Hello CEE This is a great short video. I absolutely liked it. However, is it possible to query robot to know how much of eccentricity exist between our COM and COS so as to know by how much alignment is needed to possible remove or manage the twist. Am honestly surprised that node 248 deflects more laterally than vertically, however in a bid to correct that we need to apply releases to columns. I think this strategy has it own issues as it tends to alter the fundamental period of the structure massively under lateral loads. I probably need to pay more attention to your FE class for deeper understanding Thanks
Hi there Engr. Nii, I need to check out if there is a way to give a clear indication to both mass and stiffness centers. The horizontal deflection being larger than the vertical was a surprise to me, too, TBH. The application of releases was not a suggestion to be a solution. It does not work without having shear walls to stabilize the structure in the lateral direction. If you do not provide any lateral load carrying elements, your structure would be - ultimately - unstable. That is what I meant by (House of Cards) Thnx again for your comment, and stay tuned for more content, CEE
Hello CEE, TBH, I had mainly expected that lateral loads were always necessary in order to have lateral displacements on a structure (but it is clear to me now and in hindsight [thanks to your work] that I was critically WRONG in having such expectations). Thank you so much for showing that even under gravity loads, members stiffness and members connections can (and do) impact on the structure ability to displace laterally. The “understanding or the academic value” of the “released” structure (i.e., 13:24 Fourth Case: Non Sway) is amazing to me, WoW! I do get it that for this (fourth case) scenario to be practically viable and for the structure to not be a “house-of-cards”, one additionally does need to incorporate lateral loads resisting elements [such as shear walls, and/or bracings] on the entire structure. I can see this scenario working practically well in structural steel framed buildings where “pinned” connections are perhaps even more desirable and more economical to have compared to fixed or moment connections. Bracing of steel structure is also something of a “normal” thing to do… My questions for the case of a concrete framed building are as follows: a) How would one mimic (for construction details purposes) a “pinned” or a “non-rigid” connection between a beam and a column in a reinforced concrete framed building for example? Is this even practical? b) Shear walls and core walls are also something of a “normal” thing to do in concrete framed buildings. Is it enough to expect (I am now uneasy on expectations) that because the building is provided with shear walls and/or with core walls; the lateral loads will “automatically” be “attracted” and re-distribute themselves to the more relatively stiffer shear walls and /or core walls from the columns even though the connections between the beams and the columns may not be practically “pinned”? c) If “automatic” re-distribution in (b) above is expected, would this not require the columns to first absorb the forces and to displace accordingly (possibly at the risk of getting damaged first in the process) before the forces can re-distributes to the shear or to the core walls? I have enjoyed this video and I am looking forward to the next CEE videos. Keep well and kindest regards, DK
Hi there Engr. DK, basically this video was inspired by the countless "senior year students" coming to me wondering why they are getting twisting structures when only vertical loads are applied. In hindsight, having the academic example might not be the best thing I did, but well, I am glad you found it helpful. Indeed, and I tried to be 100% clear about it being a house of cards that should not be done. Only reason I did it is to show that the tilting stops at that point. You are right, with a simple edit from my side, you see: in steel structures, if your beams are pinned to columns, then you would need to release the beams and not the columns. The columns would be continuous, and that is the reason behind such structures still maintaining some stability in a software. oh for the pinned one, that is not really easy, but here are my thoughts on it: First: scientifically, there is no 100% pin and no 100% rigid, but always something in between. So, instead of answering your question with a yes/no, I want to do something even further, namely a check list. The more your connection conforms with the check list, the more you are "justified" to assume it rigid, and vice versa: Things that make your connection "rigidy": 1) The top part of the column was cast along with the beams in 1 concrete job (no cold/str joints there) 2) There is some inter-connection between the steel of the beam and the steel of the column 3) The existence of a haunch in that connection 4) Parts of the beam steel going into the column. Things that make your connection "pinny": 1) columns being cast, cured, then beams being cast (formation of the joint) 2) no haunch 3) interaction between beam steel and column steel non-existant You know what the interesting part is? In my first years at work (country A), all our beams were assumed to be pin supported, and lateral stability was solely provided by shear walls. At the current universtiy I'm teaching (country C), it seems the general approach is the rigidly connected columns to pins, and their practice does cast parts of the column with the beams. Finally, if you think that you might have blundered something in the past, no worries. Because I am pretty sure you have