FRAME Study of the Sofa Super Store Fire - Charleston SC, June 18, 2007. A fire occurred on the evening of June 18, 2007, in the Sofa Super Store at 1807 Savannah Highway, Charleston, South Carolina. The fire swept from the rear to the front of the main showroom extremely quickly, and then into the west and east showrooms. Nine fire fighters from the Charleston Fire Department were killed in the fire.   This catastrophic has been the subject of an extensive research by NIST and the recently (October 2010) published draft report of this incident, (extracts of the report are shown here in italic) is a good reference for testing FRAME. The purpose of this FRAME study is to check if it could have indicated the potential of a catastrophic in this building, and to see what recommendations the FRAME user should have developed from the risk assessment. The fire. Based on the NIST technical study, the following sequence of events are likely to have occurred. From an unknown source, a fire ignited in a pile of packing material and discarded furniture outside an enclosed loading dock area. The fire spread from the exterior to the interior of the loading dock, which was used for staging furniture for delivery and for repair of furniture. The fire spread quickly within the loading dock and moved into both the retail showroom and warehouse spaces. During the early stages, the fire was unable to access enough oxygen (ventilation limited), which slowed its growth. Due to the lack of sufficient air to allow complete combustion, the fire produced large volumes of partially pyrolyzed fuel in the form of smoke and combustible gases. The large volumes of unburned fuel flowed into the interstitial space below the roof and above the drop ceiling of the main retail showroom. As the interstitial space filled with unburned fuel, the hot smoke also seeped through the drop ceiling into the main showroom and formed a hot smoke layer below the drop ceiling. Up to this time, the extent of fire spread into the interstitial space was not visible to fire fighters in the store.   The fire spread to the rear of the main showroom through the holding area and ignited additional fuel in the rear of the main showroom at which time it became more visible to fire fighters in the main showroom. The fire at the back of the main showroom was still ventilation limited and its growth was slowed by the lack of oxygen. As the fire burned in the rear of the main showroom, the fire pumped more hot unburned fuel into the smoke layer below the drop ceiling. The lack of oxygen prevented the unburned fuel in the smoke layer from igniting. When the front windows were broken out or vented, additional oxygen flowed in the front windows, along the floor, and to the rear of the showroom and became available to the fire. The additional oxygen allowed the heat release rate of the fire to increase extremely rapidly and ignite the layer of unburned fuel below the drop ceiling. The fire swept from the rear to the front of the main showroom extremely quickly, then into the west and east showrooms. Intense heat from sustained burning of furniture in the main showroom weakened the roof joists and supports and resulted in the collapse of a portion of the roof over the main showroom approximately 13 minutes after flames emerged from the front windows (40 minutes after the fire department arrived on scene). Furniture and merchandise in the showrooms and warehouse continued to burn for an additional 140 minutes before the fire was extinguished.   The extremely rapid spread of fire through the main and west showrooms trapped six fire fighters in the main showroom and three fire fighters in the west showroom. Although the intense heat from the fire weakened the light weight steel trusses and led to the partial collapse of the roof, the coroner's report indicated that the fire fighters died from thermal burns and/or smoke inhalation, not from compression type injuries that would have been associated with the collapse. Description of the building and its use. The NIST report includes a very detailed description of the building and its use, which gives us the opportunity to make a rather precise FRAME calculation. Construction, compartimentation and immobile fire load. Main showroom.   The description of the 1585 m² main showroom allows us to define the dimensions of the showroom, the type of construction and the immobile fire load, and the fire resistance of the building elements. Basically it is a non combustible construction with a small amount of combustible materials. For the steel construction, FRAME recommends a maximum fs=15 min, the same value can be used for the roof construction (fd=15 min) The façade made of metal sheet with windows in the front shall be considered as without any fire rating (fw= 0). The height below the roof is h= 4.3 m. As the drop ceiling was not fire resistant, the gypsum board interior walls connecting to the roof and the interior walls do not constitute a subdivision. West- and east showrooms.   The 652 m² west and 645 m² east showroom have basically the same characteristics as the main showroom.   The concrete walls with the roll-up fire doors could be considered as compartimentation walls, but the installation of fusible link operated fire doors with the fusible link close to the door opening and more than 1.5 m below the ceiling is not a well engineered design, which is proven by the fact that only three of the seven roll-up fire doors activated and closed fully during the fire.   