Computational fluid dynamics (CFD) is often used within performance based fire safety engineering and its use has increased as available computational power has increased. However, there is still a need to improve CFD modeling to push it beyond its current usage limitations. Of course such steps must be accompanied by quality assurance by means of validation and verification. In this thesis three key problems were identified within performance based design; prediction of the mass loss rate that interact with the environment dynamically, understanding of fires in enclosures equipped with mechanical ventilation as well as taking the built environment, such as building materials, building geometry and various technical installations, into account when designing a fire scenario. In the presented work Fire Dynamics Simulator (FDS) was chosen as a modeling framework which could be expanded upon if needed to be able to perform predictions of the presented problems. A validation of FDS was done against experimental data obtained using the novel, non-intrusive technique ps-LIDAR. The built in HVAC (heating, ventilation, and air conditioning) model in FDS was validated against a series of full-scale fires using mechanical ventilation. A new pool fire sub-model, which takes reduction in oxygen concentration and external radiative heat flux into account when predicating the mass loss rate, was formulated, implemented in FDS and then verified and validated. The verified and validated models and sub-models were applied on two engineering problems; predicting fire growth related to building characteristics and predicting performance of measures against smoke spread in ventilation systems.
- Patrick van Hees, Supervisor
- Daniel Nilsson, Supervisor
- Bjarne Husted, Supervisor
|Award date||2018 Jun 15|
- Division of Fire Safety Engineering, Lund University
|Electronic ISBNs||978-91-7753-680-2 |
|State||Published - 2018|
Place: V:D, V-building, John Ericssons väg 1, Lund University, Faculty of Engineering LTH.
Name: Boulet, Pascal
Affiliation: Université de Lorraine, France