Page 533 - Forensic Structural Engineering Handbook

TIMBER STRUCTURES 14.23

10 years ago would require a load duration factor of 140 percent which exceeds the 115 percent.
However, the degree of overstress does not approach the 1.88 value where we would expect
fiber damage in these examples. The grade of the timber at and near the net section is obvi-
ously very important. We still need to compare the gross section of the bottom chord with
DF #2 values where we end up with 105 percent with 20 psf snow and 135 percent with the
previous 30 psf snow event with the gross section (plus flexure on the bottom chord) com-
pared to allowable values.
It should be noted that we are operating near the ultimate stress region of the timber with
this review. It is like LRFD without safety factors. It could be argued that once the stress
exceeded the allowable stress for a length of time exceeding that allowed for the duration
factor, the safety factor has been exceeded and that the remaining structure should be
replaced or repaired to the point that the supplementary structure can carry all the load.
When doing such a study with all the variables, it is easy to understand why we have factors
of safety. By the way, how accurately was that 30 psf snow measured and did it go away
instantly after a day? You need to use your judgment and to be as clear as possible in your
forensic report to include the accumulated fiber damage potential. However it is difficult
enough to convince owners and insurance companies to replace or repair actual damaged
members, let alone overstressed ones. This is particularly true if the damage cannot be seen.
So the method indicated here compared to the K g/C is intended to reveal actual damage or
f
d
at least indicate the potential with the same 5 percent exclusion rate used for design criteria.
You are in dangerous territory as there is no factor of safety here so you must be extremely
accurate and include secondary effects that may add to the overstress. Excessive deforma-
tion beyond that determined for creep and deflection may be a visual indicator of fiber dam-
age accumulation for flexural members, but this is not always evident.
There have been numerous structural failures where people have had an opportunity to
escape prior to collapse of a building. Loud noises and severe deformations may precede the
collapse. In concrete and structural steel this is usually due to the ductility of steel, both in
reinforcement and in steel members. In timber this is often due to the duration of load effects
of the material. When overloaded, even to the point of initial fracture, redistribution of stress
(at a very high level of stress) to the unfractured material (potentially previously undam-
aged) can sometimes provide the precious seconds needed for escape. A hypothetical stress
distribution for a beam is shown in Fig. 14.3e, where the outer fibers have accumulated fiber
damage becoming ineffective so the remaining reduced effective section becomes more
stressed, therefore lasting less time than the damaged portion at the edges. I would not want
to be under a beam with this stress distribution as it may have already fractured before it got
this far. With the accumulation of fiber damage and reduction of effective section, it is eas-
ier to understand why timber structures can fail under lesser loads than they carried previ-
ously. It is then theoretically possible that a timber member could fail some time after all of
the snow load or other live load had been removed and collapse occurs later with just dead
load. The total section accumulates fiber damage, creating a reduced effective section and
the live load is removed just before failure. The dead load now causes overstress on the
reduced effective section, continuing to accumulate fiber damage until failure. Has this hap-
pened? Yes. (Refer to Case Study 9, Wilson High School Auditorium.)

Fracture. Although the effects of “accumulated fiber damage” from long duration over-
load have been emphasized, it does not mean that timber does not fail with a short duration
overload or that it is required to have accumulated fiber damage prior to failure. It is just
that the same member without “accumulated fiber damage” will be able to withstand a
greater overload prior to failure.
Impact. Since timber does have a high tolerance to short-term loading, the ASD load
duration factor of 2.0 (with some exceptions) means that we usually ignore the effects of