During the Kobe earthquake, brittle fracture occurred frequently from ductile cracks at the toes of weld access hole in the beam-to-column connections. The post earthquake recommendation (AIJ, 2007) proposed to use cope holes comprised of two arcs two circles with horizontal haunch or non-cope hole procedures to avoid stress concentration at the toes of beam copes. In addition, the partial cutting fillet welds (PCFW) procedure (Nakagomi et.al., 2016) was developed to improve the deformation capacities in the beam end. But, it is difficult to use the non-cope profiles in the beam-to-column connections using field welding because beam web interrupts the weld line in the groove welds between column and beam bottom flange. Furthermore, using horizontal haunch might increase manufacturing cost.
This paper concerns the verification of the cope profiles which shows large plastic deformation capacities using improved weld details and the effects of partial cutting fillet welds procedure. Past test results showed that the deformation capacities were improved when the ductile cracks initiated at the weld toes of web plate rather than at the toes of beam copes and propagated to the web. The aims of this research are shown below.
To compare the initiation points of ductile cracks obtained from test and that from FE analysis.
To investigate the effect of the PCFW procedure on deformation capacities of joint.
To find the best cope profiles.
In this study, cyclic testing using beam flange-to-diaphragm joint models, which were designed to represent beam-to-column connections with through diaphragm, were conducted. The specimens consist of 22mm thick flange plates, 32mm thick diaphragm plates and 16mm thick rib plates. Figure 1 shows the overall view of the specimen. All plates were grade JIS G 3136 SN490B.
The weld access hole shape used for the specimen were conventional shape, AISC weld access hole shape and AISC weld access hole shape with PCFW procedure. The three specimens and the rest of specimens had web thicknesses of 9mm and 12mm, respectively. Shapes of weld detail were shown in Figure 2. In this study, five experiments excluding the shape of type (d) were conducted.
