The Experimental Studies of Compartment Fires for Flashover and Backdraft and A Fire Performance-Based Design for A Cleanroom of 12-in Wafer Manufacturer
本論文第二部分是以SFPE性能式設計程序來評估設計十二吋晶圓廠無塵室煙控系統，本研究使用NIST所發展的CFD火災煙控軟體FDS (Fire Dynamics Simulator)模擬無塵室中火災煙流動情形。在避難模式方面，利用動態避難模擬軟體SIMULEX模擬出晶圓廠人員避難情形以評估人員避難安全。分別使用上排煙及下排煙系統結合起火區空調系統降載至0.1m/s及關閉，非起火區空調系統增載至0.6 m/s的試驗設計，並針對3MW及800kw的火源，進行四個案例的模擬分析。結果發現下排煙系統在3MW的火源下，無法有效將煙侷限在起火區劃內，無法滿足財產安全標準，但仍可滿足生命安全標準，而其他三個案例則均可滿足生命及非生命安全標準。而在人員疏散方面，由SIMULEX結果顯示總體避難時間為164秒，均小於在規格式設計下，煙沉降至1.8公尺所需模擬時間，因此，模擬結果可以證明無論人員可以在晶圓製程區成功地避難，已達到人命安全的目的。|
This study consists of two parts. The first part performed five full-scale compartment (steel container) fire tests and the corresponding numerical simulations to study the flashover and backdraft phenomena. The backdraft experiments performed three full-scale room (steel container) fire tests and the corresponding numerical simulations. The measured temperatures and oxygen indexes, obtained from the failed one and the two successful ones, could identify the backdraft mechanism. To produce the backdraft phenomena, the internal decoration of the container house should use flammable materials on both walls and ceiling. From the simulation and experiment data, backdraft causes two temperature peaks. The first one is below 600□C. Then, an abrupt opening of the front door leads to a supply of a large amount of fresh air, followed by an indication of sudden temperature rise. The second peak temperature is over 600□C. Meanwhile, the deflagration will cause the gases to heat and expand within the fire space, thus forcing unburned gases out of the vent ahead of the flame front. Comparing both cases with natural gas and solid loveseat as the fuel in backdraft, it can be found that the former can achieve pre-mixture state and create easily the instant explosion wave phenomenon, however, this wave will disappear immediately. On the other hand, the solid loveseat used as the fuel in this study produced backdraft within 30-50 seconds after opening door. After the occurrence of backdraft, fire will maintain a period of fully developed stage, which is consistent with the conditions in actual fires. In addition to the experiments, this study also carries out the simulations by using fire models. In the field and zone models, backdraft phenomena were identified. However, both models only considered complete combustion kinetics with oxygen, causing the overpredicted fire temperatures. The simulations did reproduce the phenomena similar to the backdraft fire. The flashover experiments performed the fire performance evaluations for both water-ceiling (a brand name) and common-used wooden ceiling. They are summarized as follows. (1) The flashover occurs in the first fire test, whereas the second experiment using the special fire-proof water-ceiling can successfully suppress the occurrence of flashover. It indicates that the fire-performance properties of ceiling material can significantly determine whether the flashover occurs or not; (2) It is also found that in the first experiment a peak above 600□C in the temperature curve is reached, and then it turns into the fully developed stage. The first temperature peak marks the transition of a sudden increase in temperature from the growth into the fully developed stages, i.e., flashover. In the second part, this study adopts the SFPE Performance-based design procedure to evaluate the performance of upwards and downwards desmoke systems, respectively, in a wafer fabrication zone. The tools used for this purpose include FDS (Fire Dynamics Simulator) and SIMULEX. The design fires are 800 kW (scenario 1) and 3 MW (scenario 2), respectively. Both desmoke systems in scenarios 1, 2 and 4 meet the performance criteria, whereas the downward system in scenario 3 cannot satisfy the property protection requirement. However, the downward desmoke system can still be utilized, provided it complies with some restrictions in fire protection design, for example materials used and the provision of an acceptable emergency plan. For occupant evacuation, the SIMULEX result shows that the total evacuation time is 164s. FDS simulations involving the installation of smoke curtain confirm that the evacuation time is less than the smoke layer descending time, indicating that the occupants can safely evacuate in the event of a fire.