Francesco Tamanini

Scaling parameters for vented gas and dust explosions

Results of experiments or calculations for vented explosions are usually presented by expressing a term containing the peak (reduced) pressure as a function of a vent parameter. In gas explosions, the reactivity of the system has been typically characterized through an effective burning velocity, uf. In the case of dust explosions, a normalized peak rate of pressure rise, K(=V1/3(dp/dt)max), has been used instead. Depending on the chosen approach, comparisons between systems with the same “reactivity” take different meanings. In fact, correlation formulas resulting from these two approaches imply different scaling between important system parameters. In the case of a constant-uf system, and for sufficiently large vent areas, the reduced pressure, Δpr, is approximately proportional to the square of the peak unvented pressure, Δpm. On the other hand, correlations developed for constant-K systems imply proportionality of Δpr with Δpm raised to a power between −5/3 and −1, with the exact value depending on the assumptions made on the shape of the pressure profile. While the ultimate resolution of the details of the scaling may require recourse to experiments, this theoretical analysis offers a tool for the planning of such experiments and for the interpretation of their results. The paper provides a discussion of these scaling issues with the help of predictions from an isothermal model of vented explosions.

Assessment of Commodity Classification for Sprinkler Protection Using Representative Fuels

In order to assess the classification of commodities for sprinkler protection, fire suppression tests were conducted using representative fuels. The measured heat release rates and applied water fluxes are used to calculate the critical delivered flux (CDF) and develop a suppression correlation for each tested commodity. The heat release rate (HRR) measurements together with the CDF calculations show that the CDF can be obtained within an uncertainty of ±3 mm/min. The CDF values are then combined with the measured convective HRR to project the critical sprinkler discharge flux (SDF) – the minimum water flux above which the fire will be controlled. Comparison of SDF values and full-scale sprinkler test results shows that the SDF is a reasonable choice for ranking commodities. Examination of the calculated CDF values using the suppression correlations suggests that the ranking remains the same for various fuel array heights. Furthermore, the responses to suppression by the exposed plastics are very different from those of the cartoned fuels, indicating that these two types of commodities require different treatment.

Turbulent vented gas explosions with and without acoustically-induced instabilities

The paper presents experimental results on the effect of turbulence and acoustically-induced flame instabilities on the development of vented gas explosions. Data have been obtained in a 1.35-m 3 vessel for stoichiometric methane/air (9.5%) and propane/air (4.0%) mixtures. The parameters varied in the tests include three vent ratios ( A c /V 2/3 =0.034, 0.060 and 0.134) and four levels of turbulence (rms velocity u′ =0, 2.4, 3.8 and 7.3 m/s integral scale from 0.1 to 0.14 m). The turbulence is generated by the rapid injection of the mixture in the vessel. When the vessel walls are lined with a layer of sound-absorbing material, the flame instability is suppressed and the severity of the explosion is no longer affected by the flame acceleration process and pressure fluctuations which are observed in the case of tests performed in the vessel with bare walls. The acoustic instability manifests itself as an exponentially-growing pressure oscillation at the fundamental frequency of the vessel (about 330 Hz). As the oscillations approach their maximum amplitude, higher frequencies appear in the spectrum. Estimates for the rate of growth of the oscillations show it to increase when the turbulent rms velocity exceeds about 3–4 m/s. The relative effect of the instability in enhancing the explosion is greatest the lower the turbulence, with a maximum effect for quiescent conditions. No major differences in, behavior, are apparent in the two stoichiometric mixtures tested, in agreement with published data. These results are intended to assist in the development of models for vented explosions, in practical systems.

Vented gaseous deflagrations: modelling of translating inertial vent covers

Abstract A model of explosion pressure build up in enclosures with translating inertial vent covers is presented. The previous approach, valid for inertia-free vents, is advanced by appending to it a new model of translating inertial vent cover displacement. The model and CINDY code are validated against experiments by Hochst and Leuckel (J. Loss Prev. Process Ind. 11 (1998) 89) in a 50-m 3 vessel with vertically translating covers with surface densities of 42 and 89 kg/m 2 at conditions of initially quiescent and turbulent mixtures. It is demonstrated for the first time that modelling of the vent cover jet effect is crucial for prediction of interdependent pressure-time and cover displacement-time transients, whereas air drag force and cushioning effects are negligible. The model was used further to investigate the influence of vent cover surface density on venting generated turbulence, via comparisons with experimental data of Cooper et al. (Combust. Flame 65 (1986) 1) in a 1-m 3 enclosure with vertically translating covers of various surface densities up to 200 kg/m 2 . The increase of the turbulence factor, i.e. total premixed flame front wrinkling factor, with cover inertia is obtained and explained.

Turbulent unvented gas explosions under dynamic mixture injection conditions

The paper addresses the question of the effect of turbulence on the development of gas explosions by discussing experimental results obtained in a 1.35-m3 spherical vessel. The turbulence is introduced by flow injection into the test volume from a high-pressure source. The evolution of the transient turbulence field is characterized through time-resolved velocity measurements made with a bi-directional probe specifically developed for the application. Data are presented for one (9.5%) methane-air and two (4.0 and 4.8%) propane-air mixtures for turbulence intensities up to 13 m/s, corresponding to an estimated turbulence Reynolds number of 41,900 based on the integral length scale. Values of equivalent turbulent burning velocity calculated from the initial rise of the pressure traces are similar for the three tested mixtures. Maximum rate of pressure rise data, on the other hand, show increasing reactivities of the mixtures in the order 9.5% CH4 to 4.0% and 4.8% C3H8. Furthermore, for each mixture, the maximum rate of pressure rise increases linearly with turbulence intensity. No evidence of quenching is apparent in the data, even though estimated values for the Karlovitz stretch factor indicate, that quenching conditions should have been approached at the high end of the range of turbulence tested.

