Project Topics
Get Complete Project Material File(s) Now! »
Combustion regimes of nanopowders
The study of the explosion hazard of nanoparticles is associated with modifications in the heat transfer, dispersion and reaction due to the size and surface of the elementary particles. For instance, the combustion is greatly modified when the particle size decreases to the nanoscale, as is shown in Figure 15. Bouillard et al. (2010) argument that the combustion regime of micrometric particles is controlled by the diffusion of oxygen into the particles (diffusion regime) and it is kinetically controlled for nanoparticles, explaining the differences on the combustion time.
The combustion mechanism has a direct effect on the combustion time of the reaction. For micrometric aluminum particles, the combustion time follows a 𝑑𝑝𝑛 law with n ranging from 1.5 to 2 (Beckstaed, 2002; Bouillard et al., 2010). However, further studies have shown that the values of n range between 0.3 and 1.6 for nanometric particles, indicating a faster reaction (Beckstaed, 2002; Bouillard et al., 2010). Such modifications of the combustion mechanism may generate modifications on the ignitability and explosivity of the powder as it will be shown in the following sections.
Dispersion properties of nanoparticles
The dispersion of the combustible dust is a compulsory condition to obtain a dust explosion, as shown in Figure 3. When the primary particle size diminishes, a stable dust cloud is likely to take place. However, if the elementary particle diameter reaches the nanoscale, the dust will be composed of agglomerates and aggregates. Primary particles attached by van der Waals, or other physical forces constitute agglomerates while particles bond together by chemical forces are defined as aggregates (Eggersdorfer and Pratsinis, 2014). Therefore, the generation of a homogeneous dust cloud of elementary nanoparticles will depend on the fragmentation and deagglomeration generated by the forces that produce the dust dispersion (Henry, 2013). Nanoparticles have different dispersion properties compared to micro-particles. Nanoagglomerates presents strong chemical bonds opposed to physical van der Waals forces, and in consequence, a complete dispersion of primary particles cannot be achieved for high concentrations of nanoparticles, especially greater than MEC (Wengeler and Nirschl, 2007). Similar conclusions were exposed by Eckhoff (2012, 2011), who affirms that during normal operating conditions in industrial processes, the formation of a well-dispersed nanometric primary particle cloud from bulk powder is extremely difficult because the existence of strong interparticle cohesion forces. In contrast, at laboratory conditions, a homogeneous cloud may be generated, but the delay time between the dispersion and the ignition of the cloud, and the dust concentration are decisive for the extent of the re-agglomeration. Even if it has been established that dust cloud of elementary nanoparticles is unlikely to occur, further studies should be performed in order to understand the influence of dust fragmentation and the changes of dust size distribution on the explosivity parameters.
Characterization of the initial turbulence level by Particle Image Velocimetry (PIV)
The homogeneity of the dust particles after dispersion and the turbulence level of the mixture at the moment of the ignition will influence considerably the violence of the explosion. As a consequence, an accurate determination of the turbulence properties of the system is indispensable to assure the reproducibility and the scaling of the experimental results in laboratory set-ups (Tamanini, 1990, 1998). As it has been explained in section 1.1.2, the initial turbulence related to the dust dispersion is linked to the ignition delay time between dust introduction in the device and ignition, i.e. at higher ignition delay times the turbulence of the gas flow is lower. Moreover, non-intrusive methods have been developed to estimate parameters describing the turbulence of the gas flow or particles at different ignition delay times (Cuervo, 2015; Dahoe, 2000; Dahoe et al., 2001, 2002; D’Amico et al., 2016; Murillo, 2016). For instance, the Particle Image Velocimetry has been previously used in the LRGP laboratory to determine the particles mean velocity and the root-mean-square velocity of the dust dispersion, studying the gas-particle interaction and analyzing the effects of the injection pressure or nozzle on the turbulence parameters (Cuervo, 2015; Murillo, 2016). In this work, the initial turbulence level of the gas flow in the propagation tube and 20-L explosion sphere have been investigated with the exact conditions of the explosion severity sphere and the flame propagation tube of the study.
The Particle Image Velocimetry consists in the determination of the flow or particles velocity by the detection of their position thanks to the light scattering by the solid illuminated with a laser sheet. Recording a considerable quantity of images, the measurement of the displacement of the particles Δ𝑥 at a given time interval Δ𝑡 permits the velocity estimation (Brossard et al., 2009; Pedersen et al., 2003; Thielicke, 2014). The flow patterns of a gas inside a particular geometry or around an obstacle have been studied in the past years, using tracer particles and assuming they move with the local flow velocity. For the analysis of the digital PIV recording, the evaluation region is divided into “interrogation areas”, where the local displacement vector of tracer particles between two illuminations is determined for each interrogation area by means of statistical methods (Raffel et al., 2013). It is also assumed that tracer particles within one interrogation area moves homogeneously between the two illuminations.
