Structural Engineering and Materials

Permanent URI for this community

https://hdl.handle.net/10919/11353

Browse

Browsing Structural Engineering and Materials by Author "Jain, Nonish"

Now showing 1 - 1 of 1
  • Thumbnail Image
    Developing and Validating New Bolted End-Plate Moment Connection Configurations
    Jain, Nonish; Eatherton, Matthew R.; Murray, Thomas M. (2015-07-01)
    End-plate moment connections are widely used, especially in metal buildings, between rafters (beams) and columns or at splice connections in rafters. References such as the AISC Design Guide 16 entitled “Flush and Extended Multiple-Row Moment End-Plate Connections” present design procedures, supported by physical experiments, to design these connections. It is desirable to develop and validate design procedures for additional end-plate moment connections, particularly those with larger moment capacity. In this report, four connection configurations are investigated. The selected end-plate configurations include eight-bolt extended four wide, eight-bolt extended stiffened, six bolt flush unstiffened, and twelve bolt extended unstiffened. Design procedures and some previous test data was available for the first two configurations, whereas no prior investigation was found in the literature for the latter two configurations. Design procedures including yield line analysis and bolt force models were proposed to calculate moment capacity associated with end-plate yielding, moment capacity for bolt rupture with prying action, and moment capacity for bolt rupture without prying action. Similar to the existing design approach for end-plate moment connections found in Design Guide 16, the design procedures are separated into thin end-plate behavior (the end-plate yields and then bolts fracture with prying action) and thick end-plate behavior (where end-plate yielding is prevented and bolts fracture without prying action). Design procedures found in the literature for the eight-bolt extended four wide and eight-bolt extended stiffened configurations were evaluated and modifications were proposed as necessary. Experimental data found in the literature for the eight-bolt extended four wide and eight-bolt extended stiffened end-plate configurations was analyzed and compared to moment capacities calculated using the proposed design procedures. For the eight-bolt extended four wide configuration, it was found that the experimental data from the literature corroborated the calculated moment capacities for a range of rafter depth and end-plate thickness. It was therefore concluded that no additional tests were required. For the eight-bolt extended stiffened configuration, reasonable match was found between the reported experimental data and predicted moment capacities, but the previous tested beam specimens did not exceed 36 inches depth. It was decided that two additional tests with deeper rafter sections (56 in. deep) would be useful in validating the design procedures for a wider range of rafter depth. A full-scale testing program was conducted including ten specimens that used three different end-plate moment connection configurations. Four specimens were designed for each of the two new configurations (six bolt flush unstiffened and twelve bolt multiple row extended unstiffened) such that there was a shallow rafter (36 in.) and deep rafter (60 in.) specimen predicted to exhibit both thin end-plate and a thick end-plate behavior. Also, two deep rafter (56 in.) specimens were tested with the eight-bolt extended stiffened configuration, one with thin end-plate and one with thick end-plate. The design procedures for all four investigated end-plate moment connection configurations appear reasonable. For the tested configurations, the predicted moment capacity associated with end-plate yielding was 5% smaller than the yield moment obtained during the test (conservative). The predicted moment capacities associated with bolt rupture were an average of 12% less than the experimentally obtained ultimate moment capacities. This conservatism in bolt rupture prediction was in part due to the use of nominal bolt strength in the calculations which is typically notably smaller than actual bolt strength.