Vented gaseous deflagrations with inertial vent covers: State-of-the-art and progress

Previous studies on vented gaseous deflagrations with inertial vent covers and related regulatory aspects are examined. The model of turbulent deflagration dynamics, built on energy and mass conservation principles, is developed further to take into account the influence of vent cover inertia. An engineering formula for conservative estimation of the upper limit of vent cover inertia is presented. Similarity analysis has shown that the scaling relationship between the surface density of the cover, w, and the turbulence factor, χ, is wχ3 = const, indicating a significant interrelationship between vent cover inertia and venting‐generated turbulence. Results confirm that turbulence gradually increases after vent opening begins, so that it is possible to increase vent cover inertia significantly. It is demonstrated that instead of widely accepted surface density limits of about 10 kg/m2, values of one/two orders higher, depending on the conditions, could be used for explosion protection with 100% efficiency for large‐scale enclosures.

A correlation for the impulse produced by vented explosions

Abstract This work has addressed the prediction of the impulse and the duration of vented explosions. The practical interest for this issue arises from the fact that the venting of an explosion from an enclosure produces reaction forces which must be calculated in order to properly design the structural supports. The FMRC Isothermal Vented Explosion Model was used as the foundation for the development of a new correlation, to replace those proposed by other studies that were found to be deficient. In addition to being based on a strong theoretical foundation, the correlation suggested by the FMRC model has been validated through extensive comparisons with experimental data, covering a range of dust explosions in vessel volumes from 0.64 to 95 m 3 (22.5 to 3355 ft 3 ). The perspective provided by this result has made it possible to determine the reasons and the extent of the poor performance of previous correlations. In particular, the formula in the German guideline VDI 3673, which is correct only for a particular value of the vent ratio, yields predictions for the explosion duration and, therefore, for the impulse, which are as much as five to eight times those given by the FMRC correlation. On the other hand, an early proposal made by Brunner has been found to be essentially correct. The results achieved by FMRC through this work offer another example of how the judicious use of models can be combined with experimental information to generate significant improvements in aspects of vented dust explosions which are, otherwise, inadequately handled by empirical treatments.

Engineering predictions of explosive layer formation in flammable-liquid spills

A model is described to predict the evaporation of a flammable-liquid spill and the subsequent formation of an explosive layer by mixing of the vapors with surrounding air. The model is intended for use in engineering predictions and is, therefore, designed to be easy to use and to produce answers within a very short time. To achieve these goals, simplifying assumptions have been introduced, the most important being the treatment of the flow field as one-dimensional (i.e., only variations in the vertical direction are taken into account). Special adjustments are made through submodels for effects that would not be addressed with sufficient accuracy by neglecting horizontal gradients. This is done to deal with forced ventilation and to treat the case where the vaporizing pool occupies only a fraction of the floor area of the enclosure. The model has been validated by comparison with available data from vaporizing pools and for diffusing heavy vapors; some of these comparisons are presented. Overall, agreement with experiment is found to be satisfactory and generally on the conservative side. This evaluation has confirmed the suitability of the formulation for use as a component in a comprehensive hazard-analysis approach that includes methods to predict the pressure increase from combustion of the explosive layer. Development of an integrated tool is currently under way.

Explosive Volume Formation Form The Release Of Vapors Or Liquids In Enclosures

Evaluation of the explosion hazard from accidental releases in enclosures involves estimating the flammable volume produced by mixing of the released material with air. Two specific situations are considered in the paper. In the first case dealing with gaseous jet releases, recent results are discussed on the effects of simple obstacles on the flammable volume. These results illustrate how the flammable volume can increase or decrease, depending on the values of dimensionless parameters that define the release and the obstacle. The second case considered is that of the flammable layer produced by the evaporation of a liquid pool. This particular release scenario has been studied using a computational approach based on a one-dimensional model. The paper discusses the rationale behind this approximation and the adjustments that have been found to be necessary to include multi-dimensional effects into the approximate treatment. The common thread among the correlations or calculation methods discussed in the paper is the desire to identify tools that can be of use in engineering applications.

Hydrogen combustion experiments in a 1/4-scale model of a nuclear power plant containment

Hydrogen combustion experiments were carried out in a Froude-modeled 1/4-scale facility, reproducing the geometry of a nuclear power plant containment building, to evaluate the performance of a hydrogen mitigation system based on the controlled combustion concept. Hydrogen burning was initiated by a distributed ignition system using glow plug igniters. The objective of the study was to obtain data on the combustion phenomenology and on the thermal environment produced by the burning of hydrogen as might occur during a recoverable degraded core accident. The thermal environment was characterized through measurements of three quantities: gas temperatures, gas velocities, and radiant heat fluxes. Several highly-instrumented tests were performed in the test facility with the internal geometry adjusted so as to reproduce four possible plant configurations. Following a preassigned release history, hydrogen was introduced at the bottom of the test volume through spargers and vent holes located below the surface of a body of water modeling the suppression pool existing in the plants. Two hydrogen injection histories were used during these tests, simulating two different accident recovery scenarios. The dominant combustion mode observed during the tests involved diffusion flames anchored at the surface of the pool. Other types of combustion, involving lifted diffusion flames and localized burning, were observed at low-oxygen conditions and low hydrogen injection rates, respectively. Other than mild localized initial lightoff burns, deflagrations did not occur in the tests. For oxygen concentrations down to 8%, the distributed ignition system was successful at maintaining the background hydrogen concentration below 4.5–5% by volume.