Computational and Experimental Modelling of Masonry Walling towards Performance-Based Standardisation of Alternative Masonry Units for Low-Income Housing

De Villiers, Wibke Irmtraut (2019-12)

Thesis (PhD)--Stellenbosch University, 2019.

Thesis

ENGLISH ABSTRACT: South Africa has a housing shortage estimated in excess of 2 million units. This backlog is being addressed predominantly with the construction of 40m2 low-cost, single storey, detached state subsidised houses built with conventional masonry units (CMU’s), namely concrete and burnt clay. The use of these materials has a significant negative impact on the environment and the thermal performance of conventional masonry walls is generally poor. These factors, and others, have led to the development of alternative masonry units (AMU’s) in South Africa, and internationally, with a lesser environmental impact and improved thermal performance. However, lack of standards presents a significant barrier to the uptake of AMU’s The regulatory framework governing the use of masonry in South Africa, and possible avenues through which AMU’s could gain access to the market, are explored. It is found that AMU’s could provide a reasonable and socially acceptable alternative to CMU’s in low-income housing (LIH) but the current regulatory framework does not accommodate AMU’s in a sufficiently practical manner to enable their widespread, off-the-shelf uptake. The ongoing process of the adoption of Eurocode 6 and the accompanying materials and testing standards by the South African masonry industry, facilitates the transition from prescriptive to performance-based (PB) regulation of masonry design. It is proposed that material non-specific, PB requirements for masonry units for structural application in LIH can be developed to assist the inclusion of AMU’s in the open market. To quantify PB criteria, two critical configurations of single-storey bonded masonry walls are generated, based on the deemed-to-satisfy provisions of the National Building Regulations (NBR). Subsequently, a simplified micro-scale finite element (FE) model is used to analyse these configurations under serviceability and ultimate limit state loading conditions, to serve as a performance prediction model from which PB criteria can be derived. Four masonry materials are selected for the study; conventional concrete (CON), alkali-activated material or geopolymer (GEO), compressed-stabilised earth (CSE) and adobe (ADB), representing a wide spectrum in terms of strength and stiffness. Characterisation tests of the four materials are used, together with numerical fitting to test data and data taken from literature, to generate the necessary parametric input for the FE model. The results of medium to large-scale in-plane and out-of-plane tests are used for validation of the FE model. The FE analyses revealed that for most of the load cases, the resistances of the walls failed to achieve the design load, even for the conventional CON blocks. A significant shortfall was found for the out-of-plane resistance against the wind load for all four materials, as well as structural vulnerability under seismic loading due to the geometric layout permitted by the deemed-to-satisfy rules in the NBR. These results preclude the immediate derivation of PB criteria for AMU’s but contribute significantly to the body of knowledge surrounding FE modelling of AMU’s. They also indicate that the NBR provisions for wall panel geometry require reconsideration, given the recent revision of the South African loading code. However, material non-specific PB regulation is still the recommended avenue for the standardised inclusion of AMU’s.

AFRIKAANSE OPSOMMING: Suid-Afrika het ‘n behuisingstekort van meer as 2 miljoen eenhede. Hierdie agterstand word hoofsaaklik aangespreek deur die konstruksie van 40m2, enkel-verdieping, losstaande, staats-gesubsidieerde behuisingseenhede, wat meestal met konvensionele messelwerkeenhede gebou word, naamlik beton en gebakte klei blokke. Die gebruik van hierdie materiaal het ‘n beduidende negatiewe impak op die omgewing en die termiese gedrag van konvensionele messelwerkmure is ook swak. Hierdie faktore, onder andere, het gelei tot die ontwikkeling van alternatiewe messelwerkeenhede (AME’s) in Suid-Afrika en internasionaal, met verminderde omgewingsimpak en beter termiese gedrag. ‘n Gebrek aan standaarde verhoed egter die aanvaarding en gebruik van AME’s. Die raamwerk van regulasies wat die gebruik van messelwerk in Suid-Afrika beheer, sowel as moontlike maniere om die toelating van AME’s te bewerkstellig, word ondersoek. Dit word bevind dat AME’s ‘n redelike alternatief bied tot konvensionele messelwerk in lae-inkomste behuising, wat ook sosiaal aanvaarbaar is, maar die huidige regulatoriese raamwerk akkommodeer nie AME’s op ‘n prakties uitvoerbare manier nie. Tans is die aanneming van Eurocode 6 in Suid-Afrika, met gepaardgaande materiaal- en toets standaarde, ‘n deurlopende proses en dit fasiliteer die oorgang van voorskriftelike standaarde na ‘n stelsel wat op prestasie gebaseer is vir die ontwerp van messelwerk. Dit word voorgestel dat materiaal-onafhanklike, prestasie gebaseerde (PB) vereistes ontwikkel kan word vir messelwerkeenhede vir strukturele toepassings in lae-inkomste behuising, om die gebruik van AME’s te vergemaklik. Om PB kriteria te kwantifiseer word twee kritiese konfigurasies van enkel-verdieping, gebonde messelwerkmure gegenereer, op grond van voorskrifte in die Nasionale Bouregulasies (NBR) wat as bevredigend geag word. Daarna word ‘n vereenvoudigde mikro-skaal eindige element (EE) model gebruik om die muur konfigurasies te analiseer onder diensbaarheid en uiterste limietstaat lasaanwending, om as voorspellingsmodel van die gedrag te dien, waarvan PB kriteria afgelei kan word. Vier messelwerkmateriale is gekies, konvensionele beton (CON), alkali-geaktiveerde materiaal of geopolymer (GEO), saamgeperste, gestabiliseerde grond (CSE) en adobe (ADB), wat ‘n wye spektrum aan sterkte en styfheid verteenwoordig. Die nodige parametriese data vir die EE model word verkry vanaf karaktariseringstoetse op die vier blok tipes, so wel as numeriese pas van toetsdata en data van relevante literatuur. Die EE model word gevalideer deur middel van medium- tot grootskaalse in-vlak en uit-vlak toetse. Die EE analises wys dat die mure se weerstand in die meeste lasgevalle nie die ontwerpslas haal nie, selfs vir die konvensionele CON blokke. ‘n Beduidende tekortkoming is gevind vir die uit-vlak weerstand teen die wind lasgeval vir al vier materiale. Die seimiese lasgeval dui ook op strukturele swakhede wat onstaan vanuit die geometriese uitleg wat deur die NBR voorskrifte toegelaat word. Hierdie resultate verhoed die onmiddelik afleiding van PB kriteria vir AME’s, maar dit maak ‘n beduidende bydrae tot die kennis rakende EE modellering van AME’s en dui daarop dat die NBR se voorskrifte vir muurpaneel geometrieë heroorweeg moet word, gegewe die onlangse hersienning van die Suid-Afrikaanse laskode. Materiaal-onafhanklike PB regulasie word egter steeds aanbeveel vir gestandardiseerde insluiting van AME’s.

Please refer to this item in SUNScholar by using the following persistent URL: http://hdl.handle.net/10019.1/107047
This item appears in the following collections: