Page 387 - Forensic Structural Engineering Handbook

11.18 MATERIAL-SPECIFIC FORENSIC ANALYSES

Fortunately, a number of methodologies have been formalized in recent years for quan-
titatively assessing several forms of corrosion damage in steel structures. API-579/ASME
3
FFS-1 provides methodologies for assessing general corrosion wastage or metal loss, local
metal loss (local thinning), and pitting corrosion—all corrosion mechanisms that occur in
structural steels.

ASSESSMENT OF STEEL STRUCTURES
WITH CRACK-LIKE FLAWS

Root cause analysis of steel structures with crack-like flaws is most reliably accomplished
using fracture mechanics. 3,8,17 Crack-like flaws are structural and/or metallurgical disconti-
nuities that are planar in nature, that is, flaws that are characterized by their length and
depth, with a sharp root radius along their boundaries. Typical planar crack-like flaws
include fatigue cracks, sharp notches, and weld discontinuities such as undercut, incomplete
fusion, and incomplete penetration. Discontinuities such as porosity and slag inclusions are
considered volumetric and are less severe relative to fracture initiation.
Given that crack-like flaws induce local stress fields which are not readily characterized
by customary stress analysis techniques (including the local stress concentration approach)
fracture mechanics has evolved as the only acceptable methodology for predicting the behav-
ior of cracked structures. Within the fracture mechanics framework, the failure assessment
diagram (FAD) method 3,17–20 is the most comprehensive and robust approach. This method is
particularly useful when evaluating structures, which can potentially fail by the interaction of
rapid-unstable (brittle) fracture, ductile tearing and/or plastic collapse—the modes of failure
most frequently encountered in steel structures with crack-like flaws. A commonly utilized
FAD failure locus is described by the following relationship in terms of the brittle fracture
ratio, K , and ductile (reference) stress ratio, L , and is shown in Fig. 11.8.
r
r
6
2
K = ( .10 − .014 L ( ) )( .03 + . exp[ − .065 L ( ) ]) for L ≤ L (11.4)
07
r r r r r r(max)
where L r(max) is the ductile fracture cut-off limit and varies from 1.25 for structural steels to
1.80 for austenitic stainless steels.
A FAD analysis requires the calculation of the crack-like flaw assessment point, L′ – K′,
r
r
where L′ is the limit load element defined as the ratio of the applied stress to the corre-
r
sponding stress at the limit load and K′ is the unstable (brittle) fracture constituent defined
r
as the ratio of the applied stress intensity factor, K, to the fracture toughness, K . Failure is
C
predicted to occur when the assessment point, L′ − K′, intersects or exceeds the FAD fail-
r
r
ure locus, as shown in Fig. 11.9.
Stress intensity factor solutions, K, and reference stress solutions, L, for the same flaw
3
configuration can be obtained from existing compilations or derived using the finite element
method. For example, the stress intensity and reference stress solutions for a plate with a
through-wall crack are as follows:
K = (σ + M σ )( f ) π c (11.5)
m b b w
where c = crack length
W = plate width
σ = membrane stress
m
M = membrane magnification factor
b
σ = bending stress
b
f = crack length to plate width factor
w