Concrete stress strain curve by ACI 318 3

[keewan]

I am study a case which is quite difficult so I will Open this thread to ask my questions, so please help as much as you can
1. I found a sheet calculating moments based on a model of stress strain curve(concrete) by making different layers instead of using Whitney block, and I suppose this method is adopted by ACI318, So any one have a clue,

see the attached document

Also ,if anyone who is structural engineer with concrete experience who is willing to guide me through this design process am going through(Since i feel completely lost),we can chat on private and even speak so I can get the whole picture better


Thank you all

 

[Guastavino]

It's basically a spreadsheet that calculates a more accurate response at the ultimate level by stair step integration along the curve for forces and moments. Since the code allows the whitney stress block, I use it. I've done the exercise you show above, but in predicting the response of an actual beam, I doubt it's much better than the whitney method.

As for your comment about wanting help from a structural engineer, this board isn't really for that, unless you are a practicing engineer. Since you say you are "completely lost", ethically, it would be my opinion that you shouldn't take on such a project. If you are just trying to learn about concrete, I suggest getting a book like "reinforced concrete mechanics and design" by Macgregor and wight and learn away.

 

[rapt]

Whitney stress block is a simplified model which is ok for ultimate strength of simple sections and reinforcing arrangements. If you need to calculate actual strains and stresses at any level and in any of the components of a section, you need to use a more accurate stress/strain model for the concrete. Also, if you are trying to evaluate conditions at any condition other than ultimate strength, a more accurate model is required.

We tend to use the Eurocode model as it gives a good representation of the variation in the concrete stress/strain curve for different concrete strengths.

 

[keewan]

First of all I am not bad engineer, I just want to learn, and don't worry I will not do anything I am not sure about because it is not ethical to do so, so I am here to ask and read more to become better. Back to our discussion, I read a book regarding stress strain curve and you are right, Whitney block is just for simplification and it’s more accurate to use a model for stress strain curve, as far as I know now Whitney block integration volume and force point of application will be almost the same as the model, however the model will be more accurate (while Whitney block is more easy to use), am I right????

kindly see the attached document to continue discussion

 

[keewan]

http://files.engineering.com/getfil...e-b5a3-b0befc7d917d&file=question_-_Copy.docx

 

[keewan]

This is all for now

 

[keewan]

Finally I would like to thank you all ,and I appreciate all your help ,and I am really keen if rapt can guide me more so I can find my way and clear all the doubts

 

[JoshPlumSE]

Perhaps I'm speaking up a little late on this one. But, that stress profile looks very similar to one show in the PCA notes. So, you might pick up the PCA notes for more discussion on the subject.

A few years ago RISA incorporated this stress block as a more exact option when compared to the whitney stress block. It took a lot of custom programming. A marketing feature, if you will, to demonstrate how much better the RISA solution could be compared to our competitors and how it would give a user a competitive advantage... et cetera.

What, don't remember every seeing that in our marketing literature or when speaking to our engineers? That's because we found out that the member capacities really didn't change very much. So, there wasn't anything for us to promote. It's still nice to have something to compare against. But, I wouldn't expect to get an end result that is significantly different than what the traditional rectangular stress block gives you.

 

[Ingenuity]

This paper Link is titled "Time to dump the rectangular stress block?" by Eng-Tips member IDS.

Worth a read to compare to ACI, AU and EU codes.

 

[IDS]

I'm not sure if I should admit to enjoying re-reading my own paper.

In the ACI context I think the important points are:
[ol 1]
[li]The paper references the 2005 ACI code. I don't know if there have been significant changes to the code stress block requirements since then.[/li]
[li]For pure bending or small axial load the stress block makes very little difference.[/li]
[li]For a rectangular section with concrete up to about 50 MPa a rectangular stress block will give very similar results to a more exact formulation.[/li]
[li]For very high strength concrete the ACI stress block becomes increasingly unconservative.[/li]
[li]For a circular cross-section (or similar) any rectangular stress block will be unconservative by about an additional 10%.[/li]
[li]In the most extreme case, with 90 MPa concrete and the axial load that gives the maximum bending moment to ACI provisions, the ACI results are about 30% higher than the most conservative, or 15% higher than results based on actual stress-strain behaviour.[/li]
[/ol]

As someone pointed out in a recent thread, we don't have columns failing all over the place, but on the other hand there are not a huge number of structures using very high strength concrete and designed to the limit of loads that will produce the maximum possible moment capacity.

Finally I'll reinforce the point made earlier that this is looking purely at ultimate capacity. If you are interested in moment-curvature behaviour under high loads a rectangular stress block is an over-simplification.

Doug Jenkins
Interactive Design Services

http://newtonexcelbach.wordpress.com/

 

[keewan]

JoshPlum you are right(please see the attachment)
Ingenuity and IDS I enjoyed reading the paper

Guys ,I extended the discussion for further clarification kindly see the attachment(I was not able to write here because I need to use word tables),kindly see the attachment

http://files.engineering.com/getfil...b7-8b65-e4a5e991d6ff&file=question_-_Copy.pdf

Thank you all

 

[keewan]

Hey guys, please let us continue the discussion

Thread UP UP UP

 

[keewan]

njlutzwe,,
rapt,
JoshPlum,
IDS
Ingenuity
Lets us continue guys

Thank you

 

[keewan]

maybe some of you are not capable to view from attachment
kindly see
page 1

http://postimg.org/image/7l1d1h4yn/

page 2

http://postimg.org/image/j0gj2uylf/

 

[keewan]

UP

 

[JoshPlumSE]

I've just been tied up with other things. But, don't worry.... I will look at it in a bit.

IDS's paper did jog my memory a little bit:
I now remember seeing some minor differences with columns under mostly laxial load + small amounts of bending. But, the differences between rectangular and parabolic stress blocks were still not enough for us to promote agressively. Though I don't believe we would have tested very high strength concrete. And, we would not have tested versus a Eurocode type of stress block either.

 

[JoshPlumSE]

I don't know that I will be able to explain exactly what they're doing with that table and / or spreadsheet.

Some quick thoughts:
Obviously, he is dividing the beam into 8mm layers.
The assumption is then that the stress in each layer is essentially constant. Therefore, the centroid of the force in that layer is known (i.e. mid-depth of that layer).
If you know the maximum strain and the depth of neutral axis then you can use that parabolic stress block to calculate the stress for each layer.
This in combination with the assumption of constant stress in the layer, allows you to calculate the force in that layer.
Knowing the magnitude of the force and the location of the centroid, you can then know the contribution to total moment contributed by that layer.

Now, there is no way that I can determine how he came up with the depth of neutral axis or maximum strain or such. So, I have no way of knowing how correct or incorrect this method is.

However the general method strikes me as a typical numerical method. Think back to our calculus or pre-calculus days (a long time ago for me!). We did exactly the same thing to sum up an approximate value for the "area under the curve". Then demonstrated how the limit of this method (as delta x approaches zero) will be equal to the exact area. Same concept, the accuracy will get better as the layers get smaller than 8mm and less accurate as the layers get bigger than 8mm.

 

[rapt]

Keewan,

I think we have provided sufficient information for you to investigate further using your engineering education and knowledge. If you want more information, find a good text book.

 

[keewan]

Thank you