CFD simulations of dust dispersion in the 1 m3 explosion vessel

According to standard procedures, flammability and explosion parameters for dusts and dust mixtures are evaluated in 20 L and/or 1 m³ vessels, with equivalent results provided a correct ignition delay time (60 ms in the 20 L vessel; 600 ms in the 1 m³ vessel). In this work, CFD simulations of flow field and dust concentration distribution in the 1 m³ spherical vessel are performed, and the results compared to the data previously obtained for the 20 L. It has been found that in the 1 m³ vessel, the spatial distribution of the turbulent kinetic energy is lower and much more uniform. Concerning the dust distribution, as in the case of the 20 L, dust is mainly concentrated at the outer zones of the vortices generated inside the vessel. Furthermore, an incomplete feeding is attained, with most of the dust trapped in the perforated annular nozzle. Starting from the maps of dust concentration and turbulent kinetic energy, the deflagration index K_St is calculated in both vessels. In the conditions of the present work, the K_St is found to be 2.4 times higher in the 20 L than in the 1 m³ vessel.     

CFD simulations of dust dispersion in the 20 L vessel: Effect of nominal dust concentration

Measurements of flammability and explosion parameters for dust/air mixtures require uniform dispersion of the dust cloud inside the test vessel. In a previous work, we showed that, in the standard 20 L sphere, the dust injection system does not allow generation of a uniform cloud, but rather high gradients of dust concentration are established. In this work, we used a previously validated three-dimensional CFD model to simulate the dust dispersion inside the 20 L sphere at different dust nominal concentrations (and fixed dust diameter). Results of numerical simulations have shown that, as the dust nominal concentration is increased, sedimentation prevails and, thus, when ignition is provided, the dust is mainly concentrated at the vessel walls.     

Effect of turbulence spatial distribution on the deflagration index: Comparison between 20 L and 1 m3 vessels

In this work, the effect of spatial distribution and values of the turbulent kinetic energy on the pressure-time history and then on the explosion parameters (deflagration index and maximum pressure) was quantified in both the standard vessels (20 L and 1 m³).The turbulent kinetic energy maps were computed in both 20 L and 1 m³ vessels by means of CFD simulations with validated models. Starting from these maps, the turbulent flame propagation of cornstarch was calculated, by means of the software CHEMKIN. Then, the pressure-time history was evaluated and from this, the explosion parameters.Calculations were performed for three cases: not uniform turbulence level as computed from CFD simulations, uniform turbulence level and equal to the maximum value, uniform profile and equal to the minimum value. It was found that the cornstarch in the 20 L vessel get variable classes (St-1, St-2, St-3) with respect to the 1 m³ (St-1). However, simulations performed on increasing the ignition delay time, shown that the same results can be attained only using 260 ms as ignition delay time in the 20 L vessel.     

CFD Simulation of the Dispersion of Binary Dust Mixtures in the 20 L Vessel

There are at least two main requirements for repeatable and reliable measurements of flammability and explosibility parameters of dusts: a uniform dispersion of solid particles inside the test vessel, and a homogeneous degree of turbulence. In several literature works, it has been shown that, in the standard 20 L sphere, the dust injection system generates a non-uniform dust cloud, while high gradients characterize the turbulent flow field. In this work, the dust dispersion inside the 20 L sphere was simulated for nicotinic acid/anthraquinone mixtures (with different pure dust ratios, while keeping the total dust concentration constant) with a validated three-dimensional CFD model. Numerical results show that the fields of dust concentration, flow velocity and turbulence are strongly affected by both diameter and density of the pure dusts. These different dust properties lead to segregation phenomena with the formation of zones richer in one component and leaner in the other one and vice versa, and also result in preferential paths for the solid particles inside the sphere. Overall, the obtained results highlight the need for developing a dust injection system able to overcome the shortcomings of the actual one even when testing dust mixtures.     

Application of CFD on the sensitivity analyses of some parameters of the modified Hartmann tube

