Você está na página 1de 44

SMOKE CONTROL MODELING ANALYSIS REPORT (DRAFT) SAHARA MALL EXPANSION SHARJAH, UAE

I49831 Page i 12 August 2010

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UNITED ARAB EMIRATES

SMOKE CONTROL MODELING ANALYSIS REPORT (DRAFT) SAHARA MALL EXPANSION SHARJAH, UAE

I49831 Page ii 12 August 2010

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UNITED ARAB EMIRATES

Prepared For:

Khatib & Alami CEC. King Faisal Road, Al-Faisal Building 2 Sharjah, UAE

12 August 2010

Project # I49831

SMOKE CONTROL MODELING ANALYSIS REPORT (DRAFT) SAHARA MALL EXPANSION SHARJAH, UAE

I49831 Page i 12 August 2010

TABLE OF CONTENTS INTRODUCTION ....................................................................................................... 1 APPLICABLE CODES AND STANDARDS ............................................................ 1 PROJECT DESCRIPTION ..................................................................................... 1 CODE INTENT ....................................................................................................... 2 STANDARD OF PERFORMANCE ............................................................................ 3 GOALS ................................................................................................................... 3 PERFORMANCE OBJECTIVES ............................................................................ 3 TENABILITY CRITERIA ......................................................................................... 3 SMOKE CONTROL SYSTEM OPERATION TIME................................................. 4 METHODOLOGY....................................................................................................... 5 FIRE MODEL ......................................................................................................... 5 MODEL BRIEF - FDS INPUT DATA .......................................................................... 7 CONTROL VOLUME .............................................................................................. 7 SMOKE EXHAUST VENTS AND MAKE-UP AIR ................................................... 9 BEAM DETECTORS ............................................................................................ 12 SMOKE DETECTORS ......................................................................................... 12 HEAT DETECTORS (AUTOMATIC SPRINKLERS)............................................. 13 FIRE SCENARIOS................................................................................................... 14 SCENARIO 1: AXISYMMETRIC FIRE IN OPEN ATRIUM ................................... 14 DESIGN FIRE ................................................................................................... 14 BEAM DETECTOR ACTIVATION .................................................................... 17 MODELING RESULTS ..................................................................................... 19 SCENARIO 1 CONCLUSION ........................................................................... 30 SCENARIO 2: BALCONY SPILL PLUME IN A COMMUNICATING SPACE ....... 32 DESIGN FIRE ................................................................................................... 32 SPRINKLER ACTIVATION ............................................................................... 35 SMOKE DETECTOR ACTIVATION .................................................................. 36 MODELING RESULTS ..................................................................................... 37 SCENARIO 2 CONCLUSION ........................................................................... 40 SMOKE EXHAUST RATE AND MAKE-UP AIR RESULTS .................................. 42 DESIGN CRITERIA TO BE CONSIDERED ............................................................. 45 CONCLUSION ......................................................................................................... 48

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE

I49831 Page 1 12 August 2010

INTRODUCTION This report documents the smoke control modeling analysis completed by Rolf Jensen & Associates (RJA) to assist the design team in developing an effective and efficient smoke control strategy for the Sahara Mall Atrium. The analysis is based on the architectural drawings dated June 2006, related information provided by Khatib & Alami as well as discussions with RJA during various meetings and telephone conversations. Mechanical and Architectural design changes made to these atrium after this analysis may produce varying results and may require the completion of additional evaluations of the smoke control system performance. APPLICABLE CODES AND STANDARDS The modeling analysis is based on, but not limited to, the following codes and standards: International Building Code (IBC), 2009 Edition. NFPA 101, Life Safety Code, 2009 Edition. NFPA 92B, Standard for Smoke Management Systems in Malls, Atria and Large Spaces, 2009 Edition.

PROJECT DESCRIPTION This project is an expansion to the Sahara Mall located in Sharjah, UAE which comprises of two basement floors mainly used for parking and a ground floor with two floors above grade which are mainly retail. This complex mall has floor openings at the center of the floor plan which open onto all the floors (2nd Basement through Level 2), thus creating a 5 story atrium. The atrium will include escalators for people movement between floors. The atrium will also include a water feature which extends from the 2nd floor ceiling to the 2nd basement. The escalators on the 1st and 2nd basements leading to the upper floors are enclosed and separated from the parking areas as seen in the Figures below. Hence it is assumed that no fire load will be present at the enclosed spaces in the 2nd and 1st basements. Based on the sheer size and width of the mall, it would be necessary to design the smoke control system using a sophisticated fire modeling software as a tool to develop a smoke control system which is capable of enhancing the building performance with optimal efficiency during a fire scenario.

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE
Figure 1 - 2nd Basement Level Rated Separation Figure 4 - 1st Level Floor Opening

I49831 Page 2 12 August 2010

Figure 2 - 1st Basement Level Floor Opening Rated Separation

Figure 5 - 2nd Floor Floor Opening

Figure 3 - Ground Level Floor Opening

Figure 6 - 3rd level

Fan Room location

CODE INTENT The atrium provisions in the model codes have been developed to mitigate the potential for fire and smoke spread to areas of a building beyond the compartment of fire origin as well as maintaining the atrium environment tenable in areas through which occupants egress. Prior to atrium provisions, openings between floors were treated as shafts and alternate means and methods needed to be developed in order to accommodate the openness provided by atriums within buildings. The atrium provisions incorporate other fire protection features besides smoke control. These features include passive barriers to smoke spread at the perimeter of the atrium as well as automatic sprinkler protection throughout the building. Smoke detection (i.e. smoke and beam detectors) is also required which provides early warning to the fire service and occupants as well as an automatic means for activating the smoke control system. These systems are also required to be provided with emergency power.

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE

I49831 Page 3 12 August 2010

STANDARD OF PERFORMANCE GOALS The overall goal of the atrium smoke control system being designed for this project is to maintain a tenable environment within the atrium space for sufficient time to allow occupants to exit without being exposed to smoke or hot gases that would cause injury during evacuation. This overall goal is transformed into terms of measurable performance objectives, which are based on the SFPE Engineering Guide to Performance-Based Fire Protection Analysis and Design of Buildings (SFPE Guide) and other available literature. They establish the tenability performance criteria against which the computer modeling results will be measured.

PERFORMANCE OBJECTIVES The performance objectives are in the form of specific tenability conditions that occupants of the building would be exposed to in a fire. They are: Minimize heat exposure to occupants Maintain visibility to illuminated exit signs Minimize toxic gas exposure to occupants

TENABILITY CRITERIA The performance objectives are typically evaluated using specific tenability criteria values related to visibility, air temperature and smoke toxicity. These criteria are developed from the SFPE Handbook and other sources. Visibility. Visibility is an important tenability factor since it affects an occupants ability to find a way to an exit during a fire condition. Available fire protection literature gives a wide range of acceptable visibilities. The 3rd edition of the SFPE Handbook (Section 2, Chapter 4) 1 states that acceptable values from 1.2m to 13.5m have been proposed by various fire researchers based on the building application. Lougheed 2 summarized recommendations of the visibility criterion as follows: 3-5 m for occupants familiar with a building and able to evacuate quickly, and 10 m widely accepted and used internationally. In this project, a visibility of 10 m to a lighted exit sign is proposed to be the visibility criterion based on the atrium geometry and building occupancy. This proposed visibility criterion of 10 m is at the high end of these documented values and will be applied throughout the communicating spaces of the mall. The shops are assumed be to be small spaces with 1 or 2 exits where people are able to evacuate easily and therefore 5 m visibility would be an acceptable criteria.

