Probabilistic Estimation of Fire Spreading Following an Earthquake Due to Gas Pipeline Damage

Abstract eng:
Fire following earthquakes has resulted in a considerable loss of property during the history. In San Francisco earthquake (1906) the estimated loss due to a widespread fire was about 250 million dollar of its time. The case of Tokyo earthquake of 1923 was much worse where the fire-induced loss shared 77% of the total. In recent times, Loma Prieta (1989), Northridge (1994) and Kobe (1995) earthquakes are new experiences in which ignition of fire was generally because of damages in old gas networks. It is therefore justifiable to concentrate on the performance of city gas networks, among other lifelines, when assessing vulnerability of urban areas to fire following an earthquake. From three main sources for ignition including gas leakage, damage of electrical equipments and leakage of flammable chemical substances, gas leakage is considered in this study. In fact a probabilistic model for the estimation of fire loss originated by gas pipeline damage following an earthquake can be developed using the proposed method. Linearity for correlation of loss growth on area of burned regions is assumed in the proposed model. In the so called loss model various damage scenarios and gas feeding states should be studied simultaneously. Developing the model consists of: a) assessment of damage scenarios of pipeline, b) modeling ignition of natural gas, and, c) modeling the spread of fire. New ideas are proposed in parts “a” and “b” but the method developed by Tosho is used in developing part “c”. The numerical simulation of the proposed model reads to the results which are time dependent curves illustrating how fire and the induced loss grows with time in each region after earthquake strikes. These curves can be used to determine the priority of urban areas for preventive or rescue plans. Finally, to demonstrate the applicability of the model, fire loss curves are obtained for two sample urban areas.

• Export as: BibTeX | MARC | MARCXML | DC | EndNote | NLM | RefWorks
• Add to your basket