Figure 34 shows the PIV experimental set-up for the determination of the flow velocity in the propagation flame tube (left) and the 20-L explosion sphere. The Particle Image Velocimetry experimental set-up is composed of:
1. An experimental vessel in which the flow pattern is studied. Hence, the 20L sphere was modified by Murillo (2016) in order to have visual access for the recording of the tracer movement, and to allow illuminate the particles with a parallel laser beam. Measurement was performed at the height of the electrodes, with an area of analysis of 7×7.5 cm2 with a layer thickness corresponding to the thickness of the laser light sheet, i.e. 0.25 mm.
2. A continuous wave laser allows the illumination of a plane with a parallel sheet (to the recording device) with a constant high energy and with pulsations around 5-10 ns. For this experimental set-up a Neodym-Yttrium-Aluminum-Garnet (Nd:YAG) whose laser beam has a wavelength of 532 nm has been chosen. This type of diode solid state laser is very compact and highly efficient (Thielicke, 2014).
Brief test description
The explosivity parameters of carbon blacks/methane/air mixtures were determined following the standard procedure indicated by the ASTM E1226-12a (ASTM E1226-12a, 2012), with a slight modification to test a hybrid mixture explosion (i.e. gas insertion inside the explosion vessel, which will be described later on). First, the powder sample was placed in the 0.6 L container connected to the sphere (dust reservoir). Then, the chemical ignitors of an equivalent energy are located at the center of the 20-L sphere. Generally, carbonaceous nanopowders exhibit minimum ignition energies (MIE) greater than 100 J (Turkevich et al., 2015; Vignes, 2008) whereas the MIE of methane is 0.3 mJ. Even if the MIE of methane is very low, the addition of carbon black will increase the MIE of the mixtures. Hence, the ignition energy was chosen to ignite the mixtures, without igniting the powder directly. As a consequence, the 20L sphere tests were performed with igniters delivering an energy of 100 J, which is the lowest energy that can be supplied using chemical igniters (Sobbe). A permanent spark could have been used, but the ignition delay time is more easily defined using chemical igniters. Afterwards, it must be verified that both the explosion vessel and the dust reservoir, are correctly closed before the test. The sphere is vacuumed to approximately 0.3 bars and the gas is inserted in order to reach the desired concentration (e.g. from 0.3 to 0.38 barsa to reach 8%vol.). Then, additional air is injected in order to obtain a local pressure of 0.4 barsa before the powder dispersion under 21 bara. This pressure was chosen in order to reach atmospheric pressure when adding dust entrained by the pressurized air pulse in the sphere before the ignition. The dust sample is dispersed through a standard rebound nozzle and the ignition source is activated after a specific delay. If an ignition occurs, the evolution of the pressure over the time is recorded by two piezoelectric sensors (Figure 37). The maximum explosion pressure (Pm) is identified by the maximum peak of the profile, whereas the maximum rate of pressure rise (dP/dtm) is defined by the maximum slope of the curve P vs t.
Effect of inert particle insertion
In order to test the hypothesis of a chemical contribution of the carbon blacks to the gas combustion reaction, explosion severity tests have been carried out with inert AP-D alumina particles instead of carbon blacks. As already written, AP-D 0.05 alumina has been chosen in order to obtain a comparable total surface area developed by the particles (i.e. it should be remembered that the specific surface of AP-D alumina and Corax N550 are 84 and 40 m2.g-1 respectively) at the moment of the explosion as the one of Corax N550/methane/air mixtures (Torrado et al., 2017). It should be stressed that, by changing the nature of the powder, both reactivity and radiative heat transfer can be affected. The latter point will be developed on Chapter 3.
Figure 50 shows the influence of alumina concentration on the explosion severity of methane/air explosions at an initial turbulence level of 𝑣𝑟𝑚𝑠 = 5.8 m.s-1. Similarly to the behavior observed for carbon black hybrid mixtures, the influence of the dispersed particles also depends on the equivalence fuel ratio 𝜑. As it was evidenced for carbon black particles, the explosion severity of the methane/air mixture seems to increase for fuel lean mixtures (𝜑<1) when nanoparticles are added. At 7% methane, the insertion of alumina leads to an overpressure increase from 6.5 to 7.5 bars, which is considered as significant in view of the error bars (Figure 50 – right). A similar trend is observed for the maximum rate of pressure rise; however, it is only observable for low alumina concentrations (Figure 50 – left).
