Row shear and block shear failure of connections with axially loaded screws

The present paper considers brittle failure modes inconnections with screws loaded axially perpendicularto grain. The available test data are evaluated andcompared with design proposals from Mahlknechtand Brandner (2016) for block shear, Meyer and Blass(2018) for row shear as well as with the approachonly based on an effective number of fasteners nef.Even if an effective number of fasteners nef primarilyincorporates the influence of uneven load distributionbetween the single screws in a connection, it obviouslyalso at least partly compensates for brittle failuremodes as block shear failure or row shear failure.The analytical model derived for glued-in rods takinginto account row shear failure in connections withaxially loaded fasteners arranged in rows parallelto grain and loaded perpendicular to grain also verywell predicts the load-carrying capacity of similarconnections with screws. This model may also be usedfor groups with several rows of screws, if the spacinga2 perpendicular to grain is large.The analytical model proposed by Mahlknecht andBrandner for block shear failure in connections withgroups of axially loaded screws and load componentsperpendicular to grain was modified as follows:•Shear planes perpendicular to the grain are notconsidered,•Brittle failure is either caused by tensionperpendicular to the grain in a plane defined by thescrew tips or by rolling shear in planes defined bythe outer screw rows, simultaneous load transfersvia rolling shear and tension perpendicular to thegrain are disregarded,•The tension perpendicular to the grain capacityis determined according to the German nationalannex to Eurocode 5 (DIN EN 1995-1-1/NA),•Rolling shear failure planes exceed the length ofthe connection parallel to the grain on each endby 0,75 ℓef,•It is considered that only a part of the loadcomponent perpendicular to the grain causestension perp. to grain stresses or rolling shearstresses.In order to verify the modified analytical model,ultimate test loads of axially loaded screwedconnections are compared with the results of themodel. The ultimate loads from the tests agree wellwith the model predictions. For the comparison,the model parameters screw tensile strength andwithdrawal capacity were determined separately bytests. If block shear design is disregarded, the designof axially loaded screwed connections accordingto the screws’ ETAs still leads to an adequate load-carrying capacity.