摘要
[目的]目前,由双燃料发动机组成的电力推进系统是大型液化天然气(LNG)船的主流推进方式,必须对爆炸性可燃气体进行安全可靠性的定性、定量评估,以规避潜在风险。[方法]以某双燃料电力推进大型LNG船发电机室为研究对象,对其内部不同区域的燃气(天然气)泄漏工况进行模拟分析。根据泄漏发生的形式、位置和速率等定义危险泄漏工况,选择雷诺应力模型为湍流模型,采用计算流体力学(CFD)软件Fluent对发电机室燃气供应管线的5个泄漏点进行持续泄漏模拟计算,并将泄漏扩散结果与舱室通风的流场速度分布相结合,得到不同区域发生泄漏后的天然气扩散趋势和浓度分布。[结果]根据仿真模拟结果优化了可燃气体探测器布置方案,并明确了排气风机无需进行防爆设计。[结论]研究结果可为有限空间内通风条件下的可燃气体泄漏事故分析防范提供参考,并且适用于燃烧爆炸破坏的定量评估,用以指导结构强度设计。
[Objectives]The electric propulsion dual-fuel engine is becoming dominant in newly builtLiquefied Natural Gas(LNG)carriers. To avoid the potential risks that accompany the use of flammableand explosive boil-off gas,the performance of precise safety and reliability assessments is indispensable.[Methods]This research concerns the engine rooms of large LNG carriers which are propelled electricallyby a dual-fuel engine. Possible fuel gas(natural gas)leak cases in different areas of the engine room aresimulated and analyzed. Five representative leak cases defined by leak form,leak location and leak rateare entered into a Computational Fluid Dynamics(CFD)simulation,in which the Reynolds stress modelof Fluent software is adopted as the turbulence model. The results of the leaked gas distribution andventilation velocity field are analyzed in combination to obtain the diffusion tendency and concentrationdistribution of leaked gas in different areas.[Results]Based on an analysis of the results,an optimizedarrangement of flammable gas detectors is provided for the engine room,and the adoption of anexplosion-proof exhaust fan is proven to be unnecessary.[Conclusions]These analysis methods canprovide a reference for similar gas leakage scenarios occurring in confined ventilated spaces. In addition,the simulation results can be used to quantitatively assess potential fire or explosion damage in order toguide the design of structural reinforcements.
出处
《中国舰船研究》
CSCD
北大核心
2017年第5期109-119,共11页
Chinese Journal of Ship Research
关键词
液化天然气
管道泄漏
计算流体力学
定量风险评估
Liquefied Natural Gas(LNG)
pipe leakage
Computational Fluid Dynamics(CFD)
quantitative risk assessment
