Probabilistic constitutive law for masonry veneer wall ties

Muhit, Imrose B.; Stewart, Mark G.; Masia, Mark J.

Title: Probabilistic constitutive law for masonry veneer wall ties
Creator: Muhit, Imrose B.; Stewart, Mark G.; Masia, Mark J.
Relation: ARC.DP180102334 http://purl.org/au-research/grants/arc/DP180102334
Relation: Australian Journal of Structural Engineering Vol. 23, Issue 2, p. 97-118
Publisher Link: http://dx.doi.org/10.1080/13287982.2021.2021628
Publisher: Taylor & Francis
Resource Type: journal article
Date: 2022
Description: In a masonry veneer wall system, tie strengths and stiffnesses vary randomly and so are not consistent for all ties throughout the wall. To ensure an economical and safe design, this paper uses tie calibration experimental approach in accordance with the standard AS2699.1 to investigate the tie failure load under compression and tension loading. Probabilistic wall tie characterisations are accomplished by estimating the mean, coefficient of variation and characteristic axial compressive and tensile strength from 50 specimens. The displacement across the cavity is recorded, which resulted the complete load versus displacement response. Using the maximum likelihood method, a range of probability distributions are fitted to tie strengths at different displacement histogram data sets, and a best-fitted probability distribution is selected for each case. The inverse cumulative distribution function plots are also used along with the Anderson-Darling test to infer a goodness-of-fit for the probabilistic models. An extensive statistical correlation analysis is also conducted to check the correlation between different tie strengths and associated displacement for both compression and tension loading. Based on the findings, a wall tie constitutive law is proposed to define probabilistic tie behaviour in numerical modelling.
Subject: wall-tie; strength; probability distribution; goodness-of-fit; tie constitutive law
Identifier: http://hdl.handle.net/1959.13/1464783
Identifier: uon:47106
Identifier: ISSN:1328-7982
Language: eng
Reviewed

Hits: 692
Visitors: 691
Downloads: 0