A steel bridge girder can accumulate millions of stress cycles in a single year of service. Each passing truck adds a small increment of cyclic load to every welded connection, cover plate end, and stiffener detail along the span. The steel itself does not yield under these loads — the nominal stresses are often below 30% of the yield strength. But over decades of repeated loading, microscopic discontinuities at weld toes, cope holes, and attachment points grow into detectable fatigue cracks that, if left unaddressed, can propagate through the full member cross-section and compromise the structural integrity of the bridge.
The Two Faces of Fatigue in Steel Bridges
Fatigue damage in steel bridges falls into two fundamentally distinct categories, each requiring a different analytical approach and different retrofit strategy. Understanding which type is at work on your bridge is the difference between a repair that lasts and one that fails within months.
Caused by cyclic in-plane stresses that can be directly correlated with bridge live load using conventional design theories. These stresses are predictable, calculable, and governed by the AASHTO S-N curve framework. Common locations include cover plate ends, flange splice details, and welded attachments where the stress range exceeds the detail category threshold.
Caused by secondary stresses that cannot be calculated with conventional methods — they require refined finite element analysis or field strain gauge measurement. An estimated 90% of all fatigue cracking in steel bridges is distortion-induced, typically at web gaps where connection plates are not attached to both flanges.
AASHTO Detail Categories A Through E': What Each Classification Means for Your Bridge
The AASHTO LRFD Bridge Design Specifications define eight detail categories for fatigue — A, B, B', C, C', D, E, and E' — each corresponding to a specific S-N curve that relates the applied stress range to the number of cycles to failure. These categories are not arbitrary. They were derived from thousands of full-scale fatigue tests conducted at Lehigh University and other research laboratories, establishing a statistically validated lower-bound fatigue resistance at a 97.5% probability of survival for each detail type.
The S-N Curve Framework: How Fatigue Life Is Quantified
The S-N curve is the fundamental tool for fatigue design and evaluation. On a log-log plot, stress range (S) is plotted against the number of cycles to failure (N). For each AASHTO detail category, the curve has two regions: a sloping finite-life line where the fatigue resistance follows an inverse cubic relationship with the number of cycles (N = A / S^3), and a horizontal infinite-life line defined by the constant amplitude fatigue threshold (CAFT). If the applied stress range at a detail is below its CAFT, the detail is theoretically capable of infinite fatigue life under constant amplitude loading.
Web Gap Cracking: The Most Common Distortion-Induced Fatigue Problem
Prior to the mid-1980s, bridge girders were commonly constructed with a gap between the end of the transverse connection plate and the tension flange. This practice was intended to reduce the stress concentration at the welded connection plate-to-flange detail. However, it created a different problem: the web gap region became the site of severe out-of-plane distortion under live load. As the cross-section rotates under truck loading, the relative displacement between the connected and unconnected sides of the web gap induces high localized bending stresses that can exceed yield at the web gap toe — even though the global girder stresses remain well within allowable limits.
When one girder deflects more than its neighbor under an eccentric truck load, the cross-frame or diaphragm forces the connection plate to rotate. The web gap — typically 2 to 4 inches between the connection plate end and the flange — must accommodate this rotation through out-of-plane bending of the web plate.
The most effective retrofit for web gap cracking is to weld the connection plate directly to the tension flange, eliminating the gap and preventing the out-of-plane distortion mechanism entirely. Alternative retrofits include softening the connection plate end or removing diaphragms where structurally feasible.
Non-Destructive Testing Methods for Fatigue Crack Detection
Detecting fatigue cracks before they reach critical length is the central challenge of bridge fatigue management. The National Bridge Inspection Standards require fracture-critical members to be inspected every 24 months, but the choice of inspection method determines whether a crack is found at a 1 mm length or a 50 mm length — and that difference determines whether the repair is a simple drill stop or a major structural intervention.
Retrofit Methods for Fatigue Cracks: What Works and When
The choice of retrofit method depends on whether the objective is to arrest an existing crack (repair) or to upgrade the fatigue resistance before cracking occurs (retrofit). The AASHTO-NSBA Maintenance Guidelines for Steel Bridges to Address Fatigue Cracking (G14.1-2021) provides the current state-of-the-art framework for selecting and implementing these methods.
Burr grinding the weld toe to a smooth concave profile removes the micro-discontinuities where fatigue cracks initiate. Effective for uncracked welds. Can improve fatigue life by a factor of 2 to 3. Depth should not exceed 0.02 inches below the plate surface. Air or electric die grinder with tungsten carbide burr.
Mechanical impact treatment at the weld toe introduces compressive residual stresses that counteract the tensile cyclic stresses driving crack growth. Ultrasonic needle peening (UNP) and high-frequency mechanical impact (HFMI) treatments can increase fatigue class by up to 4 categories for steels with yield below 355 MPa. IIW provides quality assurance guidelines.
Drilling a hole at the crack tip removes the high-stress concentration region and blunts the crack. Cold expansion of the drill stop (StopCrackEX system) imparts compressive residual stress around the hole, extending fatigue life by over 10 times compared to standard drill stops. Effective for small cracks where access permits.
Welding the connection plate end to the girder flange eliminates the web gap and prevents out-of-plane distortion entirely. This is the most permanent retrofit for web gap cracking. Requires careful welding procedure to avoid introducing new fatigue-critical details. Field studies show unlimited fatigue life after proper attachment.
Fatigue Life Evaluation of Existing Bridges: The NCHRP Report 721 Framework
For existing bridges, the evaluation approach differs from the design of new bridges. The AASHTO Manual for Bridge Evaluation incorporates the latest research from NCHRP Report 721, which provides updated fatigue resistance categories for riveted connections and tack welds. The MBE recommends that base metal at net sections of riveted connections in existing bridges be evaluated as Category C rather than the more conservative Category D specified for new design, reflecting the internal redundancy and load distribution characteristics of riveted built-up members.
Conclusion
Bridge fatigue cracking is not a random occurrence that must be accepted as an inevitable cost of aging infrastructure. It is a predictable, classifiable, and manageable phenomenon. The AASHTO detail category system provides a well-validated framework for understanding why certain details crack and others do not, under what stress ranges, and after how many cycles. The distinction between load-induced and distortion-induced fatigue determines the analytical method required. The choice between infinite-life and finite-life design governs the economic decision of whether to retrofit or monitor.
The most effective fatigue management strategy for any bridge inventory combines three elements: accurate detail classification using the AASHTO framework, appropriate non-destructive inspection at the right frequency with the right method, and timely retrofit intervention using proven techniques such as weld toe grinding, peening, drill stop cold expansion, or connection plate attachment. Bridges that are managed under this three-part framework consistently achieve actual fatigue lives that exceed their original design life — often by several decades.
The technology for fatigue crack detection and life extension has advanced significantly. What was once a reactive cycle of crack discovery, emergency repair, and premature replacement can now be a proactive process of condition-based monitoring, targeted retrofit, and extended service life. The bridge engineer who masters AASHTO detail categories and the full toolkit of fatigue evaluation and retrofit methods is equipped to make decisions that keep bridges in service longer, at lower cost, and with higher safety margins.
