Measurement of gas detonation blast loads in semiconfined geometry

The ignition and explosion of combustible vapor clouds represents a significant hazard across a range of industries. In this work, a new set of gas detonations experiments were performed to provide benchmark blast loading data for non-trivial geometry and explosion cases. The experiments were designed to represent two different accident scenarios: one where ignition of the vapor cloud occurs shortly after release and another where ignition is delayed and a fuel concentration gradient is allowed to develop. The experiments focused on hydrogen-air and methane-oxygen detonations in a semiconfined enclosure with TNT equivalencies ranging from 9 g to 1.81 kg. High-rate pressure transducers were used to record the blast loads imparted on the interior walls of a 1.8 m × 1.8 m × 1.8 m test fixture. Measurements included detonation wave velocity, peak overpressure, impulse, and positive phase duration. A comparison of the pressure and impulse measurements with several VCE models is provided. Results show that even for the most simplified experimental configuration, the simplified VCE models fail to provide predictions of the blast loading on the internal walls of the test fixture. It is shown that the confinement geometry of the experiment resulted in multiple blast wave reflections during the initial positive phase duration portion of the blast loading, and thus, significantly larger blast impulse values were measured than those predicted by analytical models. For the pressure sensors that experienced normally-reflect blast waves for the initial blast impulse, the Baker-Strehlow and TNT equivalency models still struggled to accurately capture the peak overpressure and reflected impulse. The TNO multi-energy model, however, performed better for the case of simple normally-reflected blast waves. The results presented here may be used as validation data for future model or simulation development.     

Influence of vacuum degree on the effect of gas explosion suppression by vacuum chamber

In order to study the influence of vacuum degree on gas explosion suppression by vacuum chamber, this study used the 0.2 mm thick polytetrafluoroethylene film as the diaphragm of vacuum chamber to carry out a series of experiments of gas explosion suppression by vacuum chamber with the vacuum degree from −0.01 MPa to −0.08 MPa. The experimental results show that: under the condition of any vacuum degree, vacuum chamber can effectively suppress the explosion flame and overpressure; as vacuum degree changes, the effect of gas explosion suppression using vacuum chamber is slightly different. Vacuum chamber has obvious influence on propagation characteristics of the explosion flame. After explosion flame passes by vacuum chamber, the flame signal weakens, the flame thickness becomes thicker, and the flame speed slows down. With the increase of the vacuum degree of vacuum chamber, the flame speed can be prevented from rising early by vacuum chamber. The higher the vacuum degree is, the more obviously the vacuum chamber attenuates the explosion overpressure, the smaller the average overpressure is, and the better effect of the gas explosion suppression is. Vacuum chamber can effectively weaken the explosion impulse under each vacuum degree. From the beginning of −0.01 MPa, the vacuum chamber can gradually weaken explosion impulse as the vacuum degree increases, and the effect of gas explosion suppression gradually becomes better. When the vacuum degree is greater than −0.04 MPa, the increase of vacuum degree can make the explosion overpressure decrease but have little influence on the explosion impulse. Therefore, the vacuum chamber has the preferable suppression effect with equal to or greater than −0.04 MPa vacuum degree.     

An overpressure-time history model of methane-air explosion in tunnel-shape space

This study investigated methane-air explosion in tunnel-shape space and developed an overpressure-time history model based on numerical results. The findings revealed that for the progressively vented gas explosion with movable steel obstacles in a 20 m long tunnel, the inner peak overpressure increased as the activation pressure of the tunnel top cover got higher but remained below 6 bar. However, as the activation pressure increased to 8 bar or higher, the peak inner overpressure remained unchanged. As the segment cover panel became wider, the peak pressure was almost unchanged, but the pressure duration and impulse declined significantly. The peak pressure and impulse increased as the tunnel length vary from 10 to 30 m. With fixed tunnel length, higher blast pressure but lower impulse was observed as the inner obstacles were closer or the activation pressure of obstacles was higher. It is also found that a local enlarged space in the tunnel enhanced the peak pressure significantly. An overpressure time history model for the tunnel with fixed top cover and enlarged end zone was established. The model considered activation pressure of vent cover, area and length of vent opening, methane concentration, number and blockage ratio of fixed obstacles was developed to calculate the overpressure and corresponding time at characteristic points of the pressure-history curve. The cubic Hermite interpolation algorithm and a specially tuned formula consisting of the power and exponential function were used to interpolate pressure values between characteristic points. The proposed model can predict both the peak pressure and the overpressure time history with acceptable accuracy.     