The aim of this work is to determine the influence of operating parameters such as the dispersion pressure, the ignition delay and height on the dust flammability. A Computational Fluid Dynamics (CFD) simulation, based on an Euler–Lagrange approach, was developed with Ansys Fluent™ and validated experimentally. Such analysis will facilitate the choice of the most conservative conditions for a flammability test. This paper is focused on a case study performed on wheat starch with the modified Hartmann tube. The dispersion process of the powder was studied with granulometric analyses performed in situ and high speed videos. Tests were performed with injections at gas pressure ranging from 3 to 6 bars and the evolution of the particle size distribution (PSD) was recorded at different ignition heights (5, 10 and 15 cm over the dispersion nozzle). The observations highlighted the presence of agglomeration/deagglomeration processes and dust segregation. Besides, a CFD simulation analysis was aimed at evaluating the impact of a set of parameters on the PSD and the local turbulence, which are closely linked to some flammability parameters. For this computational analysis, the CFD simulation was coupled with a collision treatment based on a Discrete Element Method (DEM) in order to consider the cohesive behavior of the combustible dust. Thus the results suggest performing the injection of the gases at approximately 5 bars for the flammability tests of wheat starch in order to obtain the finest PSD at a given ignition height. It is also shown that the finest PSD are obtained at 5 cm over the dispersion nozzle. However, the local instabilities and turbulence levels are so high during the first stages of the dispersion that the flame growth can be disturbed for short ignition delays. Moreover, the stabilization of the bulk of the dust cloud requires longer periods of time when the ignition sources are located at 15 cm. As a result, the recommended height to perform a flammability test is 10 cm in this case. Finally, this study proposes some tools that might improve the procedure of dust flammability testing.     

CFD modeling and simulation of turbulent fluid flow and dust dispersion in the 20-L explosion vessel equipped with the perforated annular nozzle

A three-dimensional CFD model was developed to simulate the turbulent flow field induced by dust feeding and the associated dust dispersion within the 20-L explosion vessel equipped with the perforated annular nozzle. The model was validated against experimental data for pressure and root mean square velocity.Simulation results have shown that the turbulent kinetic energy is rather uniformly distributed and its values are significantly lower than those attained with the rebound nozzle. Furthermore, the perforated annular nozzle is able to generate a uniform dust/air cloud. However, a consistent fraction of the dust remains trapped inside the nozzle and, thus, it does not contribute to the explosion process.     

Review of the DESC project

Dust Explosion Simulation Code (DESC) was a project supported by the European Commission under the Fifth Framework Programme. The main purpose of the project was to develop a simulation tool based on computational fluid dynamics (CFD) that could predict the potential consequences of industrial dust explosions in complex geometries. Partners in the DESC consortium performed experimental work on a wide range of topics related to dust explosions, including dust lifting by flow or shock waves, flame propagation in vertical pipes, dispersion-induced turbulence and flame propagation in closed vessels, dust explosions in closed and vented interconnected vessel systems, and measurements in real process plants. The new CFD code DESC is based on the existing CFD code FLame ACceleration Simulator (FLACS) for gas explosions. The modelling approach adopted in the first version entails the extraction of combustion parameters from pressure–time histories measured in standardized 20-l explosion vessels. The present paper summarizes the main experimental results obtained during the DESC project, with a view to their relevance regarding dust explosion modelling, and describes the modelling of flow and combustion in the first version of the DESC code. Capabilities and limitations of the code are discussed, both in light of its ability to reproduce experimental results, and as a practical tool in the field of dust explosion safety.     

Experimental and numerical investigation of constant volume dust and gas explosions in a 3.6-m flame acceleration tube

This paper describes an experimental investigation of turbulent flame propagation in propane-air mixtures, and in mechanical suspensions of maize starch dispersed in air, in a closed vessel of length 3.6 m and internal cross-section 0.27 m × 0.27 m. The primary motivation for the work is to gain improved understanding of turbulent flame propagation in dust clouds, with a view to develop improved models and methods for assessing explosion risks in the process and mining industries. The study includes computational fluid dynamics (CFD) simulations with FLACS and DESC, for gas and dust explosions respectively. For initially quiescent propane-air mixtures, FLACS over-predicts the rate of combustion for fuel-lean mixtures, and under-predicts for fuel-rich mixtures. The simulations tend to be in better agreement with the experimental results for initially turbulent gaseous mixtures. The experimental results for maize starch vary significantly between repeated tests, but the subset of tests that yields the highest explosion pressures are in reasonable agreement with CFD simulations with DESC.     

Numerical modelling of the effects of vessel length-to-diameter ratio (L/D) on pressure piling

Pressure piling presents a major explosion hazard in interconnected process vessels. Pressure enhancement in the secondary vessel due to the acceleration of the flame through the connecting pipe can generate a disproportionately more violent explosion than would have been expected based on the concentration of dust in the secondary vessel. Pressure piling is a very complex phenomenon that is difficult to investigate through experimentation. Advanced computational fluid dynamics (CFD) modelling is a promising route to accurately account for all the complexities associated with pressure piling.In this paper, the current state of knowledge concerning pressure piling is presented. Further, the effects of varying the length-to-diameter ratio (L/D) of the primary vessel (Vessel 1) on pressure piling was investigated using numerical modelling. The volumes and volume ratio of the interconnected vessels were kept constant while the L/D of Vessel 1 was varied from 0.5 to 15. The simulations of coal dust explosion were performed using the coalChemistryFoam solver from OpenFOAM version 5.0.1. It is hoped that the findings from this study provide insight into the effects of the geometrical design of interconnected vessels, particularly L/D, on pressure piling. Additionally, this work has implications for the optimal placement of explosion isolation devices intended to actuate before the flame front and pressure escape to downstream vessels.     

