| This dissertation describes improvement and extension of the recently developed first version of the Urban Fire Simulation (UFS1) model; and application and validation of the resulting second version, UFS2. The modifications made from UFS1 are: (1) the incorporation of a fire department suppression module; (2) the addition of spread by surface vegetation, through external walls, and down through floors; and (3) a few more minor modifications. With these additions, the UFS2 offers a complete, integrated urban fire model that includes ignition (for post-earthquake applications), spread, and suppression, and can be applied to both post-earthquake situations and the urban part of wildland urban interface fires. The UFS2 can be used to: (1) improve estimation of fire damage associated with a specific earthquake scenario or set of input ignitions in an urban or wildland urban interface area, (2) provide insights into factors that affect urban fire spread and suppression, (3) examine the effect of risk reduction measures, and (4) help identify future areas of study.;A simulation model's usefulness depends on its credibility, which can be difficult to establish. This dissertation presents two efforts to build the credibility of the UFS2 model---comparing it to the widely used Hamada model and the 2007 Grass Valley, California fire. The UFS2 model produces results that are similar in character and magnitude to those from the Hamada model for simplified conditions, and in the Grass Valley case, that compare well to observations in terms of timing, spatial pattern, and modes of spread. The comparisons also highlight the challenges in validation of urban fire models and suggest some areas for future study.;A case study application of the integrated UFS2 for a portion of Los Angeles, California is described. In addition, a parametric study of key factors to examine their roles in fire spread and interactions among them is also conducted. The results suggest that urban fire spread is highly variable and under the right combination of unlucky but possible circumstances---many ignitions, high wind speeds, and limited water availability---the losses can be very high, much higher than observed in recent earthquakes. In addition to the three factors mentioned, the locations of the ignitions (relative to wind direction and fire breaks), number of engines, and engine arrival times are shown to be important. Strong interactions are evident between wind speed and number of ignitions, and between water availability and number of engines. |
