Fire spreading in South African low-cost settlements "A physics-based model"

Moradi, Ali (2016-03)

Thesis (MEng)--Stellenbosch University, 2016.

Thesis

ENGLISH ABSTRACT: risk of fire is increasing due to urbanization, industrialization and development of cities with dense buildings. Since low cost settlements are poorly serviced and are densely built urban areas with poor service delivery, they are the most vulnerable areas to fire. In recent years, urban fire-spread simulator models have commonly been used amongst more developed countries. However, in the case of South Africa, no fire-spread model has been developed/calibrated for low cost settlements in. Therefore, this study aimed to develop, validate and verify the first physics-based fire spread model, which is specifically calibrated for low-cost settlements in South Africa. The objective of this thesis was to apply the new model in a real life case study to illustrate the results of the model and prioritising the influential fire spread factors. In order to provide a theoretical base for the development, the verification and validation of a new fire-spread model, a selected group of existing notable fire simulation models were reviewed in literature. To develop a new model, first the scope, input and output of the model was defined and described. Next, the process of the fire simulation was divided into modules, their functionality were explained and thereafter an appropriate sequence of the modules were described. Finally, the entire fire spread process was programmed in C#.net. A geographical information system (GIS) was employed to pre-process the input data as well as to provide graphical output. Verification and validation method were carried out in two hypothetical case studies. The new model was then applied to a low-cost settlement call? Imizamo Yethu which is a Metropolitan township of the city of Cape Town, South Africa, where fire is one of two top risks of the area. An area of this township was selected and input data for the new model were prepared through field measurement, information from Google earth and by determining parameters based on the environment of the area. The results were presented quantifiably and visually by a map series that showed fire spread progression across the area over time. The findings of the analysis determined that in the area of Imizamo Yethu, wind speed and the separation distance between buildings were the first and the second most influential fire spread factors. In the building factor category, fire load has the highest impact while window orientation has the lowest impact on the total burnt area?? as well as the speed of fire spread.

AFRIKAANSE OPSOMMING: Die ontwikkeling van stede met hoë digtheidsgeboue, industrialisasie asook die proses van verstedeliking, verhoog die risiko van brande. Aangesien lae-koste behuising gewoonlik baie dig beboude stedelike areas is met swak dienslewering, is sulke areas die mees kwesbare vir brande. In die afgelope paar jare, in ontwikkelende lande, word stedelike brand= verspreiding-nabootsende modelle of simulators, al hoe meer gebruik in die voorkoming van brande. In die geval van Suid- Afrika is daar egter nog geen so ‘n brandverspreidings-model ontwikkel nie; verder is geen van die bestaande modelle al toegepas in Suid- Afrika nie in ‘n poging om brande te verstaan of te voorkom nie. Met hierdie studie word beoog om die eerste fisika-gebaseerde brandverspreidings-model te ontwikkel, te verifieer en te bekragtig; ‘n model wat spesifiek aangepas en gekalibreer is vir lae koste nedersettings in Suid-Afrika. Die doelwit van hierdie tesis was om die nuut ontwikkelde fisika-gebaseerde model toe te pas in ‘n lewensgetroue situasie om die uitkomste van die model te demonstreer asook om die brandverspreidingsfaktore wat die grootste gewig dra, te prioretiseer. In die daarstelling van ‘n teoretiese basis vir die ontwikkeling, die verifikasie en bekragtiging van die nuwe brandverspreidings-model, is ‘n geselekteerde groep brand nabootsingsmodelle (simulators) in bestaande literatuur bestudeer. Om die nuwe model te ontwikkel is die strekwydte (scope), insette en uitsette van die data van die model eerstens gedefinieer en beskryf. Daarna is die proses van brandnabootsing (simulasie) verdeel in modules, waarna die funksionaliteit van die modules bespreek is; sodoende kon daar gekyk word na ‘n toepaslike volgorde van die modules. Laastens is die hele brandverspreidingsproses geprogrameer in C#.net. Die geografiese informasie sisteem (GIS) is gebruik om die insette data te proseseer so wel as om grafiese uitset data te bekom. ‘n Verifiërende en geldigheids-metode is uitgevoer in twee hipotetiese gevalle studies. Daarna is die nuwe model toegepas in a lae-koste nedersetting genaamdImizamo Yethu ‘n nedersetting in die Kaapse metropool van Suid-Afrika. In hierdie omgewing is brande een van die twee hoogste risiko’s van verstedeliking. ‘n Spesifieke gebied van hierdie nedersetting is gekies en invoerdata vir die nuwe model is voorberei deur gebiedstudies en metings, ‘Google Earth’ en bepalende grense gebaseer op die omgewing. Die resultate is kwantitatief weergegee asook visueel deur ‘n reeks kaarte wat die brand- verspreiding-stadia met tydsverloop uitbeeld. Die bevindings van die toegepasde data analise bewys dat in die area van Imizamo Yethu, windspoed eerstens en dan tweedens, die afstand wat tussen geboue bestaan, die faktore is wat die grootste invloed het op brandverspreiding. In die geboue-faktor kategorie, het brandlading/vuurvrag (fire load) die hoogste impak op brandverspreiding; daarenteen het die vensteroriëntasie die laagste impak op totale gebrand sowel as op die spoed waarmee die brand versprei.

Please refer to this item in SUNScholar by using the following persistent URL: http://hdl.handle.net/10019.1/98611
This item appears in the following collections: