double curved glu lam beam

huangyhg

double curved glu lam beam
does anyone know a good source for the design of a double curved glu lam beam?  the aitc timber construction manual says that its design method only applies to single curved glu lam beams.   
is this a reverse curve or a compound curve?
model it in risa 3d with the same design constraints.
mike mccann
mmc engineering
follow msquared48 advice!  when i have done this in the past i have used a matrix method analysis program to determine the forces in the members.  i then checked the stress by hand.  what you have to watch for when using a standard program are the boundary conditions and the member length.  when checking for lateral stability some programs will use the length between joints as the   
it is a reverse curve; it goes from a 51.5' radius curving up to a 22.5' radius curving down.  the right support is 3.5' higher then the left support.
the way i would approach the problem is to draw the beam to scale in autocad then break it into a number of short segments.  the number of sections would depend on the level of accuraccy i was looking for.  i would probably break each curved section into 10 sections.
there are a couple of things to watch for.  first of all you have to check your radial tension stresses.  with your beam you probably have both radial tension and radial compression.  the other thing you want to check closely is your support conditions.  for the glulam to behave like a curved beam one support has to behave as a roller support.  
if both supports behave as pins, resisting both horizontal and vertical force, then your glulam behaves like a radial arch.  generally unless you design the system as an arch your supports will not be stiff enough to resist the horizontal thrust.  
if you model the beam with a roller support you will get horizontal deflection of the beam at the roller.  the stiffness of your member will determine how much horizontal deflection you will have.  if the deflection is large enough you may need to detail one support as a slip connection.  if the horizontal deflection is small 1/2" to 3/4" your supports may be able to tolerate the deflection with out using a slip joint.  i always assumed that 1/2 the deflection occured at each support.
you should go to the aitc web site.  they have a really good guide on the design of curved   
depending on the assumptions made, the design will vary.  fix the ends horizontally, in flexure and vertically and the allowable load will increase.  uniform loading versus concentrated loading will change the deflection amount.  assume select timber in the outer plies for sizing the beam.