provided shear walls to that structure. Please note that: lateral loads from seismic activity will be distributed amongst the columns (!!! if rigid !!!) and shear walls based on their stiffness. Now usually shear walls have a length of more than 2.5meters, and a thickness of around 250mm, So compare their inertia 1/12 bh^3 to the inertia of your average column of 300x500 and you can see the HUGE difference, so shear walls will basically deal with the brunt of lateral forces, saving you in case you have mis-assumed your connection. Of course a disclaimer here to any reader who is gonna read this in the future: you should NOT mis-assume your connection, nor should this explanation be taken as a "license" to under-design your columns. I read points a-b and c and tried to merge them in 1 response. I am not sure if I covered all points but please feel free to reply with any further questions you have. I am thinking of actually providing lectures in Earthquake load calculation as well as wind load calculation after I finish the FEM series. My lecture series is always going to run in parallel. For the FEM series, I am thinking that once I cover most of Daryl L. Logan's book, I would call it a day and move on to the next lecture series. Anyways, once again thnx for your comments. It always initiates some really interesting discussions, it kind of remind me of myself and the things I was thinking about, for that I am thankful. Regards and stay tuned, CEE
@@CivilEngineeringEssentials , Hello Dr. CEE Thank you for the comprehensive feedback; and yes, your response does cover my questions and it also provides even more insight and I really appreciate all of this. The disclaimer is 1000% warranted (Power level over 9k ^_^). It helps a lot to realise that various countries and regions sometimes take varying approaches. I suppose that as long as the design engineer understands the nuances and the consequences of his/her assumptions [and this is where the CEE channel shines 💯] and then provide the necessary complementary mechanisms or elements to counteract and to manage those nuances and consequences, their assumptions should work. Regarding the Earthquake load calculation as well as the wind load calculation lectures; I think that it is not a bad idea at all. I would suggest for those lectures to perhaps be done parallel with the same types of loads (EQ and Wind loads) application examples in RSAP or in Etabs or in both per CEE style ^_^. The currently running RC building series in RSAP is also an ideal platform to show case the applications and the inner workings (in practice) of these types of loads. It sure would be nice to be able to have an “intuitive” understanding of the order of magnitude and of the most relevant or critical points and directions of applications of these loads in any structure from early conceptual modelling and planning stages of the structure. It is always a pleasure for me to watch any of your videos and to read your feedback [including the feedback that you give to all other CEE channel subscribers like myself on their questions and discussions], thank you for sharing your exceptional knowledge and opening up the dialogue with the rest of the world. Kind Regards, DK
Hello Dr, may I ask. To summarize, difference in building stiffness causes twisting if non-symmetrical. Practically how would we address this besides releasing the joints. Is it just by providing shear walls? I think Eng DK mentioned it some key points regarding, what happens under gravity loading when we introduce the shear walls? Many thanks.
So, in a perfect world, the columns would be symmetrical and everything is perfect. However, due to non-symmetrical stiffness distribution, twist can be included in the mix. Just to clarify, the idea to release column is not correct if your columns and beams are monolithically cast or moment-transfer is occuring. A way to address this is to "play with" the stiffnesses of the columns, i.e. make a side stronger/weaker and observe the behavior. of course, if the structure is sufficiently high, then the columns are no longer able to resist lateral forces economically, and a switch to shear walls is to be made. In that case, yes, the placement of shear walls now is controlled by both the eccentricity around the C.G. of the building as well as the torsional behavior of the building due to gravity loads. Hope I was able to address your question. If you need any further vlarifications, please feel free to tell me. Regards, CEE
Great video as always. I do see some videos in youtube for other Engineers that use moment releases for each beam ends in a build. Now I have a question, in slabs or beams, Codes require engineers to alternate Live Loads for moment envelop, does Alternating LL effect swaing or twisting in a building (for example, apply LL on on side of a buildng)? if it does, how can I make Robot try all possible casses (load Alternating) for that ?
Hi there, I am happy you liked the video. Alternating LL may cause some sway in the building. But if I remember correctly, the ASCE code let's you alternate the loads on the spans and stories, so that the total eccentricity is zero. I mean, you might have in story 1 the left span loaded, and in story 2 the right one. As for load cases in Robot, for multi-story building, it is not possible, unfortunately. Regards, CEE
If we are going to deal with a rigid frame , we need to make sure that we are willing to do the same on site , now that's the point that get me confused every time , I mean 90% of columns-slab connection are not rigid they are pinned on site so why robot by default give us it as rigid connection , I really don't know .
