approximate methods of analysis

huangyhg

approximate methods of analysis
do you ever use approximate methods such as the portal or cantilever method to check to see if the computer program is generating reasonable results?   
always.  i do not use a computer output without being able to approximate the results by hand.
i have multiple, mostly dated, resources which i employ to do this, my favorite four being:
-reinforced concrete designer's handbook by chas e. reynolds (1964 british)
-steel designer's manual, 3rd ed, (1967 british)
-structural engineer's handbook,  milo s. ketchum (1924 usa)
-structural engineering formulas, ilya mikhelson (2004, isbn 0-07-143911-0)
there is only one exception to this personal rule:  when a structure is highly complexe (such as the stadium i am currently working on) i will review the results of computer output in detail and be sure that they "feel" correct.  i will often also review the structure for what the simple span reaction/internal forces would be and ensure that my design would support them.
i will be very curious to read other's replies!
cheers,
ys
b.eng (carleton)
working in new zealand, thinking of my snow covered home...
a follow up question, if i may...
just how close do others like to be with their approximate methods to trust the computer model?  personally this depends on the complexity of the structure, however i generally like to be within +/- 15%, with a ceiling of +/- 30%.  i have been convinced of greater differences from time to time, however being outside of this range simply prompts more investigation on my end.  
i think many would be surprised just how many times i have found gross errors only after having maually computed what i "felt" should be the answer...  computers have a great aptitude for making junk look attractive!!!
cheers,
ys
p.s.  my appologies for my french spelling of "complex" in my previous post.  now if i could just stop telling people to "close" the lights.... ;0)
b.eng (carleton)
working in new zealand, thinking of my snow covered home...
i use the portal method all the time when checking frame structure output by computer and expect my analysis to be within 5% depending on the application (elastic).  
i also use moment distribution for continous beams as well as aisc's beam formulas, and their moment, shears and reactions for bridges when applicable.
for complex structures, i will break into smaller components and estimate responses.  
finally, for fea stress and dynamic problems i will usually check roarks' stress and strain book.
regards,
qshake
eng-tips forums:real solutions for real problems really quick.

yes!!!!!!!!!!!!!!!!
i have two young engineer's who can model anything in risa and quickly (perhaps too quickly) - but they also come up with some amazing results....
when you forget a bearing or use the wrong section modulus or no lateral bracing --- things get out of hand fast and they don't seem to sense that yet??
go figure.  
before i start any analsyis - i look over the whole structure and i know before hand whether i will be close.  
all engineers shouold do that!!

young structural,
i noticed milo s. ketchum has written a lot books.  can you be  more specific about the book?  what information does it have that is helpful?
for those who do a manual check of the frames, do you also do a manual check of the diaphragm force distribution?  after all, checking the frame forces is irrelevant if you are using the wrong diaphragm forces.  that can get really complex for some of the torsional load cases for wind and seismic.  
i would be curious to know how you go about checking that.
absolutely!!!!!!!!!!!
mike mccann
mmc engineering
can i ask how you go about checking the lateral load distribution for some of the more complex cases?  for example, the wind load with the torsional moments and the seismic loading with the eccentricities?
i was specfic abusementpark, i said "structural engineering manual" by milo s. ketchum...
it's his structural engineering manual that i use most often.  it also has a great deal of graphic statics in it, a method that is simple, reliable, readily understood, and sadely nearly never taught anymore.  the more we advance, the more we seem to loose touch with the simple stuff.  and graphics statics is very useful for lots of structures!
cheers,
ys
b.eng (carleton)
working in new zealand, thinking of my snow covered home...
yes always,
if you don't know what a computer package is doing, then you shouldn't be using it.  i work with a very experienced engineer/dinosaur who doesn't trust computer packages at all and is always back checking my work with emperical formulas.
good engineering is all about efficiency and there is nothing more efficient than being able to find your own errors.