Hello! Could you record a video in which a brick is tested for bending? So that the destruction and rupture can be seen on the model. I have heard that in this case Drucker-Prager Concrete Model and the Menetrey-William material model are used
@@ThanaponBuamongkolChannel I could not open the geometry file because my version of space claim is 19, so I could not get how you created the top and botton base, and as a consequence the named selection as well.
I see. The "NML_BC_Remote_Top" is the edges of geometry at the top of the steel column. And "NML_BC_Remote_Base" is the edges of geometry at the bottom of the steel column Regarding the steps of creating the named selection, try Google/RUclips search "ANSYS SpaceClaim how to make the named selection". And apply the steps (in SpaceClaim, not in Workbench) to the edges of geometry as mentioned earlier, then you are good to go.
Dear Dr. I would like to know two simple questions: a) How the beams are linked to the shell elements? b) Why top and base remote displacements behavior are set deformable?. Thanks in advance
a.) There is no beam element that is connected to the shell element in this tutorial. The finite element model consisted solely only shell elements. b.) We do not know the exact boundary condition of the test specimen. The remote displacement setup shown in this video is my own assumption and guessing of what the test condition is supposed to be and let the analysis results speak for themselves regarding the suitability of such assumption.
@@ThanaponBuamongkolChannel Thank you for your kind reply. Since the model includes beams not connected to the shell, I would appreciate some additional comments about the purpose of the beams that appear in the model. Thanks again.
See [11:36], the geometry line (not beam element) is there to only serve the point location of remote force/displacement. The location also could be manually applied as well, it is optional.
I'm a bit confused by this, by importing the geometry from the buckled beam, and then applying the non-linear analysis on the buckled geometry, how does the resultant forces relate to the original undeformed beam? Do you add the force at which it converges first in the non-linear on to the starting force in the linear buckling analysis (i.e; do linear at 1N, linear says failure at say 280kN, would your non-linear be whatever the force-convergence criterion for non-linear + 1N?). Also, what is the scale factor of 0.76 supposed to do?
Not sure I could understand all the questions, but I will try to reply to the ones I think I understand what you asked. Regarding "...by importing the geometry from the buckled beam...", in real life there is no perfect geometry. Such as there is no perfect straight column in reality. In this tutorial, the actual column geometry (which is never perfectly straight) has been measured using laser scanning. The scale factor of 0.76 is used to adjust the 1st mode from linear buckling analysis to be more in line with scanning data. While the higher mode has been neglected in this tutorial. The purpose of all the above is only to prepare the initial geometry that will be used in the nonlinear analysis. The initial geometry is the actual (or close to actual) column geometry in this case. And the analysis results already speak for themselves about whether the higher mode is supposed to be included or not.
@@ThanaponBuamongkolChannel Thanks for the response, and yeah I understand that, so the only function of piping in the eigenvalue buckled geometry is so that theres not a perfect beam? Just so that non-linear buckling can occur? As a follow up, how would this impact a comparison between the linear buckling result and non-linear result? Given the linear buckling started from a perfect geometry and the non-linear, by necessity, started from an imperfect geometry - wouldn't any direct comparisons be somewhat wrong? I was doing something similar except with a cylindrical column, which obviously without eigenvalue buckling would just compress, but when piped from the buckled geometry would deform as expected but would give me fairly absurd maximum forces when run for 10000 steps (when its supposed to be much lower than linear). I never quite understood what relevance the starting unit load analysis had on the non-linear part if all it was doing was generating deformed geometry, which is why I thought the non-linear part might have been a continuation of the linear part, i.e; conformed to this shape at 1N of eigenvalue buckling, then using this as a starting force and geometry you continue the application of force in the non-linear analysis? Apologise for the long-winded and confusing response, but I genuinely am struggling to interpret the approach and results.
@@bilbonob548 It is not possible to properly perform nonlinear analysis without introducing appropriate initial imperfection. The Eigen/Euler bucking is a theory, it is not a reality. Although the Eigen/Euler bucking theory is based on physics science and mechanics, we have to understand the "assumptions" that are being used to formulate such a theory. And the "assumptions" are not always true in real life, therefore, some adjustments need to be made before applying such a theory to properly predict and understand reality. Try to google search for something like basic structural stability and bifurcation point. Also, the video as per the link below would also help prime the basics of structural stability. The first 20 minutes should give you some idea of the difference between the bifurcation point you obtained from Eigen/Euler bucking and the load-deflection curve in reality. Video Link: ruclips.net/video/iOH6kaXX3Mk/видео.html If possible, try to practice the analysis for the load-deflection curve for a simple straight column using line elements with various amounts of initial imperfection and compared them with the known source (and observe the difference of load-deflection results obtained from various amounts of initial imperfection), this will greatly help you understand why and how to properly start the nonlinear analysis using shell or solid elements.