It depends on the construction practice. When I worked in Germany, I was surprised to see that they actually try to achieve rigid connections between column and beams. Even here where I work now, it seems they cast the final part of the column with the slab to achieve some sort of rigidity. Still, Personally, I believe that rigid connections can be achieved by interconnecting the reinforcement of columns and beams as well as maybe adding a stiffening haunch. I think it is worth a video to explain. Regards, CEE
The explanation in the end doesn't make much sense to me, it's more like a computational artifact than a real physical case. If the beams were not connected to the columns they wouldn't deflect at all. They would just fall down without any interactions at the nodes. Frankly I'm surprised Robot even shows any deflections of the beams in this case. Only the columns would compress a bit under their own weight (provided they would be connected between eachother at the nodes which I'm not even sure is the case in this scenario).
After rewatching the video it seems you're not really changing the joints actually but you just artificially remove the non-axial components of the forces at the beams. Not sure what that is supposed to demonstrate, especially that you say in your explanation that you're removing the connections at the nodes.
@@bzqp2 I agree. The reason for the explanation at the end is to show that: the reason why the structure tilts is because one side is weaker than the other, thus the columns on the weaker side become "shorter" than the columns on the stronger side. This causes the structure to tilt, because - naturally - the columns are connected via the slab. To make sure that people get the point, I showed what happens if you make a "house of cards". In that case, the tilting stops. Please note that I tried to be clear that we are doing a "house of cards" which is something that we should never attempt in reality. The reason was to show that the 2 factors contribute to the structure twisting and tilting under gravity load, namely: Discrepancy in column stiffness and the joints themselves. I think I should not have done the last example. It was kind of an academic one. The point still stands that gravity loads can cause horizontal sway movement due to what I explained before. In case any further explanation is necessary, I would be more than happy. Regards, CEE
Hello CEE
This is a great short video. I absolutely liked it. However, is it possible to query robot to know how much of eccentricity exist between our COM and COS so as to know by how much alignment is needed to possible remove or manage the twist.
Am honestly surprised that node 248 deflects more laterally than vertically, however in a bid to correct that we need to apply releases to columns. I think this strategy has it own issues as it tends to alter the fundamental period of the structure massively under lateral loads.
I probably need to pay more attention to your FE class for deeper understanding
Thanks
Hi there Engr. Nii,
I need to check out if there is a way to give a clear indication to both mass and stiffness centers.
The horizontal deflection being larger than the vertical was a surprise to me, too, TBH.
The application of releases was not a suggestion to be a solution. It does not work without having shear walls to stabilize the structure in the lateral direction. If you do not provide any lateral load carrying elements, your structure would be - ultimately - unstable. That is what I meant by (House of Cards)
Thnx again for your comment, and stay tuned for more content,
CEE
Hello CEE,
TBH, I had mainly expected that lateral loads were always necessary in order to have lateral displacements on a structure (but it is clear to me now and in hindsight [thanks to your work] that I was critically WRONG in having such expectations).
Thank you so much for showing that even under gravity loads, members stiffness and members connections can (and do) impact on the structure ability to displace laterally.
The “understanding or the academic value” of the “released” structure (i.e., 13:24 Fourth Case: Non Sway) is amazing to me, WoW!
I do get it that for this (fourth case) scenario to be practically viable and for the structure to not be a “house-of-cards”, one additionally does need to incorporate lateral loads resisting elements [such as shear walls, and/or bracings] on the entire structure.
I can see this scenario working practically well in structural steel framed buildings where “pinned” connections are perhaps even more desirable and more economical to have compared to fixed or moment connections. Bracing of steel structure is also something of a “normal” thing to do…
My questions for the case of a concrete framed building are as follows:
a) How would one mimic (for construction details purposes) a “pinned” or a “non-rigid” connection between a beam and a column in a reinforced concrete framed building for example? Is this even practical?
b) Shear walls and core walls are also something of a “normal” thing to do in concrete framed buildings. Is it enough to expect (I am now uneasy on expectations) that because the building is provided with shear walls and/or with core walls; the lateral loads will “automatically” be “attracted” and re-distribute themselves to the more relatively stiffer shear walls and /or core walls from the columns even though the connections between the beams and the columns may not be practically “pinned”?
c) If “automatic” re-distribution in (b) above is expected, would this not require the columns to first absorb the forces and to displace accordingly (possibly at the risk of getting damaged first in the process) before the forces can re-distributes to the shear or to the core walls?
