This thesis aims to find a new weld sizing criterion for the steel construction industry in New Zealand. Current standards, such as NZS 3404, ANSI/AISC 360-16, and Eurocode 3 use a factor of 0.6 to calculate weld capacity from the weld metal’s ultimate tensile strength (UTS). This difference between weld capacity and UTS is thought to have arisen from the need for a large factor of safety to ensure welds perform correctly during an earthquake. The events in Christchurch proved that this criterion was able to work as intended. Several papers have been published by P. Dong from University of Michigan, and alongside other researchers, they investigate a new definition of weld shear strength by using a traction stress-based method. This new method not only allows realistic angles of weld fracture to be investigated, but also different weld geometries such as partial penetration butt welds. Ongoing research at HERA is showing how this welding technique is a more economical option than larger fillet welds with similar performance. For this thesis a range of sample types were statically tested until failure. UTS of several weld metals was found and then compared with transverse shear results to see if 0.6 is indeed correct. It was found that if the results from the standardized transverse shear samples was used, this ratio could be increased to 1.0. But if the results from cruciform joint samples was used, which still load the weld in a transverse direction but with a higher stress concentration, required the ratio to be 0.8 for welds that could be welded with a single pass, and decreased further to 0.75 for large welds with 3 passes. Two types of partial penetration butt weld (PPBW) geometries were compared to a comparatively sized fillet weld. These tests showed the PPBWs were the best performers, with all PPBWs surviving testing compared to only 33% of fillet welds.
The increasing prevalence of mixed-material buildings that combine concrete walls and steel frames in New Zealand, coupled with a lack of specific design and detailing guidelines for concrete wall-steel beam connections, underscores the need for comprehensive research to ensure that these structures behave as intended during earthquakes. Bolted web plate connections, commonly found in steel framing systems, are typically used to connect steel beams to concrete walls. These connections are idealised as pinned during design. However, research on steel framing systems has shown that these connections can develop significant stiffness and moment resistance when subjected to large rotations during seismic loading, potentially leading to brittle failure when used in concrete wall to steel beam applications. This thesis was written to understand the seismic performance of concrete wall-steel beam bolted web plate connections, providing experimental evidence, numerical modelling insights, and design recommendations to address critical gaps in current design practices. The study is divided into three phases. First, a review of 50 concrete wall-steel frame buildings in Auckland and Christchurch was conducted to understand current design practices and typical connection details. The findings revealed significant variation in design and detailing practices and a lack of specific guidelines for concrete wall-steel beam connections. Second, an experimental programme was conducted on four full-scale concrete wall-steel beam sub-assemblages, each incorporating variations in connection detailing. The tests were designed to quantify the rotation capacity of concrete wall-steel beam connections, identify failure modes and investigate the effectiveness of potential connection improvements. Results demonstrated that concrete wall-steel beam bolted web plate connections designed using current design standards and following existing practices are vulnerable to non-ductile failure characterised by concrete breakout. However, using slotted holes in the web plate and bent reinforcing bar anchors instead of headed stud anchors improved connection rotation capacity. Third, a numerical model of a case study building was developed on OpenSeesPy, with different connection conditions assumed based on the experimental results. Pushover and time history analyses were conducted to evaluate the implications of different connection conditions (pinned vs non-pinned) on global building response and local member demands. The findings revealed that using non-pinned connection conditions does not significantly affect the global building response and shear and bending moment demands on lateral load-resisting elements. However, doing so generates overstrength moments on the connections that induce different actions on out-of-plane concrete walls connected to steel beams. Synthesising findings from all three phases, this thesis concludes with a proposed design procedure for concrete wall-steel beam connections based on a capacity design approach to ensure ductile failure modes and suppress brittle ones. Key recommendations include selecting appropriate bolt hole geometry and anchorage, providing sufficient rotation capacity, and accounting for connection overstrength in global analyses.
A non-destructive hardness testing method has been developed to investigate the amount of plastic strain demand in steel elements subjected to cyclic loading. The focus of this research is on application to the active links of eccentrically braced frames (EBFs), which are a commonly used seismic-resisting system in modern steel framed buildings. The 2010/2011 Christchurch earthquake series, especially the very intense February 22 shaking, which was the first earthquake worldwide to push complete EBF systems fully into their inelastic state, generating a moderate to high level of plastic strain in EBF active links, for a range of buildings from 3 to 23 storeys in height. This raised two important questions: 1) what was the extent of plastic deformation in active links; and 2) what effect does that have to post-earthquake steel properties? This project comprised determining a robust relationship between hardness and plastic strain in order to be able to answer the first question and provide the necessary input into answering the second question. A non-destructive Leeb (portable) hardness tester (model TH170) has been used to measure the hardness, in order to determine the plastic strain, in hot rolled steel universal sections and steel plates. A bench top Rockwell B was used to compare and validated the hardness measured by the portable hardness tester. Hardness was measured from monotonically strained tensile test specimens to identify the relationship between hardness and plastic strain demand. Test results confirmed a good relationship between hardness and the amount of monotonically induced plastic strain. Surface roughness was identified as an important parameter in obtaining reliable hardness readings from a portable hardness reader. A proper surface preparation method was established by using three different cleaning methods, finished with hand sanding to achieve surface roughness coefficients sufficiently low not to distort the results. This work showed that a test surface roughness (Ra) is not more than 1.6 micron meter (μm) is required for accurate readings from the TH170 tester. A case study on an earthquake affected building was carried out to identify the relationship between hardness and amount of plastic strain demand in cyclically deformed active links. Hardness was carried out from active links shown visually to have been the most affected during one of the major earthquake events. Onsite hardness test results were then compared with laboratory hardness test results. A good relationship between hardness from onsite and laboratory was observed between the test methods; Rockwell B bench top and portable Leeb tester TH170. Manufacturing induced plastic strain in the top and bottom of the webs of hot rolled sections were discovered from this research, an important result which explains why visual effects of earthquake induced active link yielding (eg cracked or flaking paint) was typically more prevalent over the middle half depth of the active link. The extent of this was quantified. It was also evident that the hardness readings from the portable hardness tester are influenced by geometry, mass effects and rigidity of the links. The final experimental stage was application of the method to full scale cyclic inelastic tested nominally identical active links subjected to loading regimes comprising constant and variable plastic strain demands. The links were cyclically loaded to achieve different plastic strain level. A novel Digital Image Correlation (DIC) technique was incorporated during the tests of this scale, to confirm the level of plastic strain achieved. Tensile test specimens were water jet cut from cyclically deformed webs to analyse the level of plastic strain. Test results show clear evidence that cyclically deformed structural steel elements show good correlation between hardness and the amount of plastic strain demand. DIC method was found to be reliable and accurate to check the level of plastic strain within cyclically deformed structural steel elements.
