The 14 th World Conference on Earthquake Engineering October 12-17, 2008, Beijing, China URBAN UTILITY IGNITION MODEL, A PROBABILISTIC APPROACH FOR MODELLING FIRE FOLLOWING EARTHQUAKE M.R. Zolfaghari 1 , E. Peyghaleh 2 , L. Golmoradi 3 and Gh. Nasirzadeh 3 1 Assistant Professor, Dept. of Civil Engineering, K. N. Toosi University of Technology, Tehran, Iran, 2 PhD candidate, Dept. of Civil Engineering, K. N. Toosi University of Technology, Tehran, Iran 3 MSc student, , Dept. of Civil Engineering, K. N. Toosi University of Technology, Tehran, Iran mzolfaghari@catrisks.com , e_peyghaleh@yahoo.com ABSTRACT: This paper presents the methodology proposed for modelling ignitions caused by urban utility network following earthquakes. In this model ignition following earthquake is modelled probabilistically, being dependent on several parameters such as strong ground motions, volume of utility network, type of gas pipeline, type of power network and urban building density. In order to consider uncertainties associated with the parameters controlling urban utility ignitions, an analytical approach using logic tree and Monte Carlo simulation process is proposed here. A GIS-based computer program has been also designed and developed in this work which can estimate ignition probability for urban utility network following earthquakes. As a pilot study, the utility network for a city district in northern Tehran, the capital city of Iran is modelled in this work. Preliminary results for a scenario earthquake as well as probabilistic earthquake scenarios are shown in this paper. KEYWORDS: Fire, Earthquake, Ignition, Urban, Utility, Probabilistic 1. INTRODUTION Experiences from past earthquakes showed that urban utilities are among city infrastructures damaged by earthquakes. These facilities are vulnerable to most seismic hazards such as fault rupturing, liquefaction, landslide and above all strong ground motions. Excessive damages to certain utilities can themselves pose secondary hazards, examples are flooding, electrification, environmental pollution, etc. Fire Following Earthquake is major threat for cities with high pressured natural gas distribution network or air-drawn electrical transfer and distribution network. During an earthquake such systems can cause ignitions and generate wide spread fire following big earthquakes. Flammable materials such as automobile fuels, trees and even road asphalt, once exposed to such ignitions, can turn into big fires. Fire Following Earthquake (FFE) consists of many simultaneous and catastrophic fires which could result in widespread economic damages and loss of life. Example of such historical cases is the 1906 San Francisco Earthquake with destructive consequences. More recent examples are the 1994 Northridge and the 1995 Kobe Earthquakes. Like other fire pattern, the FFE process consists of three main phases; ignition, spread and suppression. With regard to FFE hazard and risk modelling, most of researches in recent years have focused on modelling fire spread modelling and less attention is paid to modelling the sources of ignitions. Statistical correlations made between strong ground motion and ignition frequencies are mostly used as a mean to simulate this phase of an FFE model. Mizuno et al (1978) developed the first IFE models based on statistical analyses of FFE damage data from earthquakes in Japan. Scawthorn (1986) followed this approach and expanded this concept to develop probabilistic post-earthquake fire ignition and spreading model. Such models have been used to study the FFE pattern in jurisdictional scale and to estimate the aggregated economic FFE-related losses on regional scales. FFE damage data obtained from US earthquakes in twentieth century are used to model ignition mean rate as a function of seismic intensity and urban population density. Eidinger (2005) investigated the effects of gas