Experimental study on the feasibility of explosion suppression by vacuum chambers

In view of the invalidity of suppression and isolation apparatus for gas explosion, a closed vacuum chamber structure for explosion suppression with a fragile plane was designed on the base of the suction of vacuum. Using methane as combustible gas, a series of experiments on gas explosion were carried out to check the feasibility of the vacuum chamber suppressing explosion by changing methane concentration and geometric structure of the vacuum chamber. When the vacuum chamber was not connected to the tunnel, detonation would happen in the tunnel at methane volume fraction from 9.3% to 11.5%, with flame propagation velocity exceeding 2000 m/s, maximum peak value overpressure reaching 0.7 MPa, and specific impulse of shock wave running up to 20 kPa s. When the vacuum chamber with 5/34 of the tunnel volume was connected to the flank of the tunnel, gas explosion of the same concentration would greatly weaken with flame propagation velocity declining to about 200 m/s, the quenching distance decreasing to 3/4 of the tunnel length, maximum peak value overpressure running down to 0.1-0.15 MPa and specific impulse of shock wave below 0.9 kPa s. The closer the position accessed to the ignition end, the greater explosion intensity weakened. There was no significant difference between larger section and smaller vacuum chambers in degree of maximum peak value overpressure and specific impulse declining, except that quenching fire effect of the former was superior to the latter. The distance of fire quenching could be improved by increasing the number of the vacuum chambers.     

Numerical simulation of blast wave oscillation effects on a premixed methane/air explosion in closed-end ducts

This paper investigated the effects of blast wave oscillation on the overpressure of premixed methane/air explosions by numerical simulations and experiments. The AutoReaGas 3D code and a duct were used in the study. The oscillation induced by the repeat reflection of a blast wave in a closed-end duct was observed by high-speed camera. There was an oscillation zone in the blast wave which exhibited some saw-toothed characteristics. This explained why the overpressure in closed-end ducts was higher than that in open-end ducts. In addition, some of the peak overpressure was even higher than the C–J pressure of 1.76 MPa. The peak overpressures at the ducts' ends were higher than at other measurement points for 5 m, 10 m, and 15 m long ducts. This was mainly due to the reflection of the blast wave. The oscillating period increased with increasing duct length, and could be calculated by t = 0.0003 + 0.00198 L. However, the duct's diameter had no significant influence on the oscillation's period. The amplitude increased with increases in the duct length, except for the case of a 20 m long duct, and increased with decreases in the ducts' diameter. In addition, the peak overpressures in ducts of the same length/diameter ratio were similar. The peak overpressure increased with the increase of the length/diameter ratio, and the maximum value of the peak overpressure in the ducts had the same trend. However, the overpressure did not increase when the length/diameter ratio reached 250.     

Small-scale physical explosions in shock tubes in comparison with condensed high explosive detonations

A series of small-scale experiments involving physical explosions in a 1.6 l pressure vessel was carried out. Explosions were initiated by spontaneous rupture of an aluminium membrane on one side of the vessel at a pressure in the range 1–1.2 MPa. The pressure waves released were measured at different distances along two separate shock tubes, one 10 m long and 200 mm in diameter (closed at one end by the high pressure vessel) and the other 15 m long and 100 mm in diameter.TNT equivalency was used for predicting the blast wave characteristics after vessel rupture. TNT equivalency was used because equations for prediction of peak pressure and impulse of the blast wave in 1-D geometry after detonations of condensed explosives are known. Some experiments with an equivalent amount of real explosive were carried out for comparison with the theoretical and experimental data obtained. The applicability of the TNT equivalency method presented for calculations of maximum pressure and shock wave impulse generated after rupture of the pressure vessel in 1-D geometry is discussed.     

Characteristic overpressure–impulse–distance curves for the detonation of explosives, pyrotechnics or unstable substances

A number of models have been proposed to calculate overpressure and impulse from accidental industrial explosions. When the blast is produced by explosives, pyrotechnics or unstable substances, the TNT equivalent model is widely used. From the curves given by this model, data are fitted to obtain equations showing the relationship between overpressure, impulse and distance. These equations, referred to here as characteristic curves, can be fitted by means of power equations, which depend on the TNT equivalent mass. Characteristic curves allow determination of overpressure and impulse at each distance.     

