Need a book? Engineering books recommendations...

Return to index: [Subject] [Thread] [Date] [Author]

Re: Site specific response spectra

[Subject Prev][Subject Next][Thread Prev][Thread Next]
Dear Chris,

I have been analysising two storey to seven storey masonry builidngs as part
of my research work.  Its purpose is to try and get a handle on what a shear
wall sees in a real building.  The building model I used was based on
Benedettis work at Milan.  

It is interesting to note that one collapsed when they were tuning the table.

I have been using a number of earthquake traces,  but the one that is most
interesting is Nahanni.  It is a horrendous event.

What I have found is a very strong selection of the dominant mode in the two
principal directions.  That's waht one would expect from the mathematics so
I am not trying to teach about egg sucking.

The problem is in an event like Nahanni there is so much energy around right
across the spectrum that most buildings will find energy at their natural
modes.  It may be less than the peak ground acceelration,  but 0.1g is
enough to stuff brickwork  and their are very few frequencies at Nahanni
that are less than that level.

The Fast Fourier transform of acceleration trace from nodes in the shear
wall shows a very strong peak at the dominant frequancy of the direction of
the shear wall.

So depending on where you live and how long since your area has had a major
event I would be loath to scale back any loads that can get into the
building, unless you have a basement.  Rutherford and Chekenes recent report
for Abrams on the LAMB project comments on this point.
 
By the way the UBC loads are implicitly minimum design loads.  For
Californai the method is probably ok,  but for Eastern North America, I
would be extreemly loath to design a building for less than a good
Intraplate event say Nahanni.

Yes I am conservative,  make no apologies for that.  Your problem is picking
the buildings in America that will last a millenium and be hit by a New
Madrid 7+ or series.  The problem is of course cost in terms of human life
and buildings.  This is of course always teh problem with codes.  it would
be interesting to known teh retrun period for teh design of teh white house.
Then again the old british saying probably applies 

The King is dead long live the King.  (Dennis I have no idea who coined this
phrase so please if you know dinna tell me.)


The other thing is it is much more imporatnt to get a good bond on the
masonry.  If you are using calcite white bricks Sugo's recent work in
Australai definately shows that that brick needs lime in the moratr to
achieve bond.  A 1:1:6 mix is best.

Good luck

john nichols





Chris Palmateer wrote: 
 
>Due to the occupancy of the structure, a site specific analysis was  
>performed by the geotechnical engineer...for short period structures,  
>the accelerations are greater than the normalized response spectra  
>given in Fig. 16-3 of the UBC...we are now designing the "structure"  
>for this increased base shear...Do "elements" of the structure also  
>need to be scaled up by the same factor?   
 
I'll give you two answers: 
 
1. Since the UBC allows you to "scale down" the results to the code required 
base shear, as far as I can see there is no code requirement to "scale up"
the 
element forces.  The sole purpose of the site specific response spectrum in 
this case appears to be to determine a more accurate distribution of the
base 
shear within the structure.    
 
2. However, if the intent of obtaining a site specific response spectrum is
to 
design for the potential of higher than code prescribed forces,
amplification 
of code values for elements should be considered.  Since you state that you 
are designing for increased base shear, it appears that it is your intent to 
design for forces beyond code prescribed forces.  (This is not uncommon for 
critical structures.)  In this case, I would consider the code element
forces 
as minimum forces, and I would calculate the acceleration at each level of
the 
structure from a dynamic analysis using the site specific response spectrum.
 
Then I would base the acceleration of each element on these results.  (I
would 
divide the calculated acceleration by the R-factor for design.)   
 
Hope this helps.