Tadahisa Jin, Visibility and Human Behavior in Fire Smoke, SFPE Handbook, 3rd edition, Section 2 Chapter 6, pp. 2-47. 2 G.D. Lougheed, Basic Principles of Smoke Management for Atriums, Construction Technology Update No.47, Published by Institute for Research in Construction.

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE

I49831 Page 4 12 August 2010

Air Temperature. Surrounding air temperature is a potential way by which the heat from fire and smoke can affect occupant evacuation ability. The 3rd edition of SFPE Handbook (Section 2, Chapter 6) 3 states that 60C has found to be the highest temperature at which 100% water-vapor saturated air can be breathed. As the combustion process and potential fire suppression can produce water vapor, 60oC is identified as the tenability limit of the surrounding air temperature in the fire environment. Carbon Monoxide. Carbon monoxide (CO) is one of the measures of the smoke toxicity that occupants are exposed to during evacuation. CO exposure can cause occupant incapacitation by combining with hemoglobin in blood to form carboxyhemoglobin (COHb), which reduces the amount of Oxygen (O2) delivered to the brain and other body tissues. The COHb concentration likely to cause incapacitation depends on the activity of the victim. The 3rd edition of SFPE Handbook (Section 2, Chapter 6) 4 gives the relationship between the time to incapacitation and CO concentration exposed for a 70 kg human at different levels of activity. For light work (e.g. walking) and heavy work (e.g. slow running), the time to incapacitation can be more than 30 minutes when exposed to a CO concentration of 1,000 ppm (parts per million). In this analysis, a single point value of 500 ppm as a volume fraction or 5.0E-04 mol/mol as a concentration factor is proposed to be the CO concentration criterion, which is conservative. Table 1 shows the tenability criteria used in this analysis. In each fire scenario, each criterion will be evaluated within the computer fire model at a height 2 m (6 ft) above the floor, which is assumed to be at occupant breathing height, in accordance with IBC Section 909.8.1.
Table 1 - Tenability criteria used in each scenario

Tenability Parameter Visibility Air Temperature Carbon Monoxide

Tenability Criterion > 10 m during egress in the atrium > 5 m during egress in the retail area < 60C during egress < 500 ppm during egress

SMOKE CONTROL SYSTEM OPERATION TIME In accordance with the requirement of IBC Section 909.4.6, all portions of the smoke control system shall be capable of continued operation after detection of the fire event for a period of not less than either 20 minutes or 1.5 times the calculated egress time, whichever is less. For this analysis we will assume that the fire will continue for a period of 20 minutes upon detection of the fire.

David A. Purser, Toxicity Assessment of Combustion Products, SFPE Handbook, 3rd edition, Section 2 Chapter 6, pp. 2-127. 4 David A. Purser, Toxicity Assessment of Combustion Products, SFPE Handbook, 3rd edition, Section 2 Chapter 6, pp. 2-102.

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE

I49831 Page 5 12 August 2010

METHODOLOGY FIRE MODEL The fire modeling for this project was completed by using Fire Dynamics Simulator (FDS, version 5.0) developed and maintained at the National Institute of Standards and Technology (NIST). It is a computational fluid dynamics (CFD) model designed for low speed and thermal driven flow, with an emphasis on the numerical simulation of the heat and smoke spread from a fire. There is substantial documentation of this software and additional large scale and full scale fire testing which was done to validate the model. More information on the model validation can be found in the FDS Users Guide that can be downloaded from the NIST website: http://www.bfrl.nist.gov/info/software.html The FDS model uses a text file for the input data. The location of structural elements, as well as their physical properties, is mapped to define the volume and geometry of the space. The volume is subdivided into small cubes, some of which represent portions of solid obstructions and others that represent air. In addition to the input data file, there is a properties database that contains the physical characteristics of various materials such as the combustion characteristics of building materials and contents. This file is incorporated by reference in the input data file and provides information necessary to complete the computational process. More information on the input file parameters can be found in the FDS Users Guide. The following variables can be included as inputs in the FDS model and are addressed in the analysis: 1. 2. 3. 4. 5. 6. 7. 8. 9. Ambient temperature of the space Building geometry and combustion properties Specific combustion properties of design fires The effects of suppression Activation time of the smoke control system Duration of the simulation needed to collect useful data Grid resolution (computational accuracy) Location and type of measurements (data collection) Design fire assumptions as discussed in the following sections

In addition to the raw data output, the FDS model has several graphic output formats which are helpful in visualizing the data, such as slice files, Plot3D files, and thermocouples (THCs). The visualization is done in the program SmokeView specifically developed for viewing the output data from FDS. More information on viewing the output files can be found in the SmokeView Users Guide which can be downloaded from the NIST website.

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE

I49831 Page 6 12 August 2010

Slice files are color coded sheets or slices through the control volume which allows the user to visually observe conditions at the specified plane. Slice files can be defined at any horizontal or vertical plane within the computational domain during the definition of the model input. This will provide results at this specified plane. Slice files can be used to show a variety of output including visibility, temperature, volume fractions of various gas products, etc. It allows the user to observe distributions and changes of a specific output parameter at the specified plane within the model space, for the duration of the simulation. Plot3D files show similar output as slice files, and can be used to evaluate a similar range of output. However, the Plot3D files can be used to view conditions at any horizontal or vertical plane within the modeled space, at specified time intervals. For example, one can set the FDS model to output temperature or visibility information every 30 or 60 seconds, allowing one to view the temperature or visibility results anywhere within the modeled space at this time interval. Thermocouples (THCs) are a spot-type representation of the output data that can be plotted on a graph. The thermocouple can record data including temperature, visibility, and CO concentration at a specific location and is used for these purposes in this analysis.

In this analysis, slice files and Plot3D files are used in evaluating gas temperature, visibility, and carbon monoxide concentration. Thermocouples (THCs) are used to track the volume flow rate of the smoke exhaust vent and make-up air vents.

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE

I49831 Page 7 12 August 2010

MODEL BRIEF - FDS INPUT DATA CONTROL VOLUME The control volume used in this analysis will be the main core area of all floors including the atrium floor openings, retail corridors and shops and excluding the two basement floors and the side shop areas as shown in Figure 7. These side shop areas will not affect the calculation since it is considered that occupants will be able to evacuate these shops immediately upon fire alarm signal.
Figure 7 - Control Volume used in the Model

The geometry for the model is shown in Figure 8. The atrium construction is approximated using the rectilinear coordinate system used in FDS and the balconies are modeled along with other obstructions caused by interior partitions throughout the building. A computational grid of 0.5 x 0.5 x 0.5 meters in the x, y, and z planes is used within the control volume. This computational grid is represented graphically in Figure 9. All geometric features used in the model, shown in must conform to these units of measurement. The result is approximately 1,679,296 cells over which the calculations are conducted.

