeps to 100, the Maximum to 200.nHow many elements are through the thickness? You didn't show your mesh.n

[biomeriel]

this is the cross section of the model, the parts moving from outside to inside are: Socket, Liner, Thigh ,spacer, bone and implant stem. A zoom in on the liner:nni also just tried to run the model and got this error, n*** ERROR ***                          CP =      7.438  TIME= 01:48:51n Internal node 270794 on target element 254893 should not be the same   n node in element 4920 Please send the data leading to this operation to n your technical support provider, as this will allow ANSYS, Inc to      n improve the program.     nn    but my node number is at 252,576, so i dont know want node 270,794 could be, i even have a mesh numbering to compress the nodes. nLastly, increasing the substeps that much will surely create a very long solve time? nthank you so much for your help! nMerieln

[peteroznewman]

Hi Meriel,nVersion 19.1 is four years old, so ANSYS has probably fixed the error that occurred. Can you upgrade to the latest version?nYes, increasing substeps will surely create a very long solve time. I have had models that ran for days. It helps if you have a computer with a fast clock, like more than 3.6 GHz, many cores, like more than 8 cores, a fast drive, like an SSD, and enough RAM so that the model solves incore and not out-of-core.nOne way to speed things up if the load is only axial is to use an axisymmetric model. That is a 2D radial slice. There will be far fewer elements so the solution will be much faster.n

[biomeriel]

Hi Peter,nnUnfortunately i cannot upgrade my ansys as i'm using the university's licence and this is what they have.My project is due in a week, two weeks ago my hard drive died and i lost my models, i'm not sure what why but this time i cant seem to get them working again. I dont have time to run models for multiple days. nFor the 2D model, i have moments applied in each axis, so the loading on the model is not symetrical. I was under the impression that this means i cannot run this project in 2D? is that a false assumption?