The high predictive performance of fire detector models is essentially required for the reliable design of evacuation safety using the fire modeling. The main objective of the present study is to measure input information in order to predict the accurate activation time of photoelectric smoke detector adopted in fire dynamics simulator (FDS) recognized a representative fire model. To end this, the fire detector evaluator (FDE) which could be measured the device properties of detector was used, and the input information of Heskestad and Cleary's models was obtained for a spot-type photoelectric smoke detector. In addition, the activation times of smoke detector predicted using default values into FDS and measured values in the present study were quantitatively compared. As a result, the Heskestad model could result in an inaccurate the activation time of photoelectric smoke detector compared to the Cleary model. In addition, there was a distinct difference between the default values used into FDS and the measured values in terms of device properties of smoke detector, and thus the activation time also showed a significant difference.

The high predictive performance of fire detector models is essentially needed to assure the reliability of fire and evacuation modeling in the process of Performance-Based fire safety Design (PBD). The main objective of the present study is to measure input information in order to predictive the accurate activation time of fixed temperature heat detectors adopted in Fire Dynamics Simulator (FDS) as a representative fire model. To end this, Fire Detector Evaluator (FDE) which could be measured the device properties of detector was used, and the spot-type fixed temperature heat detectors of two thermistor types and one bimetal type were considered as research objectors. Activation temperature and Response Time Index (RTI) of detectors required for the fire modeling were measured, and then the RTI was measured for ceiling jet flow and vertical jet flow in consideration of the install location of detectors. The results of fire modeling using measured device properties were compared and validated with the experimental results of full-scale compartment fires. It was confirmed that, in result, the numerically predicted activation time of detector showed reasonable agreement with the measured activation time.

This is the first fire in Japan of a large wooden gymnasium with a large-scale glue-laminated timber structure which was totally burnt with a rapid flashover (FO). The fire deserves close attention for not only rapid FO in a large enclosure but also availability of extremely good documentation on the fire. The extent of carbonization of the members that was collected when the building was dismantled and its structural property were measured. A series of burning tests of the combustibles stored in the source of the fire, conducted to clarify why the fire spread so rapidly in the gymnasium where a FO had been least anticipated and to make guidelines to prevent fire from spreading rapidly.

Investigation ofthe properties of live fire load is conducted on ］ arge book stores from the interest on performance based fire safety design ． The study fecuses not only t翫e rcport of fire load， surface area to mass ratiQ but also dじvelopment of a simpl 面 cd site investigation methedology of 員re load of book stores where extremo difficulty is anticipated fbr the few yearly hoiidays available fbr such inves【igation ． In this repo 「【 ， 1h 巳 characteristics oflive fire Ioad were grasped fbr fire safety design ofl ロrge book stores and a methodology f［｝rthe measurement of fire load was developed to reduce the effbr 【 of investigation ． This proposed method was essentially developed from the distinction of the charact じ ristics of live fire load［ ike the arrangement and shape of book cases ． The results reveal thaUhe fire load is high 且 y dependent on the height and material ofbookcases ， This paper also indica 【 es thanhe prescrib じd total heal release for book store in bui］ dingcode standards ofJapan ， 960MJ1 ， is not on the dangerous side estimation because the investiga 宦 ion has resulted in total heat release over twice largor ， and the sur 魚 ce ar じa 卩 er unit weigh し rati 〔 ｝ is smaller than 1／ 4 the previous repor に s on o 吐 hcr occupancies ． These suggest that a fire in a large bQok s ！ ore may proceed generally notably more slowly and Iast longer than in other building occupancies if not ex 【 inguished ． The proposed simplified investigation mcthod is fc ) und to reduc ¢ manpower necessary fbr the site investigation by 80％ and enable si【 e investiga1ion in short time befbre or after the business hours ．