equivalent static load procedure and modal response spectrum 3

[Saedhalteh]

What is the difference between defining the seismic load as static load using the equivalent lateral force procedure and as a dynamic load using the modal response spectrum procedure ?? or when I should use these different procedures , I get confused about this topic in general the static load analysis and dynamic load analysis . I would really appreciate it if someone can help me with .

 

[r13]

The applicability of each method is indicated in the building code. The ELF is the most simple for use, and the slope of sophistication is going up from here on. Read some text books regarding design for seismic force will help.

 

[Saedhalteh]

Do you recommend any book that clarify the dynamic loads analysis in simple way ?

 

[r13]

Instead of rushing out to buy a costly textbook and find little use later, I suggest to go through course like this Link, or seminar Link.

 

[r13]

Book: "The Seismic Design Handbook", by Farzad Naeim. One of the best book I've ever read.

By the way, if you are satisfied with gaining basic concept of the aforementioned analytical methods, you can read through the code, with attention paid to the commentaries. Afterward, you can even google the terms that unfamiliar to you, and were not well explained in the code, BUT beware the credibility of those source materials you are reading though.

 

[Saedhalteh]

Really appreciate your help. Many thanks.

 

[JoshPlumSE]

A very brief discussion of the difference between the two:

1) The equivalent lateral load procedure assumed that the first (or fundamental) mode of vibration of the structure will dominate the seismic response of the structure. If this is the case, then the procedure can generally be used. This is true even if you different "fundamental" modes for each direction. This would be the case for a structure that has a braced frame in one direction and a moment frame in the other.

2) The response spectra method is generally used when a single mode cannot adequately capture the response of the structure in a particular direction. Reasons for this are usually spelled out in the code:
a) A stiffness irregularity between adjacent floors
b) A mass irregularity in the structure.
c) A torsional irregularity.

I'm probably missing something else. But, the general concept is what I said above, that ELF procedure assumes a single mode dominates the response in each direction. If this assumption is not valid, then you need to use a different method....

 

[JLNJ]

The ELF method is simple and (usually) conservative.

Sometimes you must perform a dynamic analysis, but it may work in your favor when it comes to designing members. You may well get more favorable lateral forces and smaller members.

 

[rscassar]

@Saedhalteh

Equivalent static load only considers a single mode and you get and upside-down triangular load distribution whereas modal response analysis is a response of all modes. The closer the mass participation of the first mode gets to 100% the closer the load distribution is to the equivalent static method. Dynamic analysis may result in larger forces, equivalent static method requires applying the forces 100% in one direction and 30% in the orthogonal direction. Dynamic analysis you apply 100% in both orthogonal directions to capture the maximum demand.

Which building code and software are you using?

Just watched a good video on this the other day Link

 

[JoshPlumSE]

Dynamic analysis you apply 100% in both orthogonal directions to capture the maximum demand.

What code are you using? My guess is you are are not fully understanding it, or are slightly mis-interpreting it. For Response Spectra Analysis, if you are combining load from both directions, you typically use an SRSS combination of the two directions.... Which is statistically pretty similar to a 100% + 30% combination.

Note: Some of the older UBC / IBC codes allowed a 100% + 30% combination for response spectra loading as well. But, I believe they found that folks were not understanding this correctly and were doing an SRSS combination of 100% and 30%. Not good.

An example of the difference, Say you get an axial force of 10 kips in a column for each load direction:

100% + 30% ===> 13 kips total axial force.
SRSS ===> sqrt (10^2 + 10^2) = 14.14 kips
SRSS of 100% and 30% ==> sqrt (10^2 + 3^2) = 10.44 kips.
​

 

[rscassar]

Using Australian code, and it is not very informative. The Australian code equivalent static method is much the same of other international codes, but Dynamic analysis section is a 1-pager which basically says either 'Modal Response' or 'Time History' are suitable analysis methods.

Engineer's in my locale take different approaches to this. I have spoken with people who when defining their response spectrum case will apply two cases:

Case 1= 1.0*gravity_accelertion_X-dir + 0.3*gravity_acceleration_Y-dir
Case 2= 0.3*gravity_accelertion_X-dir + 1.0*gravity_acceleration_Y-dir

But my understanding when defining a response spectrum load case the correct application is:

Case 1= 1.0*gravity_accelertion_X-dir + 1.0*gravity_acceleration_Y-dir

 

[hardbutmild]

An interesting discussion about combining seismic effects. Maybe I'll go offtopic, but I've always done it like this.

This makes sense only when talking about modal analysis so I do a modal analysis and combine all the effects for one direction (for example x direction), usually with CQC. Then I do the same for Y direction (some modes might activate 0 mass in y direction so they'll give 0 moment). In the end I combine x and y direction with 1,0*X +- 0,3*Y and the other way around. This might be wrong, but I don't think it matters much.

Why? Because it's usually a very rough estimation anyway. For example, if we're talking about RC modulus of elasticity changes significantly during a structure's life (usually 3-4 times). Also, the value given by the codes is usually underestimation of MoE. Cracking is different if there was a smaller earthquake before the design one, and it depends on a lot of things. We usually use the same stiffness reduction for all the elements which is certainly not the case. Non-loadbearing elements are rarely considered in design. Usually dynamic soil parameters, such as stiffness are not known, but evaluated (at least where I'm from). Local conditions, such as small hills are rarely considered. All this can influence the fundamental period (and all the rest) significantly.
I think I read an article by Fajfar once where he and his team made a simple experiment of RC cantilever with a lumped mass at the top. They then gave all the data (exact values of MoE, fc', connection details, direction, intensity and shape of earthquake etc. at the time of test) to something like 10 different experts and asked them to evaluate the response. I remember that the dispersion of the results was very high.

Why then worry about what is statistically the best way to combine actions? Maybe I'm wrong, but I think that detailing and rational design is the most important aspect of seismic design.

 

[JoshPlumSE]

Rcassar -

See my previous post. Response spectra combinations from multiple directions should be a "statistical" combination method. Like an SRSS (square root sum of the squares) or the 100% + 30% method. So, while I'm not explicitly familiar with the Australian seismic code, I believe your understanding is incorrect.

 

[rscassar]

@JoshPlum

Thanks for the insights. I will have a look into it.