In addition to this poor design, the report indicates that the doors should remain open to comply with the egress requirements : "However, if the paths through the fire door openings are assumed to be acceptable for egress due to their "normally open" status, and the space is treated as one large showroom, then the showroom areas are equipped with an adequate number of exits with adequate separation distance, adequate exit width, and exits that are within the required travel distance." Under these circumstances, the FRAME user should consider the space of the three showrooms as a belonging to the same compartment. There were no openings in the roof or upper third of the walls that can be taken into account for the ventilation factor v.   Holding area. The 15 m² holding area is connected to the showroom. The roll-up fire door located on the access through the breezeway to the warehouse could have been a fire barrier, but it was shunted by the loading dock. The loading dock. The loading had an area of 210 m², had a combustible construction and 9m² of fiberglass skylights. The main showroom was not fire- proof separated from the loading dock, it is part of the same compartment. Its construction will slightly increase the immobile fire load density of the compartment The lack of a fire proof separation between holding area, the loading dock area and the showrooms means that these areas have to be considered as belonging to the same compartment of 3107 m² . The warehouse and repair shops.  The warehouse was 36.9 m wide and 39.8 m deep for an area of 1470 m2. It had a non combustible, non fire rated construction (fs=15 min, fw= fd= ff= 0) .   The repair shops were communicating with the warehouse through non-fire stop doors. This means that the repair shops shall be considered as secondary activities for the warehouse. The repair shops had no outside doors, and one employee who was trapped inside it, was eventually rescued by the fire fighters who were able to create an opening in the metal wall and extract the trapped employee.   The loading dock area was adjacent both to the warehouse and the showrooms with sheet metal walls. The time line showed that the fire spread from the loading dock area into the warehouse and the showrooms almost simultaneously. This means that there was no fire separation between the warehouse and the showrooms, and that the whole site was in fact one fire compartment.   The NIST report indicated that at the time of the loading dock addition (1996-1998), upgrading of the fire resistance of these partition walls would be necessary to avoid that the warehouse and showrooms should be considered as a single compartment. Mobile fire load. The estimated total energy content for the showrooms and warehouse was up to 480 GJ and 840 GJ, respectively. The energy content of the loading dock area was estimated at 100 GJ, or approx. 480 MJ/m². Total energy content of the structure and contents was estimated at up to 1450 GJ. This corresponds with a fire load density of 150 MJ/m² for the showrooms and 570 MJ/m² for the warehouse.   The NIST report appears to use rather low values for the energy content of the furniture . The average energy content of a sofa, chair, mattress, dresser, table, rug is only 11.6 MJ/kg where the FPE handbook ed.3 app. C4 suggests a net value of 18 tot 21 MJ/kg for wood materials. But even when we increase the energy content values by some 50 % to be more in line with the FPE database, the real fire loads appear to be much less than the FRAME recommended values based on the sprinkler hazard classes (1500 MJ/m² and 7500 MJ/m² respectively). The FRAME user can use the exact fire load when it is available, and using the recommended values will result in a more severe risk assessment.   Horizontal geometry (factor g)  The theoretical length of the whole building can be found on the Store floor plan with dimensions and is app. 91 m. Total area is 4650 m² and the equivalent width is 51.1 m. The warehouse compartment alone is some 40 m long for an area of 1550 m², and the showrooms compartment is 75 m long for an area of 3100 m² Accessibility (factor z).  This description and the aerial photos show that the whole building was accessible from 3 sites. But when the warehouse and the showrooms are calculated as separate compartments, accessibility is limited to 2 sides only. Ventilation (sub factor k). The photos in the NIST report show that there were no openings in the roof and upper third of the walls of the showroom building. The 3 skylights of the loading dock area do not bring any change here.   The aerial photos of the warehouse show some lighting strips representing 1/18th or 5.5 % of the roof area. The after fire pictures indicate that these lighting strips were indeed destroyed by the fire and this can be considered as ventilation openings for the warehouse calculation.   Height floor to ceiling.  The showroom has a lower ceiling (4.3 m) than the warehouse (8.8 m). Instead of using the average height, it is more indicated in this case to use the lower ceiling height of the showrooms (4.3 m) for the calculation for the whole building, as this corresponds better with the fire situation. For the separate calculations, the ceiling heights of each compartment can be used   Occupant Load.  