Realistic Wind Loads on Unreinforced Masonry Walls
Twenty full-scale unreinforced masonry walls were constructed and tested to failure in the Structures Laboratory at the University of Saskatchewan. The focus of the testing related to two primary objectives. The first objective was to study the effects that the support conditions of the walls had on their behaviour. The masonry wall specimens tested spanned vertically under the application of out-of-plane loads. Ten of the full scale walls were tested with support conditions that modeled ideal pinned connections at the top and bottom of the wall, while the remaining half of the walls were tested with nominally “pinned” supports that were similar to the supports typically encountered in practice. The second objective was to determine the effects that dynamic loads had on the behaviour of the walls. Half of the masonry specimens for each group of support conditions were loaded laterally with monotonically increasing quasi-static loads representative of the effects of uniform wind pressure, while the remaining specimens were loaded laterally with dynamic time histories that varied randomly in a manner that was representative of real “gusty” winds. The research was therefore done to determine the influence of load and connection type on the behavior of the masonry walls. When comparing the effects of the support conditions, it was found that the walls constructed with realistic support conditions were able to resist larger out-of-plane loads, with greater ductility than the walls that had ideally-pinned supports. Specifically, the realistically-pinned walls required an average moment (of both the statically and dynamically loaded walls) that was 63% larger to cause mid-height cracking than the average mid-height moment required to cause mid-height cracking in the ideally-pinned walls. After mid-height cracking occurred, the realistically-pinned walls exhibited reserve capacity, resulting in additional strength, such that the ultimate moment capacity of the realistically-pinned walls was 140% greater than the ultimate strength of the ideally-pinned walls, where the ultimate strength was the capacity of the wall at mid-height cracking. As a result, the ductility of the realistically-pinned walls was also significantly larger than that of the ideally-pinned walls. Specifically, the ductility ratio of the realistically-pinned walls was 70 (where the ductility ratio is defined as the displacement at the ultimate load divided by the displacement at mid-height cracking), while the ductility ratio of the ideally-pinned walls was unity (the ultimate load coincided with formation of the mid-height crack). The results of the dynamically and quasi-statically loaded walls were harder to evaluate. In comparing the ideally-pinned walls it was found that the specimens that were loaded dynamically had an average moment capacity that was approximately 10% larger than the walls that were loaded quasi-statically, which was found to be statistically significant at the 90% level. However, the results from the realistically-pinned walls were not as conclusive. At mid-height cracking the dynamically loaded walls had an average moment capacity that was 24% lower than the quasi-statically loaded walls, which seems to contradict with the data from the ideally-pinned walls and from the literature suggesting that dynamic strengths should be higher. At the ultimate condition, the dynamically loaded walls had an average strength that was 12% larger than the quasi-statically loaded walls; however, these comparative results were not statistically significant at the 90% confidence level. It was also found that the dynamic loading failed the wall specimens as a result of sustained, large amplitude “gusts” rather than at the largest instantaneous peak load. The displacement behaviour of the walls was generally independent of the method of loading, but, rather, largely dependent on the support conditions. The collapse of the wall specimens were all initiated when they reached a geometrically unstable displaced shape that was fairly consistent for a given support configuration, regardless of the type of load that was applied. Lastly, results from a numerical model suggested that the dynamically loaded walls exhibited higher apparent stiffness properties as compared to the quasi-statically loaded walls. The difference in the apparent stiffness between the dynamic and quasi-static specimens decreased with increasing damage levels until the dynamic stiffness converged to the static stiffness near the collapse of the walls.
DegreeMaster of Science (M.Sc.)
DepartmentCivil and Geological Engineering
SupervisorSparling, Bruce F.
CommitteeBoulfiza, Mohamed; Feldman, Lisa; Fleming, Ian; Hoeppner, Colin
Copyright DateAugust 2014