ICF10B

(i) A large deflection theory analysis of stresses in crack – prone areas of breathing webs. (ii) An analysis, based on Fracture Mechanics, of the propagation of cracks in breathing webs. (iii) An ultimate load theory for fatigue – cracked webs subjected to predominantly shear. The experimental investigation consists of several series of tests on steel plate girders under the action of many times repeated load, the geometrical characteristics of the test girders and the regime of the cyclic load being varied in them. Two series of test girders were tested, by M. Zornerová and M. Škaloud, at the Institute of Theoretical and Applied Mechanics of the Czech Academy of Sciences in Prague. (i) One series of girders with slender webs, whose depth–to-thickness ratio λ = 250. (ii) Another series of girders with less slender webs, whose depth–to-thickness ratio λ = 175. This depth– to-thickness ratio was chosen so as to be lower than the maximum slenderness λn,b max of so-called non-breathing webs discussed above. The former group to date comprised 71 test girders, the latter 22 girders, but experiments of other girders in both groups are running. As it was demonstrated during static tests on the ultimate load behaviour of shear girders that these characteristics were fundamentally influenced by the flexural rigidity of girder flanges, one half of the test girders were equipped with very flexible flanges (50 x 5 mm) while the other half had more rigid flanges (50 x 10 mm). Two other series of girders were tested at Klokner Institute of the Czech Technical University in Prague. In the first one, the web thickness tw = 4 mm so that the web depth–to-thickness ratio λ = 200 (this means, that these girders were designed so as to have as good as non-breathing webs); in the other, tw = 2,5 mm and λ = 320. Like the girders tested in the laboratory of the Institute of Theoretical and Applied Mechanics, even the above girders had two kinds of flanges being thin (tf = 8 mm) and for the others being thicker (tf = 20 mm). The former series comprised 12 girders and is regarded as completed; the latter to date consisted of 13 girders, but other experiments are under way. So, if looking at the general details of all the girders tested, it can be seen that the depth–to-thickness ratios of their web panels (which is beyond any doubt one of the main factors affecting web breathing) were of 175, 200, 250 and 320. Accordingly, they ranged from so-called non-breathing webs to very slender ones. All the test girders were subjected to a point load F = 2 P, located at mid-span and cycling between a minimum value Fmin = 10 kN and a maximum value Fmax, which varied from test to test. Both of the two web panels were then exposed to combined shear and bending, with the influence of shear predominating. In addition to usual deflection and strain measurements, the main objective of the experiments was to study the initiation and propagation of cracks in the webs and the fatigue failure mechanism of the test girders. This was achieved largely by way of visual inspection checks of the breathing webs, aided by a magnifying glass. In several tests the acoustic emission method was also applied. The strain measurements were used mainly to determine the zones of stress concentration in the crack-prone areas of the breathing webs, this being also supported by a theoretical study based on the finite element method. Before every test, the initial curvatures of both web panels were cautiously gauged, and so was the plastic residue (including fatigue crack) in the web sheet and girder flanges after the termination of the experiment, this rendering it possible to study the collapse mechanism of the test girder. As stated above, the main intendment of the experimental investigation was to study the initiation and propagation of fatigue cracks in the breathing webs, and their impact on the failure mechanism of the test girders. It was seen that the fatigue cracks arose at the toes of the fillet welds connecting the web sheet panels with their boundary members ( ie flanges and stiffeners). They initiated at the regions of maximum principal surface stress ranges and grew, with the number of load cycles increasing, along the boundary members and often diagonally across the corner of the web as well, so as in the end to “cut” the tension diagonal in the buckling web sheet. The initiation and the character of crack propagation was as follows: (i) Either a crack appeared in the diagonal tension band and rapidly advanced in a direction approximately perpendicular to that of the tension band. But still, this situation occurred only twice among the almost one hundred and twenty tests performed in Prague to date, and was probably owing to the presence of a manufacturing imperfection in the web material at the zone of crack initiation

RkJQdWJsaXNoZXIy MjM0NDE=