Hello people. I am new to FEA, and using Ansys to solve a few textbook problems to validate the solution. Please bear with me as the doubts might sound trivial. So, I am trying to simulate the stress in a circular cantilever beam of diameter 100mm, where an axial tensile force of 4000N is applied. By calculation, the normal stress along the axis should come to be around 0.51N/mm^2. However, ANSYS gives the normal stress and the von-mises stress as 0.75 and 0.63 N/mm^2 respectively. Can someone tell me where I am going wrong? Thanks in advance.
Compare the color of your model and the color bar, the normal stress and von-Mises stress (at the face where you applied the force), is about 0.48-0.52 N/mm2. That means the calculation of ANSYS at that face was right. You can also use scope tool to find the mean value at that face.
However, the maximum stress value is 0.75 and 0.63 N/mm^2, indicating that the calculation was wrong at somewhere else inside your model. My suggestion is that you should look closely, find where it is and tune your model (for example, change meshes).
Yes correct. Thanks for clarifying it to me. I am really just starting out, so my ideas about the error-sources of ANSYS are not very developed.
The maximum and minimum stresses are near the fixed end, I guess that's how it will be in real life too.
FEM is a proximate method and All software programs ( staad, sap 2000, Ansys,......) are dipend on several parameters for accuracy results
Like element shape, number of elements, degree of freedom and other....
You can review chapter: accuracy of FEM
In any book and you can obtained, how make correct results
Let me make two comments to your simulation.
1.) You have modelled the bar using 3d (hexahedral) elements. It is subject to a single force in the cross section in the direction of the bar's axis (is the load application -by the way- exactly in the centre of gravity?). As a consequence the normal stress distribution in the cross section of force application will be rather nonuniform, right below the point of force application the normal stress will be larger, in some distance to that point stress will lower. This is, for my understanding, what can be seen in the stress output showing values in an interval of approx. [0.40, 0.76] MPa. (The fact that the non-uniform normal stress distribution cannot be seen might be due to the display options you have chosen, averaged values?) In some distance to the cross section of force application, normal stress will be homogeneous over the cross section and will show the precalculated value of approx. 0.5 MPa you have mentioned.
2.) If you carry out your stress analysis using a 1d truss element, the cross section exists as a numerical quantity but is not discretized. Hence nonuniform stress distribution in the section of load application cannot occur, the numerical output will yield constant stress everywhere.
Two final comments. We should be aware, that the 3d model you used as well as the 1d model sketched in 2.) are "models" and in that they are approximations; the 1d model neglects the load application in the cross section, your 3d model assumes a point force, where a more realistic model should consider the force acting not on points but rather distributed on an area.
Beyond that particular example you considered, stress concentrations at cross sections of load application are an important aspect in the design of structures (e.g. in hinges, connections) and their computation. If not properly considered they can induce failure of structures.
Have a great day ahead!
- Badges
- Science topic