Experiment-based investigations of magnesium dust explosion characteristics

An experimental investigation was carried out on magnesium dust explosions. Tests of explosion severity, flammability limit and solid inerting were conducted thanks to the Siwek 20 L vessel and influences of dust concentration, particle size, ignition energy, initial pressure and added inertant were taken into account. That magnesium dust is more of an explosion hazard than coal dust is confirmed and quantified by contrastive investigation. The Chinese procedure GB/T 16425 is overly conservative for LEL determination while EN 14034-3 yields realistic LEL data. It is also suggested that 2000-5000 J is the most appropriate ignition energy to use in the LEL determination of magnesium dusts, using the 20 L vessel. It is essential to point out that the overdriving phenomenon usually occurs for carbonaceous and less volatile metal materials is not notable for magnesium dusts. Trends of faster burning velocity and more efficient and adiabatic flame propagation are associated with fuel-rich dust clouds, smaller particles and hyperbaric conditions. Moreover, Inerting effectiveness of CaCO₃ appears to be higher than KCl values on thermodynamics, whereas KCl represents higher effectiveness upon kinetics. Finer inertant shows better inerting effectiveness.     

Effect of ignition delay on explosion parameters of corn dust/air in confined chamber

This paper presents the explosion parameters of corn dust/air mixtures in confined chamber. The measurements were conducted in a setup which comprises a 5 L explosion chamber, a dust dispersion sub-system, and a transient pressure measurement sub-system. The influences of the ignition delay on the pressure and the rate of pressure rise for the dust/air explosion have been discussed based on the experimental data. It is found that at the lower concentrations, the explosion pressure and the rate of pressure rise of corn dust/air mixtures decrease as the ignition delay increases from 60 ms; But at the higher concentrations, the explosion pressure and the rate of pressure rise increase slightly as the ignition delay increases from 60 ms to 80 ms, and decrease beyond 80 ms. The maximum explosion pressure of corn dust/air mixtures reaches its highest value equal to 0.79 MPa at the concentration of 1000 gm⁻³.     

Experimental investigation and modelling of aluminum dusts explosions in the 20 L sphere

An experimental investigation was carried out on the influences of dust concentration, particle size distribution and humidity on aluminum dust explosion. Tests were mainly conducted thanks to a 20 L explosion sphere. The effect of humidity was studied by storing the aluminum particles at constant relative humidity until the sorption equilibrium or by introducing water vapour in the explosion vessel. The tested particles sizes ranged from a volume median diameter of 7 to 42 μm and the dust concentrations were up to 3000 g m^?3.Among other results, the strong influence of the particle size was pointed out, especially when the Sauter mean diameter is considered. These results stressed the predominance of the specific surface area on the mass median particle diameter.The effect of water on aluminum dust explosion was decoupled: on the one hand, when water adsorption occurs, hydrogen generation leads to an increase of the explosion severity; on the other hand, when the explosion of dried aluminum powder occurs in a humid atmosphere, the inhibiting effect of humidity is put forward.A model based on mass and heat balances, assuming a shrinking core model with chemical reaction limitation, leads to a satisfactory representation of the pressure evolution during the dust explosion.     

Quantifying the effect of strong ignition sources on particle preconditioning and distribution in the 20-L chamber

Computational fluid dynamics is used to investigate the preconditioning aspect of overdriving in dust explosion testing. The results show that preconditioning alters both the particle temperature and distribution prior to flame propagation in the 20-L chamber. A parametric study gives the fluid pressure and temperature, and particle temperature and concentration at an assumed flame kernel development time (10 ms) for varying ignitor size and particle diameter. For the 10 kJ ignitor with 50% efficiency, polyethylene particles under 50 μm reach 400 K and may melt prior to flame propagation. Gases from the ignitor detonation displace the dust from the center of the chamber and may increase local particle concentration up to two times the nominal value being tested. These effects have important implications for explosive testing of dusts in the 20-L chamber and comparing to larger 1-m³ testing, where these effects may be negligible.     

