Issue 48

V. M. G. Gomes et alii, Frattura ed Integrità Strutturale, 48 (2019) 304-317; DOI: 10.3221/IGF-ESIS.48.30 312 obtained from static tests are lower than specimens without coating due the presence of soft zinc layer that tends to act like a lubricant. The high value of slip factor in zinc coated specimens may be caused by the surface treatment processes that the specimen was subjected before galvanizing process, and/or the effect of the decrease of the thickness that increases the coefficient of friction, which according to Picard [14] is the main reason for the variation of friction coefficient values. Lastly, specimens with zinc plus paint coating resulted in more similar values between static and slip tests. Surface Finish Without Coating Zinc coating Zinc plus paint coating Monotonic/Slip Tests EN 1090 Preloaded Bolts (70% Fu ) 2+2 bolts Slip Factor µ slip AVG 0.28 0.31 0.14 Charact. Slip µ c 0.16 0.20 0.09 Monotonic /Static Tests Snug Tight Bolts (25%×70% Fu ) 1+1 bolts 2mm 0.18 0.09 0.06 3mm 0.22 0.10 0.09 4+4 bolts 2mm 0.23 0.11 0.13 3mm 0.14 0.11 0.14 Preloaded Bolts (70% Fu ) 1+1 bolts 2mm 0.17 0.15 0.08 3mm 0.16 0.14 0.10 4+4 bolts 2mm 0.13 0.11 0.09 3mm 0.14 0.13 0.10 Average friction coefficient from static tests 0.17 0.12 0.10 Table 5 : Slip factors ( µ slip AVG ) obtained from the slip and static/monotonic tests. The load-displacement experimental results from the static monotonic tests are showed in Fig. 6. From Fig. 6 one verifies that the increase of thickness induces greater failure loads as expected, however the increasing of failure loads is greater on specimens with 4+4 bolts than in specimens with 1+1 bolts. The comparison of the geometries show a large difference in the maximum displacements due to the different failure modes verified. The 4+4 bolted connections failed in the net cross-section of the first two holes, while the 1+1 bolted connections presented shear fracture and shear with splitting fractures for S355MC and S350GD steel grades, respectively. Regarding the preload level applied on bolts, Fig. 6A shows that the preload level is the important factor to increase the sliding load, however it is also possible to see that the increase of preload has some influence on ultimate failure load (Fig. 6B). Fig. 6 also includes a comparison between experimental and numerical curves and there are some differences on initial behaviour corresponding to the sliding of joint. These differences are due to the complexity of the friction phenomenon and therefore a detailed analysis is necessary to characterize the friction behaviour correctly. From Fig. 6A, we may verify that S355MC specimens present a sticking/sliding effect as well as S350GD with zinc coating plus painting; specimens made of S350GD with zinc coating the sticking/sliding effect is not clear. This aspect is extremely important to the real representation of friction behaviour of the joint. However, it should also be highlighted that the associated errors for friction coefficient did not influence the final behaviour for lower bolt preload levels; but the friction coefficient may have a significant influence for higher bolt preload levels, if the thickness is small enough, like the specimens with 2mm thickness. From Fig.6B we can verify that for specimens with 3mm thickness, the failure load is greater for uncoated steel. However, for specimens with 2mm thick plates, the failure load is greater for specimens with zinc coating. In addition, it verifies that the difference between ultimate loads of specimens with zinc plus paint coating and only zinc coating is very small, allowing to conclude that the paint do not influence the strength of the connections. The numerical results were very satisfactory on predicting the ultimate loads, with a maximum error of 8.68% on specimen made of S350GD, with 2mm thick plates and preloaded bolts. About experimental tests, the maximum deviation of results occurred for specimen S355MC (uncoated) with 3mm thick plates and 1+1 configuration and preloaded bolts. It is also possible to verify that the greatest CoV was verified for specimens with preloaded bolts, reinforcing the idea that for higher preload values, the dispersion of results is greater. Tab. 6 compares the experimental and estimated failure loads where the maximum values of CoV and Error are in bold figures.