BUILDING FIRE SAFETY: NUMERICAL SIMULATION AND EVACUATION PLANNING

Proceedings of BS2015: 14th Conference of International Building Performance Simulation Association, Hyderabad, India, Dec. 7-9, 2015.

BUILDING FIRE SAFETY: NUMERICAL SIMULATION AND EVACUATION PLANNING

Vatsal Sanjay, Arup Kumar Das

Department of Mechanical and Industrial Engineering, IIT Roorkee-247667, India

ABSTRACT
Fire dynamics and fire propagation study reckon attention due to the role they play in evacuation planning and minimizing the loss of lives and properties in case of a fire breakout. For this paper, Fire Dynamics Simulator (FDS) is used to simulate a fire situation involving an air conditioning device in a two-room domain. The rooms are connected through a door. In our simulation, a window Air-Conditioner is modelled as a heat source with constant Heat Released Rate Per Unit Area (HRRPUA) to signify a fire source which in real life could be a result of AC malfunctioning. HRRPUA of AC is varied and its effects on soot flow pattern, burning rates of materials,
temperature contours and gauge pressure of both the rooms are studied.

Further, the Available Safe Evacuation Time (ASET) is calculated for the above cases. Using the ASET values and soot flow pattern, certain design changes in the rooms are suggested.

Windows are placed at those spots on the wall where the soot hits first or shows a tendency to flow.

Similarly, the exit doors are planned based on the safe zones and the ASET data. For the modified geometry, ASET is calculated which is observed to be higher than those of above cases. Reasons for this difference are suggested. Feasibility of safe evacuation is discussed for all the above cases including the calculation of Required Safe Evacuation Time (RSET). Present simulation and its findings will help in designing buildings and aid safety engineers to recognise and assess the risk of fire originating from air conditioning devices. Engineers could improvise and bring appropriate changes in their designs for systems and buildings of similar dimensions.

INTRODUCTION
Fire is one of the major hazards featuring among the causes leading to loss of properties and lives. A major chunk of non-natural deaths worldwide features fire as one of the prominent reasons [1]. Fire could be initiated due to pantry related causes including gasleakage, over-heating of cooking medium or electrical related causes like short circuits, overloaded circuits, leakage currents and electric sparks. In US alone, 2300 air-conditioners fires are reported annually. Each contributes to property loss of over $10,558. 86% of these fires are initiated due to mechanical failure or malfunctioning [2]. With increasing ownership of airconditioners in India [3] and air-conditioners working at full capacity during summer [4], the chances of malfunctioning due to over-loading, clogging of drainage, improper wiring or short-circuiting have also increased. A review of accidents involving fires show that the causalities are caused mainly due to suffocation (lack of oxygen), toxic smoke inhalation or direct exposure to flame [5]. These factors have been considered while calculating ASET.

In 1982, Markatos et al. outlined the need of numerical simulations in their paper [6]. They argued that apart from the experience we gain from real fires, the major dependence was on the scaled physical models since a full-scaled model would require prohibitive human and financial resources. However, it is not possible to achieve complete similarity between the real and scaled model and a compromise is invariably reached.

With the advancement in the field of digital computers and development of mathematical methodologies for fire modelling, we now have the flexibility of aptly simulating fire behaviours in different enclosed configurations, therefore overcoming the constraints in experimental and theoretical approaches [7].

Ta-Hui Lin et al. have highlighted the remarkable correlation between the full-scaled experiments and the numerical simulation [8]. In 2012, followed by a series of fire test burns carried out by Fire and Rescue New South Wales in discarded furnished room, a CFD based large eddy simulation was carried out. G.H. Yeoh et al. published the above results in their paper [9] and established the fact that if the pyrolysis and combustion rate of fuel are appropriately modelled then numerical simulations gave reasonable temperature and flow predictions. Although there have been studies validating the numerical simulations; studies with parametric variations are still unexhausted.

Owing to the sudden nature of fire breakout, we cannot guess its intensity beforehand. There appears to be very less experimental work on electric fires caused due to air conditioner malfunctioning. In this paper, we have modelled the air conditioner as a heat source with constant HRRPUA, which is varied in subsequent cases to cover a wide range of possibilities, which could be encountered during fire hazard.