axial load for rafters

huangyhg

axial load for rafters
when dealing with pitched rafters does anyone here consider the axial load generated by dead weight, snow, and live loadings?  usually, these type beams are handled using a horizontal plane method... which is theoretically incorrect.  true, you arrive at the same values for moment but you actually have an axial load per foot of beam as well.  
are the stress interactions typically analyzed or only in beams of a larger span, say 40 or more feet?
thanks for the input.
yes.  the outward thrust on the walls is one component you have to check, and to do so requires resolving all the forces in the rafter.  you will also need it for designing the connection to the collar tie (beam).
thanks for the reply.  sure, the thrust definitely needs to be taken into account.  i was just curious if you guys always consider the combined bending and axial load interaction for the   
if you have flat bearing at the top and bottom of the rafter, a strap over the ridge with a35's at the bottom should suffice.  i realize that this is against the principles of statics, but if the bearings top and bottom are flat, how is the horizontal component to be generated at the supports, and hence in the rafter, unless through frictional forces, the strap and a35 resisting such? so the question remains, if a strap is provided at the top only, technically no connector is needed for horizontal forces at the bottom, as it is removed at the top.  
collar tie connections are nice if the floor joists and rafters align, but this is not normally the case unless you space the collar ties at 48" on center.
mike mccann
mmc engineering
if the rafters are pitched at some angle, then gravity loads will generate some axial forces due to traction. likewise if your relying on frame action, then a horizontal thrust at the supports will generate compression in your rafter. also if you have roof bracing, the rafter will be behaving like a vertical chord in a truss and will again develope axial forces.
if you have underpurlins or a load bearing ridge board then the support at the top of the wall can be treated as a roller and you have no lateral component of the reaction.
you do have a small axial load in this case but it can be treated as negligible.
for a truss or a roof containing collar ties, i would always take the axial load into account.
for the common case where the lower support is pinned in both directions, and the upper support is free in the vertical direction, i use the interaction formulas for compression and bending, no matter what the length of the beam or joist.  i use the compressive force at the point of maximum moment.
before getting too crazy with checking compression, remember that the rafters will be braced continuously along their top by the roof sheathing.  they will not be allowed to have gross section buckling so the allowable compressive stress will be much higher than for a column.  so the effect of compression will likely be negligible compared to the bending.  the unbraced bottom side will still be in tension because the bending stresses are much higher than the compressive stress due to axial load.
when checking your combined loads, compare the design you get with what is provided in the tables in the international residential code.  these tables can help verify that your design procedures meet standard practice.  just make sure when doing your comparison that you meet all the assumptions made in the use of the irc tables (ceiling joist location and connections, non-load-bearing ridge, etc.).