@@bilbonob548 Also, beware of the "Snap-Through" behavior because the maximum structural strength could be much higher than what is obtained from linear Eigen buckling as the structure enters the post-buckling state. Try to google search something like Snap Through nonlinear analysis for more information in this regard.
@@bilbonob548 The magnitude of the initial imperfection does directly impact the load-deflection curve. We can not use any random amount of the initial imperfection and expect to capture the actual structural behavior.
@@ThanaponBuamongkolChannel I'm glad you responded to my question. I learned from several videos, especially from video tutorials from you. I want to ask about beam extracted profile 1 (top and bottom). What is its function and why have different lengths? Is the profile a roundbar? Thank You.
@@ThanaponBuamongkolChannel In linier buckling analisys on C table : setup, solution & result are Not Check mark. Is it ok? And why Z component is -5mm on remote dispalcement top (on 27:35)? I'm sorry to much asking to you. this video is very meaningful for my thesis. And I learn step by step 😀
@@tutioktavia6421 Regarding [27:35], if you mean what should be the initial value for your first run, it could be anything. With trial & error, you will eventually get a suitable value. For example, when the initial value is too low, you notice that the structural strength could be higher by observing the load-deflection curve. Or in the case, the initial value is too high, the structure exhibit failure before reaching the specified value. In this case, you can either continue the analysis after the last failed attempt by using restart analysis or lower the initial value and run the analysis again. Both methods give the same final results. PS. Regarding your other questions, I didn't have a response for now as I'm not sure I understand what you asked.
Am I correct you are referring to [30:59] that the analysis stopped at a load factor of 0.72 rather than 1.00? If so, there is nothing to do with convergence. The solutions already converged several times along the way up until reaching a load factor of 0.72. Not sure if this answers your point.
Hello!
Could you record a video in which a brick is tested for bending? So that the destruction and rupture can be seen on the model. I have heard that in this case Drucker-Prager Concrete Model and the Menetrey-William material model are used
Very good! Could you do a video showing how you created the remote bases for the BC please?
It was already shown here [10:56].
Or is it the named selection (how-to) that you refer to?
@@ThanaponBuamongkolChannel I could not open the geometry file because my version of space claim is 19, so I could not get how you created the top and botton base, and as a consequence the named selection as well.
I see. The "NML_BC_Remote_Top" is the edges of geometry at the top of the steel column. And "NML_BC_Remote_Base" is the edges of geometry at the bottom of the steel column
Regarding the steps of creating the named selection, try Google/RUclips search "ANSYS SpaceClaim how to make the named selection". And apply the steps (in SpaceClaim, not in Workbench) to the edges of geometry as mentioned earlier, then you are good to go.
Hi Dr
Thanks for new video. Can you do a nonlinear coupled thermal analysis of RC beam using Ansys Workbench?
Thank you
Dear Dr. I would like to know two simple questions: a) How the beams are linked to the shell elements? b) Why top and base remote displacements behavior are set deformable?. Thanks in advance
a.) There is no beam element that is connected to the shell element in this tutorial. The finite element model consisted solely only shell elements.
b.) We do not know the exact boundary condition of the test specimen. The remote displacement setup shown in this video is my own assumption and guessing of what the test condition is supposed to be and let the analysis results speak for themselves regarding the suitability of such assumption.
@@ThanaponBuamongkolChannel Thank you for your kind reply. Since the model includes beams not connected to the shell, I would appreciate some additional comments about the purpose of the beams that appear in the model. Thanks again.
See [11:36], the geometry line (not beam element) is there to only serve the point location of remote force/displacement. The location also could be manually applied as well, it is optional.
@@ThanaponBuamongkolChannel Thanks a lot. Best regards
I'm a bit confused by this, by importing the geometry from the buckled beam, and then applying the non-linear analysis on the buckled geometry, how does the resultant forces relate to the original undeformed beam? Do you add the force at which it converges first in the non-linear on to the starting force in the linear buckling analysis (i.e; do linear at 1N, linear says failure at say 280kN, would your non-linear be whatever the force-convergence criterion for non-linear + 1N?). Also, what is the scale factor of 0.76 supposed to do?
Not sure I could understand all the questions, but I will try to reply to the ones I think I understand what you asked.