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE
Figure 8 - Sahara Mall Geometry Model

I49831 Page 8 12 August 2010

Figure 9 - Computational grid of 0.5 x 0.5 x 0.5 meters in the x, y, and z planes

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE

I49831 Page 9 12 August 2010

SMOKE EXHAUST VENTS AND MAKE-UP AIR Smoke Exhaust Vents. The overall smoke control concept of this atrium involves extracting smoke from the top of the atrium space and supplying make-up air both mechanically and naturally from the ground floor. The smoke exhaust vents will be located above ceiling of the 2nd floor. The fans will be located in a mechanical room on the level 3 and will include ducts that will pass horizontally across the ceiling space and exhaust from the ceiling void as shown in Figure 10. Twelve sections located as part of the ceiling voids on level 2 (as seen in blue color around the atrium perimeter in Figure 11) will be fitted with 10 cm width openings on each side length for exhaust. This arrangement will not be provided in the model due to its intricate details which can exceed the model cell limitations of 2,000,000 cells. However, a 6 cm width opening is distributed around each of the ellipses as shown in Figure 12. These ellipses are in a shape of steps with each step being fitted with a 6 cm grille. Based on the Fire Dynamic Simulators limitations, the openings on each of the ellipses will be considered. However, given the complexity and detail of including 6 cm in the model, a 25 cm opening width will be considered. This value is an approximation and will be used as an alternate method in terms in approximating the aggregate area of openings provided on the drawings (6 cm openings around the ellipses and the 10 cm openings around the void sections) and the ones provided in the model (25 cm opening around the ellipses).
Figure 10 - Atrium Section showing the exhaust system configuration

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE
Figure 11 - Exhaust Vents as shown on K & A plans

I49831 Page 10 12 August 2010

Figure 12 - Exhaust Vents as shown in the Mode

Exhaust Vents

25 cm openings between each roof top

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE

I49831 Page 11 12 August 2010

Mechanical Make-up Air. Per the coordination meeting with the architect and mechanical engineers of K & A on 29 July 2010, it was determined for the current architectural/Mechanical design that the only way to supply make-up air from the ground floor was through the HVAC diffusers located on the ceiling of the ground level as shown in Figure 13 below. According to the drawings sent to us and dated 04 August 2010, there are 27 make-up air diffusers located in the ceiling voids on the ground level in the retail corridor. In case of a fire scenario, a signal from the fire alarm will automatically switch the control valve of the main duct allowing make-up air to be provided through all the diffusers in the retail areas and around the atrium floor opening. Natural Make-up Air. Natural make-up air will also be provided on the ground floor in addition to the mechanical make-up air. One automatic sliding door located at the ground floor retail communicating spaces will receive a signal from the fire alarm to open to the exterior in case of a fire, allowing natural make-up air to flow into the building. This will be applied in the event of beam smoke detection in the atrium, smoke detection by smoke detectors and sprinkler activation within the retail and communicating spaces. The feasible rate at which the total make-up air would be provided was calculated by the mechanical engineer of record to be approximately 75 m3/sec. It is recommended per international building code practices that make-up air be designed at 75 percent to 95 percent of the exhaust to avoid positively pressurizing the large-volume space. Meanwhile, a maximum make-up air velocity is recommended to be 1.02 m/s (200 ft/min) so as not to deflect the fire plume significantly. The remaining percentage is assumed to be make-up air resulting from cracks, openings, etc.
Figure 13 - Model showing the natural and mechanical make-up air

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE

I49831 Page 12 12 August 2010

BEAM DETECTORS The model includes smoke laser beam detectors to predict the activation time. A delayed fan configuration resulting from smoke detection transport lag time, signal processing from the fire alarm and fan ramp up time to final speed was added to the activation time. This detector activation with a delayed fan configuration had to be determined to identify critical events in the simulation including the time at which the ventilation systems would start the exhaust mode operation, the time at which makeup air would start and other factors such as the time at which occupant notification would take place for egress. The proposed design includes beam-type smoke detectors with a default sensitivity setting of 33 percent obscuration. The beams are located at the atrium opening of Level 2. The transmitter and receiver are positioned such that they cross interfere to cover most of the atrium opening.
Figure 14 - Beam Detectors within the model

SMOKE DETECTORS Smoke detectors have been provided along the retail communicating space at the ceiling of the ground floor to identify the activation time of when the smoke exhaust vents and make-up air diffusers would start for a balcony spill fire scenario as shown in Figure 15 below. An obscuration factor of 3.28%/m was taken based on the FDS value for a photoelectric type detector. An uncertainty factor has also been used to assess delay and activation times as indicated in the Balcony Spill Plume Scenario below. The type and characteristic of the smoke detectors have been identified to be photoelectric type and a chamber obscuration quantity with: Alpha E=1.80 Beta E=-1.00 Alpha C=1.00 Beta C=-0.80

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE

I49831 Page 13 12 August 2010

HEAT DETECTORS (AUTOMATIC SPRINKLERS) Heat type detectors have also been used in the model to simulate the time to activation of sprinklers for a controlled balcony spill fire scenario as shown in Figure 15 below. An uncertainty factor has also been used to assess delay and activation times as indicated in the Balcony Spill Plume Scenario below. The heat detectors have been placed on the ceiling of the ground floor with the following characteristics: Initial Temperature=23.00 deg C Activation Temperature=74.00 deg C RTI=90.00
Figure 15 - Smoke and Heat Detectors within the model for a Balcony Spill Plume

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE

I49831 Page 14 12 August 2010

FIRE SCENARIOS Two fire scenarios will be examined in an effort to determine the most severe yet reasonable design fire in terms of the fire and life safety hazard to evacuating occupants. A non-sprinkler controlled axi-symmetric plume in the open area of the atrium and a sprinkler controlled balcony spill plume condition caused by a potential fire in the communicating space. Scenario 1 Axisymmetric fire in an open atrium Scenario 2 Balcony Spill Plume in a communicating space.

SCENARIO 1: AXISYMMETRIC FIRE IN OPEN ATRIUM The first fire scenario to be considered will be a large, non-sprinkler controlled fire in the open atrium area. The floor openings extend from Basement 2 to Level 2 and will connect 5 levels. The two basement levels will be excluded from calculation since they are enclosed and separated from the parking spaces. The basements will also include a water feature and it is assumed that no fire loads will be present within the basement atrium levels. The fire location will be on the ground level at the center of the atrium. A large fire within the open main atrium area will most likely be a non-sprinkler controlled fire since the main atrium space is +20.0 meters high and roof level sprinklers (if provided) will likely have little effect on the fire. The code does not require automatic sprinkler protection at the ceiling of the atrium if the atrium height does not exceed 16.764 m per IBC Section 404.3.

DESIGN FIRE For the purpose of developing input parameters for a fire model simulation, a design fire must be considered with an emphasis on establishing a heat release rate (HRR) curve. The expected use of the atrium is assumed to be circulation area housing kiosks, seating areas with chairs and sofas, and seasonal decorations. A severecase combustible load has been determined to be typical combustibles that may be present on the ground floor in this type of atrium. The potential fuel package will possibly include upholstered seating (sofas), groups of tables and chairs, and seasonal decorations, including trees or plants. A severe-case combustible load has been determined to be typical combustibles that may be present on the bottom floor in this type of atrium in the tower. These potential fuel packages include groups of chairs, upholstered seating (sofas), and seasonal decorations including Christmas trees. Table 2 identifies measured peak heat release rates for these types of fuel loads based on fire tests.

