Small-scale flame acceleration and application of medium and large-scale flame speed correlations

https://doi.org/10.1016/j.ijhydene.2016.10.042Get rights and content

Highlights

  • The maximal flame speed and overpressure were studied at small-scale.

  • The mixture reactivity was studied for the most obstructed configuration.

  • Models from literature were adapted to correlate with small-scale experiments.

  • Models were modified to consider the geometry and reactivity of the initial mixtures.

Abstract

The flame acceleration plays a major role on the explosion effects. Then, it is of importance to understand the flame acceleration process and to predict explosion effects in open and congested areas for industrial safety reasons. In this aim, small-scale deflagration experiments were performed in cylindrical congested volumes of hydrogen – air mixtures varying from 1.77 L to 7.07 L. The influence of the reactivity was studied since the equivalence ratio of hydrogen – air mixtures were ranging from 0.5 to 2. The congestion was realized with varying numbers of grid layers and configurations. Experimental results, in term of flame speeds, were compared to results from correlations of the literature. Correlations were also adapted to the small-scale and modified to take into account the volume and the reactivity of the combustible mixtures.

Introduction

It is well known that the overpressure generated during a deflagration process increases with the flame velocity [1], [2]. When the deflagration occurs in a congested area, the obstruction in the path of the flame propagation can accelerate the flame due to the turbulence generated [3]. That is to say that the overpressure resulting of the deflagration can be greater in an obstructed area as compared to a free field. Numbers of experiments were achieved in order to better understand the process of flame acceleration and to predict this effect in an industrial context.

Several experiments were performed in obstructed tubes to analyze the mechanism of acceleration. For example, with methane-air mixtures in a 7 meters long and 0.3 m × 0.3 m square cross-section tube [4], in a 3.66 m long and 0.076 m × 0.076 m square channel [5] or more recently in a 4.25 m long and 0.162 m internal diameter [6], [7] where the effect of obstacle separation distance and blockage ratio BR (20%, 30% and 40%) on methane-air (10% by vol.) mixture were studied. Boeck et al. [8] have studied the flame propagation with a single obstacle (50% BR) in a 0.4953 m long square channel (0.0762 × 0.0762 m2) with a stoichiometric H2/O2 mixture. Flame acceleration of hydrocarbon-air and more specifically hydrogen-air mixtures were studied in three 11 m long tubes with 0.05, 0.15 and 0.30 m cross-sectional diameters [9], in an 11 m long tube and 0.05 m cross-sectional diameter [10] for example. Beauvais et al. [11] have discussed about the parameters influencing turbulent flame acceleration process, using experiments carried out with hydrogen-air mixtures (from 10% to 26% by vol.) in a 6 m long tube and 0.066 m internal diameter with 32% and 69% BR. Cross et al. [12] have studied DDT and the detonation propagation limits of hydrogen-air (17%–60% by volume) in a 6.1 m long, 0.1 m diameter tube filled with obstacles (44% BR) at atmospheric pressure. Indeed, hydrogen is of interest since it is widely used in the chemical industry, the fertilizer industry, the sector of petrochemical for example and it is in the containments of nuclear reactors in case of incident. Moreover, it is seen as an important energy carrier for the future. For example, experiments were performed in semi-confined layers with hydrogen-air mixtures, in a 2.5 m long rectangular channel with varying width of 0.3, 0.6 and 0.9 m [13]. Experiments at medium and large-scale involving hydrogen have been also undertaken. For example, Grune et al. [14] have studied the effect of obstacles in a cubical volume of 0.166 m3 filled with a stoichiometric H2 – air mixture. At large-scale, Sherman et al. [15] carried out experiments for nuclear reactor safety with H2 - air mixtures (from 12% vol. to 30% vol. H2 in air) in 30.5 m long rectangular obstructed channel and 1.83 m × 2.44 m cross-section (136 m3). Dorofeev et al. [16] have studied the flame acceleration with lean hydrogen - air mixtures (from 9.8% vol. to 14% vol. H2 in air) in a channel of 34.6 × 2.3 × 2.5 m coupled with a “canyon” of 10.55 m × 6.3 m × 2.5 m and a last channel of 24.75 m × 2.3 m × 2.5 m (total volume of 430 m3). Royle et al. [17] performed experiments with several hydrogen concentration (0% vol. to 100% vol.) in methane-air mixtures in a 3 m × 3 m × 2 m congested volume (18 m3). Groethe et al. [18] have realized experiments with mixtures composed of 15% vol. to 30% vol. H2 in air inside a hemispherical volume of 300 m3 filled with obstacles. Lowesmith et al. [19] have studied the flame acceleration of methane – hydrogen – air mixtures in a 26.2 m × 2.8 m × 3 m volume (220 m3) using rows of obstacles.

