Non-stationary seismic response and amplitude-dependent stiffness evolution of a post-tensioned LVL building under recorded ground motions of the 2011 Canterbury earthquake sequence

This paper documents and physically interprets the in-situ seismic response of the two-storey EXPAN/STIC Office Building - a full-scale, in-service post-tensioned Laminated Veneer Lumber (Pres-Lam) structure located at the University of Canterbury (Christchurch, New Zealand) - based on the acceleration records acquired by its permanent monitoring system during four moderate-amplitude earthquakes of the 2011 Canterbury sequence. The monitoring system, deployed at the end of March 2011 (approximately five weeks after the 22 February 2011 mainshock), captured the response of the building under peak ground accelerations between 0.10 g and 0.21 g. To extract physically interpretable indicators of the non-stationary dynamic behaviour, the acceleration records were processed through a chain combining a single-station roof-level rotational Horizontal-to-Vertical Spectral Ratio (HVSR) and the half-power-bandwidth method—with a closed-form correction of the Konno–Ohmachi smoothing bias—for the modal frequencies and the equivalent viscous damping ratio, and the Stockwell Transform for the time–frequency description of the non-stationary response. Inter-storey drifts, storey shears and equivalent secant stiffnesses were derived from the recorded accelerations by means of a fully documented signal-processing chain. The recordings consistently exhibit a transient reduction of the fundamental frequency during the strong-motion phase, compatible with the partial activation of the rocking and re-centring mechanisms of the post-tensioned system, followed by a recovery towards the pre-event values during the unloading phase; no statistically significant drift of the dynamic parameters is observed across the four events (exact Mann–Kendall test for 𝑛 = 4). A simple kern-limit mechanical reading of the gap-opening interface, with the post-tensioning force inferred from installation records, predicts a decompression drift 𝛿dec,𝑋 ≃ 0.040% that is exceeded by all four recorded events, supporting the rocking-driven interpretation. Within the limits of an observational case study—and bearing in mind that the dataset does not include the 22 February 2011 mainshock and therefore does not characterise the response at the design Ultimate Limit State—these results contribute to the very limited body of recorded seismic data available worldwide for in-service post-tensioned timber buildings.