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE
Table 3 - Peak Heat Release Rates

I49831 Page 15 12 August 2010

Fuel Package 3-Cushion Upholstered Sofa1 20 Kg Scotch Pine Tree1 8 Metal Frame Chairs with Cushions1 12 Stackable Polypropylene Chairs with Steel Frames2 Traditional Sofa3

Peak Heat Release Rate 3200 kW 5300 kW 2400 kW 2300 kW 3120 kW

(1) Klote, J.H., Milke, J.A., Principles of Smoke Management, Society of Fire Protection Engineers (2002) (2) Babrauskas, V., Heat Release Rates, The SFPE Handbook of Fire Protection Engineering, 3rd Edition 2002 (3) NFPA 92B. Smoke Management Systems in Malls, Atria, and Large Areas, 2009 Edition

In accordance with the review of potential fuel packages that may be present on the ground floor of the atrium, the peak heat release rate of a 20kg Scotch Pine Christmas Tree of 5300 kW Heat Release Rate is selected to be the design fire size which represents a challenging design fire for the selected fire location. This peak heat release rate is an average of the heat release rate values for the variety of individual fuel packages listed above. Since it is assumed that the fire is nonsprinkler controlled, this heat release rate value is used for smoke exhaust calculation to be conservative. The fire growth is assumed to be a t-squared medium growth rate fire where the constant alpha () equals 0.01172 kW/sec2. It is a typical fire growth rate for light combustible materials often found in atria and is conservative with respect to detection time. The simplified fire growth rate formula of Q = t 2 is used as the fire growth curve until it reaches the peak heat release rate. It is also assumed that the fire will maintain the peak heat release rate after it reaches the peak value. It is a conservative assumption since the fire normally decays quickly after reaching the peak heat release rate. Figure 16 shows the selected design fire size curve, which is inputted into the FDS model of this atrium space. The peak HRR of 5.3 MW occurs approximately 672 seconds after the start of the simulation. The incipient stage is relevant in this analysis because it could influence the detection time. Detection prior to the onset of severe conditions can be a valuable asset as part of the life safety performance of the building. The final data point is at 1,200 seconds (20 minutes) to reflect the duration of smoke control.

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE
Figure 16 - Design Fire Size Curve for an Axisymmetric Fire

I49831 Page 16 12 August 2010

In addition to specifying the heat release rate curve of the design fire in FDS, combustion properties related to the materials involved with the design fire must also be assigned. The two primary values that impact the amount of smoke produced are the soot yield, which is the mass of smoke or soot produced for a given mass of fuel consumed, and the heat of combustion. The heat of combustion is critical, as it relates to the amount of fuel that needs to be burned or consumed to produce a given heat release rate. The fire geometry has been taken to be 0.5 x 0.5 x 0.5 cubic meters with the upper surface simulated as the burner while the other sides being inert. The materials involved in a potential fire are assumed to be a combination of typical materials that might be found in this type of atrium. To provide a conservative fire representation with respect to soot production, which has a direct impact on visibility through the smoke, the materials for the design fires are assumed to be a relatively equal mixture of cellulosic materials (wood, paper, cardboard) and synthetic materials (plastics, rubber, fabrics, polyurethane foams). Cellulosic fires typically have low soot yields while synthetic materials typically have high soot yields. Based on the soot yield values and heat of combustion given by Tewarson 5 , an average soot yield of 0.05 g/g (grams of soot produced per gram of fuel burned) and an average heat of combustion of 18,250 kJ/kg are used in the FDS model.

Archibald Tewarson, Generation of Heat and Chemical Compounds in Fires, SFPE Handbook, 3rd Edition, Section 3 Chapter 4, pp. 3-111.

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE

I49831 Page 17 12 August 2010

BEAM DETECTOR ACTIVATION Analysis of beam detector activation is accomplished using the horizontal slice files of extinction coefficient and integrating across the beam length to determine the percent obscuration of light. This is done using the following equation:

I / I 0 = e KL
Where; I / I 0 = ratio of light which passes through the smoke compared to the original intensity of light emitted, K = extinction coefficient, and L = path length of the beam. The formula used to calculate the extinction coefficient in the FDS model is:

K = K m * YS *
Where; Km = constant, = density of smoke (air), and Ys = soot yield. The default value used in FDS for Km is 7.6 m2/g. This value is the generally accepted default value for most materials and for a number of factors such as the mass size distribution, the ratio of particle diameter to the wavelength of light, particle density, etc. Another widely used item in engineering is light obscuration, Sx in %, which is used to describe the visibility in a smoky enclosure. The definition of the light obscuration is given as follows:

S x = 100(1 I / I 0 )
OB = S x / L = (1 e KL ) / L ; Where OB is the obscuration per meter
Using a beam length of 27 meters and a setting of 33 percent obscuration, the extinction coefficient is computed to have a value of approximately 0.03 m-1. This calculation assumes a relatively constant value throughout the distance between the beam transmitter and receiver, which is the normal assumption for such calculations. In reality, the smoke properties will vary with the amount of mixing as predicted by the FDS model. Assuming the extinction coefficient, and resulting obscuration, the predicted time of actuation is approximately 45 seconds and adding 60 seconds for signal processing, the time of notification will results in a 105 second for the activation of the fans.

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE

I49831 Page 18 12 August 2010

Figure 17 - Horizontal illustration of extinction coefficient on level 2 of the atrium at 45 seconds. The obscuration scale is taken to be 0 to 0.03 m-1.

Red Color Contour indicates a 0.03 m-1 obscuration level.

In this simulation, the extinction coefficient increases substantially with time over the volume of the atrium. The value at 45 seconds approaches 0.03 m-1 over most of the volume monitored by the beam detectors in FDS and activation can be predicted with confidence. Therefore, the conservative value of 45 seconds will be used as the time of detector activation, and adding 1 minute for signal processing, the time of notification is 105 seconds.
Figure 18 - Obscuration level at activation time

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE

I49831 Page 19 12 August 2010

MODELING RESULTS As discussed in the section titled Standard of Performance, the tenability criteria defined for this scenario are: visibility >10m, air temperature <60oC, and CO concentration <500 ppm, evaluated 2 m above the walking surface on each occupied floor. The visibility tenability criteria limit was slightly exceeded for this atrium space due to the small vent openings around the perimeter of the atrium that allows smoke to accumulate and descend at 2 m above the highest walking level for the 20 min egress duration (i.e. at 16 m). Although the vent openings were selected to be 25 cm which is more than what the original design includes (6 cm openings around each of the roof levels and the 10 cm openings around the ceiling voids at level 2), it is predicted that the results will be much severe if the original design was not able to be maintained properly (i.e. velocities around the perimeter opening and exhaust rate). The air temperature and CO concentrations remained an order of magnitude below the defined tenability criterion limiting value in the occupiable walking space within the atrium. Heat Release Rate The actual HRR determined by the model is plotted in Figure 19. The HRR from the FDS model simulation tracked almost identical to the input. The FDS simulation models the combustion process during the simulation, and numerous factors may influence the actual HRR such as air entrainment, insufficient oxygen, re-radiation, etc. Differences of 100-200 KW are usually not significant. The good correlation is probably the result of the open area around the fire which ensures there are no restrictions in the flow of oxygen to the fire.
Figure 19 - Heat release rate of design fire as input into and output by the FDS model.

