load required to reduce unbraced length

huangyhg

load required to reduce unbraced length
i am designing an aluminum tube column with a steel tube insert to help increase the capacity. the two columns are the same length and are adequately connected along the whole length. the member has lateral wind loads and therefore acts as a beam column.i am designing so that all of the axial load is taken by the steel tube and all of the lateral loads with the aluminum. in order to further increase my steel tubes allowable axial load i would like to say that it is braced at mid span by the aluminum. i would then apply a lateral force to the aluminum column to account for this bracing.  does anyone know of a common procedure to determine the force required to brace a column, to effectivly reduce its unbraced length. some individuals i have talk to mention using some percentage of the axial load, but no one is certain of common practice.  please respond with anything that you know about this and any references (books or magazines) that you might know of.  thank you
unless your aluminum "jacket" is significantly larger and stiffer than the steel column, you have a non-convergent condition.  as you brace the steel column with a connection to the aluminum column, the aluminum column will deflect under the load thus decreasing the reaction (bracing) against the steel, thus allowing more lateral movement of the steel, putting more load on the aluminum, causing more deflection of the aluminum, less bracing of the steel.......
secondly, attaching the aluminum column to the steel column will create a corrosion condition intolerable to the steel column, which will be hidden by the aluminum jacketing column.  this could be catastrophic without warning if the interior steel column fails under high axial load due to loss of section from corrosion, and you won't have any indication of the problem since the aluminum will have hidden the corrosion from view.
re-think your design approach.
ron
i have to agree with ron.  you are asking for serious trouble from a corrosion standpoint with this design.
i don't think this is a non-convergent condition.  at some point, it will reach equilibrium.  however, ron is on the right track regarding stiffness as a parameter in bracing.  traditionally, engineers have used 2% of the column axial force as a brace design force.  however, the traditional methods of bracing use kickers or other axial-force members that are very stiff.  if you want to use a more flexible bending   
beware of the greatly differing (aluminum's c.t.e is ~2x steel's c.t.e.) coefficients of thermal expansion.  this could cause significant thermal stresses in addition to your design loads.  a preliminary calculation shows that if your steel and alumimum are of equal areas, the thermal stress is about 5 ksi tension in the steel and 5 ksi compression in the aluminum for a 100 °f temperature change.
also, as already noted, is the very relevant concern for galvanic corrosion.   
how are the columns "adequately connected along the whole length"?
the aluminum column is considerably larger and stiffer than the steel tube.  i don't believe a non-convergent condition existis here.  as for corrosion the steel is galvanized and the aluminum is painted, the structure is also inside in a controlled enviroment diminishing the thermal effects.  they are 8 feet long and are attached together with (2) screws at 12" o.c.
considering your last post, why not just re-compute "r" as a composite and check buckling.  maybe you don't need lateral bracing.
ron
ps..corrosion will still occur, even in the environment you described.
i would second ron's point that corrosion will still occur.  since you are screwing them together, the steel screws will connect the aluminum directly with the raw steel that the screw is exposed to, creating a nice battery.
in terms of convergent/non-convergent, you must perform a pdelta analysis to verify its stability, or at least perform the necessary calcs to check it.  however, i agree that 8 foot long columns probably won't be subject to lateral instability.  
i attended the yura/helwig seminar in 1998 and they presented a brace concept that would seem to apply here.  for bracing, yura emphasized that the brace must have adequate stiffness as well as strength.
the required stiffness was given as beta = 4 x ps / l where beta is the required stiffness (kips/inch)
ps is the service axial load (in this case - in your steel column)
l is the length (in this case - 4 feet to the brace point)
the required strength was given as fbr = 0.004 x ps
the above was for asd.  for lrfd, he gave
beta = 2 x pu / (phi x l)
fbr = 0.004 x pu
he indicated that the above was supposed to be coming in the next aisc specification.  be cautioned that the above was a proposed concept, not yet adopted by any code.
so the force you apply laterally to the aluminum column would be the brace strength force fbr.  separately, you would want to check the stiffness of the aluminum column by placing a unit load of 1 laterally at the midheight of the aluminum column and verifying that the deflection is less than l / (4 x ps).  you also would want to include (at least i would) any additional pdelta force due to the lateral deflection of both columns (they will move together) developed from wind and axial load happening together.
i know it doesn't answer the question entirely but have you tried a concrete filled tube to increase the axial capacity of the aluminium?  this may also assist if fire protection is required.
nb
the concrete-aluminum reaction is even stronger than the steel-al one.  never let aluminum touch concrete.
witt55,
  why are you choosing this approach?  why not just size a steel