ShearWall - Temperature Design in ETABS

[rummaan17]

Dear all,

I am trying to design shearwalls through ETABS with temperature load applied over shell. At various location, spandral section fails in Shear due to temperature and piers (sometime in shear, mostly in flexure).  (See Attached Image)

Certainly all the problem in Shearwalls are due to temperature. I don't want to increase cross section of spandral or pier at some location just due to temperature load case as it will appears non-uniform with rest of the wall. 

I have seen stiffness modifier affect distribution of forces and also rigid/semi rigid daiphragm assumption. 

 

Can anybody guide how to properly design the shear wall with temperature load applied in ETABS or share any similar experience. Thanks in Advance.    

[Ayesha]

Please summarize your temperature loading. Why do you think wall shoul not fail?

[rummaan17]

There is alot of debate as what stiffness modifier to use F11=F22 for temperature. Previously, slab was getting high temperature reinforcement which was fixed using this value of F11=F22. Walls still have F11=F22=0.7 which is also a reason for high temperature load. 

[WR1]

You should not get moment from temperature load in ETABS. Because temperature is applied as axial load not as gradient like in SAP.

 

Yes walls will fail in F11, F22 (axial force, tension mostly). But remember for tension, walls will be cracked and to account for cracking you have to apply a very low value of axial stiffness. I would say to start with F22 and F11 as 0.25 and if it still fails you can even reduce it to like 0.10 (tension will be carried only by steel) plus some portion by tensile capacity of concrete.

 

This is true if cracks are not your concern like in water tanks etc.

Imagine what is the goal of your calculations? failure mode? service or ultimate? Ultimate right? Ok, now at ultimate stage what is the predicted cracking mode? How much cracking do you expect? For axial tension, I expect fully cracked concrete with section capacity only equal to n.As. So nAs/Ag is the modifier you should use for f11 and f22 in tension.

 

For compression, i dont expect cracks so section is mostly less cracked (only from shear or flexure).

[rummaan17]

Dear Mr.Rana, for negative temperature when we have compression will it have same modifier ?

[UmarMakhzumi]

@Rana,

Good comments. I have got a few questions about the subject that I believe will make this topic more detailed.

8 hours ago, Rana said:

You should not get moment from temperature load in ETABS. Because temperature is applied as axial load not as gradient like in SAP.

I am assuming that the discussion here is for temperature design for the final finished building and we are not discussing temperature affects cobsidering sequential construction (which could be an interesting case- general comment). Here are the questions:

Temperature loads generally produce tension or compression. For the current case, I assume there would be some slabs connected to shear wall. Considering that the building is insulated at all level and temperature is maintained inside the building once it would be operational, temperature differential would exists at top slab (roof) , which would likely induce moment in the wall. Why isn't that being considered.

Also, the discussion above talks about temperature induced load in shear wall as axial load. I would like to know the source of axial tension in shear wall if temperature in building is maintained?

8 hours ago, Rana said:

Yes walls will fail in F11, F22 (axial force, tension mostly). But remember for tension, walls will be cracked and to account for cracking you have to apply a very low value of axial stiffness. I would say to start with F22 and F11 as 0.25 and if it still fails you can even reduce it to like 0.10 (tension will be carried only by steel) plus some portion by tensile capacity of concrete.

I am not sure what would be the best way to consider temperature loads concurrent with other loads, but what I have done is run a separate load combination as temperature load and add that load combination to other load combos (RISA 3D). I design steel structures for Oil and Gas industry so genre is different. In my opinion, the affect of temperature on shear wall should be considered in combination with other working loads. The design would be similar to that of a beam column subjected to axial load. I am not sure if additional reduction is stiffness is justified based on isolated consideration of temperature load. What do you think ?

Thanks.

[rummaan17]

Dear Mr.Umar,

We have to submit our design for peer review which includes one international and one local consultant. I was amazed when they asked to apply load on overall structure. Even in my opinion its only outersurface that is most prone to temperature change, as here in middle east mostly building inside is temperature controlled. 
Though we also design slab for temperature change for all level, which i don't find just.   I was getting high temperature stress even in slab which was then modified when I started using 0.25 modifier. The temperature reinforcement decreased drastically. I still consider SAFE as a better option to design as it consider temperature difference between top and bottom surface. But still shear walls remain a question, as if I decreased stiffness modifier of forces F11=F22 it will transfer all loads to column. 