Visibility

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE

I49831 Page 20 12 August 2010

Figure 20 and Figure 21 illustrate the visibility conditions within the atrium at different times in the analysis. The output is shown using vertical and horizontal slice files passing through the middle of fire and and at 2 m above the highest walking space of the 2nd level respectively. The colored contours in each figure show areas with a visibility of 10 m or less due to low visibility conditions. As shown in the figures below, the visibility is maintained at 10 m except for some areas where visibility is around 8 m. This is due to the size of the openings located on the ceiling that do not allow enough smoke to be exhausted. The velocities through these openings are also low compared to the exhaust vents in this case which lessens the smoke momentum for suction. The solution would be to create more vent areas and increase the velocities around the openings which enable smoke flow movement for better exhausting. The simulated model does not take into account the rest of the shop areas around the perimeter due to the sheer size and volume of the model that cannot be incorporated into FDS due to the grid cell limits. Currently the boundary of the model is acting as a wall thus impacting and descending the smoke. This is to say that if the spaces around the boundary are included in the model, the smoke will flow and spread to these areas thus requiring more time to accumulate and descend. Therefore, the model with the fire located on the center is considered to be capable of maintaining a tenable condition of visibility for occupants evacuation if proper design changes are taken into considerations.
Figure 20 - FDS modeling results for visibility within the Atrium with an axisymmetric fire located at the ground floor at 1200 seconds. Figure below show horizontal slice files through the open atrium at 16 meters above ground level. The colored contours show areas where the visibility is 10 meters or less. The scale is taken to be between 0 and 10 m.
Model boundaries (represented by blue arrows) assumed as a wall, thus smoke impact and descend. Should the areas of the mall be taken into consideration, smoke would flow to these areas and wont descend immediately.

Areas with low visibility

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE

I49831 Page 28 12 August 2010

Figure 21 - FDS modeling results for visibility within the Atrium for an axisymmetric fire located at the ground floor at 1200 seconds. Figure below show vertical slice files through the open atrium at 16 meters above ground level. The colored contours show areas where the visibility is 10 meters or less. The scale is taken to be between 0 and 10 m.

Level 2 Level 1 Ground

Temperature Figure 22 and Figure 23 illustrate the temperature output slice files. The outputs are shown using vertical and horizontal slice files. The results show that temperatures in the communicating spaces open to the atrium are less than the defined tenability criteria.
Figure 22 - FDS modeling results for air temperature within the Atrium 2 m above level 2 at 1200 seconds for an axisymmetric fire. The figure shows the horizontal slice file at 2 meters above level 2 in the open atrium space. The temperature scale is set between 0 and 60 oC.

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE

I49831 Page 29 12 August 2010

Figure 23 - FDS modeling results for air temperatures within the atrium 2 m above level 2 at 1200 seconds for an axisymmetric fire. The figure shows the vertical slice files through the fire and the open atrium space. The temperature scale is set between 0 and 60 oC.

Level 2 Level 1

Ground

The Plot3D in Figure 24 below shows the tenability criteria for temperature at 1200 seconds. This 3D plot shows the temperature variations within the atrium where we can get an idea of what the temperature profile looks like at any x, y and z location. As seen in the figure, the high temperatures are shown by the colored contours at the center of the fire location and smoke rise. These temperatures tend to cool down as they float along the ceiling. The smoke along the ceiling is exhausted thus cooling the temperatures and keeping the temperature below 60C.
Figure 24 3D Temperature profile at 1200 seconds showing the different temperatures within the atrium at this duration. The temperature scale is set between 0 and 60 oC. Similar profiles can be obtained for the visibility and CO Concentration.

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE

I49831 Page 30 12 August 2010

Carbon monoxide (CO) Concentration Figure 25 illustrates the CO concentration output slice files. The outputs are shown using vertical slice files. The results show that the CO concentrations in the communicating spaces open to the atrium are significantly less than the defined tenability criteria.
Figure 25 - FDS modeling result for CO concentration within the Atrium; 2 m above level 2 at 1200 seconds for an axisymmetric fire. The figure shows the horizontal slice file through the open atrium space. The concentration scale is set between 0 and 500 ppm. (i.e. between 0 and 5x10-4 mol/mol)

SCENARIO 1 CONCLUSION Based on the above analysis, the following conclusions can be made regarding the performance of the building under the Axisymmetric Fire Scenario.

The beam type smoke detector system should activate in this scenario in about 105 seconds based on the 33% obscuration on Level 2 of the atrium including a fan delay configuration. The temperature and carbon monoxide levels are significantly less than the established tenability limits in the atrium. Visibility on the 2nd level within the atrium approaches the limits of tenability and consideration should be taken to increase the areas of the openings leading to exhaust as well as increasing the velocities at these openings to ensure that smoke movement and flow is being exhausted and not becoming stagnant and descending. The system shall be tested and verified to ensure proper functionality.

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE

I49831 Page 31 12 August 2010

Given the results of the analysis for this scenario, the performance of the smoke control system can be acceptable if the proper design changes as mentioned above are taken into consideration.
Table 4 Summary of Axisymmetric Fire Scenario

ITEM Scenario Design Fire Fire Size Fire Curve Fire location Smoke Zones Smoke Exhaust Rate Make-up Air (Mechanical & Natural) Beam Smoke Detector Activation Smoke Exhaust Velocity at Inlet Make-up Air Velocity at Outlet Visibility Temperature CO Concentration

DESCRIPTION 1 Axisymmetric Fire (Non-sprinkler controlled fire) 5.3 MW Steady State occurs after 672 seconds at 5.3 MW Fire located at the Ground Level at the center of the atrium. The geometry of the fuel is taken to be 0.5 x 0.5 x 0.5 meters. 1 zone for the entire mall with no fire shutters. 125 m3/sec total for 5 vents located at the ceiling of Level 2 75 m3/sec for 27 vents located at the ceiling of the Ground Level 55 seconds for two cross interfering beams with 33%/m obscuration located at the atrium opening of the 2nd Level in addition to 60 second delay for startup of Exhaust Vents and Make-up air diffusers. 3 3.5 m/sec 1 1.2 m/sec > 10 m; passed in most areas, but proper design changes should be considered for compliance in all areas. < 60 deg C; passed < 500 ppm; passed

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE

I49831 Page 32 12 August 2010

SCENARIO 2: BALCONY SPILL PLUME IN A COMMUNICATING SPACE The second scenario considers a sprinkler controlled fire in the communicating space in the retail areas. This space will be similar to a balcony created by the floor above. This scenario assumes a fire in the retail area of the communicating space on the Ground Level with a ceiling directly above. Based on the ceiling height of 5 m, we anticipate that the sprinklers will control or suppress the fire. For this scenario, it is assumed that the fire will grow until the calculated sprinkler activation time. After the sprinklers activate, we have conservatively assumed that the fire will be maintained at steady state conditions. This potential fire will result in smoke spilling out from under the balcony created by the floor above and into the open atrium space. This scenario relates most closely to the balcony spill plume case described in Section 6.2.2 of NFPA 92B.

DESIGN FIRE The fire will be located in the retail area on the ground level. The following assumptions have been made to determine the sprinkler activation time: 1. The initial room temperature is 23 oC; This is based on the mechanical engineer input for HVAC building design. 2. The spacing of the sprinkler is 4.6 m; therefore the distance from a detector to the center axis of the fire is 3.23 m; See Figure 26.
Figure 26 - Sprinkler (Heat Detector) spacing.

