Fatigue in welded beams and girders

TLDR: In this paper, the authors have made a series of observations about the effect of different types of fittings on the success or failure of the fittings in the presence of stiffeners.

Abstract: RESULTS ARE SUMMARIZED OF TESTS MADE TO DEMONSTRATE THE EFFECT OF DETAILS ON THE FATIGUE BEHAVIOR OF WELDED FLEXURAL MEMBERS. TEST CONDITIONS ARE RELATED TO ACTUAL SERVICE CONDITIONS. DETAILS SUCH AS SPLICES, STIFFENERS, COVER PLATES AND ATTACHMENTS, ALL OF WHICH CAN BE EXPECTED TO PRODUCE REDUCTIONS IN THE FATIGUE STRENGTH OF THE BASIC MEMBER, WERE INCLUDED. TESTS WERE CONDUCTED ON SPECIMENS FABRICATED FROM ASTM A73 STEEL. MANUAL ARC WELDING WITH NUMBER 7016 ELECTRODES WAS USED IN THE FABRICATION OF MOST OF THE SPECIMENS FOR THE TESTS. THE WELDS WERE DEPOSITED WITH REVERSED POLARITY, IN THE FLAT POSITION, AND WITH 5/32- IN. ELECTRODES FOR THE ASSEMBLY OF THE BASIC SECTION AND 1/8-IN. ELECTRODES FOR THE ATTACHMENT OF STIFFENERS. ALL ELECTRODES WERE STORED IN A DRYING OVEN TO PREVENT ABSORPTION OF MOISTURE IN THE ELECTRODE COATING. THE BEAMS WERE TESTED IN A 200,000-LB FATIGUE TESTING MACHINE. AFTER FAILURE OF THE TEST MEMBERS, THE ACTUAL DIMENSIONS OF THE CROSS-SECTION AT THE FAILURES WERE MEASURED AND THE STRESSES AT THESE LOCATIONS CALCULATED. TESTS ON THE FATIGUE BEHAVIOR OF I-BEAMS SHOW THAT ONE FACTOR WHICH HAS AN EFFECT ON THE FATIGUE RESISTANCE OF THE WELDED BEAMS IS THE RATIO OF THE THICKNESS OF THE FLANGE TO THAT OF THE WEB. IN ADDITION TO THE VARIATION IN FATIGUE LIFE WITH THE FLANGE-TO-WEB THICKNESS RATIO, A DIFFERENCE IN THE MODE OF FAILURE WAS OBSERVED. THE POINT OF FAILURE INITIATION WAS GENERALLY THE SAME--IN A WELD CRATER OF THE FILLET WELDS AT THE WEB-FLANGE JUNCTION. A SUMMARY OF THE FATIGUE STRENGTHS OF THE VARIOUS TYPES OF SPLICED MEMBERS IS PRESENTED IN A TABLE. THE FATIGUE STRENGTH OF THE SPLICED BEAMS CAN BE MADE TO APPROACH OR EQUAL THAT OF A PLAIN PLATE WITH A TRANSVERSE BUTT WELD. HOWEVER, THE FATIGUE STRENGTH OF THE SPLICED MEMBERS AT 2,000,000 CYCLES IS ONLY ABOUT TWO-THIRDS AS GREAT AS THE FATIGUE STRENGTH OF THE PLAIN PLATE MATERIAL. RELATIONSHIPS WERE USED TO DETERMINE THE MAXIMUM PRINCIPAL TENSILE STRESS AT THE POINT WHERE FAILURE INITIATED IN STIFFENERS. IT APPEARS THAT WHEN STIFFENERS ARE NOT WELDED TO THE TENSION FLANGE OF A FLEXURAL MEMBER THAT IS SUBJECTED TO REPEATED LOADS, THE SHEAR IN THE WEB OF THE MEMBER SHOULD BE CONSIDERED. IT IS CONCLUDED THAT THE MOST EFFECTIVE MEANS OF INCREASING THE FATIGUE RESISTANCE OF THE BEAMS OF VARIABLE SECTION IS TO VARY THE FLANGE IN WIDTH OR THICKNESS. IT CAN BE SEEN THAT WHEN THE CURRENT SPECIFICATIONS ARE CORRECTLY APPLIED, WELDED BEAMS AND GIRDERS OF HIGHWAY BRIDGES, IF PROPERLY FABRICATED, SHOULD HAVE ADEQUATE FATIGUE CAPACITY. IN VIEW OF THE WEB FAILURES OBSERVED, FURTHER CONSIDERATION SHOULD BE GIVEN TO THE EFFECT OF SHEAR ON THE FATIGUE BEHAVIOR OF MEMBERS WITH THIN WEBS.