I have enjoyed this video and I am looking forward to the next CEE videos.
Keep well and kindest regards, DK
Hi there Engr. DK,
basically this video was inspired by the countless "senior year students" coming to me wondering why they are getting twisting structures when only vertical loads are applied.
In hindsight, having the academic example might not be the best thing I did, but well, I am glad you found it helpful. Indeed, and I tried to be 100% clear about it being a house of cards that should not be done. Only reason I did it is to show that the tilting stops at that point.
You are right, with a simple edit from my side, you see: in steel structures, if your beams are pinned to columns, then you would need to release the beams and not the columns. The columns would be continuous, and that is the reason behind such structures still maintaining some stability in a software.
oh for the pinned one, that is not really easy, but here are my thoughts on it:
First: scientifically, there is no 100% pin and no 100% rigid, but always something in between.
So, instead of answering your question with a yes/no, I want to do something even further, namely a check list. The more your connection conforms with the check list, the more you are "justified" to assume it rigid, and vice versa:
Things that make your connection "rigidy":
1) The top part of the column was cast along with the beams in 1 concrete job (no cold/str joints there)
2) There is some inter-connection between the steel of the beam and the steel of the column
3) The existence of a haunch in that connection
4) Parts of the beam steel going into the column.
Things that make your connection "pinny":
1) columns being cast, cured, then beams being cast (formation of the joint)
2) no haunch
3) interaction between beam steel and column steel non-existant
You know what the interesting part is?
In my first years at work (country A), all our beams were assumed to be pin supported, and lateral stability was solely provided by shear walls.
At the current universtiy I'm teaching (country C), it seems the general approach is the rigidly connected columns to pins, and their practice does cast parts of the column with the beams.
Finally, if you think that you might have blundered something in the past, no worries. Because I am pretty sure you have provided shear walls to that structure.
Please note that: lateral loads from seismic activity will be distributed amongst the columns (!!! if rigid !!!) and shear walls based on their stiffness.
Now usually shear walls have a length of more than 2.5meters, and a thickness of around 250mm,
So compare their inertia 1/12 bh^3 to the inertia of your average column of 300x500 and you can see the HUGE difference, so shear walls will basically deal with the brunt of lateral forces, saving you in case you have mis-assumed your connection.
Of course a disclaimer here to any reader who is gonna read this in the future: you should NOT mis-assume your connection, nor should this explanation be taken as a "license" to under-design your columns.
I read points a-b and c and tried to merge them in 1 response. I am not sure if I covered all points but please feel free to reply with any further questions you have.
I am thinking of actually providing lectures in Earthquake load calculation as well as wind load calculation after I finish the FEM series. My lecture series is always going to run in parallel. For the FEM series, I am thinking that once I cover most of Daryl L. Logan's book, I would call it a day and move on to the next lecture series.
Anyways, once again thnx for your comments. It always initiates some really interesting discussions, it kind of remind me of myself and the things I was thinking about, for that I am thankful.
Regards and stay tuned,
CEE
@@CivilEngineeringEssentials , Hello Dr. CEE
Thank you for the comprehensive feedback; and yes, your response does cover my questions and it also provides even more insight and I really appreciate all of this.
The disclaimer is 1000% warranted (Power level over 9k ^_^). It helps a lot to realise that various countries and regions sometimes take varying approaches. I suppose that as long as the design engineer understands the nuances and the consequences of his/her assumptions [and this is where the CEE channel shines 💯] and then provide the necessary complementary mechanisms or elements to counteract and to manage those nuances and consequences, their assumptions should work.
Regarding the Earthquake load calculation as well as the wind load calculation lectures; I think that it is not a bad idea at all. I would suggest for those lectures to perhaps be done parallel with the same types of loads (EQ and Wind loads) application examples in RSAP or in Etabs or in both per CEE style ^_^.
The currently running RC building series in RSAP is also an ideal platform to show case the applications and the inner workings (in practice) of these types of loads.
It sure would be nice to be able to have an “intuitive” understanding of the order of magnitude and of the most relevant or critical points and directions of applications of these loads in any structure from early conceptual modelling and planning stages of the structure.
