Date of Award


Degree Type


Degree Name

Master of Applied Science (MASc)


Civil Engineering


John C. Wilson




The goal of seismic response modification is to elicit more favourable behaviour from structures during earthquakes. A popular means for this modification is through the use of energy dissipation devices (EDDs), which effectively lower the seismic demand on main structural components by absorbing a portion of the seismic input energy. This energy is often absorbed through the yielding of metallic components within these devices. Examples of traditionally used metals are steel and lead which exhibit bilinear hysteretic behaviour. The introduction of shape memory alloys (SMAs) into the realm of seismic energy dissipation has occurred fairly recently and has occurred in large part due to their exceptional characteristics for earthquake engineering applications, including hysteretic properties. SMAs exhibit flag-shaped hysteretic behaviour creating the potential for relatively simple yet effective EDDs.

In this study, the non-linear time history analysis of 1, 2, 4, and 8 storey chevron-braced steel frame models was carried out using the program Ruaumoko®. The braces in each frame were modelled to emulate EDDs exhibiting flag-shaped hysteresis. For comparison, the same frames were analyzed with braces modelled to emulate EDDs exhibiting bilinear behaviour with full hysteresis loops. Each frame was subjected to an ensemble of ground motions to determine their mean responses as measured by several response indices. A parametric approach was adopted to determine the effect of various parameters on response and the parameter values resulting in advantageous behaviour. Tension-compression (T-C) and tension-only (T-O) braces were studied to illustrate their effect on response as well. Pushover tests on the frames were also performed, using the program SAP2000®, to help validate the results.

The behaviour of the frames in this study was generally governed by the backbone curve of the brace hystereses defined by post-yielding stiffness parameter, α, normalized yield force, η, and normalized stiffness, θ. Increasing these parameters generally decreased interstorey drift, residual interstorey drift, column ductility demand, beam ductility demand, and absorbed energy. However, this also had the effect of increasing acceleration and base shear. The energy dissipation parameter, β, had little effect on response. In general, the responses of the frames with flag-shaped hysteretic braces, as measured by interstorey drift, column ductility demand, beam ductility demand, base shear, and absorbed energy, were very similar to those frames with the bilinear hysteretic braces - in both magnitude and trend. However, the frames with bilinear braces tended to have lower accelerations, and the frames with flag-shaped braces tended to have lower residual drifts. The T-C braces were generally more effective than the T-O braces in enhancing seismic performance. However, the frames with T-O braces had a resultant lower lateral stiffness - suggesting that T-O braces of sufficient stiffness could be used to garner similar responses offered by T-C braces.

McMaster University Library

Files over 3MB may be slow to open. For best results, right-click and select "save as..."