View non-flash version
probability of event occurrence (risk). For our purposes here, the focus is limited to assessing the potential of an initiation re to ignite interior nish materials, the abil- ity of those materials to propagate ame spread once ignition occurs, and the contri- bution of the interior materials to the total heat release rate (vulnerability and conse- quence assessment). For our example here, the source fire consists of several bags of trash on the floor against a wall of the vehicle. Such a fuel load is representative of a mixture of combustible materials contained within a relatively small fixed volume, as may be typical of items carried on to a pas- senger rail vehicle. This is modeled as a fire HRR of approximately 500 kW, which results in an incident heat flux to the wall of 50 kW/m 2 over a rectangular area of 1 m in height and 0.6 m in width. In this scenario, the wall is considered to be the initial material burning. Flame spread to additional surfaces (ceiling, floor, and seats) is incorporated through the use of CFD analysis. Material ammability characteristics Once the source re has been character- ized, the next step is to determine whether the interior lining materials are vulnerable to ame propagation and re development. Material properties can be obtained from cone calorimeter tests. e data can then be used as part of a rst order go/no-go screening tool that uses the b-parameter to help evaluate whether di erent vehicle lining materials are likely to support ame spread at vari- ous heat ux levels. b = 0.01 E? ? 1 ? Where, E?= average heat release rate per unit area (HRRPUA) (kW/m) tig = time to ignition (s) tbo = the burnout time (s) For any representative heat ux, if the b-parameter is greater than zero, the ame is predicted to accelerate and spread; but if the b-parameter is less than zero, the ame is predicted to decelerate and eventually extinguish itself. Table 1 shows average HRRPUA values for GRP A and GRP B at cone incident heat uxes of 15, 20, 50, and 75 kW/m with the respective b-parameters. GRP B has posi- tive b-parameter values for all heat flux levels, indicating that it has a tendency to spread ame. is is especially true given heat ux values in excess of 20 kW/m 2. GRP A exhibits negative b-parameters given a heat ux of 20 kW/m 2 or less. Even at a heat flux level of 75 kW/m 2, the b-parameter calculated for GRP A is less than that cal- culated for GRP B given a much lower heat ux (20kW/m 2).Flame spread tool If the b-parameter screening analysis indi- cates that a material has a propensity to propagate a me spread, additional vulner- ability assessment is warranted. To support this, a simplified upward flame spread model was developed to represent the ini- tial ame spread on vehicle lining materials. e model uses only parameters measured in the cone calorimeter. The initial source fire imposes a heat ux on the wall, q?e. If the heat ux is suf- cient to ignite the wall material, the ame extends up the wall, in turn emitting a ame heat ux, q?f to the virgin fuel above. Figure 2 illustrates characteristics of upward ame spread. e overall ame height, xf, results from the burning of the wall material. e pyrolysis height is represented by xp. When the fuel is considered spent or used up, it can no longer support a ame and a burnout front develops, indicated by x b. Heat release rates are calculated from the spread results to better understand the re hazards. As a simplifying assumption, the pre-heat and spread areas were based on expected burn patterns. In this case, assuming a source fire located directly www.sname.org/sname/mt July 2012 Material Heat ux (kW/m 2)Average HRRPUA (kW/m 2)b-parameter GRP A1592.1 -0.60 2098.4 -0.28 50145 0.31 75158 0.50 GRP B15166 0.20 20206 0.64 50272 1.60 75236 1.33 TABLE 1: CALCULATED B-PARAMETER VALUES FIGURE 2: UPWARD FLAME SPREAD MODEL FIRE SPREAD tigtbo