ANSYS Tutorial: Nonlinear analysis of Reinforced Concrete Beam and compare with test results
ฝัง
- เผยแพร่เมื่อ 23 พ.ค. 2020
- Link for load test data, import geometry, APDL command and graph :
drive.google.com/drive/u/0/fo...
---------------
My ANSYS Tutorial Playlist :
• ANSYS tutorial for pra...
Remark:
1. SOLID65 is meshing dependent. It is recommended to follow the mesh size as shown in this video for your first run. After you succeed with your 1st run, you may play with your own mesh size later on.
2. For the explanation of SOLID65, please refer to the link below.
th-cam.com/video/0JiKNWt095I/w-d-xo.html
3. Regarding CEINTF, please refer to the slide in the timestamped link below.
th-cam.com/video/-vHl5ipqEr0/w-d-xo.html
4. My playlist as per the link below covers the majority of daily used in structural engineering. Hopefully, this will help you figure out which part of the command that needs to be modified for your own structure.
th-cam.com/play/PLYX5ua4SRf6zibnuG3WM43gLJkUVBUgkx.html
5. This video is verified on ANSYS workbench 2019, therefore as long as you are using WB2019 you will be fine. You can scroll down the comments for guidelines if you are stuck somewhere.
For ANSYS workbench 2020 on the other hand, somehow WB2020 has trouble with meshing the straight LINK element. In case you are using WB2020, please replace the longitudinal rebars with the following link below.
drive.google.com/drive/u/0/folders/1DDDh5pT2nGZ2rLCX7b97LUHLXNKKLBm0
And also make sure all rebars are set to "LINK" and meshing order set to "LINEAR" before solving.
You are a benefit to the structural engineering community
i want to convey my thanks, this video is very helpful for my thesis research.
thank you very much sir/madam. Have a nice day!
Wonderful video. Thank you
thank you for your great effort, and i hope you perform an example of nonlinear analysis of reinforced concrete beam exposed to fire (thermal analysis)
Hi, I really appreciate the effort and clarity of this tutorial.
Meanwhile I'm working on Ansys 18.2 and model type does not contain link/truss, is there a way to make it appear?
Thanks for the video can you please tell how to strenghtened this beam by applying cfrp at the bottom and side face of the beam
In my thesis given force in KN can I add displacement instead of force?
Thank you for the video!
Did you take into account the self-weight of reinforced concrete in this calculation?
Good question. There is no gravity applied (self-weight) in this tutorial. By including it should improve the accuracy of the results for like 5%-8% which is considered worthy to be included.
Sir how did you get apdl command for both steel and concrete together
Thank You
How to get that last graph load v/s deflection in ansys
hello sir, how to join concrete with steel rebar in ansys?
Sir how to find udl on beam using ansys this is about point load but how to get udl
Hello! Thank you for the tutorial. I have a question, I've followed your tutorial step by step, but at the end, when checking the solutions, I've noticed that the values of my Force Reaction are very very low, something around 3 Newtons only. Do you know what could have happened?? My deformation is similar to yours, 72mm while yours is 74mm, but your force reaction was almost 160000 Newtons, while mine is 3 N. I've checked every step but I can't find the error. PLEASE HELP ME!!
Try to suppress SOLID65 APDL command and run the linear elastic, if problem still persist try to fix the linear elastic model to work properly before turning back on SOLID65 for nonlinear analysis.
Thank you for the video. In engineering data, you added bilinear isotropic hardening values. The tangent modulus value you entered was 20. I use 500MPA steel. What value should I use?
The tangent modulus of 20 MPa is from the reference document.
For design purposes, the code of practice such as ACI 318 does not incorporate strain hardening (except for seismic design), therefore, in theory, this refers to tangent modulus of 0 MPa (zero). For nonlinear FEM, it is preferred to put some small tangent modulus solely for numerical stability. 20 MPa is also a good number.
Thanks...
I have two doubts. What is the characteristic compressive strength of concrete used here?
