150 likes | 161 Views
This study presents a CFD model for simulating flame length and radiative heat flux of cryogenic hydrogen jet fires. It includes experiments on vertical cryogenic hydrogen jet fires and discusses the effects of turbulence models, angular discretization, air humidity, hood inclusion, and extraction velocity in the simulation results. The research emphasizes the importance of detailed experimental measurements and highlights the sensitivity of simulations to environmental factors. The study was validated against experimental tests, showing good accuracy in predicting radiative heat flux. Further research is recommended to expand the model's applicability to high-pressure cryogenic releases.
E N D
7th International Conference on Hydrogen Safety11-13 September 2017, Hamburg, Germany Thermal radiation from cryogenic hydrogen jet fires D. Cirrone, D. Makarov and V. Molkov HySAFER Centre, University of Ulster, Newtownabbey, BT37 0QB, UK cirrone-dmc@email.ulster.ac.uk, dv.makarov@ulster.ac.uk, v.molkov@ulster.ac.uk
Introduction Cryo-compression is a competitive technique when storage of large quantities of hydrogen is needed. Understanding of consequences of potential accidents with ignited cryogenic releases of hydrogen is fundamental to protect life and prevent property loss. Sandia National Laboratories (SNL) observed that, for a fixed mass flow rate, the decrease of release temperature causes (Panda and Hecht, 2016): • Longer flame length; • Higher radiative heat flux from a jet flame. This study aims at development and validation of a CFD model to simulate flame length and radiative heat flux for cryogenic hydrogen jet fires.
SNL experiments Description of validation tests Vertical cryogenic hydrogen jet fire experiments (Panda and Hecht, IJHE, 2016). 5 tests covering the coldest jets were selected for the study. Radiative heat flux measurements. Flame length measurements. Radiometers (RD) location based on (Panda,2016, private communication)
CFD Model Numerical details
Modelling of release source Mass flow rate: notional nozzle theory versus experiments “Slightly” under-expanded jet (>1.9 bar). Release source is modelled using the Ulster’s notional nozzle theory. Abel-Noble EOS is applicable to cryogenic releases for P ≤ 6 bar. The maximum deviation is about 10% and it is given for test 5.
Simulation results Effect of turbulence model The region 1300-1500 K was adopted as indication of the flame length. Realizable κ-ε shows the best agreement with experiment exp: 0.66 m
Simulation results Effect of DO angular discretisation The angular divisions’ refinement for radiation model affects considerably the simulated radiative heat flux. Simulations result to be independent of further refinement over 10x10 angular divisions. Results are not sensitive to the increase of pixels number. 3x3 pixels
Simulation results Effect of air humidity The effect of water vapour presence in air must be assessed. Mass fraction equal to 0.008 was assumed, which corresponds to relative humidity equal to 74% and temperature 288 K. The presence of water vapour has a significant effect Importance of detailed experimental measurements
Simulation results Effect of hood inclusion in the geometry The laboratory was equipped with an exhaust gases system. Therefore, combustion can be affected by the volumetric flow imposed at the hood. The removal of the combustion products by forced ventilation caused a decrease of radiative heat flux, reaching over 30% reduction.
Simulation results Effect of hood extraction velocity The volumetric flow rate of the ventilation system was adapted to the released mass flow rate of hydrogen (0.1-0.7 g/s). The range of variation of the volumetric flow rate is 5100-7650 m3/h, corresponding to extraction velocity 7.0 m/s and 10.5 m/s respectively. Max overestimation +36 % V=10.5 m/s : deviations are contained in the range ±14%
Simulation results Definition of model set-up and validation Experimental radiative heat flux is predicted with ±15% accuracy; only exceptions are the measurements at 5th sensor in tests 1-2. +17.5% +29% = 7.0 m/s = 7.0 m/s = 7.0 m/s = 10.5 m/s = 10.5 m/s
Simulation results Flame Length Simulations of tests 4 and 5 resulted in slight overestimation of flame length. This can be due to the mass flow rate overestimations by, respectively, 5% and 10% of the release source modelling .
Concluding remarks Simulations of cryogenic hydrogen jet fires were conducted to develop and validate a predictive CFD model for assessment of thermal hazards. Three turbulence models were compared and realizable κ-ε model demonstrated the best performance in reproducing experiments. DO model discretisation strongly affect simulations. 10x10 was the minimum number of angular divisions providing convergence. The simulations are sensible to experimental parameters such as humidity and exhaust system volumetric flow rate, highlighting the importance of accurate and extended publication of experimental data. The simulations were validated against 5 experimental tests with release pressures 2-5 bar abs and temperatures 48-82K. For all the tests experimental radiative heat flux at 5 sensors along the jet flame was predicted within ±15% accuracy, with few exceptions. Further research should be conducted to extend the domain of the CFD model applicability to high pressure cryogenic releases.
Thank you for your attention! Acknowledgements The authors are grateful to Dr. P.P. Panda (SNL) for providing experimental data indispensable to conduct this numerical study. Support of EPSRC through SUPERGEN H2FC Hub and SUPERGEN Challenge, and support of Fuel Cell and Hydrogen 2 Joint Undertaking through NET-Tools project are highly appreciated!