It is always a pleasure for me to watch any of your videos and to read your feedback [including the feedback that you give to all other CEE channel subscribers like myself on their questions and discussions], thank you for sharing your exceptional knowledge and opening up the dialogue with the rest of the world.
Kind Regards, DK
Hello Dr, may I ask. To summarize, difference in building stiffness causes twisting if non-symmetrical. Practically how would we address this besides releasing the joints. Is it just by providing shear walls? I think Eng DK mentioned it some key points regarding, what happens under gravity loading when we introduce the shear walls? Many thanks.
So, in a perfect world, the columns would be symmetrical and everything is perfect.
However, due to non-symmetrical stiffness distribution, twist can be included in the mix.
Just to clarify, the idea to release column is not correct if your columns and beams are monolithically cast or moment-transfer is occuring.
A way to address this is to "play with" the stiffnesses of the columns, i.e. make a side stronger/weaker and observe the behavior.
of course, if the structure is sufficiently high, then the columns are no longer able to resist lateral forces economically, and a switch to shear walls is to be made.
In that case, yes, the placement of shear walls now is controlled by both the eccentricity around the C.G. of the building as well as the torsional behavior of the building due to gravity loads.
Hope I was able to address your question. If you need any further vlarifications, please feel free to tell me.
Regards,
CEE
Great video as always.
I do see some videos in youtube for other Engineers that use moment releases for each beam ends in a build.
Now I have a question, in slabs or beams, Codes require engineers to alternate Live Loads for moment envelop, does Alternating LL effect swaing or twisting in a building (for example, apply LL on on side of a buildng)? if it does, how can I make Robot try all possible casses (load Alternating) for that ?
Hi there,
I am happy you liked the video. Alternating LL may cause some sway in the building. But if I remember correctly, the ASCE code let's you alternate the loads on the spans and stories, so that the total eccentricity is zero. I mean, you might have in story 1 the left span loaded, and in story 2 the right one.
As for load cases in Robot, for multi-story building, it is not possible, unfortunately.
Regards,
CEE
If we are going to deal with a rigid frame , we need to make sure that we are willing to do the same on site , now that's the point that get me confused every time , I mean 90% of columns-slab connection are not rigid they are pinned on site so why robot by default give us it as rigid connection , I really don't know .
It depends on the construction practice. When I worked in Germany, I was surprised to see that they actually try to achieve rigid connections between column and beams.
Even here where I work now, it seems they cast the final part of the column with the slab to achieve some sort of rigidity.
Still, Personally, I believe that rigid connections can be achieved by interconnecting the reinforcement of columns and beams as well as maybe adding a stiffening haunch.
I think it is worth a video to explain.
Regards,
CEE
I find it so unnatural to have the RC column pinned Am I right to assume you are not in a seismic region, Eng Ali ?
The explanation in the end doesn't make much sense to me, it's more like a computational artifact than a real physical case. If the beams were not connected to the columns they wouldn't deflect at all. They would just fall down without any interactions at the nodes. Frankly I'm surprised Robot even shows any deflections of the beams in this case. Only the columns would compress a bit under their own weight (provided they would be connected between eachother at the nodes which I'm not even sure is the case in this scenario).
After rewatching the video it seems you're not really changing the joints actually but you just artificially remove the non-axial components of the forces at the beams. Not sure what that is supposed to demonstrate, especially that you say in your explanation that you're removing the connections at the nodes.
@@bzqp2 I agree. The reason for the explanation at the end is to show that: the reason why the structure tilts is because one side is weaker than the other, thus the columns on the weaker side become "shorter" than the columns on the stronger side.
This causes the structure to tilt, because - naturally - the columns are connected via the slab.
To make sure that people get the point, I showed what happens if you make a "house of cards". In that case, the tilting stops.
Please note that I tried to be clear that we are doing a "house of cards" which is something that we should never attempt in reality. The reason was to show that the 2 factors contribute to the structure twisting and tilting under gravity load, namely: Discrepancy in column stiffness and the joints themselves.
I think I should not have done the last example. It was kind of an academic one. The point still stands that gravity loads can cause horizontal sway movement due to what I explained before.
In case any further explanation is necessary, I would be more than happy.
Regards,
CEE
Another point: RSA shows deflections even for an unstable structure because it basically uses an extremely weak spring to model a release.