Attenuation of a point blast shock wave in the dusty air

The behavior of the blast impulse initiated by a point blast in the dusty air is investigated theoretically. It is shown that the jumps of parameters at the shock front in the dusty air follow other regularities in comparison with the case of an ideal gas, beginning from the distance of three dynamic radii, so at ten dynamic radii the difference in overpressure exceeds 60%. When the air heterogeneity is taken into account, substantial gradual changes of wave profile come over and the total blast wave impulse can't be determined by the front overpressure only. The known far asymptotic law takes no place in the point blast flow at the volume dust densities ρ₂₀ > 3·10^?3 kg/m³. In contrast to the ideal gas, the shock front discontinuity vanishes in the dusty air at a finite distance from its origin and the blast wave eventually turns into a dispersive wave without discontinuity. The wave structure changing is studied in the process of the shock wave transformation into the dispersive wave.     

Numerical simulations on the attenuation effect of a barrier material on a blast wave

This paper numerically modeled previous experimental results and quantitatively revealed the attenuation effect of a barrier material on a blast wave. Four fluids were considered in the present study: the detonation products, water, foamed polystyrene, and air. These fluids were modeled by Jones-Wilkins-Lee (JWL), stiffened gas, and ideal gas equations of state. A mixture of water and foamed polystyrene was used as a barrier to encircle a 0.1 kg mass of spherical pentolite, and the interface problem between the barrier and the blast wave was investigated. The simulation parameters were the radius and the water volume fraction of the barrier. To elucidate the effect of the barrier, we conducted two series of numerical simulations; one without a barrier, and another with a barrier of 50 or 100 mm in outer radius and 0–1 in the water volume fraction. Peak overpressure, positive impulse, and pressure history all agreed well with the previous experimental results. We focused on the energy transfer from high-pressure detonation products to other fluids. The sum of the kinetic energies of the detonation products and the barrier induced by the blast wave could quantitatively estimate the attenuation effect of the blast wave and was minimized when the water volume fraction was 0.5, as was the case in the previous experiment.     

Vented confined explosions in Stramberk experimental mine and AutoReaGas simulation

Research Mining Institute, Inc., Ostrava-Radvanice, in cooperation with Dept. of Theory And Technology of Explosives of University of Pardubice and Klokner Institute of CTU in Prague, has performed three series of experiments examining methane–air mixture explosions and their impact on 14 and 29 cm thick wall. The project was named ‘Modeling Pressure Fields Effects on Engineering Structures During Accidental Explosions of Gases in Buildings’ and was sponsored by Grant Agency of Czech Republic (project No. 103/01/0039). The project is aimed at deeper understanding of pressure field effect upon the structures. Methane-air mixture explosion was used to generate the blast wave. The geometrical configuration of the environment resembled a room of an average size, such as larger kitchen. Preliminary simulations were made by AutoReaGas code (Century Dynamics and TNO). The design phase was followed by tests in an experimental mine in Stramberk. Two masonry dams were build in the mine, with cross-section areas of 10.2 m² and longitudinal distance of 5.7 m, creating an explosion chamber with a volume of 58 m³. Two vent openings with an adjustable free cross-section were used to control the maximum overpressure inside the chamber. The concentration of methane-air mixture was approximately 9.5% (vol.) and the volumes of the clouds were 5.25, 10.2 and 15.3 m³ respectively. The generated blast wave overpressures inside the chamber ranged between 1 and 150 kPa. According to experimental results a calibration of the code was performed. After the calibration it is possible to make relatively accurate simulations in similar geometry and to calculate the pressure loading of the structure at any spot in the simulated space. This paper describes the experiments performed and compares experimental and computational results.     