Sprinklers Fire

3. The sprinkler response time index (RTI) is 90 (m/s)1/2, which is the maximum permitted for standard response sprinklers in retail areas. This is conservative. 4. The sprinkler activation temperature is 74 oC (165 oF);
2 5. The analysis assumes a medium growing t2 fire ( = 0.01172 kW/sec ).

6. The ceiling height is approximately 5 m for the ground floor as shown in Figure 27 below.

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE

I49831 Page 33 12 August 2010

Figure 27 -Atrium Section showing height of ceiling at ground level

5m

Based on these assumptions, the sprinkler activation time is estimated using calculations provided in the SFPE Handbook Third Edition Section 4 of Chapter 1 Heat Detection Power-Law Fires. The sprinkler activation time is calculated to be 320 seconds as shown in the following section below. The maximum heat release rate based on the sprinkler activation time can be estimated by:

Q = t 2 = 0.01172 320 2 1,200 kW = 1.2 MW

(NFPA 92B Appendix Section C.1)

Therefore, the peak HRR of 1.2 MW will occur approximately 320 seconds after the start of the simulation as part of the sprinkler activation analysis as indicated. The incipient stage is relevant in this analysis because it could influence the detection time. Detection prior to the onset of severe conditions can be a valuable asset as part of the life safety performance of the building. The final data point is at 1,200 seconds (20 minutes) to reflect the duration of smoke control required by the building code. In addition to specifying the heat release rate curve of the design fire in FDS, combustion properties related to the materials involved with the design fire must also be assigned. The two primary values that impact the amount of smoke produced are the soot yield, which is the mass of smoke or soot produced for a given mass of fuel consumed, and the heat of combustion. The heat of combustion is critical, as it relates to the amount of fuel that needs to be burned or consumed to produce a given heat release rate. The fire geometry has been taken to be 0.5 x 0.5 x 0.5 cubic meters similar to the axisymmetric scenario with the upper surface simulated as the burner while the other sides being inert.

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE
Figure 28 - Design Fire Size Curve for a Balcony Spill Plume

I49831 Page 34 12 August 2010

For this scenario, the materials involved in a potential fire are assumed to be a combination of typical materials that might be found in this type of atrium. To provide a conservative fire representation with respect to soot production, which has a direct impact on visibility through the smoke, the materials for the design fires are assumed to be a relatively equal mixture of cellulosic materials (wood, paper, cardboard) and synthetic materials (plastics, rubber, fabrics, polyurethane foams). Cellulosic fires typically have low soot yields while synthetic materials typically have high soot yields. Based on the soot yield values and heat of combustion given by Tewarson 6 and average soot yield of 0.05 g/g (grams of soot produced per gram of fuel burned) and an average heat of combustion of 18,250 kJ/kg are used in the FDS model.

Archibald Tewarson, Generation of Heat and Chemical Compounds in Fires, SFPE Handbook, 3rd Edition, Section 3 Chapter 4, pp. 3-111.

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE

I49831 Page 35 12 August 2010

SPRINKLER ACTIVATION The time to automatic sprinkler activation is simulated as part of the FDS model. The sprinklers have been replaced with heat detectors with the same properties as the sprinkler bulb. The activation temperatures of the bulb have been set to be 74C. Although heat detector activation times may be accurate, suppression algorithms in FDS are not relied upon as a simulation of actual conditions without validation studies particular to each project. The 0.5 meter grid is less effective for such detailed analysis since the heat detectors in the model are physically located directly at the ceiling. Detector spacing can be physically input as 3.5 meters and not as 3.6 meters as specified by code. Despite the limitations, the resulting detector activation time was modeled using conservative assumptions to be about 320 seconds. The heat release of the design fire is controlled to a steady rate not exceeding 1.2 MW after heat detectors (i.e. sprinkler) activation. The equations that were used to calculate the sprinkler activation time are shown below. This value was compared against the heat detector (sprinkler) activation time from FDS. The calculated value of 320 seconds thus is more stringent and therefore has been used.

c = c
r * t 2 f = 0.813 1 + H
* u2 =

Fire Intensity convective coefficient Reduced arrival time of heat front at a point r/H

u
1

Reduced gas velocity A 5 1 5 H 15 c 3 2 T 2 T = A 5 a c 5 H 5 Change in reduced gas temperature g T2*


g c p Ta
r D = 0.126 + 0.210 H

A=

The calculation was performed using an iterative process to calculate the activation time. For each time step, the following equations were performed until the chosen temperature is reached. The result is a sprinkler activation time of 246 seconds as well as the corresponding heat release rate of the prescribed t-squared time of 704 kW. A 30% uncertainty was included which is mainly used in case of high ceilings. The total activation time was found to be 320 seconds with a HRR of 1200 kW.
* t2 =

t A
1 5

* * t2 t2 f T = D * 2

Non-dimensional time for heat front to reach detector


5

Non-dimensional change in smoke detector temperature

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE

I49831 Page 36 12 August 2010

Td =

1 e Y T T2* 1 T2* Y
* u2 * 2

Change in detector temperature where Y is equal to:

Y=

3 u * 4 u2

(T )

T * t 2 * D RTI t 2

Q = t2

Heat Release Rate

SMOKE DETECTOR ACTIVATION The smoke detectors have been included in the model to simulate the time at which the smoke control system including exhaust vents and make-up air will activate. The smoke detectors have been placed at a distance 5 m above the ground level and on the ceiling with a default obscuration setting of 3.28%/m as provided in FDS. Smoke was detected at 40 seconds based on the FDS model. A 60 seconds additional time was added to compensate for fire alarm signal processing. The total time was calculated to be 100 seconds. This estimate is conservative considering a smoke detector is installed directly above the fire location.
Figure 29 - Smoke Detector Obscuration Levels

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE

I49831 Page 37 12 August 2010

MODELING RESULTS As discussed in the section titled Standard of Performance, the tenability criteria defined for this scenario are; visibility >10 m for the open atrium space (> 5 m is also acceptable for retail shops), air temperature <60oC, and CO concentration <500 ppm, evaluated 2 m above the walking surface on each occupied floor. The worst case tenability criterion for this atrium space is the visibility, due to the the large atrium volume configuration that results in significant air entrainment. The air temperature and CO concentrations remained an order of magnitude below the defined tenability criteria value in the occupied walking or occupiable space within the atrium for the egress duration of 20 min. Heat Release Rate The actual HRR determined by the model is plotted in Figure 30. The HRR from the FDS model simulation tracked almost identical to the input. The FDS simulation models the combustion process during the simulation, and numerous factors may influence the actual HRR such as air entrainment, insufficient oxygen, re-radiation, etc. Differences of 100-200 KW are usually not significant.
Figure 30 - Heat release rate of design fire as input into and output by the FDS model.