Temperature load is applied in conjunction with other load combination as stated by ASCE 7-10 with loads factor not less than 1.   Thats all I know about it uptill now. 

 

[UmarMakhzumi]

7 hours ago, rummaan17 said:

Dear Mr.Umar,

We have to submit our design for peer review which includes one international and one local consultant. I was amazed when they asked to apply load on overall structure. Even in my opinion its only outersurface that is most prone to temperature change, as here in middle east mostly building inside is temperature controlled. 
Though we also design slab for temperature change for all level, which i don't find just.   I was getting high temperature stress even in slab which was then modified when I started using 0.25 modifier. The temperature reinforcement decreased drastically. I still consider SAFE as a better option to design as it consider temperature difference between top and bottom surface. But still shear walls remain a question, as if I decreased stiffness modifier of forces F11=F22 it will transfer all loads to column. 


Temperature load is applied in conjunction with other load combination as stated by ASCE 7-10 with loads factor not less than 1.   Thats all I know about it uptill now. 

 

Rummaan17,

Thanks for the summary. It does provide some useful background on the topic. 

Temperature loads should be applied but only to areas exposed to temperature differential (Roof slab or periphery members if exposed). Lets get some feedback from @Rana as he might has more experience on temperature load application in commercial design and see where this goes.

Thanks.

[WR1]

1. I forgot to mention and as is Rummann understood correctly, that reason for using reduced stiffness for axial tension is that we will get less tension from analysis (due to less stiffness/restraint).

2. Negative temperature will produce tension near restraint (like slab supported on heavy walls).

3. What is the use of building? Is it a car park where you need to apply temperature to all the structure? Otherwise as Umar and you already pointed out, its the top slab where you need to consider differential temperature between top and bottom slab and thats not possible with ETABS 9 atleast.

4. Yes, tension should be checked on appropriate combination with other loads and that is 1.2D+1.6L+1.2T plus other loads if you have. If still you are getting tension that means it will be carried by modulus of rupture + steel only.

5. I would recommend you to use reduced stiffness of 0.25 and then also subtract the tension capacity of concrete (equal to half of 0.62(f'c^0.5)xthicknes in Mpa), half because slabs are restrained as per ACI.

6. For compression, cracking modifier should be increased but walls are quite stronger in compression. So just use 0.25. This is your separate model for temperature analysis. Keep it separate from lateral analysis where you might be using 0.7 or 0.35 modifiers for shear walls. Or it will affect the distribution of lateral forces to columns as you pointed out. But keep in mind, that is the real situation, if temperature cracks walls before earthquake hits, walls would be very weak transferring lateral forces to columns. Others can shed light on this more.

[rummaan17]

Thanks alot Mr.Rana . Indeed very useful. This thing was a pain for me since I have started applying temperature. 

At my last peer submission I finally realize that we should keep temperature model separate. There is also one more thing I figured out 

I was using Rigid daiphragm assumption at start which was resulting in very high axial forces in ETABS spandral. (ETABS doesn't design for spandral axial forces and designer has to check manually).  I discussed with CSi Support about this and they told to use semi-rigid as temperature has no point with rigid daiphragm.

Now, I use separate model with semi-rigid daiphragm for temperature which has decreased my axial load in spandral but has affected walls alot in terms of shear and flexure. 

Now what I understood.  That i will keep a separate model with all WALLS and SLABS modifier =0.25 just for temperature analysis. Hopefully it will solve my problem. 

 

[WR1]

Ofcourse, you should use semi-rigid diaphragm for temperature analysis or other axial loads.

[UmarMakhzumi]

13 hours ago, Rana said:

5. I would recommend you to use reduced stiffness of 0.25 and then also subtract the tension capacity of concrete (equal to half of 0.62(f'c^0.5)xthicknes in Mpa), half because slabs are restrained as per ACI.

6. For compression, cracking modifier should be increased but walls are quite stronger in compression. So just use 0.25. This is your separate model for temperature analysis. Keep it separate from lateral analysis where you might be using 0.7 or 0.35 modifiers for shear walls. Or it will affect the distribution of lateral forces to columns as you pointed out. But keep in mind, that is the real situation, if temperature cracks walls before earthquake hits, walls would be very weak transferring lateral forces to columns. Others can shed light on this more.

I think there are some items that need to be addressed with the above approach. The ULS state being checked should be done for only one stiffness modifier. The basis aren't consistent when we are using different modifiers for temperature load and others for active loads as both relate to finding the most critical ULS condition under combined loads. Use of separate model can be justified for either where code requires a special design condition to be checked like Dual Systems or when a different limit state is being checked like serviceability.

