Implementation of the Fire Beam Element (FBE) in OpenSees for the analysis of structures in fire

Abstract

ENGLISH ABSTRACT: Fire safety design for buildings has become an increasing concern worldwide, especially after disasters such as the World Trade Centre collapse and the more recent Grenfell Tower re. In such cases, structural re engineering plays a crucial role in the inherent re resistance of a structure. Traditionally, prescriptive methods were used to achieve re resistance. These methods, however, are largely based on singleelement tests in a standard re and therefore do not consider critical global building behaviour. Consequently, performance-based design methods are becoming accepted as the more scienti c approach to the design of structures in re. The Fire Beam Element (FBE) method is such a performance-based approach that simpli es a structure into a skeletal frame consisting of only beam and column elements. The work in this thesis employed the FBE method, modelled it in the OpenSees software environment, and subsequently applied it the context of global structures with restraint. To validate the developed model, benchmark case studies were sourced from literature and conducted in order to ensure that the model behaved accordingly. These case studies comprised a steel beam with high axial and bending forces, two simplysupported composite beams and a 2D steel frame, respectively. The second case study con rmed that even composite cross-sections can be modelled using a single beam element the model produced results similar to that of more a complex 3D shell model. The nal case study demonstrated that complex behaviour such as non-linear heating, restraint and buckling were captured by the model the latter two phenomena have not been considered in previous FBE work. A sensitivity analysis of the FBE model found that the e ect of increasing the mesh size of a cross-section does not increase the computational run time exponentially, thereby illustrating the developed models ability to handle more complex analyses. Furthermore, the e ciency of the FBE model allowed for a temperature sensitivity study to be conducted, and as temperature changes have signi cant e ect on structural behaviour, this would potentially be useful in design. In future, if the FBE method is developed in 3D, it could be linked to the Slab Panel Method, and presented as a possible analysis tool for structural re design of entire buildings.

AFRIKAANSE OPSOMMING: Brandveiligheidsontwerp in geboue is wêreldwyd 'n toenemende bekommernis, veral na rampe soos die Wêreldhandelsentrum-ineenstorting en die meer onlangse Grenfelltoring brand. In sulke gevalle speel strukturele brandontwerp 'n deurslaggewende rol in die inherente brandweerstand van 'n struktuur. Tans gebruik struktuuringenieurs hoofsaaklik voorskriftelike metodes in die ontwerp van brandbestandheid. Hierdie metodes word egter hoofsaaklik gebaseer op enkel-element toetse in 'n standaard vuur en beskou dus nie kritiese globale bougedrag nie. Gevolglik word prestasiegebaseerde ontwerpmetodes aanvaar as die meer wetenskaplike benadering tot die ontwerp van brandende strukture. 'n Voorbeeld van so 'n prestasie-gebaseerde ontwerp benadering is die Brand-Balk Element (BBE) metode. Die BBE vereenvoudig 'n struktuur deur dit in 'n skeletale raam te verander wat net uit balke en kolomme bestaan. Die werk in hierdie tesis het die BBE metode gebruik, gemodelleer in die OpenSees sagteware-omgewing en daarna is dit in die konteks van ingeklemde strukture toegepas. Om die ontwikkelde model te valideer, is standaard experimente van literatuur verkry en uitgevoer om te verseker dat die model homself dienooreenkomstig gedra. Drie sulke gevallestudies was uitgevoer: 'n staal balk met 'n hoë aksiale krag en buigkragte, twee saamgestelde eenvoudig-opgelegde balke en 'n 2D staal raamwerk. Die tweede eksperiment het bevestig dat selfs saamgestelde dwarssnitte met 'n enkele balk element gemodelleer kan word die model het resultate soortgelyk aan dié van meer komplekse 3D-dopmodelle gelewer. Die nale maatstafeksperiment het getoon dat komplekse gedrag soos nie-lineêre verwarming, ingeklemdheid en buiging deur die model vasgevang word. Die laasgenoemde twee verskynsels is nie in vorige BBE werk oorweeg nie. 'n Sensitiwiteitsanalise van die BBE model het bevind dat die e ek van die vergroting van die maaswydte van 'n deursnit nie die model se berekeningstydperk eksponensieel verhoog het nie, en sodoende is die ontwikkelde modelle se vermoë om meer komplekse ontledings te hanteer geïllustreer. Die eenvoudigheid van die BBE model laat toe om verskillende temeratuur pro ele te toets en is dus gerie ik vir ontwerp doeleindes. In die toekoms, as die BBE metode in 3D ontwikkel word, kan dit in samewerking met die Paneelmetode gebruik work, en aangebied word as 'n moontlike analise-instrument vir strukturele brandontwerp van hele geboue.