Visibility The vertical slice files show that smoke spreads to the back end of the retail corridor space on the ground level at various times resulting in low visibility conditions, however, this low visibility condition is limited to the back end of the corridor. Moreover, due to the long corridor leading to the atrium opening in the retail on the ground level along with the shops located in the middle of the corridor acting like

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE

I49831 Page 38 12 August 2010

barriers, the travel of smoke through the corridor is interrupted which results in the smoke descending and thus creating low visibility condition. The back end of the corridor creates a dead-end for smoke travel along with make-up air from the diffusers pushing the smoke downwards, creates low visibility. It is considered that this local failure within such a short time will not affect occupants evacuation given the proximity of the exits found within the mall. Additionally it is anticipated that automatic sprinklers will limit the smoke and fire to the shop of fire origin and smoke will not spread significantly beyond the shops of fire origin. Therefore, the model is considered to be capable of maintaining tenable conditions for the duration of occupants evacuation. Figure 31 and Figure 32 illustrate the visibility conditions within the atrium at 1200 seconds in the analysis. The output is shown using horizontal and vertical slice files passing at 2 m above the highest walking space of the ground level and through the fire respectively. As shown in the figures below, the visibility is below 10 m in the back end of the retail corridors. This can be justified for a maximum of 5 m to be used as described in the Standard of Performance by Lougheed 7 who summarized recommendations of the visibility criterion as follows: 3-5 m for occupants familiar with a building and able to evacuate quickly. The shops are assumed be to be small spaces with 1 or 2 exits where people are able to evacuate easily and therefore 5 m visibility would be an acceptable criteria.
Figure 31 - FDS modeling results for visibility for a balcony spill plume located at 2 m above ground level at 1200 seconds. The colored contours show areas where the visibility is 5-7 m. These areas include shops where occupants can egress in a short time.

Areas with low visibility

G.D. Lougheed, Basic Principles of Smoke Management for Atriums, Construction Technology Update No.47, Published by Institute for Research in Construction.

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE

I49831 Page 39 12 August 2010

Figure 32 - FDS modeling results for visibility within the Atrium for a balcony spill plume fire located at the ground floor. Figure below shows horizontal slice files through the open atrium at 2 meters above ground level at 1200 seconds. The colored contours show areas where the visibility is 10 meters or less,

2m above ground floor

Temperature Figure 33 illustrates the air temperature output slice files. The outputs are shown using vertical and horizontal slice files. The results show that the air temperatures in the communicating spaces open to the atrium do not exceed the defined tenability criterion.
Figure 33 - FDS modeling result for air temperatures within Atrium at 1200 seconds. The figures show the horizontal slice file at 2 m above the ground level. Temperatures are shown to be around 36 oC which is acceptable. Temperature scale is set between 0 and 60 oC.

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE

I49831 Page 40 12 August 2010

Carbon monoxide (CO) Concentration Figure 34 illustrates the CO concentration output slice files. The outputs are shown using vertical slice files. The results show that the CO concentrations in the communicating spaces open to the atrium are less than the defined tenability criteria.
Figure 34 - FDS modeling results for CO concentration within the atrium 2 m above level 2 at 1200 seconds. The figure shows the horizontal slice file through the open atrium space. The concentration scale is set between 0 and 500 ppm (i.e. between 0 and 5x10-4 mol/mol)

SCENARIO 2 CONCLUSION Based on the above analysis, the following conclusions can be made regarding the performance of the building under the Balcony Spill Plum Fire Scenario.

The smoke detector system should activate in this scenario in about 100 seconds including fan delay configuration. The temperature and carbon monoxide levels do not come close to threatening tenability in the atrium. Visibility on the Ground level within the atrium approaches the limits of tenability at any time during the 20 minute simulation. The visibility is between 5-10m however per Lougheeds 8 research, a 5 m visibility limit would be acceptable for the shops given that occupants can easily egress from that space.
8

G.D. Lougheed, Basic Principles of Smoke Management for Atriums, Construction Technology Update No.47, Published by Institute for Research in Construction.

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE

I49831 Page 41 12 August 2010

For Smoke Exhaust, consideration should be taken to increase the areas of the openings leading to exhaust as well as increasing the velocities at these openings to ensure that smoke movement and flow is being exhausted and not becoming stagnant and descending. The system shall be tested and verified to ensure proper functionality. Given the results of the analysis for this scenario, the performance of the smoke control system is acceptable.
Table 5 - Summary of Balcony Spill Plume Scenario

ITEM Scenario Design Fire Fire Size Fire Curve Fire location Smoke Zones Smoke Exhaust Rate Make-up Air (Mechanical & Natural) Smoke Detector Activation Heat Detector/Sprinkler Activation Smoke Exhaust Velocity Make-up Air Velocity Visibility Temperature CO Concentration

DESCRIPTION 2 Balcony Spill Plume (sprinkler controlled fire) 1.2 MW Steady State occurs after 320 seconds at 1.2 MW Fire located at the Ground Level near the retail communicating corridor. The geometry of the fuel is taken to be 0.5x0.5x0.5 meters. 1 zone for the entire mall with no fire shutters. 125 m3/sec for ten vents located at the ceiling of Level 2 75 m3/sec for 27 vents located at the ceiling of the Ground Level 40 seconds activation with 3.28%/m obscuration in addition to 60 seconds delay for startup of Exhaust Vents and Make-up air diffusers 320 seconds; assumed steady state conditions met thereafter. 3 3.5 m/sec 1 1.2 m/sec > 5 m; passed but proper design changes should be considered for compliance in all areas. < 60 deg C; passed < 500 ppm; passed

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE

I49831 Page 42 12 August 2010

SMOKE EXHAUST RATE AND MAKE-UP AIR RESULTS The smoke exhaust rate is selected to be 264,800 CFM (125 m3/s) based on the evaluation of modeling results with varying exhaust rates. The model was first tested with a 100 m3/sec exhaust rate for both axisymmetric and balcony spill plume scenarios, however it was determined that low visibility originated around the retail area of the level 2 for axisymmetric scenario (< 10 m visibility) as well as for the balcony spill plume at ground level near the back end of the retail space (< 5m visibility). The 100 m3/sec scenario almost passed for an axisymmetric fire and failed for a balcony spill plume. Thus a higher smoke extract rate (125 m3/s) was selected. The anticipated activation times for the smoke control system for the two fire scenarios is as follows:

For an Axisymmetric Fire: 45 seconds upon the detection of the smoke by beam detectors and a delay for fan configuration of 60 seconds. The total is 105 seconds for the activation of the smoke control system. For a Balcony Spill Plume: 40 seconds upon the detection of the smoke by smoke detectors and a delay for fan configuration of 60 seconds. The Total is 100 seconds for the activation of the smoke control system.
The velocity of the smoke exhaust at the vents was found to be in the range of 3 to 3.5 m/sec. As shown in below in Figure 35, the velocities of >3 m/sec in the middle represent the smoke generated by the 5.3 MW axisymmetric fire on the ground level. The smoke is shown traveling up and spreading radially across the ceiling, thus creating high velocity profiles along the ceiling. The smoke spread along ceiling and exhaust vents decrease the smoke velocity as it approaches the vent inlets. Therefore design changes should be taken as described in the previous section.
Figure 35 Velocity vectors of Smoke at 1200 seconds for an Axisymmetric Fire. The velocity scale has been set between 0 and 3 m/sec.

Level 2

Velocity Vectors Exhaust Vents

Level 1

Ground

Fire Load

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE

I49831 Page 43 12 August 2010

Figure 36 - Velocity vectors of Smoke at the Inlet Vents at 1200 seconds for an Axisymmetric Fire. The velocity scale has been set between 0 and 3 m/sec.

