Prediction model of the flow properties inside a tube during hydrogen leakage

• Published in: Journal of loss prevention in the process industries 2019-11, Vol.62, p.103955, Article 103955
• Main Authors: Nagase, Yuri, Sugiyama, Yuta, Kubota, Shiro, Saburi, Tei, Matsuo, Akiko
• Format: Article
• Language: English
• Subjects: Hydrogen leakage; Mass flow rate; Numerical simulation; Prediction model; Pressure distribution; Safety engineering
• ISSN: 0950-4230
• DOI: 10.1016/j.jlp.2019.103955

We numerically investigated high-pressure hydrogen leakage from transportation facilities, focusing on the steady mass flow rate and pressure distribution in a tube during the leakage. We studied steady leakage from a square opening in a square duct as well as leakage from a ruptured cylindrical tube with unsteady closure of a cutoff valve from fully open. A prediction model for the mass flow rate and pressure distribution inside the tube was proposed; such a model would help prevent physical hazards during an accident. We considered changes in the physical quantities according to the fluid dynamics occurring inside the tube. The flow properties were divided into two phases: (i) the unsteady expansion wave propagating inside a tube filled with hydrogen and (ii) the acceleration of hydrogen due to the reduction in the cross-sectional area between the tube and the leakage opening. To close the prediction model, we introduced contraction coefficient models depending on how the hydrogen leakage occurred. The mass flow rate and pressure drop during the leakage estimated by our prediction model showed good agreement with numerical simulation results when the contraction coefficient model was appropriately chosen. This model is considered highly applicable to the construction condition of pressure sensors, the operating conditions of a valve, and the prediction of mass flow rate during an accident. •Numerical simulations of hydrogen leakage from a tube were conducted.•Pressure decrement inside the tube and mass flow rate during leakage depended on the initial pressure and opening area.•A prediction model that considers two flow features inside the tube is proposed.•The two features are the expansion wave inside the tube and reduction of cross-sectional area between tube and opening.•The prediction model agreed well with numerical results when an appropriate contraction coefficient model was used.