The explosion of non-nano iron dust suspension in the 20-l spherical bomb

The understanding of dust explosion is still incomplete because of the lack of reliable data and accurate models accounting for all the physic-chemical aspects. Besides, most of the experimental data available in the current literature has been accumulated on the 20-l spherical bomb tests, which gives coarse results for the pressure history that cannot be easily converted into fundamental combustion parameters. Nevertheless, the large amount of experimental data available in the spherical bomb is attractive. In this work, the explosion of non-nano iron dust in the standard spherical vessel is analyzed, aiming at evaluating the burning velocity from the theoretical point of view and the simple experiments performed by the standard explosion tests. The choice of iron is of relevance because its adiabatic flame temperature is below the boiling temperature of both the reactants and oxidized gaseous, liquid, or solid (intermediate and final) products and for the negligible particle porosity, which instead is typical of organic dust. Therefore, a non-nano iron dust explosion can be reconducted to a reduced mechanism since heterogeneous (surface) combustion may be determinant, and the diffusion mechanism for oxygen is the only relevant. The laminar burning velocity is strongly dependant on the particle diameter, whereas little effects are due to the dust concentration. The reported final value was found in agreement with typical limiting laminar burning velocity, adopted for the estimation of flammability limits.     

Experiment-based investigations on the effect of ignition energy on dust explosion behaviors

Explosion behaviors of typical light metal and carbonaceous dusts induced by different ignition energies were investigated based on systematic experiments in a Siwek 20 L vessel. Comparative analysis reveals that the explosion mechanism of carbonaceous dust is the volatile combustion, whereas the mechanism for light metal dust mainly features the surface heterogeneous oxidation. Influences of ignition energy on severity and flammability limit are much more significant for carbonaceous dust than light metal, especially for the powder with less volatile. An innovative approach was introduced to derive flame thickness from the pressure–time trace. The relation between explosion induction time and combustion duration of ignitor was also analyzed. Results show inappropriate ignition energy will cause under-/over-driving in the thermodynamic/kinetic characteristic measurements. In this way, a dimensionless parameter pressure ratio was introduced to evaluate the under-driving, while two methods by using flame thickness and induction time respectively, were proposed to evaluate over-driving. To improve the accuracy of dust explosion tests, authors advocate that explosion severity determination should be conducted at the critical ignition energy. Moreover, a comparison between the European and Chinese flammability limit determination procedures was also conducted, indicating that EN 14034-3 is suitable for light metal but not for carbonaceous, while GB/T 16425 appears to be slightly conservative for both carbonaceous and light metal dusts.     

Fast and tiny: A model for the flame propagation of nanopowders

To avoid the influence of external parameters, such as the vessel volume or the initial turbulence, the explosion severity should be determined from intrinsic properties of the fuel-air mixture. Therefore, the flame propagation of gaseous mixtures is often studied in order to estimate their laminar burning velocity, which is both independent of external factors and a useful input for CFD simulation. Experimentally, this parameter is difficult to evaluate when it comes to dust explosion, due to the inherent turbulence during the dispersion of the cloud. However, the low inertia of nanoparticles allows performing tests at very low turbulence without sedimentation. Knowledge on flame propagation concerning nanoparticles may then be modelled and, under certain conditions, extrapolated to microparticles, for which an experimental measurement is a delicate task. This work focuses on a nanocellulose with primary fiber dimensions of 3 nm width and 70 nm length. A one-dimensional model was developed to estimate the flame velocity of a nanocellulose explosion, based on an existing model already validated for hybrid mixtures of gas and carbonaceous nanopowders similar to soot. Assuming the fast devolatilization of organic nanopowders, the chemical reactions considered are limited to the combustion of the pyrolysis gases. The finite volume method was used to solve the mass and energy balances equations and mass reactions rates constituting the numerical system. Finally, the radiative heat transfer was also considered, highlighting the influence of the total surface area of the particles on the thermal radiation. Flame velocities of nanocellulose from 17.5 to 20.8 cm/s were obtained numerically depending on the radiative heat transfer, which proves a good agreement with the values around 21 cm/s measured experimentally by flame visualization and allows the validation of the model for nanoparticles.     

Methods for more accurate determination of explosion severity parameters

Accurate determination of explosion severity parameters (p_max, (dp/dt)_max, and K_St) is essential for dust explosion assessment, identification of mitigation strategy, and design of mitigation measure of proper capacity. The explosion severity parameters are determined according to standard methodology however variety of dust handled and operation circumstances may create practical challenge on the optimal test method and subsequent data interpretation. Two methods are presented: a statistical method, which considers all test results in determination of explosion severity parameters and a method that corrects the results for differences of turbulence intensity. The statistical method also calculates experimental error (uncertainty) that characterises the experimental spread, allows comparison to other dust samples and may define quality determination threshold. The correction method allows to reduce discrepancies between results from 1 m³ vessel and 20-l sphere caused by difference in the turbulence intensity level. Additionally new experimental test method for difficult to inject samples together with its analysis is described. Such method is a versatile tool for explosion interpretation in test cases where different dispersion nozzle is used (various turbulence level in the test chamber) because of either specific test requirements or being “difficult dust sample”.     