Pressure characteristics and dynamic response of coal mine refuge chamber with underground gas explosion

Coal mine refuge chambers are new devices for coal mine safety which can provide basic survival conditions after gas explosion. In order to simulate the propagation of underground methane/air mixture blast wave, and check structural safety of coal mine mobile refuge chamber, an underground tunnel model and a refuge chamber model have been established based on explicit nonlinear dynamic ANSYS/LS-DYNA 970 program. Results show that the reflected wave pressure on the impact surface was about two times higher than that on the incident one. The relationship between the pressure fields of the chamber was analyzed. The maximum pressure of gas explosion reached about 0.71 MPa, and the pulse width was 360 ms. The maximum absolute displacement and stress occurs at the main door center and the connection of stiffeners and the front plate, respectively. The entire coal mine mobile refuge chamber was in elastic state and its strength and stiffness meet the safety requirements. The cabin door, the front plate and the connecting flange at cabin back as well as the stiffeners on each side were the most critical components. Suggestions were put forward for the refuge chamber.     

The acceleration of flames in tube explosions with two obstacles as a function of the obstacle separation distance

The separation distance (or pitch) between two successive obstacles or rows of obstacles is an important parameter in the acceleration of flame propagation and increase in explosion severity. Whilst this is generally recognised, it has received little specific attention by investigators. In this work a vented cylindrical vessel 162 mm in diameter 4.5 m long was used to study the effect of separation distance of two low blockage (30%) obstacles. The set up was demonstrated to produce overpressure through the fast flame speeds generated (i.e. in a similar mechanism to vapour cloud explosions). A worst case separation distance was found to be 1.75 m which produced close to 3 bar overpressure and a flame speed of about 500 m/s. These values were of the order of twice the overpressure and flame speed with a double obstacle separated 2.75 m (83 characteristic obstacle length scales) apart. The profile of effects with separation distance was shown to agree with the cold flow turbulence profile determined in cold flows by other researchers. However, the present results showed that the maximum effect in explosions is experienced further downstream than the position of maximum turbulence determined in the cold flow studies. It is suggested that this may be due to the convection of the turbulence profile by the propagating flame. The present results would suggest that in many previous studies of repeated obstacles the separation distance investigated might not have included the worst case set up, and therefore existing explosion protection guidelines may not be derived from worst case scenarios.     

An experimental study on the temperature dependence of CO2 explosive evaporation

The need for transportation and storage of CO₂ in bulk quantities is likely to increase in the near future. The handling of CO₂ on such scale gives rise to a number of technological challenges and safety aspects. The accidental rupture of a vessel containing liquefied CO₂ may lead to a Boiling Liquid Expanding Vapour Explosion (BLEVE). Whether explosive evaporation of liquefied CO₂ is also possible at storage temperatures below the homogeneous nucleation temperature 271 K (?2 °C) is unclear.This article describes the results of 12 experiments with 40 L CO₂ cylinders at various temperatures to investigate the temperature dependence of explosive evaporation. The cylinders were opened with linear shaped charges to simulate a near instantaneous rupture, and blast was measured at various locations. The observed blast could be clearly attributed to explosive evaporation. The results show that below the homogeneous nucleation temperature, BLEVE blast does not disappear abruptly, but instead follows a gradual decay. Predictions with a numerical BLEVE blast model overestimate the observed blast peak overpressure and impulse, but qualitatively show a similar behaviour. The energy lost by the acceleration of the cylinder parts is a possible reason for overestimations of the model.The consequence of the test results is that for accident scenarios with CO₂ at low temperatures a BLEVE should not be neglected in hazard assessments. Future large scale bulk storage will take place at a 10⁵ times larger volume than the cylinders applied in the current small scale experiments. We expect that the blast-reducing effects of a tank shell will disappear at such scale.The research leading to these results has received funding from the European Community's Seventh Framework Programme (FP7/2007–2013) under grant agreement n° 241381.     

Propane-air mixture deflagration hazard with different ignition positions and restraints in large-scale venting chamber

In order to better assess the hazards of explosion accidents, propane-air mixture deflagrations were conducted in a large-scale straight rectangular chamber (with a cross-section of 1.5 m × 1.5 m, length of 10 m, and total volume of 22.5 m³). The effect of initial volume, ignition position, and initial restraints on the explosion characteristics of the propane-air mixtures was investigated. The explosion overpressure, flame propagation, and flame speed were obtained and the computational fluid dynamics (CFD) software was used to simulate the flame-propagation process and field flow for auxiliary analysis. The hazards of large-scale propagation explosion under weak and strong constraints were evaluated and the different phases of flame propagation under weak and strong constraints were discriminated. Results indicate that the hazards caused by propane deflagration under weak constraint are mainly caused by flame spread. And the maximum overpressure under strong constraint appeared at the front part of the chamber under the large-scale condition, which is consistent with the previous small-scale test. Moreover, the simulations of flame structures under weak and strong constraint are in good agreement with experimental results, which furthers the understanding of large-scale propane deflagration under different initial conditions in large-scale spaces and provides basic data for three-dimensional CFD model improvement.     