When I opened mechanical, the long rebars appeared to be meshed. Once I inserted mesh sizings and updated the mesh, only the long rebars are not getting meshed. What to do?
The solution for "The long rebar not getting meshed" already provided in the pinned comments. Actually, it got meshed but the number of meshing is limited to one which make it looks like not getting meshed.
Hi Sir, when I switch to the Model tab, my reinforcements (longitudinal bar and stirrups) do not appear in the bottom of geometry tree. I can see only concrete/solid and impactor. What do you think could be the reason for this? Please help me urgently. Thank you for understanding.
If you can see all the lines in SpaceClaim and you have already assign section properties to them, you suppose to be able to see those lines in the Mechanical module.
Would you mind to explain what the SOLID65 and CEINTF intended to do to the model? and if I want to make different beam model, which part of command must be changed?
For the explanation of SOLID65, please refer to the link below.
th-cam.com/video/0JiKNWt095I/w-d-xo.html
Regarding CEINTF, please refer to the slide in the timestamped link below.
th-cam.com/video/-vHl5ipqEr0/w-d-xo.html
Regarding your last question about how to apply the workflow to the various structural elements, please refer to my playlist as per the link below, those cover majority of daily used in structural engineering. Hopefully, this will help you figure out which part of the command that needs to be modified.
th-cam.com/play/PLYX5ua4SRf6zibnuG3WM43gLJkUVBUgkx.html
@@ThanaponBuamongkolChannel thank a lot for fast responding. Sorry for inconvenience, I am new in FE modelling
hey man, thank you a lot for the video!
Let me ask you something, if I wanted to model a carbon fiber reinforcment at the bottom of this beam. What would I have to do with Solid65? Thank you!
If the concrete beam is subjected to non-reversal loading then you could simply attach the carbon fiber (solid or layered shell) element to the concrete surface. SOLID65 is not an option if the concrete beam is subjected to reversal loading because the material model can not be damaged/degraded.
And if the delamination is neglected, then the bond contact could also be applied. In practice, it is easier to consider delamination separately, not in FEM.
@@ThanaponBuamongkolChannel Thank you a lot for the response! I just started learning how to use ANSYS and I'm at a very beginner level yet. I attached a solid (carbon fiber element) to the bottom of the beam with the data of the composite, but I don't know if I did it right. Basically I'm trying to reinforce the beam you showed in your video with a layer of carbon fiber at the bottom.
Would you mind if I sent you the file of my project just to see if I did it right or wrong? If that's not a problem, of course.
Thank you so much for the response and the video!
@@ThanaponBuamongkolChannel This is the file I mentioned in my previous comment. If you have the time to tell me if I did it right or wrong, I'd really appreciate it! Thank you again for the help!
drive.google.com/drive/folders/1li-ruoKPIKnOPxydz8BP-mfqsJqmUVE2?usp=sharing
For anyone who try to re-model several times but the solution could not be finished or converged. Try to check the unit. In the Mechanical module, from the Home tab, select "Unit" and then choose Metric (mm, kg, N, s).
Hello! Thank you for the video..I'm using ANSYS 2021 R1
In Solution information/force convergence, my substep, disection part is different from the video, and in the case of longRebar, it appears that there is no force (MAX=MIN). LongRebars didnt show the same behavior with concrete. Do i use the different Rebar file? Thank you
I used fixed Rebar instead but I taked error message : The solver engine was unable to converge on a solution for the nonlinear problem as constrained...
If you have trouble with rebar, please try the steps shown in the pinned comments (the one with the hard meshing behavior setting at the line element). Although this tutorial was made with WB2019, by fixing the meshing behavior of the line element, the workflow should be completely fine up to WB2022 for the least.
hi thank you for this video . ı have a question about this study, my longitudinal bars are not included in the system and ı have this warning message( line bodies with type set to link/truss or cable may not be able to solve if large deflection is turned off in the analysis settings for the model) how can ı fix this problem? thnk you for your helping.
