Search

Fire hazard studies of clean room facilities indicated that significant losses due to fire may occur in the semiconductor industry. The present study reports the results of full-scale wet bench fire tests conducted (1) to assess the fire hazards of existing wet bench materials not meeting the listing requirement of NFPA 318, (2) to assess the impact on wet bench fires of engineered materials with improved flammability characteristics, and (3) to compare the observed fire behavior with the results of simpler small- and intermediate-scale fire tests using the same materials. The full-scale wet bench fires were observed to be consistent in terms of chemical heat release rate, fire propagation, and smoke generation with the results of the small- and intermediate-scale test results. The simpler fire tests are incorporated in FM Approvals 4910 certification for materials to be used in semiconductor fabrication facilities. The small-scale test protocol is also standardized in NFPA 287. Copyright © 2008 John Wiley & Sons, Ltd.

The fire propagation apparatus (FPA) is the bench scale fire calorimeter that was recently described in its updated version in ASTM E 2058. The apparatus was originally developed in the USA by Tewarson and co-workers from the mid 1970s, under the name '50 kW lab-scale flammability apparatus', and is therefore still known in Europe as the 'Tewarson apparatus'. The paper focuses on the experience achieved so far with the first modern version of the apparatus implemented in Europe (France). Part I in this series of articles reports on the main results achieved during the commissioning period of the apparatus. In a first step, preliminary experiments were carried out in order to check and calibrate different sub-equipment of the calorimeter. The results are principally presented for the load cell system and the infrared heating system which are essential pieces of sub-equipment. In a second step, a set of fire tests using methane or acetone as fuel was carried out in order to check and calibrate the overall working of the calorimeter in well-fire conditions. The performance of the calorimeter was also checked when it operates in under-ventilated fires. Relevant testing procedures and potential technical problems are discussed. A set of recommendations are derived from the early learning obtained at the INERIS fire laboratory in order to check the consistency of the results obtained from bench-scale fire tests. These recommendations are thought to be applicable to all types of bench scale fire calorimeters.

The paper focuses on the detailed characterization of ventilation controlled fires of industrial products that are likely to govern accidental fire scenarios in fire resistant enclosures. Results regarding under-ventilated fires of substances that are not polymers are presented to illustrate the capability of the fire propagation apparatus (FPA) to qualify such types of fires. Based on results from heptane fire tests in both well- and under-ventilated fire conditions, a set of recommendations was previously provided in order to check the validity of the experimental results. The application of these recommendations is illustrated for the selected liquid substances containing hetero-atoms. It emerges that the fire propagation apparatus assesses quite easily the performance of well-controlled fires in both well- and under-ventilated conditions. Another major outcome of our work is that the potential of the FPA has the capability to address fire issues outside the conventional use of the equipment, in particular to qualify the burning behaviour of chemicals on the full spectrum of ventilation conditions.

The extent of flame propagation of electric cables in the FM Global intermediate-scale parallel panel test (PPT) using a 60 kW sand burner has been correlated with a dimensional fire propagation index (FPI,)[(m/s(1/2))/(kW/m)(2/3)] derived from ignition and vertical propagation tests in small-scale ASTM E-2058 Fire Propagation Apparatus (FPA). Two distinct types of fire behavior were observed in the PPT. The cables that did not propagate beyond the vicinity of the ignition burner flames have a FPI equal to or less than 7 [(m/s(1/2))/(kW/m)(2/3)], whereas cables that propagated to the top of the 4.9 m parallel panels had a FPI equal to or greater than 14 [(m/s(1/2))/(kW/m)(2/3)]. All the plenum rated (Ul-910 or NFPA 262 tested) cables tested in this study did not exhibit flame propagation in the PPT, had FPI values of 7 [(m/s(1/2))/(kW/m)(2/3)] or less and generated considerably lower smoke than those cables that propagated in the PPT. Based oil this study, it is suggested that FPI measured in ASTM E-2058 FPA may be used for screening cables for UL-910 or NFPA262 tests, thus requiring substantially less cable to be supplied for testing.

Experimental investigations were made to assess the fire behaviour of rice husk particleboard, a potential substitute for wood particleboards. Releveant fire properties, namely non-combustibility, ignitability, fire propagation index, surface spread of flame classification, specific optical density of smoke generated, flammability and flame penetration were determined in accordance with existing standard methods. In this paper the salient results obtained are discussed.

Extension of the British Standard Fire Propagation test (BS 476 Part 6) as a combustion mode for measuring smoke and toxic gas production from small panels of UPVC building products has been investigated. Hydrogen chloride gas concenatrations in the test room have been measured continuously with a method developed to ensure high collection efficiency from the fire gases. Using this method, the influences of test conditions (room temperature and humidity, room surface finish, reactivity of fire box interior) on HCl gas concentrations were examined to explain the fate of HCl gas after its evolution. The relative humidity (RH) in the room was the most critical factor influencing HCl gas concentrations during tests. Although HCl evolution early in the test was influenced little by RH and closely matched smoke production, maximum HCl gas concentrations in the room attained later varied markedly with RH. Generally, no more than one-third of the HCl expected theoretically remained airborne under any test condition. Further investigations suggested that this loss and the effect of RH are associated with HCl plate-out on room surfaces. Using test conditions set to minimize the influence of RH, HCl evolution from a series of UPVC building products was investigated. HCl concentrations in the test room did not exceed incapacitating levels for those products where less than a certain quantity of UPVC was combusted. This observation is discussed in relation to French regulations, which limit the total quantity of chlorine in some synthetic materials within building compartments.

No standard method has been developed for measureing the evolution of specific toxic gases from building lings when involved in fire. The British Fire Propagation test (BS 476 Part 6) operated in an instrumented room has been proposed for this purpose previously but has not found general acceptance. It is considered further in this report, which investigates the movement and measurement of smoke and specific fire gases under different conditions of room stirring and the effect of the latter on fire propagation indexes. Stiring has been found to have no statistically significant effect on fire propagation indexes provided that the effects of this on calibration of the apparatus are taken into account. Stirring also had little effect upon smoke production per se. Under unstire conditions smoke and toxic gases stratify in the same layer early in the test, and measurement of their production at any single room location will be subject to the location, the way the room influences stratification and how the room is instrumentee, as well as by the prpduct performance. Under stirred room conditions smoke and toxic gases are evenly distributed and product performance can be assessed more simply from concurrent measurements of fire, smoke and toxic gas parameters. The latter procedure is proposed for obtaining relative data on building linings and for examination in further studies for correlation to room and corridor burns.

The Fire Propagation Test (BS 476 Part 6) has been used with the discharge of combustion products into a small test room, to provide a fire hazard assessment of some polyvinylchloride floor covering systems, and includes measurements of fire propagation, smoke, CO and HCI production. Particular attention has been given to the change in the HCI concentration of the Room atmosphere during the test and the influence of relative humidity and the nature of the surface linings of the room. The results have been discussed as a contribution to the understanding of the role of HCI in toxicity of fire atmospheres. Some approaches for improving further the reproducibility of the test method are indicated and support is given to the Japanese proposal that such a test be standardized to enable comparison of fire hazard potential of lining systems used in buildings.