Second comment that I have is that temperature gradient across a member would not only cause axial loads but as well as curvature/ bending. If ETABS doesn't allow application of temperature gradient then these loads need to be applied to reflect the actual loading condition.

22 hours ago, rummaan17 said:

We have to submit our design for peer review which includes one international and one local consultant. I was amazed when they asked to apply load on overall structure. Even in my opinion its only outersurface that is most prone to temperature change, as here in middle east mostly building inside is temperature controlled. 

You can challenge your peer review team as long as you can defend your position clearly. Do you have load development document created for this project that summarizes basis of all loads?

Thanks.

[rummaan17]

2 hours ago, UmarMakhzumi said:

You can challenge your peer review team as long as you can defend your position clearly. Do you have load development document created for this project that summarizes basis of all loads?

 

Dear Mr.Umar, Yes we have created some document for their timely response to our design but I guess we dont have any offical document from start that could justify this load application technique.   Also Mr.Umar, certainly my review team never cease to surprise me as they argue over things that even doesn't make sense to me, who is a newbie in this field. Since, they are approving authority and we are tough on schedule we try to sort it out as they want.


Secondly, i have some similar concerns. This 0.25 modifier track back to reference from code ?   if any please guide me. I started using 0.25 modifier as I found this value on certain discussion on forums and after applying it I found the reinforcement to be just. but as you quoted, using two modifier at ULS state for two different load case ?  how can we back it up.

Any addition would be highly appreciated. Thanks in advance. 

[UmarMakhzumi]

@rummaan17,

I hear you and that's why we have this forum.

All these questions are legit.  @Rana might be able to shed better light because of his experience in commercial design. If not, we can all agree on most rational conclusion and run numbers to do sanity checks.

Cheers

[WR1]

First-off, I am not sure about your system, is it a moment resting frame, dual frame or shear wall frame interaction?

I assume it is a building frame system where lateral loads (all) are taken by shear walls (in-plane) and all gravity (mostly ~ 90%) is taken by columns.

The above assumption can be checked by checking reactions after applying 0.7,0.35 modifiers to walls and 0.7 modifiers to columns.

This is your routine design..Right?

Now comes the question of temperature loading, for that please update what is the temperature magnitude you are using? What is the use of building? It is heated/un-heated structure? For example if its a parking structure where under-sides of beams also have temperature loading, slab stresses will be minimal (almost zero if equal values). Also for parking structure, the first slab will have huge stresses due to fixity of base.

And as @UmarMakhzumi pointed out, you also need to check bending under temperature gradient (manually by using C and T approach for a single reinforced section or by computer using complex analysis).

Anyway, In my view, I hope that authorities would not go for that. But for normal temperature analysis, I would recommend you to make a 1 page A4 calculation, stating the design philosophy for temperature loads and put it as appendix or part of your report.

In that report you would mention,

1. use of semi-rigid option

2. use of reduced modifiers for tension cracking for walls (even for columns if they are showing tension). You can use 0.25 and if it not working reduced further. It would be nAs/Ag for the members where you havee tension.

3. show formula for modulus of rupture for restrained slabs and subtract that capacity from ETABS values.

4. Assign piers or average manually the axial force over whole length of wall piers to reduce magnitude further.

 

Make sure that reinforcement provided for temperature is same on both sides for slabs/walls and same on all sides for beams/columns.

 

Now next question is regarding lateral load RE-distribution.

If you reduced the stiffness of walls and columns both by same amount, that would not affect re-distribution of lateral forces.

But if you just reduced stiffness of walls, columns will start taking more lateral loads (depends on relative stiffness of wall/columns, I do not know the dimensions) so As @UmarMakhzumi pointed out, you should this design too. Because axial stiffness of walls/shells = in-plane stiffness of walls/shells so changing axial stiffness will also change the bending stiffness for lateral loads and re-distribution will happen.

So check for this condition too, you might need to increase some columns that again depend on dimensions and framing system. In that case, read Tranath book on tall buildings chapter 3 on lateral loads to see what you actually need to increase, columns or beams?

Please update us what happens.

[rummaan17]

Thanks alot Mr.Rana for taking time to explain it in such detail. i will act accordingly and let you know. 
My last concern if you could answer, this 0.25 value is coming from code ?