It looks like you may have an issue with meshing the straight LINK element. Please take a look at the pinned comment above for a possible solution.
Another question please, I need to know the spacing between stirrups in mm one by one I will try to model the beam with designmodeler
Please refer to figure 3.2 in Wolanski's document which included in the link in the description.
the same problem i want to model in abaqus, kindly i request you to make the same modeling in abaqus
Hello, thank you for your helpful tutorial.
I found a problem with my model. It seems that my longitudinal bar can't deform to the z-direction and only able to stretch to the x-direction. However, my stirrups have similar deformation. Do you know what might cause this? I tried to simulate my model twice and I obtained similar results.
I used ANSYS 2020 R2 Academic for this simulation.
I can send you the picture via email if you're interested.
Thank you so much
Try to recheck all the rebars are set to "Link" instead of "Beam" and also meshing element order shall be set to "Linear".
Named Selection is also a critical part as it directly tied to the APDL command. The element grouped in the Named Selections and its name needs to be exactly the same as shown in the video for APDL command to properly working.
If the problem still persists, skip this tutorial for now and try the following RC Column tutorial instead, it is probably the easiest to model among all of my tutorials.
th-cam.com/video/RHI1fuIAHGs/w-d-xo.html
Once you succeed with the RC column, you might be able to figure out what went wrong with your RC Beam.
Someone got the same issue as yours and the problem solved. Please scroll down the comments for solutions.
Hey, I tried to run same setup using ANSYS R2 2020, but instead of using autocad dwg for steel, I added them manually using design modeller. However my solution fails to converge always. I followed all the steps in the same way as shown in video. Can you tell me if anything else needs to be adjusted?
Yes, you can either use the Design modeler or ANSYS Spaceclaim to prepare geometry.
According to your question, I would suggest you follow the tutorial until you can get your 1st successful run. Once you got your 1st run, you may begin to modify the geometry one by one, e.g. replacing rebar geometry with your own geometry. This way, it is easier to track in case something wrong happened.
And since you use WB2020, please scroll down the comments for the solutions to fix the issue regarding LINK meshing.
@@ThanaponBuamongkolChannel Yes, another issue i have with dwg file is that i have 2017 version autocad hence unble to open the files. Can you share me dwg which is compatible for me? 2013 or 2017 one. Thanks for reply!
what ansys version do i need for this modelling?
You can find Workbench 2019 in ANSYS Release 2019.
Great lesson! I have several questions:
1. Why only one mesh element through the width of the beam?
2. You have physical rebars that are tied together with concrete, but in command script you use Solid65 options to smearing rebars through the element? How is it possible to use both of these approaches?
Thank you for your video
One element through the width is just for demonstration, anyone can play with mesh size once they succeed with their first run.
In the case that all rebars have been explicitly modeled, smeared rebars are not required.
However, in some situations, a small amount of smeared rebar will help with numerical convergence. It is just my old personal habit to put some small smeared rebars in the minor axis which is not something that you need to do as long as you can achieve solution convergence.
So, from the following APDL:
tb_rebarratio_x = 0.0
tb_rebarratio_y = 0.0018
tb_rebarratio_z = 0.0018
meaning that there is no smeared rebar along the global X-axis (equivalent to no redundant smeared rebar with longitudinal rebars).
and there are smeared rebar of 0.0018 (volume ratio) along the global Y/Z-axis (redundant with stirrups).
The other videos will show that smeared rebars are not something that always needed with the following APDL:
tb_rebarratio_x = 0.0
tb_rebarratio_y = 0.0
tb_rebarratio_z = 0.0
@@ThanaponBuamongkolChannel Thank you for your quick response! Ok, as I understand this problem has troubles with convergence. I have seen a lot of tutorials of that problem and everybody solving that problem in different ways.
I still do not understand one thing in your tutorial. You fixed supports in a way that they cannot rotate around the bending axis. Does it mean that it is not a hinged beam anymore, but clamped?