Therefore, the inert properties of the alumina suggest that the increase of the explosion severity at fuel lean mixture may not be caused by changes in the chemical combustion reaction, but by other factors as the deformation/stretching of the flame surface due to the dispersion of nanoparticles or the modification of heat transfers (especially radiation). Regarding the stoichiometric and rich mixtures, no significant effect can be seen on the explosion severity when alumina nanoparticles are dispersed in the gas/air cloud before the explosion (e.g. 995 and 1055 bar.s-1 for gas and the alumina hybrid mixture respectively at 9%v. of CH4).
Radiative heat transfer contribution
Nevertheless, the flame stretch is probably not the only cause and previous experiments have also shown that the radiative heat exchange is apparently modified by the presence of nanopowders (Torrado et al., 2016). Figure 53 shows the influence of carbon black nanoparticles on the radiative heat transfer in methane/air explosions. An increase on the heat radiation transfer is obtained when 6 g.m-3 of Printex EX2 are added into the system. The alumina emissivity being lower than that of carbon (0.07 for alumina compared to 0.97 for carbon particles), the negative impact of the particles radiation, observed for instance in Figure 44 for rich mixtures, is probably less perceptible on the maximum overpressure compared to carbon black hybrid mixtures. The explosivity decrease observed for carbon blacks/methane/air mixtures in Figure 50 could also be related to an increase of the heat radiation transfer. This assertion will be verified with flame propagation tests (Chapter 3).
Table of contents :
CHAPTER 1: Dust nanoparticles – gas hybrid mixture explosions
1.1. Dust explosions
1.1.1. Explosion parameters
1.1.2. Turbulence
1.1.3. Dust dispersion characteristics
1.1.4. Radiative heat transfer
1.1.5. Particle size and shapes
1.2. Nanoparticles explosion
1.2.1. Combustion regimes of nanopowders
1.2.2. Dispersion properties of nanoparticles
1.2.3. Explosion severity
1.2.4. Ignition sensitivity
1.3. Hybrid mixture explosion
1.3.1. Ignition sensitivity
1.3.2. Explosion severity
1.4. Burning velocity
1.4.1. Laminar burning velocity
1.4.2. Flame stretch
1.4.3. Expanding spherical flames
1.4.4. Turbulent combustion
1.4.5. Laminar flame velocity measurements
1.4.6. Dust and hybrid mixture flame velocity measurements
Nomenclature
CHAPTER 2: Effect of nanoparticles dispersion on the explosion severity parameters
2.1. Introduction
2.2. Dust characterization
2.2.1. Particles properties
2.2.2. Characterization of the initial turbulence level by Particle Image Velocimetry (PIV) 63
2.3. Explosion severity parameters
2.3.1. 20L sphere test
2.4. Chemical contribution evaluation
2.4.1. Reaction product analysis
2.4.2 Effect of inert particle insertion
2.5. Radiative heat transfer contribution
2.6. Summary of the effect of different parameters
2.7. Conclusions
2.8. Conclusions (Français)
Nomenclature
References
CHAPTER 3: Study of the burning velocity of nanoparticles/methane/air hybrid mixtures
3.1. Introduction
3.2. Flame propagation study: methodology
3.2.1 Experimental set-up
3.2.2 Calculation of burning velocity of unstretched flames
3.2.3 Video analysis
3.3. Flame propagation measurements
3.3.1. Carbon black/methane/air mixtures
3.3.2. Alumina/methane/air mixtures
3.3.3 Influence of the radiative transfers
3.4. Burning velocity from severity tests
3.4.1. Thin flame propagation model
3.4.2. Burning Velocity Results
3.5. Flame detection using Schlieren Images
3.5.1 Brief description of Schlieren method
3.5.2 Schlieren images for flame velocity estimation – Brief state of art
3.5.3 Results of flame velocity tests using Schlieren set-up
3.6. Conclusions
3.7. Conclusions (Français)
Nomenclature
References
CHAPTER 4: One dimensional model of flame propagation of methane/air/dust hybrid mixture
4.1. Introduction
4.2. Model Equations and Hypotheses
4.2.1 1-D Flame System
4.2.2 Mass and Species Balance
4.2.3 Energy Balance
4.2.4 Numerical Scheme, Boundary and Initial Conditions
4.3. Results of Numerical Simulations
4.3.1. Methane/Air Flame
4.3.2. Hybrid mixture Flame
4.3.3 Hybrid mixture Flame – Contribution of the chemical reaction of the solid
4.5. Conclusions
4.6. Conclusions (Français)
Nomenclature
References
