A number of localized failures have developed in cantilevered supports of
highway signs, luminaries and traffic signals over the past ten years. Failures due to
fatigue crack growth around welded structural details have occurred in socket
connections within New Jersey, Iowa, Florida, Wisconsin, California, Massachusetts and
Wyoming. Many of these failures have resulted from the interaction of the wind and the
structures, resulting in numerous applied stress cycles.
Accordingly, fatigue tests were conducted on full-scale welded aluminum light
poles containing through plate and shoe base socket connections. Through plate and shoe
base socket connections are used to anchor aluminum light poles to a break away
foundation.
Initiative for the study was a lack of available data for aluminum structural
details, and the result of detrimental changes to specifications governing the design and
proportioning of welded aluminum luminaire supports subjected to fatigue loading. As
such, fatigue tests were conducted on light poles containing both shoe base and through
plate socket connection details order to study the fatigue behavior and determine a lower
bound resistance suitable for design in terms of S-N curves. Parametric studies using the
finite element method were conducted on both detail types in order to understand the
nature of the local stress fields governing fatigue behavior and how changes in geometry
affect the local stresses.
As stress range is the primary parameter used to describe fatigue strength, hole
drilling strain measurements were utilized to examine whether the light pole details
contained significant tensile residual stresses normally assumed to exist with welded
construction.
Fatigue tests revealed relatively low strengths for the through plate socket
connections as compared to the shoe base details. Near the constant amplitude fatigue
limit, the difference in strength was nearly a factor of 3.5. Residual stress measurements
revealed the existence of compressive residuals stresses on the surface of the tubes for
both types of details measured to be close to -18 ksi. Results of the parametric study of
the through plate socket connection showed a 30% reduction in the longitudinal stress on
the surface of the tube by increasing the base plate thickness from 1 to 2 inches. For a 1
inch base plate thickness, additional bending through the tube wall and elevated
longitudinal stresses were observed opposite to the bolt location on the tension side of the
pole. Attempts to stiffen the through plate socket connections using triangular plate
stiffeners resulted in elevated longitudinal stresses at the tip of the stiffeners for short
stiffeners that contradict current AASHTO specification fatigue categories for such type
of structures. SEM examination of typical fracture surfaces showed the existence of fine
striations, secondary cracking and a region of ridges and grooves.
Future recommendations from this study includes (1) fatigue tests on steel and
aluminum through plate socket connections with the proposed 3 inch base plate
thickness, (2) Long term monitoring of Signal Sign structures in the State of Ohio to
validate pressure and load recommendations for vortex, galloping, truck induced and
wind load effect and (3) Vibration tests on luminaire support structures to measure the
first modal natural frequency for such types of details.