The design occupant load is not mentioned in the NIST report. NFPA 101 Life safety Code indicates an occupant load factor of 1 pers/2.8 m² for mercantile occupancies. This would apply to the showrooms. For the warehouse a much lower occupant load factor would apply. For furniture stores, the occupant load factor for mercantile occupancies is an overestimation, as the number of visitors per m² is usually lower than e.g. in a supermarket. As the fire did not occur during the opening hours, the exact occupant load is not an issue, but it is part of the risk assessment. The occupants are to be classed as visitors, i.e. mobile persons needing guidance. It is assumed that there is a clear evacuation plan an that there is no danger for panic. These are optimistic assumptions.   Exits.  There were two sets of double doors near the center of the front (north) wall, which were the only entrances on the front of the structure. The two front doors were flanked on either side by small and large glass display windows. There were no exits or windows on the rear wall of the main showroom. In addition to the six fire doors leading to the other showrooms, there was an additional fire door which was located in the rear southwest corner of the main showroom that connected to the warehouse. Also, in the rear southwest corner of the main showroom was a non-fire roll up door that provided access to the loading dock area.   On the NIST floor plan, a total of 12 outside doors are indicated, each door is taken as an exit unit for the FRAME calculation..   Content factor.  The content is easily replaceable and the total value is estimated to be about 5 million USD, with an inflation factor of 1.4 between 2007 and 2000. The dollar/ euro rate is estimated to be 0.75. These assumptions do not affect the result of the risk assessment. Dependency factor.  The average value for most businesses can be used (d=0.3) . For the warehouse alone the lower value for storage buildings can be used.   Water supply.  The water supply and the hydrant system are considered to be adequate. Fire control.  Based on the observations in the NIST report, the "normal protection" is substandard: no fire alarm (although required by NFPA 101) , inadequate number of extinguishers, no hose stations, and probably only a few persons that can use extinguishers (which they did).   Fire Department Classification.   The US organization of fire departments is somehow different of the European (British) type classification, which is used as the basis for FRAME. The description of the Charleston Fire Department in the report indicates that the responding fire stations can be classified as: Professional crewed station ( day time crewed, night time retained ).   Escape Protection. It is assumed that the exit paths were properly indicated and that emergency lighting was installed.   Salvage factor  The only organizational protection assumed is the possibility of a swift relocation of the activities. FRAME calculations. With the data found in the NIST report, a series of FRAME calculations can be made. As the building lacks proper compartimentation between warehouse and showrooms, the first calculation to make is for the building as a whole. It might be interesting to make also separate calculations for warehouse and showrooms, on the assumption that a correct compartimentation was available. This brings us to 3 possible FRAME calculations: a) The whole building of 4650 m² size b) the showroom annex loading dock of 3100 m² c) the warehouse annex repair shops of 1550 m² Risk assessment for the whole building. The software allows us to study 3 variants within the same calculation. For the whole building, the following 3 variants were calculated:   a) the reference case is the situation as it was on the day of the fire : 18 June 2007 b) with sprinkler protection as required by the applicable codes c) an alternative based on the analysis of the reference case with preventive measures, more exits and automatic fire detection. The summary sheet gives us the following results: Risk for: Reference Variant 1 Variant 2 Property R 4,85 0,97 1,23 Occupants R1 7,19 2,42 0,94 Activities R2 2,81 0,50 0,97 Summary Report  Potential Risk P 2.27 2.27 2.24 P1 1.26 1.26 1.24 P2 1.90 1.90 1.88 FRAME reports for P P-REF  P-V1 P-V2 Acceptable Risk A 0.67 0.67 1.20 A1 0.22 0.22 0.75 A2 0.85 0.85 1.10 FRAME reports for A A-REF A-V1 A-V2 Protection Level D 0.70 3.49 1.51 D1 0.79 2.36 1.76 D2 0.79 4.45 1.76 Frame reports for D D-REF D-V1 D-V2 As could be expected by the events, the reference case with the as-was situation results is a high and totally unacceptable risk level. Whole building with sprinklers.   The second calculation uses sprinkler protection (and fire alarm system) as required by the building codes referenced in the NIST report. The report indicates that the building codes might also require a separation of the paint repair shop. NIST does not refer to any NFPA code for life safety or good fire safety practice. No reason is given for this, maybe Charleston did not use these codes as reference. The FRAME calculation for this variant gives an interesting result: The property and activities risk level is OK, but the occupants risk is still considered as being high and unacceptable (R1=2.42). This can easily be explained as no preventive action is taken to improve the situation. In FRAME, the acceptable risk A considers the potential ignition sources and the evacuation time as components of the level of exposure. The value obtained A1=0.22, (and also A=0.67 and A2=0.85) in the reference case , means that there is a very high level of exposure, resulting in an unacceptable high overall risk level.   Adding sprinkler protection does not reduce the level of exposure, it limits the effect of a fire. A basic rule in risk management is that prevention comes before protection. This means that the first thing to do is to reduce the level of exposure.   