Fatigue cracking has been noted in a number of riveted, coped stringer-to-floorbeam connections on truss bridges in Washington state. The fatigue cracking has raised uncertainty regarding the future performance and life expectancy of these bridges. This laboratory experimental research study evaluated the expected performance of these critical connections to better understand the causes of the fatigue cracking and to compare and evaluate several damage limitation methods for this critical location.
Sixteen test specimens were built and tested. During the initial testing, the goal was to develop fatigue cracks in the region of the cope. The cracks were observed, and the rate of crack growth was closely monitored. The results of this work showed that the rate of initial visible cracking is dependent on the surface condition of the cope. Rough or notched copes develop cracks after a very few cycles. Once the crack has developed, the rate of crack growth does not depend on the surface condition of the cope.
A damage limitation method (DLM) was applied to most specimens. Three methods--the hole drilling method, the drilled hole with inserted bolted, and the bolt removal method--were evaluated. After the DLM was applied, the specimen was retested to determine whether and when further crack growth occurred. The rate of crack growth was closely monitored, and the relative effectiveness of alternative DLMs were determined.
A design and evaluation procedure was developed for predicting the time required for crack development and for selecting appropriate connection stiffnesses for application of the DLMs. This method utilizes load spectra developed in previous field investigations and methods for translating the load spectra into an effective stress and number of cycles for the fatigue evaluation. Models are provided for evaluating connection stiffness and for analyzing the effect of the connection stiffness on the stringer moments.