Exhaust vents

25 cm opening with Low Velocities are seen. Consideration should be taken to increase these velocities and provided more smoke flow and movement through th i

Figure 37 - Velocity of Smoke at the Inlet Vents at 1200 seconds for a Balcony Spill Plume. The velocity scale has been set between 0 and 3.5 m/sec.

Exhaust Vents

High velocity profiles under the beam

Beam

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE

I49831 Page 44 12 August 2010

The make-up air rate (both mechanical and natural) is anticipated to be 75 m3/sec through the diffusers located at the ground floor ceiling and the sliding door that opens to the exterior. It is assumed that each of the 27 diffusers will produce close to the same volume rate of 2.67 m3/sec. Figure 38 and Figure 39 illustrate the make-up air velocity variations for a balcony spill plume scenario. The air velocity vectors are seen to be in the range of 2-2.5 m/sec however these velocity profiles result from the impact of the fire induced flow on the smoke and make-up air entrainment given the proximity of the fire. The make-up air velocity vectors are shown to be in the range of 1-1.2 m/sec depending on the size of the diffuser provided.
Figure 38 - Make-up Air Velocity at 1200 seconds for Balcony Spill Plume. The velocity scale has been set between 0 and 5 m/sec.

Velocity Vectors make-up air Velocity Vectors of smoke Fire Load

Figure 39 - Smoke Spread on the ground level.

Make-up air diffusers

Shops

Fire Load

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE

I49831 Page 45 12 August 2010

Figure 40 - Make-up Air Velocity at 1200 seconds for Balcony Spill Plume. The velocity scale has been set between 0 and 2 m/sec.

Air Velocity Profiles due to impact of smoke and make-up air

Make-up air diffusers

Fire Load

DESIGN CRITERIA TO BE CONSIDERED The following design criteria should be considered by the Design Team and Contractors to ensure that the smoke control system can function properly in accordance with the input and output considerations that were assumed in the model. Exhaust Vents. The 6 cm openings around each of the roof levels and the 10 cm openings around the ceiling voids at Level 2 were not included in the model. This is due to the cell resolution level that cannot be established given the models sheer size in FDS. Thus a 25 cm opening was assumed as a substitute to be an average of the 50 m2 total area of these openings. It is recommended however that the 6 cm and 10 cm function properly to allow smoke through and not debilitate the smoke flow. The model as seen at 600 seconds in Figure 41 showed signs of smoke accumulation and descent of the smoke layer due to the small openings provided around the perimeter of each roof level. Although 25 cm openings were incorporated in the model, the 6 cm and 10 cm could imbalance the system and produce more accumulation and smoke descent if not properly maintained. Currently the exhaust vents will likely have a 3 m/sec average velocity. Exhausting from the closed space on the roof will result in higher air velocities around the 6 cm and 10 cm perimeter openings. Consideration should be taken into account for the velocity of these openings.

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE

I49831 Page 46 12 August 2010

Figure 41 Atrium section showing smoke accumulation and descent at 600 seconds balcony spill plume scenario.

Level 2 Level 1 Ground

Smoke accumulation and descend

Figure 42 - Smoke accumulation and descend due to small opening areas leading to exhaust and low velocities through these openings for axisymmetric fire scenario.

Make-up Air Diffusers. Although the original design includes diffuser provided vertically, the model currently shows diffusers provided horizontally parallel to the ceiling as shown in Figure 43. This difference is due to the limitation on cell resolution which can be input into FDS given the sheer size of the project and number of cells used. In terms of functionality, the difference between the two is the direction of air that is introduced. Although the FDS model provides make-up air directed downwards towards the fire and the rise of the smoke, this is considered to be much more stringent than having make-up air directed horizontally with the flow of the smoke along the ceiling as provided in the original design. Therefore the design criteria that need to be considered as per the original design does not affect the results, however it is important to note that the velocity of the make-up air should be in the range given at the openings of the diffusers.

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE

I49831 Page 47 12 August 2010

Figure 43 - Comparison between what was provided in the original design and what was used in the model to compensate for level of details that provides challenges to compute in FDS.

Smoke Control System Activation. The smoke control system is intended to be activated by beam detectors, smoke detectors throughout the mall areas or by sprinklers. The alarm system should be designed and installed in accordance with NFPA 72. Fans. Fans shall have 1.5 times the number of belts required for the design duty with the minimum number of belts being two. Motor driving fans shall not be operated beyond their nameplate horsepower (kilowatts) as determined from measurement of actual current draw, and shall have a minimum service factor of 1.15 per IBC Section 909.10.5 Ducts. Duct materials and joints shall be capable of withstanding the probable temperatures and pressures to which they are exposed. Ducts shall be lead tested to 1.5 times the maximum design pressure. The measured leakage shall not exceed 5 percent of the design flow. Ducts shall be supported directly from fire resistance rated structural elements fo the building by substantial, non-combustible supports per IBC, Section 909.10.2. Power System. The smoke control system shall be supplied with two sources of power. Primary power shall be from the normal building power systems, Secondary power shall be from an approved standby source complying with IBC Chapter 27. The standby power source and its transfer switches shall be in a room separate from the normal power transformers and switch gears and ventilated directly to and from the exterior. The room shall be enclosed with not less than 1-hr fire barriers. Detection and Control systems. Fire detection systems providing control input or output signals to mechanical smoke control systems or elements thereof shall be equipped with a control unit complying with UL 864 and listed as smoke control equipment. Control systems for mechanical smoke control system shall include provisions for verifications. Verification shall include positive confirmation of actuation, testing, manual override and the presence of power downstream of all disconnects and through a preprogrammed weekly test sequence, report abnormal conditions audibly, visually and by printed report per IBC, Section 909.12.

SMOKE CONTROL MODELING ANALYSIS REPORT SAHARA MALL EXPANSION SHARJAH, UAE

I49831 Page 48 12 August 2010

CONCLUSION This document outlines the results of the smoke control modeling analysis for the atrium space of the Sahara Mall Expansion. This analysis has been completed by RJA on behalf of Khatib & Alami to evaluate the acceptability of the proposed smoke control strategies, including exhaust rates, locations of operable door openings to introduce natural make-up air for the modeled atrium spaces, exhaust vents and make-up air diffusers. The performance criteria establish pass-fail limits for which the design must maintain compliance. For the evaluated atrium space the modeling results demonstrate that the proposed smoke control strategies described in this report are capable of maintaining tenable environments during occupants evacuation based on the design fire size and fire location. For the evaluated communicating space in the retail section, the modeling results have shown that the proposed smoke control strategy including 125 m3/sec exhaust rate is generally able to maintain a tenable environment during egress. It is suggested that draft curtains be provided for for better results in containing the smoke and limiting smoke spread into the shops as well as maintaining the vent openings around the ceiling with uniform velocities and exhaust rates such that there will be no smoke accumulation and smoke layer descent which could imbalance the system. Sincerely, ROLF JENSEN & ASSOCIATES, Inc. Prepared by: Bassel M. Mehio, LEED AP Associate
BMM/SRS:bmm

Reviewed by: Shamim Rashid-Sumar, P.E., LEED AP Director of Middle East Operations

2010 Rolf Jensen & Associates, Inc. All Rights Reserved

Você também pode gostar