Loss of containment: experimental aerosol rain-out assessment

https://doi.org/10.1016/S0304-3894(02)00039-0Get rights and content

Abstract

We measured the temperature profiles and rain-out spatial distribution for flashing water jets generated from a pilot scale experimental setup. This allowed us to define the transitions between three types of jets (stable, mechanically fragmented, flashing). The present experimental data when compared to other authors’ data show that the transition to flashing type occurs at lower superheat when the orifice length increases, and that homogeneous aerosol behavior could be a valid assumption for mechanically fragmented jets whereas it is not for flashing jets.

Introduction

We consider the accidental release of a toxic or flammable pressurized liquid which gives rise to a two-phase jet. We would like to predict the jet mass fraction which deposits on the ground as a liquid (rain-out fraction) because this fraction does not directly contribute to the atmospheric dispersion cloud. Aerosol rain-out prediction is now far from accurate despite its role in the loss of containment event. The liquid mass flow entrained with the gas as an aerosol is often assumed (see for example [1]) to have the same value as the vapor mass flow obtained from a flash calculation. This can lead to a severe prediction error. More recent works [2], [3], [4], [5], [6] include a physical approach but the predictions remain unverified. A better knowledge of the initial fragmentation of the liquid phase would probably be useful.

Liquids discharged from a high pressure zone to a zone where pressure is lower than the equilibrium one have been studied by several authors like Bushnell and Gooderum [7], and Brown and York [8]. They studied the shattering temperature Tsh, i.e. the temperature at which a coherent jet becomes almost completely atomized.

From their experiments with water out of nozzles for We<12, Bushnell and Gooderum [7] deduce that “the value of (TshTeb)/Tsh=0.1 may be approaching the upper limit of the amount of superheat that can be tolerated by a liquid jet before it shatters”. “Negligible effect of the nozzle diameter and the velocity indicates that the aerodynamic forces were indeed secondary during the atomization process”.

Brown and York [8] considered that the growth rate of bubbles determines the shattering effect. From Plesset and Zwick [9], and Forster and Zuber [10], the radius of a bubble follows the relation:r=r0+Ctwhere C is the bubble growth rate constant given byC=Cp(T−Teb)hfgρfρgπDthThey observe that “a jet of large diameter may shatter at a superheat for which a smaller jet does not shatter” (note that this observation seems to be in contradiction with the conclusions of Bushnell and Gooderum [7]). This leads them to use the Weber number (which measures the momentum exchange with air to surface tension ratio) as a parameter. They represent their experimental data for transition from a coherent to a shattered jet on a chart with We and C as coordinates. On this chart, the transition line for water is almost the same for sharp-edged and not too rough orifices. “Higher values of the Weber number permit shattering to occur with less superheat”. Both Brown and York [8] and Bushnell and Gooderum [7] used small orifices (generally less than 2 mm ID). They performed their jet’s observations at a few diameters downstream.

Our objective is to enlarge their observations to situations more representative of accidental flashing liquid discharge in order to examine the validity of the assumptions in some existing models [2], [3], [4], [5], [6].

Section snippets

Experimental

We generated water flashing jets from a pilot scale set-up (0.230 m3, 0.1–1.3 MPa) [11] through a 1.8, a 8 mm ID orifice (Fig. 1(a)) and a 4 m long pipe of 8 mm ID. Inlet temperature was varied from 313 to 453 K, and the sub-cooling varied from 50 to 600 kPa. Twelve capture basins (Fig. 1(b)) allowed to weigh the liquid rained out under the jet as a function of the distance from the source. A thermocouple also measured the final temperature of the jet. The 1.8 mm ID orifice gives rise to three types of

Discussion

As suggested by Brown and York [8], we plotted our experiments in Fig. 7 (nozzles) and Fig. 8 (pipe), using the growth rate constant C [9], [10] and the Weber number as co-ordinates. Note that we use the velocity of the liquid phase (before flashing) in the definition of the Weber number for a jet out of a pipe.

Fig. 7 shows that for jets at low Weber number after an orifice whatever the diameter, the transition to flashing occurs when C is between 0.085 and 0.088 m s−1/2 (i.e. 38K<T0−Tsat<40K).

Conclusion

In the introduction to the RELEASE program Johnson and Woodward [2] asserted: “the liquid release models available in 1986 could not adequately predict the complicated processes occurring during the release of a superheated liquid”. We saw here that the RELEASE model does not lead to a sufficient solution.

Our new experimental data demonstrates that different types of jets have to be considered and that some models for rain-out give reasonable predictions. We are now looking for an improvement

Acknowledgements

The financial support from “Conseil Régional Rhône-Alpes” is gratefully acknowledged.

References (11)

  • M. Epstein et al.

    J. Loss Prev. Process Ind.

    (1990)
  • A. Papadourakis et al.

    J. Loss Prev. Process Ind.

    (1991)
  • F.P. Lees, Loss Prevention in the Process Industries, Butterworhts, London, 1986, p....
  • D.W. Johnson, J.L. Woodward, RELEASE: a model with data to predict aerosol rainout in accidental releases, CCPS Concept...
  • C.J. Wheatley, SRD Report R410,...
There are more references available in the full text version of this article.

Cited by (3)

  • A novel method to characterize morphological dynamics and determine flashing jet regimes

    2021, International Journal of Multiphase Flow
    Citation Excerpt :

    Fine droplets suspend in the air and form vapor cloud (Calay and Holdo, 2008; Deaves et al., 2001; Webber and Kukkonen, 1990). Large droplets rain out and develop a liquid pool (Cleary et al., 2007; Hocquet et al., 2002; Witlox et al., 2007; Witlox and Harper, 2013, 2014; Witlox et al., 2010; Woodward and Papadourakis, 1995). The vapor cloud of flammable (e.g., LNG (liquefied natural gas) (Baalisampang et al., 2019; Zhang et al., 2020), LPG (liquefied petroleum gas) (Brzezinska and Markowski, 2017; Demichela et al., 2004), liquid hydrogen (LH)) or oxidant materials (Palazzi et al., 2014) (e.g., liquefied oxygen (LO)) provide beneficial conditions for jet fire or explosion (Makhviladze and Yakush, 2005).

  • Toxic industrial chemical (TIC) source emissions modeling for pressurized liquefied gases

    2011, Atmospheric Environment
    Citation Excerpt :

    The experiments, in spite of some criticisms (Ramsdale and Tickle, 2000), provide valuable data with which to validate rainout models. Apart from these experiments, very few rainout data exist, except for the Rohm & Hass (R&H) MMA data (Lantzy et al., 1990) and the Hocquet et al. (2002) and Bigot et al. (2005) water data. Chemicals were released 1.22 m above the liquid capture pans in CCPS, 1.73 m in R&H experiments, and 1.5 m in the Hocquet et al. and Bigot et al. experiments.

  • Rain-out investigation: Initial droplet size measurement

    2005, Journal of Loss Prevention in the Process Industries
View full text