Flame behaviors and pressure characteristics of vented dust explosions at elevated static activation overpressures

Dust explosion venting experiments were performed using a 20-L spherical chamber at elevated static activation overpressures larger than 1 bar. Lycopodium dust samples with mean diameter of 70 μm and electric igniters with 0.5 KJ ignition energy were used in the experiments. Explosion overpressures in the chamber and flame appearances near the vent were recorded simultaneously. The results indicated that the flame appeared as the under-expanded free jet with shock diamonds, when the overpressure in the chamber was larger than the critical pressure during the venting process. The flame appeared as the normal constant-pressure combustion when the pressure venting process finished. Three types of venting processes were concluded in the experiments: no secondary flame and no secondary explosion, secondary flame, secondary explosion. The occurrence of the secondary explosions near the vent was related to the vent diameter and the static activation overpressure. Larger diameters and lower static activation overpressures were beneficial to the occurrence of the secondary explosions. In current experiments, the secondary explosions only occurred at the following combinations of the vent diameter and the static activation overpressure: 40 mm and 1.2 bar, 60 mm and 1.2 bar, 60 mm and 1.8 bar.     

A quantitative correlation of evaluating the flame speed for the BST method in vapor cloud explosions

In recent decades, vapor cloud explosions (VCEs) have occurred frequently and resulted in numerous personnel injuries and large property losses. As a main concern in the petrochemical industry, it is of great importance to assess the consequence of VCEs. Currently, the TNT equivalency method (TNT EM), the TNO multi-energy method (TNO MEM), and the Baker-Strehlow-Tang (BST) method are widely used to estimate the blast load from VCEs. The TNO MEM and BST method determine the blast load from blast curves based on the class number and the flame speed, respectively. To quantitatively evaluate the flame speed for the BST method, the experimental data is adopted to validate the confinement specific correlation (CSC) for the determination of the class number in the TNO MEM. As a bridge, a quantitative evaluation correlation (QEC) between CSC correlation and the flame speed is established and the blast wave shapes corresponding to different flame speeds are proposed. CFD software FLACS was used to verify the quantitative correlation with the numerical models of three geometrical scales. It is found that the calculated flame speeds by the QEC are in good agreement with the simulated ones. A petrochemical plant is selected as a realistic scenario to analyze the TNT EM, TNO MEM, BST method and FLACS simulations in terms of the positive-phase side-on overpressure and impulse at different distances. Compared with the flame speed table, the predicted overpressure from BST curves determined by the proposed QEC is closer to that from FLACS and more conservative. Furthermore, the predicted results of different methods are compared with each other. It is found that the estimated positive-phase side-on overpressure and impulse by the TNO MEM are the largest, and the estimated impulse by the TNT EM is the smallest. Moreover, the estimated overpressure and impulse are larger in the higher reactivity gas.     

The influence of the charge-to-mass ratio of the charged water mist on a methane explosion

To study the influence of the charge-to-mass ratio of a charged water mist on a methane explosion, the induction charging method was used to induce charge on a normal water mist; a mesh target method was employed to test the charge-to-mass ratio of its droplets. The propagation images, propagation average velocities, and overpressures of a methane explosion suppressed by charged water mist were analysed. The influence of the charge-to-mass ratio of the suppressant water mist on a methane explosion was studied. Results show that the explosion temperature, propagation average velocity, and peak overpressure deceased more obviously with charged water mist than ordinary water mist. With increasing charge-to-mass ratio, the suppression effect of the charged water mist underwent a significant increase. Under experimental conditions, compared with ordinary water mist, when the charge-to-mass ratio was 0.445 mC/kg and the mist flux was 4 L, the minimum flame propagation average velocity was 3.456 m/s, with a drop of 2.37 m/s (40.68%), and the peak overpressure of the methane explosion was 10.892 kPa, with a drop of 10.798 kPa (49.78%). The suppression effect is considered from the changes of the physico-chemical properties of the water mist as affected by the applied charge-to-mass ratio.     

