The present paper describes a practical method for analysing and estimating the remaining fatigue life of a 'cut-short' web stiffener detail widely used in highway steel box girder bridges constructed in Ontario, Canada between 1960 and the early 1980s. The paper also discusses the design used to repair steel box bridges fabricated with this stiffener detail. Since 2000, over 20 bridges in southern Ontario have been retrofitted using these details. Steel box girder bridges with composite concrete decks were commonly fabricated in Ontario with a 'cut-short' welded stiffener detail that was used to connect intermediate vertical cross bracing to the girder webs. The bracing members, consisting of single angles, were attached to the ends of the stiffeners with fillet welds. The vertical web stiffener would usually be terminated above the girder bottom flange with a 20 - 50 mm web gap. Often, the fillet welds attaching the stiffener and the web would terminate short of the stiffener end, resulting in an even longer web gap. This detail introduced both in-plane and out-of plane bending into the portion of the web between the end of the stiffener and bottom flange. As a result, several steel box girder bridges have exhibited fatigue cracking problems at the end of the stiffener welds after less than 20 years of service. The cumulative fatigue effect on a bridge is influenced by the volume of truck traffic as well as truck size, configuration and weight. The following approach was used by the authors to estimate the remaining fatigue life of un-cracked stiffener details fabricated in two steel box girder bridges located at the Highway 406 interchange of the Queen Elizabeth Way in Ontario, one of the busiest highways in Canada. A third bridge had already experienced cracking problems and was used to validate the approach: (1) Historical traffic volumes were obtained from available traffic survey records. (2) The volume of truck traffic was determined either from available traffic data or from an assumed percentage based on existing data from other similar Ontario Provincial highways. (3) The cumulative number of truck/axle loads that exceeded the allowable fatigue threshold value for the detail was calculated. (4) The cumulative number of load cycles from trucks that resulted in fatigue damage to the web was estimated with consideration for the load effects caused by multi-lane loadings. (5) A finite element analysis of the fatigue prone transverse stiffener detail using an average' damaging fatigue truck loading was used to estimate the maximum fatigue life of the web stiffener detail based on Miner's rule. (6) From the model results, an 'average' total stress range was applied to estimate the fatigue life of the detail. The remaining service life was estimated by subtracting the maximum life from the cumulative number of damaging load cycles since construction. The results of the analysis compared closely with the actual observed fatigue life of the stiffener detail for the three bridges that were investigated. The current paper also discusses design details that may be considered for repairing bridges with similar cracking problems as well as those considered to be susceptible to future fatigue cracking.