The boundary condition of ANSYS model in this tutorial is not a perfect roller. The perfect roller has been demonstrated in following video which clearly shown in the experimental setup picture.
th-cam.com/video/RHI1fuIAHGs/w-d-xo.html
@@ThanaponBuamongkolChannel can you make an analysis of a simple RC beam-column connection subjected to earthquake loads?
Thank you for video. I want to ask why do we choose the value Z=-60mm in the condition (impactor)? Could I choose the other value for example -40,-50,-70..?
The method in this tutorial is the deflection controlled method. The goal is to continue increasing deflection until the beam fails. You can definitely use any value as the start value. It could be -40mm at first and after you get your first run you can extend to -50mm, -60mm, and so on. Or in the case that -40mm is too much, then you need to reduce to say -30mm for example, and the trial-error process continues until we get the full structural response.
To help the above process easier, in ANSYS, there is a feature called "Restart analysis", which will allow the user to keep continuing the analysis until we feel like to stop. The "Restart analysis" method is not yet included in any of my tutorials at this moment.
PS. The reason I use -60mm right at the start in this tutorial just because to make this video as short as possible.
@@ThanaponBuamongkolChannel Thank you very much. I do step by step following your instruction in ANSYS 2020 R1 (student license). The deflection in the middle beam is a little different (-77.274mm- you: 74.754mm). But I still fell happy.
@@trandangbao4044 Hi, i have some problem, i tried following the tutorial, but it seems my longitudinal rebar is not attached to concrete. do you have any idea ?
@@ThanaponBuamongkolChannel i have some problem, i got the displacement around 77mm, but the axial force of my longitudinal bar doesn't match with the video. Any idea why ??
This video is verified on ANSYS workbench 2019, therefore as long as you are using WB2019 you will be fine. You can scroll down the comments for guidelines if you stuck somewhere.
For ANSYS workbench 2020 on the other hand, I never tested it. But judging from the feedback I get from this and other videos, I suspect that WB2020 might have trouble with meshing the straight LINK element.
In case you are using WB2020, please try to replace the longitudinal rebars with the following link below. Please let me know if it works or not.
drive.google.com/drive/u/0/folders/1DDDh5pT2nGZ2rLCX7b97LUHLXNKKLBm0
How to know element type is solid 65 or 186 or any other in ansys workbench?
You can verify the element type that you created from Workbench by accessing Mechanical APDL.
Sir, I want to know why the final output force turned out to be twice the impactor? Hope for your reply
Someone already asked the same question and get answered. Please scroll down the comments for the explanation.
@@ThanaponBuamongkolChannel Thank you, sir. I get it
In the wolanski reference paper, MISO is used to calculate the F-D results, however in our ADPL, we have used MELAS. I tried to run the same setup using MISO, but the solution does not converge properly. Can you explain how to understand this behaviour?
You can literally replicate Wolanski's workflow in the old ANSYS version. I had done those too, a long time ago which I can't recall which version.
What's happening with SOLID65 nowadays is because it's architecture is not quite in line with the new ANSYS development. Without proper maintenance from the developer, the future of SOLID65 is only going downhill, features no longer properly working.
You may be able to find references showings good results using ANSYS's modern concrete model such as Drucker Prager concrete, Menetrey Willam, Microplane including CPT. Some of them directly supported by Workbench, hence, APDL is not required. But when you try to apply those models to real-life structure, you will be able to notice that none of those are able to give the performance that comes close to what SOLID65 could offer.
Those modern concrete models are just simply not yet fully mature and reliable enough for real-life structures.
Thanapon Buamongkol Yes, even ANSYS developers are not providing much reference in regards to newer material models, looks like we just to have figure it out ourselves how to make those newer models work. Are you planning to make videos on newer models next?
Can we apply same APDL and modify it to use with micro plane models and try to calibrate the results?
My previous approach had to apply command to every line body which was very time consuming, I have learnt a lot from your videos, so thanks a lot 😃.
@@ThanaponBuamongkolChannel Can you provide me links to some literature regarding MELAS usage? I cant find much on google.