Duct-venting of dust explosions in a 20 L sphere at elevated static activation overpressures

Ducts are often recommended in the design of dust explosion venting in order to discharge materials to safe locations. However, the maximum reduced overpressure increases in a duct-vented vessel rather than in a simply vented vessel. This needs to be studied further for understanding the duct-venting mechanism. Numerous duct-vented dust explosion experiments were conducted, using a 20 L spherical chamber at elevated static activation overpressures, ranging from 1.8 bar to 6 bar. Duct diameters of 15 mm and 28 mm, and duct lengths of 0 m (simply venting), 1 m and 2 m, were selected. Explosion pressures both in the vessel and in the duct were recorded by pressure sensors, with a frequency of 5 kHz. Flame signals in the duct were also obtained by phototransistors. Results indicate that the secondary explosion occurring in the duct increases the maximum reduced overpressure in the vessel. The secondary explosion is greatly affected by the duct diameter and static activation overpressure, and hence influences the amplification of the maximum reduced overpressure. Larger static activation overpressure decreases the severity of the secondary explosion, and hence decreases the increment in the maximum reduced overpressure. The secondary pressure peak is more obvious as the pressure accumulation is easier in a duct with a smaller diameter. However, the increment of the maximum reduced overpressure is smaller because blockage effect, flame front distortion, and turbulent mixing due to secondary explosion are weaker in a narrow duct. The influence of duct length on the maximum reduced overpressure is small at elevated static activation overpressures, ranging from 1.8 bar to 6 bar at 15 mm and 28 mm duct diameters.     

Experiments on the influence of pre-ignition turbulence on vented gas and dust explosions

Experiments were performed on the influence of pre-ignition turbulence on the course of vented gas and dust explosions. A vertical cylindrical explosion chamber of approximately 100 l volume and a length-to-diameter ratio (l/d) of 4.7 consisting of a steel bottom segment and three glass sections connected by steel flanges was used to perform the experiments. Sixteen small fans evenly distributed within the chamber produced turbulent fluctuations from 0 to 0.45 m/s. A Laser-Doppler-anemometer (LDA) was used to measure the flow and turbulence fields. During the experiments the pressure and in the case of dust explosions the dust concentration were measured. In addition, the flame propagation was observed by a high-speed video camera. A propane/nitrogen/oxygen mixture was used for the gas explosion experiments, while the dust explosions were produced by a cornstarch/air mixture.It turned out that the reduced explosion pressure increased with increasing turbulence intensity. This effect was most pronounced for small vents with low activation pressures, e.g. for bursting disks made from polyethylene foil. In this case, the overpressure at an initial turbulence of 0.45 m/s was twice that for zero initial turbulence.     

Experimental study on explosion characteristics of ethanol-gasoline blended fuels

In the present work, a series of experiments have been performed to analyze the explosion characteristics of ethanol-gasoline with various blended ratios (0%, 5%, 10%, 15%, 30%, 50%, 70%, 80%, and 100%). A vented rectangular vessel with a cross-section of 100 mm × 100 mm, 600 mm long and a 40 mm diameter vent on the top is used to carry out the experiments. The flame propagation is recorded by a phantom high-speed camera with 5000 fps, while the histories of the explosion overpressure are measured by two PCB pressure sensors and the explosion sound pressure level is obtained by a CRY sound sensor. The results indicate that the maximum overpressure and flame propagation speed increases linearly as the blended ratio increases when the initial volume of blended fuel is 1.0 mL; While the change of explosion overpressure and flame propagation speed shows a trend of decreasing at first and then increasing as the concentration increases to 1.8 mL. It is also found that the peak of the sound pressure level exceeds 100 dB under all tests, which would damage the human's hearing. What's more, relationships between explosion overpressure and sound pressure level are examined, and the change of the maximum overpressure can be reflected to some extent by the measurement of the maximum sound pressure level. The study is significant to reveal the essential characteristic of the explosion venting process of ethanol-gasoline under different initial blended ratios, and the results would help deepen the understanding of ethanol-gasoline blended fuels explosion and the assessment of the explosion hazardous.