Methodology for assessing the seismic performance of unreinforced masonry single storey walls, parapets and free standing walls

Abstract

In the early 1980s US researchers subjected full-scale unreinforced masonry (URM) face-loaded wall specimens to earthquake motions. These specimens representing a wall element spanning between two adjacent floors. They found that a single horizontal crack tended to form near mid-height of the test specimens and another crack formed at the test bed floor, and that the walls were able to sustain large displacements normal to the face of the wall, comparable with the wall thickness. This ability to withstand large displacements without collapse resulted in the walls having a significant post cracking seismic resistance. The term “dynamic stability” was used to distinguish this type of behaviour from the behaviour that might have been expected from static force calculations.

Subsequently this concept was used to develop the current New Zealand Society of Earthquake Engineering (NZSEE) Guidelines for the assessment of face-loaded walls.  Previous research by the author indicated that these current Guidelines were often very non-conservative. This previous research also led to the development of a more reliable methodology for assessing the earthquake behaviour of face-loaded URM.

In this study the assessment methodology is extended to cover face-loaded single-storey walls, freestanding walls supported only by the ground, and parapets.

The methodology was extended and developed using the results obtained when these types of wall were modelled on a computer. The computer model was verified by comparing the walls’ predicted behaviour with the results obtained from a wall specimen recently tested on shake table by Australian researchers.

Design charts are provided to enable rapid design office assessment of face-loaded wall elements.

---

Technical Abstract

An assessment methodology, developed in previous research that can be used to predict the seismic stability of a cracked face-loaded unreinforced masonry (URM) multi-storey wall was extended in this study to cover single-storey walls, freestanding cantilever walls and parapets.

The methodology makes use of both the acceleration and displacement response spectra for an earthquake motion. The acceleration spectrum is used to predict the earthquake intensity that will just open the joint cracks in the wall element. The displacement spectrum is used to predict the earthquake intensity that will generate wall displacements equal to the displacement at which the wall element becomes unstable. Modification factors are applied to allow for the effect of the wall element boundary conditions and to allow for amplification of the earthquake motion due to flexibility in any building structure or diaphragms that support the wall element.

The methodology was principally developed from the results of inelastic dynamic analyses of computer models of face-loaded URM walls supported by flexible shear walls and flexible or yielding floor and roof diaphragms. Parameters examined included interaction between the parapet and its supporting wall and the effect of building and/or diaphragm flexibility. The effects of long acceleration pulses in the ground motion, which may occur in the near-fault zone during an earthquake, were also evaluated.

The analyses indicated that the earthquake intensity required to collapse a face-loaded wall element, as indicated by the computer modelling, is generally predicted conservatively by the proposed methodology. However, when the earthquake motion contains a near-fault long duration pulse, the methodology is not conservative and tends to predict the mean earthquake intensity required to collapse the wall element. The methodology also becomes non-conservative for near-fault EQC motions if the building and/or diaphragms have more than moderate flexibility.

Design charts are provided to enable rapid design office assessment of face-loaded wall elements in terms of the current proposed revision to the NZS4203 Loading Standard Basic Seismic Hazard Spectra (ie, DR 2270.4/PPC3). Similar design charts could be prepared for other earthquake records or code design spectral intensities using the proposed methodology.

Laboratory testing of pre-cracked face-loaded wall specimens under simulated seismic loading has recently been carried out in Australia. Good agreement was obtained when inelastic dynamic modelling was used to predict the time-history of the mid-height displacement of one of the Australian test specimens. This agreement should improve confidence in the assessment methodology that was derived using similar computer models.