Can you please share the revit model too? I wanna learn how to model in revit and export to Ansys
I made a video tutorial for preparing the REIVT model for FEM, check out my REVIT playlist. Part 6.
Hello sir I'm doing the same things as you until the 15:28 minute of the video. But , link / truss option does not appear in model type tab for longitudional bars. What do you think could be the problem? Please help me. Best Regards.
Which ANSYS version you are using, and what is the drop-down list shown instead of the link/truss?
Dear friend, i solved the problem. the problem was caused by the version difference. Now I will apply them to the beam related to my thesis. Thank you so much.
what type of this file (11FullBeam_Concrete_Impactor.sat) ? and can you upload this program?
You can directly open or import the SAT file with SpaceClaim.
How did you figure out the usage of MELAS hardening ? I thought that it didn't work with SOLID65?
It was written somewhere in the manual. In case you want insight, try to search from help for something like a support material model for SOLID65 or support element for MELAS.
I personally don't like MELAS because there is a better choice in the past. SOLID65 got lesser maintenance from the developer, some features that work in the past no longer works anymore.
Thanapon Buamongkol Do you have any other preference for SOLID65 like MISO or KINH? Also have you worked on other material models like micro plane and MW?
I did some cube tests with the Microplane model, and its performance is incredible. The compressive degradation and tension softening are by far superior to the SOLID65 concrete model.
However, for real-life RC/PT concrete structure, its parameters are quite skeptical and really hard to verify and need some more time for the model to be fully explored. In the meantime, SOLID65 is still the most reliable for the practical structural engineer.
Thanapon Buamongkol yes, the parameters for elastic micro plane model have very limited resources and CPT model has too many parameters to consider for calibration. I am doing cyclic loading analysis using ANSYS and was wondering if you have any guidance or reference for me?
I can not say much in the research field because I am a practical structural engineer. ANSYS is not an option for us in Performance-Based-Design (PBD) for a seismic design that requires nonlinear-time history analysis.
Under monotonic static loading, on the other hand, ANSYS is considered one of the best for Disturbed-Region nonlinear FEM for a reinforced concrete structure. This is when SOLID65 is really shining.
As far as I can see, the chance of proper application of SOLID65 for reverse loading is quite slim. The main reason is that its kinematic hardening support seems to be crippled by the developer plus the compressive degradation is prohibited. Those two properties are crucial to capture proper hysteresis.
Hello!
I want to solve this problem by applying force instead of a given displacement, but ANSYS doesn't even start calculating. Please tell me what needs to be done to calculate this variant of the problem?
Try something like this:
After complete the run with displacement control, just replace the displacement with force should be fine.
If the problem still persists, try to perform linear elastic analysis by suppressing the SOLID65 and fix the model before continuing nonlinear analysis.
@@ThanaponBuamongkolChannel I solved my problem, thanks!
Hi sir
why does the load displacement curve not start at zero?
It indeed starts from zero. To complete the load-deflection curve, you can either add (0,0) point in the spreadsheet to complete the curve or add the initial load step in ANSYS with zero amount of controlled displacement.
@@ThanaponBuamongkolChannel please can you make a video about that.
@@ThanaponBuamongkolChannel Can you explain more this point please. how we can add the initial load step in ANSYS with zero amount of controlled displacement?
And other question. I see in 30: 30 the graph. The time 1 of step 1 the displacement in the middle of beam is -74.754 mm=2.943072 inch OK. But the value of the force is -1.556e5 N= -349800.29277 lbs. The experiment result is approximate 16000lbs, 16382 lbs (Wolanski's thesis). It means the results of you are not aggrement.
The ANSYS model in this tutorial is a full-beam model. While the model in the reference paper is half beam model with symmetry. Therefore, the total reaction of the full-beam model of 1.556e5 N (1 N = 0.2248 lb) will become 1.556e5*0.2248/2 = 17,490 lb in the load-deflection curve.