Visualization and analysis of dispersion process of combustible dust in a transparent Siwek 20-L chamber

Dust explosion severities are closely associated with dust dispersion behaviors. To characterize the dispersion process of dust cloud, visualization experiments were conducted by using a transparent Siwek 20-L chamber. Dispersion processes of typical carbonaceous dust were recorded by a high-speed camera and, with the image processing technique, the qualitative analysis based on the transmission of dust cloud was carried out. Results have evidenced the three consecutive stages of dust dispersion process: the fast injection stage of dust particles, the stabilization stage and the sedimentation stage of dust cloud. The motion of dust particles and the variations of dust cloud in space and time can be clearly distinguished. In the stabilization stage, the good uniformity of dust dispersion is achieved when the deviation of transmission data at different locations reaches to the minimum value. Under different nominal dust concentrations, the time periods for dust dispersion stabilization are found to be significantly different, suggesting that different dust concentrations should correspond to different ignition delay in order to accurately measure the explosion characteristics in the Siwek 20-L chamber. Moreover, it is found that the decrease trend of transmission with increasing nominal dust concentration will become gradually leveling off, different from the inversely proportional relationship according to the Bouguer's law, and this indicates that the actual dust concentration will be lower than the nominal concentration or the dust cannot be fully dispersed at the case of high dust concentration. According to the experiment, when the nominal dust concentration exceeds to 1000 g/m³, the transmission will no longer vary visibly.     

Validation of a new formula for predicting the lower flammability limit of hybrid mixtures

A correlation of the lower flammability limit for hybrid mixtures was recently proposed by us. The experimental conditions including ignition energy and turbulence which play a primary role in a gas or dust explosion were at fixed values. The sensitivity of such experimental conditions to the accuracy of the proposed formula was not thoroughly discussed in the previous work. Therefore, this work studied the effect of varying the ignition energy and turbulence intensity to the formula proposed in our previous paper. For ignition energy effect, results from methane/niacin mixture demonstrated that the MEC and LFL will not be affected by changing ignition energy. There is no distinguishable difference among gas explosion index (K_G) and dust explosion index (K_St) derived from tests with every ignition energy (2.5 kJ, 5 kJ and 10 kJ) in a 36 L vessel. The proposed formula is independent of ignition energy. For turbulence effect, the proposed formula can have a good prediction of the explosion and non-explosion zone if the ignition delay time is within a certain range. The formula prediction is good as the ignition delay time increases up to 100 ms in this work. Propane/niacin and propane/cornstarch mixtures are also tested to validate the proposed formula. It has been confirmed that the proposed formula predicts the explosion and non-explosion zone boundary of such mixtures.     

Investigations into the influence of dustiness on dust explosions

A new safety characteristic the “dustiness” according to VDI 2263 – part 9 (Verein Deutscher Ingenieure, 2008) is investigated. Dustiness means the tendency of a dust to form clouds. The paper deals with the influence of the dustiness on vented dust explosions. In order to look into the effects of the dustiness on dust cloud formation and explosion properties experiments and simulations in a vertical dust dispersion glass tube apparatus were carried out.Preliminary explosion experiments showed that the dustiness has an influence on the reduced explosion pressure in a vented 75 L test apparatus. Dusts with comparable p_max and K_St values and different dustiness were tested. Dusts with higher dustiness produced higher overpressures, despite comparable safety characteristics. In order to verify the results for applications in the process industries further tests with different settings are planned as well as industrial scale experiments. Characteristics of the dust such as particle size, density, specific surface area and particle shape, which influence the dispersibility, have been determined experimentally.The Euler/Lagrange and the Euler/Euler approaches are compared for simulating an exemplary dust/air mixture. Especially sedimentation and the ability of the approaches to simulate the tendency of dust to stay airborne were investigated. The Euler/Lagrange approach is better suited for simulating local dust concentrations, particle size distributions and particle forces. It could be used to point out regions of high dust concentrations in a vessel. With the Euler/Euler method it is possible to achieve fast solutions for one specified diameter, but the simulated dust/air mixtures are always more homogenous than in reality. ANSYS CFX version 13 was used in all simulations.