Modelling Support in Structures 3

[Eren Jaeger]

Hi!

I was thinking about the proper way to model the support of structures. I have 3 cases in mind assuming this is for reinforced concrete:

Case 1:
Model the superstructure and put fixed or pinned support at grade level (based on engr's judgment). Support was put at grade level because soil was assumed to give zero displacement on columns.
Model the columns, footing tie beam below the grade level and put fixed or pinned support (based on engr's judgment) at the location of bottom of footing.

Case 2:
Pinned support here will be placed at footing tie beam level instead of grade level. This was told to me by my superior before which was also told to him by his professor/structural engineer during his college years. Columns below footing tie beam will now have their fixed or pinned support at the bottom of footing.

Case 3:
Whole structure from top to bottom of footing will be in one model. My problem here is that at the grade level, columns will be allowed to have lateral deflection. Is that okay?

What are your thoughts on these? Are they okay or acceptable based on your experience? I just want to know what is the proper and best way to design structures. I think there is no one method in terms of modelling but how do you do yours? Thank you.

 

[HTURKAK]

Yes... this is the common approach.. we simply put fixed or pinned supports . However , support level should be top of the footing rather than grade. The tie beams should connect the bottom of columns rather than bottom of footings. The modelling of support and foundation can be support fixity , winkler springs, half space and element modelling of the supporting ground.


In case of perimeter grade beams, for the perimeter columns , the fixed or pinned support level will be top of the perimeter grade beam for one direction. If the interior columns supported at a lower level, the support level should be top of tie beams or footing which ever is applicable. The support level should be decided using engineering judgement together with foundation system.


I could not get your point. You mean columns will be allowed to have lateral deflection at grade level ?.

 

[human909]

My problem here is that at the grade level, columns will be allowed to have lateral deflection. Is that okay?

Really depends on what is surrounding them. Earth? Thin pavement? Heavy slab?

 

[milkshakelake]

I wouldn't model the footings and tie beams. It would need non-linear analysis if you have uplift, which would make computations longer. With current technology, it's better to take the results and design foundations separately.

 

[Eren Jaeger]

thank you HTURKAK. I am wrong saying supports will be at bottom of footing. but when designing footings, do you transfer the reactions generated at the top of footing to the bottom of footing which you will then compare with the bearing capacity?
2nd question, I am just curious, in your design practice, do you put the top of footings at the same level or is it the bottom of footings that you put on the same level?

is it because columns should be design for forces at the face of the support (in this case grade or tie beam) so that's why supports should always be at the top of grade or tie beam and not at the center? If that's really the case, then is it also your practice to transfer the column shear at the top of grade or tie beam to the center of beam to become an addtional bending moment for beam? 3rd question sir, do you include your grade and tie beams in your model?
could you also clarify what you mean sir with "one direction"? let's say that if my grade beam supporting a certain column runs along x-axis then my support for that column will be horizontal x-axis and vertical only (assuming it is to be pinned)?

What I mean sir is that, since the whole structure from top to top of footing will be in one model (unlike in case 1 wherein I have 2 separate models), support will now be only at top of footing. in my model, at grade or ground level, there will lateral deflections since there's nothing to restrain the columns horizontally in my model (assuming I have no grade beams). do you put horizontal supports to account for restrain provided by ground slab?

sorry sir if I have many questions.

 

[Eren Jaeger]

thank you human909,
if in case it is earth, is it okay for my columns to have lateral deflections on my model? is that a realistic scenario since earth will most likely not provide horizontal restrains on columns?

how about if it is thin pavement or heavy slab?
Am I right than in heavy slabs, to take that into account in my model, I can put horizontal pinned supports at grade or ground level?

 

[HTURKAK]

Eren Jaeger ,

Below find my responds ,


The transfer of the reactions generated at the top of footing to the bottom , will be necessary in case of horizontal loading , shear at the bottom of the column. Consider a footing for a retaining wall... In general the reactions should be transferred to the bottom of footing ( or centerline for a thick raft )for geotechnical bearing stress check and structural design.
The choice of the same level footing top or bottom level depends on several factors. I preferred same bottom level for most of the cases.


The tie beams in general have nominal, code specified sizes and designed for the 10% of the largest column axial load connected. The grade beams may have substantial depth depending on the GFL and frost depth. The frost depth in my town was 1600 mm and the grade beams in general have 1200 mm or more depths . Regarding the interior tie beams , sometimes the use of the same depth with perimeter grade beams will be more practical to support interior walls and to get a grid foundation having spread footings at nodes. In this case, the modelling of grade beams with line elements together with superstructure is a good option.


The modelling of supports in general is chosen (support fixity model ; fixed for rigid and pin for nominally connections ). Both supports restrained to lateral movements . The structural system should be self standing and should not need horizontal thrust from the ground , for vertical loadings. The horizontal loads ( wind, seismic, active soil thrust...) are resisted by friction primarily and passive thrust of surrounding soil. If the columns connected with slab on grade, the column bottoms will have horizontal restraints at slab level (similar to tie beams ). Regarding the choice of pin or fixed support, i remember BS 5950 ,for the modelling of pin or rigidly connected of a steel column to foundation, if base rotational stiffness = 4EIc /h rigidly connected, and if base rotational stiffness = 0.4EIc /h pin connected .

 

[human909]

Am I right than in heavy slabs, to take that into account in my model, I can put horizontal pinned supports at grade or ground level?

By the sound of it the answer would be yes in many cases. But then scale matters. Also what about columns that may not have a slab on one side due to penetrations. For smaller structures its is probably easiest and in many cases not unconservative to just laterally restrain the sub grade levels.

However if you go building 200m towers with deep basements and footings then the slab at grade not going to enjoy restraining the columns if they want to move.

 

[Eren Jaeger]

If the columns connected with slab on grade, the column bottoms will have horizontal restraints at slab level (similar to tie beams ).

About the horizontal restraints provided by slab on grade, do we need to always consider them? or are there cases where we can simply ignore them? I am just curious.

 

[Eren Jaeger]

However if you go building 200m towers with deep basements and footings then the slab at grade not going to enjoy restraining the columns if they want to move.

Is there a way to not make the slab on grade restrain the columns? Maybe using compressible fillers?

 

[HTURKAK]

This is scale issue.. consider columns with very large axial loads..a simple 6 in. SOG cannot impose or dictate the position of the columns..

Mr human909 (Structural) well explained the case and I voted with pink star his respond.

 

[Eren Jaeger]

This is scale issue.. consider columns with very large axial loads..a simple 6 in. SOG cannot impose or dictate the position of the columns..

Is this where using sealants or backer rods in actual construction becomes useful where we can separate columns from slabs? or even if not using them for columns with very large axial loads, that doesn't matter anymore since as what you mentioned sir HTURKAK, 6" sog cannot impose or dictate the position of columns? so in cases like this, in modelling, no horizontal restraints should be put?

 

[Eren Jaeger]

thank you sir HTURKAK, milkshakelake and human909 for sharing your knowledge. I really enjoyed learning from you.