Each year 50,000 groove welds are made in production
of steel bridges. Under the requirements of AASHTO/AWS D1.5, Bridge Welding
Code (the Code), welds are tested by non-destructive examination (NDE). The NDE
acceptance criteria prescribed by the Code have effectively ensure excellent
performance of welds starting in the late 1950s. However, these criteria are
based on workmanship and not do not explicitly consider the factors that
influence whether a given flaw is or is not critical. For example, the current
AWS criteria do not consider the magnitude of the stress range, mean stress,
toughness of the material, height, or location of the flaw within the
weld. This results in unnecessary
repairs and may also permit some flaws that should be repaired to remain. Improved
criteria are based on fitness for purpose and align with actual performance
needs and both ensure that the condition of the weld is suitable for
performance and that the repairs made to welds in production are needed.
Further, these criteria can be used for existing welds when questions arise
about the welds’ condition and their suitability for continued performance.
Fitness for purpose was defined
by Wells as follows (as quoted by Barsom and Rolfe): “Fitness for purpose is
deemed to be that which is consciously chosen to be the right level of material
[that is, having the appropriate fracture toughness, e.g., KIC,
CTOD, J-integral, CVN, etc.] and fabrication quality [that is, the appropriate
loading or stress level for the application], having regard to the risks and
consequences of failure; it can be contracted with the best quality that can be
achieved within a given set of circumstances, which may be adequate for some
exacting requirements, and needlessly uneconomic for other that are less
demanding.” The approach of “the best quality that can be achieved” describes
the approach currently utilized for the fabrication of steel bridges. The
current weld acceptance criteria is generally workmanship-based, not fitness
for purposed-based. This leads to increased inspection costs, unnecessary weld
repairs, delivery delays and other undesirable economic consequences.
Previous attempts to establish
a fitness for purpose weld acceptance criteria for bridge groove welds has
resulted in acceptance criteria that are more rigorous than the current
workmanship-based criteria. While the predictions have a theoretic basis, they
cannot explain why the vast majority of bridges in service, fabricated to
workmanship-based standards, performance acceptably. Valid fitness-for-purpose
weld acceptance criteria must consider at least the following:
· the actual stresses on the welded connection
· the actual toughness of the deposited weld (with particular focus
on the toughness at the stress raiser)
· the actual stress concentration factor created by the
· the actual orientation of the stress concentration relative to the
The repetition of the term
“actual” is deliberate; simplifying compromises can significantly impact the
validity and value of any fitness for purpose model.
A fitness for purpose model for
steel bridges must consider the whole range of codes and standards that are
associated with modern bridge design and construction. The design of many
bridges is governed by fatigue or deflection, not by strength. Accordingly,
stresses in the welded connection are often significantly less than the tensile
capacity of the weld or surrounding steel. The details of welded connections
are routinely selected for fatigue resistance, which is directly associated with
reductions in stress concentrations (which also aids in fracture resistance).
Mature welding codes govern acceptable welding filler metals, base metals, WPS
qualification, welder qualification, and fabrication techniques. NDE personal
are similarly qualified and procedures controlled. NDE technology has advanced
and is capable of detecting and defining weld imperfections in more precise
ways, including the number, shape and orientation. The fitness for purpose
approach as proposed must examine all of these factors that are associated with
modern steel bridge welding in a holistic manner. The resulting
method will be compared to a database of defects and performance
Acceptance criteria for performance in service should
align with performance expectations that result from the design. In design, fatigue
is addressed by aligning loads and frequency of loading to codified fatigue
performance curves based upon 97.5% survival. Further, in almost all girder and
other member design, short attachments corresponding to Category C or C’
control the design reducing the stress at the groove welds in the structure,
and these lower stresses increase the flaw tolerance of the weld. Acceptance
criteria should consider both the design reliability and the actual demand that
results from design practices.
Acceptance criteria should also consider the
reliability of the NDE methods used for inspection. Ideally the judgement of
whether a flaw can be tolerance should be based on the flaw’s location, size,
shape, and nature – i.e., whether it is crack-like. Inspection techniques vary
in how accurately they can measure these characteristics but are improving over
time with advance in technology and data collection.
Literature Search Summary
Criteria of Complete Joint Penetration Steel Bridge Welds Evaluated Using
Enhanced Ultrasonic Methods”, NCHRP Report 908, Connor, Crowley, Fish,
Schroeder, Washer, 2019
b. API fitness for
service publications (details to follow)
c. Eurocode fitness
for service publications (details to follow)
Major tasks for this research include the following:
a. Establish the actual performance demand of welds that includes consideration of hypothetical conditions and how bridge welds have performed in service over the past six decades.
b. Establish the performance impact of actual weld flaws that occur in bridge welds and use these for establishing acceptance criteria instead of simply using an idealized modeled crack and calibrate with experimental fatigue test results.
c. Establish the real stress ranges seen by bridges in-service and use these ranges for establishing acceptance criteria instead of assuming the welds will see design stress ranges.
d. Evaluate new math models that account for effects of plasticity, three dimensional strains, residual stresses and their changes over time, and load histories.
e. Develop new acceptance criteria considering what is learned in a through d and ensure this is calibrated to service.
f. Evaluate current NDE methods for their suitability in measuring the new criteria. As needed, make adjustments to the proposed criteria that are needed for each method.