In order to improve industrial safety, methods were developed from experimental results to predict explosion effects due to the flame acceleration in congested areas. Among them, two different types of methods can be distinguished: methods based on blast curves like the TNO multi-energy method [20] and the Baker-Strehlow-Tang method [21] or analytical method like models proposed by Grune et al. [14] or Dorofeev [22]. Indeed, Grune et al. [14] have developed a method from medium-scale experiments with a stoichiometric hydrogen-air mixture. The method proposed by Dorofeev [22] was based on medium and large-scale experiments (hydrocarbon-air mixtures and hydrogen-air mixtures).

To our knowledge, no papers deals with the flame acceleration in an open and congested area at small-scale with volumes inferior to 0.01 m3. Then, in order to add data for the prediction of explosion effects, this paper will first present the experimental results for small-scale deflagrations in obstructed areas. The influence of the obstruction on the flame speed and the maximal overpressure was investigated with hydrogen – air mixtures in initial cylindrical volumes varying from 1.77 L to 7.07 L. Lean and rich mixtures were studied, since the equivalence ratios were ranging from Φ = 0.5 to Φ = 2. The second section of this paper will concern the application of the model of Grune et al. [14] and the model of Dorofeev [22] our small-scale experiments. Experimental results and predicted results in terms of flame speeds will be compared. Finally, the last part will present some models adaptations to take into account the volume, the geometry of the initial confinement and the reactivity of the mixtures.

Section snippets

Experimental setup

Experiments were performed on a 2.85 m long and 1.32 m width workbench (Fig. 1). An obstructed cylindrical volume, called flame acceleration device, was placed on the workbench.

It consisted in a cylinder made with n layers stainless steel grids with a diameter of 0.5 mm. Each grids were made of 2⋅δm = 6 mm square cells, with δm the half of the grid cell size. Consequently, the area blockage ratio of the obstacle, which is defined as the ratio of area blocked by obstacles to the cross-sectional

Varying obstacles number at a fixed equivalence ratio

All tests were conducted with an equivalence ratio Φ = 1 for all configurations reported in Table 1. A general trend (Fig. 3) shows that the maximal flame speed increases with the number of layers for both heights (h0 = 0.1 m and h0 = 0.2 m). This evolution of the flame velocity according to the number of obstacles follows the same trend as Grune et al. [14] observations.

The geometry of the most obstructed initial confinement (n = 89) is three times wider than high (d0 = 0.3 m and h0 = 0.1 m)

Models of the literature

It is of importance to predict the effects of explosions for industrial safety. Several authors have published analytical models to quantitatively determine parameters of deflagrations in obstructed volume like the flame speed, the overpressure and the impulse for example. Two analytical models, Grune et al. [14] and Dorofeev [22], have been applied and modified in order to correlate with our small-scale experimental results.

Conclusion

In order to add data for the prediction of explosion effects, experiments were performed to study the influence of the congestion on the flame speed and the overpressure at small-scale when a deflagration occurs in hydrogen-air mixtures. To our knowledge, no papers deals with flame acceleration in an open and congested area at laboratory-scale with volumes inferior to 0.01 m3. The experimental study was divided in two parts, a first part concerning the influence of the number of obstacles

Acknowledgements

The authors acknowledge the support of the French Agence Nationale de la Recherche (ANR) under reference ANR-10-CSOG-008.

References (28)

Cited by (5)

  • Numerical studies of premixed hydrogen/air flames in a small-scale combustion chamber with varied area blockage ratio

    2020, International Journal of Hydrogen Energy
    Citation Excerpt :

    As a result, a suitable numerical model would need to be able to predict how the obstacles and turbulence levels within the chamber influence the flame structure [20,21,25]. Large and small scale experiments have been previously carried out for hydrogen/air explosions and compared them to substitute hydrocarbons [16,17,26–29]. These experiments can be used to validate numerical models for further use.

  • Numerical study of vented hydrogen explosions in a small scale obstructed chamber

    2018, International Journal of Hydrogen Energy
    Citation Excerpt :

    There has been a significant amount of experimental work regarding the explosion features of H2. Besides medium or large sale explosion tests [15], a few lab-scale experimental studies [16–19] have been conducted in order to gain insights into the factors that influence the hydrogen-explosion behaviours such as vent area [18], vent burst pressure [16] and the position of ignition sources [17,18]. Computational fluid dynamics (CFD) is evolving as an efficient and reliable alternative to experiments and has been used in safety-related research in buildings and off-shore processing plants [20–23].

  • Combustion of hydrogen–oxygen microfoam on the water base

    2017, International Journal of Hydrogen Energy
    Citation Excerpt :

    Thus gaseous hydrogen-based mixtures are characterized by high probability of deflagration-to-detonation transition (DDT) [15–19] leading to the formation of detonation waves [20–23]. As a rule the origins of DDT are related with the phenomenon of flame acceleration [24] which can be determined by a set of factors such as combustible mixture composition [25], its volumetric distribution (e.g. in the form of concentration gradients [26]), channel geometry [27,28] including obstacles [29–31] and perforation of the wall [32,33]. One of the leading roles in flame acceleration and DDT in gaseous hydrogen-based mixtures belong to acoustic and compression waves defining the flame and flow evolution inside closed vessels [34–36].

View full text