parapet design - asce 7 - roof pressures

huangyhg

parapet design - asce 7 - roof pressures?
per section 6.5.12.4.4 of asce 7-02 two wind load cases shall be considered for parapet design.  the first involves positive wall pressure to the front surface  and negative corner zone roof pressure to the back surface.   
the commentary describes this principle of using roof pressure as "based on the idea that the zone of suction caused by the wind stream separation at the roof eave moves up to the top of the parapet when one is present.  this the same suction which acts on the roof edge will also act on the back of the parapet wall"
using both windward and leeward wall pressure makes sense to me, but windward and roof suction seems very conservative and hard to visualize conceptually occuring. does anyone have any references regarding this issue?  (i searched other post but wasnt' able to find anything)
please let me know your thoughts.  thanks!
   
if using both the windward and leeward pressures makes sense, then you need to pick a value for the leeward pressure.  since the roof edges have a suction value defined it's logical to use the same value at the parapet.  what else would you use?  the suction shouldn't change by that much due to the orientation of the surface.
i would tend to differ with your opinion.  i don't believe the concept of using roof uplift for the leeward side of a parapet is 'logical'.  i have had this discussion with many other engineers within my office and all think this system seems far too conservative and the logic behind it is not entirely clear.
this is primarily due to the reason you mention.  roof uplift loads are based on being perpendicular to the roof surface.  simply assuming this same load applies at a 90 degree angle seems very odd.
i visualize wind moving over the parapet and 'pulling' it similiar to the same way it would pull the leeward side of a building as it moves over the roof edge.  this is why it would seem more logical to me to use the leeward suction load.  this is, however, typically a much lower load than that of the roof uplift.        
any others have thoughts on this topic?  
agree with jedclempett - there is a dead zone of suction directly behind the parapet that pulls equally on the back side of the parapet and the roof.  suction doesn't care which orientation the surface is (think of internal pressures - if you pull a vacuum in a box the side walls and top and bottom all see that same vacuum).  
thanks willis - great input.
see following link for a visual
if you can get a copy of asce 7-05, the commentary section c6.5.11.5 gives a decent explanation.  according to the commentary, the 2002 edition was based on a rational method but the 2005 edition was based on research results.
leeward pressure is positive + negative wall pressures. windward uses positive wall with roof uplft. seems reasonable to me until you calculate the pressures which seem very high to me. the mwfrs loads on a parapet seem very high and result in diaphragm problems if the parapets are very high. i don't think they have done much testing on parapets and these values are an upper bound.
i agree that the mwfrs parapet pressures are extremely high. for a single story building, the parapet pressures for a 5' parapet can effectively double the wind force to the bracing system. this doesn't seem right.
i designed a tilt wall project a while back that had big parapets. the numbers basicaly said that a wall panel will fly. it was hard to explain to the client what was happening. we have started adding a sloped roof area on the back side of the parapet to reduce loads for this condition. i don't think the results are realistic but you have to follow the code.