Return to index: [Subject] [Thread] [Date] [Author]

Seismic design of pole building

[Subject Prev][Subject Next][Thread Prev][Thread Next]
I am designing a pole building in northern California (Mendocino
county, near Hwy 101).  The poles are already in place, but the client
would like me to design the structural system for gravity loads and
lateral loads.

The poles are tapered wood poles, approx 12"-14" in diameter at the
base, and are embedded 8 feet in the ground.  The structure is two
stories above the ground (raised floor diaphragm and roof diaphragm)
with floor-to-floor heights approx 10 feet.  There are 3 rows of 8
poles each, with the rows 16 feet oc and the poles 8 feet oc in each

The client would like to keep the interior spaces as open as possible,
and therefore would like to avoid shear walls.  The proposal is to use
knee braces at each pole in each direction.  If the braces are at 45
degrees, then each leg would be approx 2.5 feet long to keep the
braces above 7.5 for adequate headroom.  Whether wood or steel braces
are to be used is an open question (see below).


1.  WHAT IS Rw?

    This building does not seem to fit into one of the accepted
    structural systems listed in 1994 UBC table 16-N.

    Strictly speaking, this is a moment frame system since the poles
    are required to resist lateral forces through bending, shear,
    and axial forces.  However, Item 3.4 lists only steel or concrete
    ordinary moment-resisting frames (not timber).  If steel braces
    are used, then this would be some sort of mixed system anyway.

    Item 2.4.c. (Ordinary braced frame of timber in a building frame
    system) seems like a close match, but Rw=8 seems too high for the
    actual lateral system proposed (see below for discussion).

    Item 1.4.a. or 1.4.c. (Braced frame of where bracing carries
    gravity load) seems like a close match too, but this is for a <<
    Bearing wall system>>.  However, using Rw=4 (for timber) or Rw=6
    (for steel) seems like a good idea (again, see below).


    This lateral system seems to have little ductility if
    constructed with wood braces. It may have some with steel braces
    (see below).

    I do not have any experience with wood braced frames or knee
    braces resisting lateral loads as the primary LFRS.  If a wood
    brace were to buckle, it seems that the connections at each end
    would fail if the connection were comprised of more than one bolt,
    due to the rotation at the ends.

    Steel knee braces may perhaps be made to work, using connections
    at each end similar to that used in special concentric braced
    frames where a gusset plate is allowed yield when the brace
    buckles (in the plane formed by the brace and pole).  However, I
    don't think <<exposed>> steel braces are what the client had in
    mind for a pole house in the country!

    In any case, it seems prudent to use a shear demand on the pole
    that would correspond to yielding of the brace, using conservative
    material properties.


    How do I account for drift without a soils report?  How do I know
    how much the building will deflect under <<real(?)>> earthquake
    loads with such a "flexible" foundation.  The pole formula in
    Chapter 18 (1994 UBC 1806.7) will give me the moment capacity of
    the pole at the foundation (assuming the worst soil) but it does
    not say what the rotation is under that load.  I realize
    determining the rotation is complex (read: non-linear), but I
    don't know where to begin (other than hiring a soils engineer).
    Perhaps some of you will think I'm making this too complicated,
    but I don't want to ignore the drift requirements.  The whole
    intent is to avoid collapse due to P-delta effects, isn't that

    Put another way, how can I model the superstructure reasonably
    accurately without knowing how much the pole will rotate at the

Based on the above discussion, my inclination is to use Rw=4, C=2,
and to design as much ductility into the knee braces as possible
(whether wood or steel).  I am tempted to use Rw=3 for an inverted
pendulum, but the "real" C value would then most likely be lower
than 2 anyway, due to the long period of the structure.  By using this
approach, most of the elements will remain elastic anyway, assuming a
built-in factor of safety of at least 3 or 4 anyway (for the
connections and for the pole values for axial, shear, and moment).
Remember, this is relatively far from active faults.

There are some other aspects to this this building (that I have not
described) that make it even more complicated.  I will mention those
only if it seems necessary to do so.  I think I have described enough
to generate some comments.

I would also appreciate any references (Internet or otherwise)
regarding this type of construction.  I know this type of construction
is used all over the US, but I am particularly concerned with adequate
seismic performance.


Mark Thomas Swingle, SE
Oakland, CA

PS  I will not be able to check my email for about a week, so I will
have to respond with clarifications (if required) and thank yous at
that time.  Thanks in advance for your comments.