A numerical procedure for estimating an area-wide post-earthquake fire loss of an unspecified group of buildings is proposed. The procedure consists of three parts, namely the fire behavior simulation using the two-layer zone model, the evacuation simulation of occupants using the network-based model and probabilistic estimation of fire extinguishment by fire-fighting. The sub-procedures considers the damage of fire safety equipment systems (sprinkler system, smoke exhaust system, fire alarm system, indoor fire hydrants, fire extinguishers, and fire doors) using the event tree after seismic shaking in order to consider their availability following an earthquake. The composition data of each building (floor area of rooms, number of partitions, connection type of rooms), which are generally unknown but prerequisite of the fire loss analysis, are estimated using a statistical model obtained by analyzing floor plans of 300 existing buildings of commercial use. Fire loss analysis of an individual building for the validation of the building composition estimation is conducted. The result shows that the proposed procedure captures the overall features of the human loss and smoke spread area although it may involve a certain error. As a case study, post-earthquake fire losses in three commercial areas are analyzed under a JMA (Japan Meteorological Agency) seismic intensity of 6-upper. The result shows that the total floor area can be used as a parameter generalizing the expected loss caused by a fire.

In performance-based structural fire safety design, various calculation models are used to predict the structural behaviors of building elements exposed to fire. Although many of such calculation models require material properties of building elements as input data, it is often technically and economically difficult to obtain such material properties. A potential alternative is to develop a method to translate the fire resistance time of a building element assessed by existing fire furnace testing to the time for which the element can endure in arbitrary compartment fire conditions.

In this paper, we checked and explored the possibility of such a translation method of fire resistance time between fire furnace tests and ventilation controlled compartment fires and tried to identify the data that are indispensable in performance-based structural fire safety design. It is found that ‘equal heat absorption’ hypothesis is not appropriate as the translation basis. Instead, we propose a practicable and direct translation method of fire resistance time at which the temperature becomes critical temperature as one of the fire resistance criteria. In this method, it is not necessary to know the properties of the building element tested although it is necessary to have the thermal property of compartment boundary.

In performance-based structural fire safety design, various calculation models are used to predict the structural behaviors of building elements exposed to fire. Although many of such calculation models require material properties of building elements as input data, it is often technically and economically difficult to obtain such material properties. A potential alternative is to develop a method to translate the fire resistance time of a building element assessed by existing fire furnace testing to the time for which the element can endure in arbitrary compartment fire conditions.

In this paper, we checked and explored the possibility of such a translation method of fire resistance time between fire furnace tests and ventilation controlled compartment fires. It is found that ventilation controlled compartment fires cannot be expressed by a closed form formula, e.g., MQH formula. Instead, we propose a practicable translation method of fire resistance time, in which it is not necessary to know the properties of the building element tested although it is necessary to have the thermal property of compartment boundary.