| It is well known within the engineering community that the load distribution within-a-row of bolts is not even. This variation in bolt loading makes it difficult to predict the overall connection resistance. A research study was undertaken to predict the capacity of multi-bolted connections through a finite element model based on the within-a-row behaviour. The initial numerical model was developed by Tan and Smith, which was adapted by Shin through finite element methods.;High and low range predictions of the ultimate failure load were obtained by varying the initial bolt-to-hole gap for single- and double-row three- and four-bolted connections. Changing the gaps in the model cause approximately a 15% difference in the resistances. The predictions were also compared to experimental averages. The mean of the predicted resistances, for the double-row three- and four-bolted connection, is 16% higher than the experimental average and over 20% higher for the single-row bolted connections. This is believed to be due to averaging the load-slip spring function, instead of varying the ultimate point.;Lastly, a comparison between the experimental capacity of a single-row and double-row three-bolted connection with 5d row spacing, showed that the currently used reduction factor (JR) of 0.8 is not required.;Shin's finite element model of a single-row three-bolted connection was improved upon and the springs were modified to have one common equation that represents the bolt's load-slip behaviour for either a single- or double-row connection, with a different ultimate slip depending on either a 7d end distance (e) or 4d bolt spacing (sb). The spring's load-slip relation was created from experimental tests involving double-row one-bolted connections. |