The first way to this, is to isolate the secondary hazardous activities such as woodwork, paint shop and flammable liquid storage by housing them in a separate room with fire proof walls and doors. In the Sofa Super Store this could be done with a slightly different layout and fire doors for the repair shops and for a dedicated flammable liquid storage room. The result of these modifications can be seen in the second variant calculation as the a factor goes down from a=0.48 to a=0.20.   The second way to improve the egress capacity to reduce the exposure for the occupants. In the reference case, the calculated number of distinct exit paths is 1.5 , taking into account the maximum capacity of 120 person per exit unit. In reality this means that some people might be trapped inside the building. In the second variant , the number of exit units is increased from 12 to 18 (which can be done by adding exit doors) and the t-factor goes down from 0.48 to 0.20.   With both these preventive actions (at a limited cost) the overall risk level is already improved, but additional protection will still be necessary.   The building codes required sprinkler protection, which for a building like the Sofa Super Store is an expensive investment : In the showrooms sprinklers are to be installed above and below the false ceiling, and in the warehouse, the high rack storage demands a high capacity system. Its is not evident that the municipal water system can supply the required pressure and flow and a dedicated fire pump and water storage tank may be required. Because of these technical requirements, the investment in sprinkler protection may be economically unaffordable for a rather small business unit like the Charleston Sofa Super Store.   But what alternative could be recommended by a FRAME approach ? With the occupants safety in mind, the most obvious improvement is the installation of a combined manual and automatic fire detection and alarm system linked to an automatic call to the fire department. The basic protection shall be improved with the installation of an adequate number of fire extinguishers on the premises. In the second variant, these improvements result in higher value for the N, S and U factors and the overall risk level goes down to 1.23 for property, 0.94 for the occupants and 0.97 for the activities.   The addition of an adequate number of fire hose stations inside the building is the final step to obtain a good level of protection (R=1.01) for the property too. Risk assessment with compartimentation. According to the NIST report, the building codes impose a maximum area of 1161 m² for a non sprinklered building without allowance for frontage (i.e. easy access to the buildings' perimeter). Assuming that the original building (the main showroom of 1585 m²) was code compliant at the time of its construction, we can make a FRAME calculation to find the originally "built-in" fire safety level. To make this calculation, we have to make some assumptions: the fire load is that recommended for an OH1 type risk, the accessory activities are not present, the building is protected by a manual operated fire alarm system, fire extinguishers, the number of exit units is adequate for an occupant load of 1 person per 3 m² (mercantile occupancy). Note that in Europe, a hose system would also be required as standard protection, but such requirement cannot be found in US codes, e.g. NFPA 101. The results of this calculation are: R = 1.60, R1= 2.01 and R2= 1.18, which are still unacceptable high fire risk levels according to the FRAME philosophy. In this case, the orientation value Ro = 1.33 indicates that an automatic fire detection system would be required. The first variant includes this recommendation and the installation of a hose system inside the building. The area of the building has also been increased to include the east and west showrooms. The result is a satisfactory level of fire safety for the showrooms : R= 1.00 , R1= 0.56 R2= 0.70 .   For the second variant calculation, the same options and the data of the warehouse compartment were used. A fire separation with the loading dock and repair shops was assumed and the recommended high value for the fire load was also used as design data. The result is also a satisfactory level of fire safety for the warehouse: R=0.91 , R1=1.01 and R2= 0.40   Based on these FRAME calculations, the recommended fire protection concept for the Sofa Super Store should have been:  1. Provide a fire separation between the loading dock - repair shops- breezeway area and both the warehouse and the showrooms, e.g. with concrete block walls and fire doors in these walls protection the openings (re-use the existing fire doors).  2. Add some 50 % egress capacity for the showrooms (more exits to the outside).  3. Install an automatic fire detection and alarm system in the all areas with a link to the fire brigade, e.g. based on the NFPA 72 standard.  4. Provide an adequate number of fire extinguishers per NFPA 10 standard and an Class II fire hose system per NFPA 14. This recommendation is in line with the current practice in most European countries, but is not compliant with the US building codes and fire standards, where sprinkler protection is often foreseen as a "master key" solution, disregarding the cost of investment.   FRAME offers the possibility to combine preventive measure (hazard separation) with less expensive alternatives such as automatic fire detection for medium sized buildings to obtain an equivalent level of fire safety.   Summary of the alternative calculations with compartimentation.   Erik De Smet PRINT  THIS PAGE  (pdf)