Regarding "...by importing the geometry from the buckled beam...", in real life there is no perfect geometry. Such as there is no perfect straight column in reality. In this tutorial, the actual column geometry (which is never perfectly straight) has been measured using laser scanning. The scale factor of 0.76 is used to adjust the 1st mode from linear buckling analysis to be more in line with scanning data. While the higher mode has been neglected in this tutorial.
The purpose of all the above is only to prepare the initial geometry that will be used in the nonlinear analysis. The initial geometry is the actual (or close to actual) column geometry in this case. And the analysis results already speak for themselves about whether the higher mode is supposed to be included or not.
@@ThanaponBuamongkolChannel Thanks for the response, and yeah I understand that, so the only function of piping in the eigenvalue buckled geometry is so that theres not a perfect beam? Just so that non-linear buckling can occur? As a follow up, how would this impact a comparison between the linear buckling result and non-linear result? Given the linear buckling started from a perfect geometry and the non-linear, by necessity, started from an imperfect geometry - wouldn't any direct comparisons be somewhat wrong? I was doing something similar except with a cylindrical column, which obviously without eigenvalue buckling would just compress, but when piped from the buckled geometry would deform as expected but would give me fairly absurd maximum forces when run for 10000 steps (when its supposed to be much lower than linear). I never quite understood what relevance the starting unit load analysis had on the non-linear part if all it was doing was generating deformed geometry, which is why I thought the non-linear part might have been a continuation of the linear part, i.e; conformed to this shape at 1N of eigenvalue buckling, then using this as a starting force and geometry you continue the application of force in the non-linear analysis? Apologise for the long-winded and confusing response, but I genuinely am struggling to interpret the approach and results.
@@bilbonob548 It is not possible to properly perform nonlinear analysis without introducing appropriate initial imperfection. The Eigen/Euler bucking is a theory, it is not a reality. Although the Eigen/Euler bucking theory is based on physics science and mechanics, we have to understand the "assumptions" that are being used to formulate such a theory. And the "assumptions" are not always true in real life, therefore, some adjustments need to be made before applying such a theory to properly predict and understand reality.
Try to google search for something like basic structural stability and bifurcation point. Also, the video as per the link below would also help prime the basics of structural stability. The first 20 minutes should give you some idea of the difference between the bifurcation point you obtained from Eigen/Euler bucking and the load-deflection curve in reality.
Video Link: ruclips.net/video/iOH6kaXX3Mk/видео.html
If possible, try to practice the analysis for the load-deflection curve for a simple straight column using line elements with various amounts of initial imperfection and compared them with the known source (and observe the difference of load-deflection results obtained from various amounts of initial imperfection), this will greatly help you understand why and how to properly start the nonlinear analysis using shell or solid elements.
@@bilbonob548 Also, beware of the "Snap-Through" behavior because the maximum structural strength could be much higher than what is obtained from linear Eigen buckling as the structure enters the post-buckling state.
Try to google search something like Snap Through nonlinear analysis for more information in this regard.
@@bilbonob548 The magnitude of the initial imperfection does directly impact the load-deflection curve. We can not use any random amount of the initial imperfection and expect to capture the actual structural behavior.
Could you make tutorial special for this geometry? This is the firs time i study about ANSYS. Thank u
Try to RUclips search for "ansys spaceclaim tutorial". There are already many good tutorials in this regard.
@@ThanaponBuamongkolChannel I'm glad you responded to my question. I learned from several videos, especially from video tutorials from you. I want to ask about beam extracted profile 1 (top and bottom). What is its function and why have different lengths? Is the profile a roundbar? Thank You.
@@ThanaponBuamongkolChannel In linier buckling analisys on C table : setup, solution & result are Not Check mark. Is it ok? And why Z component is -5mm on remote dispalcement top (on 27:35)? I'm sorry to much asking to you. this video is very meaningful for my thesis. And I learn step by step 😀
@@tutioktavia6421 Regarding [27:35], if you mean what should be the initial value for your first run, it could be anything. With trial & error, you will eventually get a suitable value. For example, when the initial value is too low, you notice that the structural strength could be higher by observing the load-deflection curve.
Or in the case, the initial value is too high, the structure exhibit failure before reaching the specified value. In this case, you can either continue the analysis after the last failed attempt by using restart analysis or lower the initial value and run the analysis again. Both methods give the same final results.
PS. Regarding your other questions, I didn't have a response for now as I'm not sure I understand what you asked.
@@ThanaponBuamongkolChannel Thank you so much. i appreciate your response. Have a nice day.☺
It has not converged
Am I correct you are referring to [30:59] that the analysis stopped at a load factor of 0.72 rather than 1.00?
If so, there is nothing to do with convergence. The solutions already converged several times along the way up until reaching a load factor of 0.72.
Not sure if this answers your point.