Actually, the flexural strength of beam could be estimated by hand. It will help figure out that "load" in the load-deflection curve is actually equal to load from only one impactor, not the sum of two.
The green arrow's location in the "Force Reaction" in ANSYS can be misled. The value is the sum of those two impactors reactions.
@@ThanaponBuamongkolChannel Thank you very much.
Do you have a command to show the crack & Crush plot?
Unfortunately, this workflow does not compatible with and not support the "PLCRACK" APDL command.
Please keep your video in more resolution.
Dear Sir, you can find the way to show cracks in workbench through the commands in the following video th-cam.com/video/Vb-xVXdY3Go/w-d-xo.html
@@SSV380 thank you sir
Hello Mam, Can you kindly help me in analyzing a 7 storeid RCC building using ANSYS workbenc?,it will be a great help for my thesis.THANKS in ADvance.
Use Etabs.
Sir, is this Solid 65 element?
Yes, this tutorial based on SOLID65 concrete model.
Sir how to use COMBIN 39 element between Concrete and steel node in ANSYS APDL. Please give me reply
If you are looking to model the bond slipping between concrete & rebar or a concrete spalling model to capture rebar buckling behavior. You can create COMBIN39 by defining concrete nodes and rebar nodes.
The concrete node and rebar can be different node numbers with the same coordinate. Define material properties and real constants then use E Command to create a COMBIN39 element that is linked between concrete and rebar.
@@ThanaponBuamongkolChannel Thank You for your reply sir. I have all the real constants for COMBIN 39 elements,i now need to assign the COMBIN 39 element between Concrete and Steel node for which the nodes Shouldn't be coincided,it should have some very finite distance between two nodes...But in my Case Concrete and Steel node are Coinciding,how can I obtain a finite distance between these two nodes in APDL
Are your "Coincided Node" referring to which one between:
1) Two nodes with different node numbers which both have the same X, Y, Z coordinates.
or
2) One node that is shared and connected to both concrete and rebar elements.
And what kind of behavior you are modeling, is it bond slipping only or the combination of bond slipping and concrete spalling?
@@ThanaponBuamongkolChannel I have chosen equal element size for meshing Concrete block and Re bar..So both these are sharing Common node..And My model is to analyse the Bond Behaviour of RC under the influence of Corrosion for which I have to consider both Bond Slip and Concrete spalling
พอจะมีไฟล์ ไหมครับ ผมทำตามแต่ก็ยัง error solve ไม่ได้เลยครับ
The solver engine was unable to converge on a solution for the nonlinear problem as constrained. Please see the Troubleshooting section of the Help System for more information.
Please try the following, see if it could help you pass the error.
After you have followed every step and the model refuse to run, try to recheck all the rebars are set to "Link" instead of "Beam" and also meshing element order shall be set to "Linear".
Named Selection is also a critical part as it directly tied to the APDL command. The element grouped in the Named Selections and its name needs to be exact same as shown in the video for APDL command to properly working.
Although meshing size and substep size can vary, it is recommended to follow the exact same meshing size/substep size as shown in the video for your first run. Once you succeed with your first run and get the complete results, you can play with meshing size/substep size later on in combination with restart analysis to continue the solution after failing the convergence.
After trying all the above, if the model still does not run may be better to restart from scratch and follow the steps again in case something has been previously overlooked.
@@ThanaponBuamongkolChannel ขอบคุณครับ เมื่อกี้ลองทำเหมือนทุกอย่างก็ยังไม่ได้เลยครับ ยังไง เดี๋ยวลองอีกรอบครับ
@@assadakorn For anyone who try to re-model several times but the solution could not be finished or converged. Try to check the unit. In the Mechanical module, from the Home tab, select "Unit" and then choose Metric (mm, kg, N, s).
Madam could you share with us the thesis of Buckhouse please
Wolanski's is all I have at the moment.
@@ThanaponBuamongkolChannel thank you for your response
"in RC Structural analysis, how to know the concrete is cracked or not? if cracked how deep".