Axel Schönbucher

Electronically produced equidensities from time exposures and instantaneous photographs in the investigation of pool flames

Abstract Information contained in the negatives of time exposures and instantaneous photographs of pool flames can be quantitatively evaluated by the use of electronical equidensitometry. In this method, up to 15 colored lines of constant optical density (equidensities) are produced. N -hexane, methanol, n -butane and liquefied natural gas, in pools of varying diameter, as well as a natural gas jet flame, are investigated without influencing the combustion processes. Equidensities obtained from time exposures, instantaneous photographs, and a series of high-speed frames are evaluated. The mathematical relationship between the optical density of the emulsion and the spectral radiance L λ emitted by the flame in the visible region is then derived. Thus the quantitative significance of the equidensities as lines of L λ = const. becomes apparent. The equidensities obtained from time exposures (long-time equidensities) can be related to the time-averaged values of flame properties such as radiance, temperature, flame shape, flame height, and soot concentration. A characteristic dependence on fuel type exists. Furthermore, the minimum sampling time for statistical combustion processes taking place in the flames is easily measured. The equidensities obtained from instantaneous photographs (short-time equidensities) afford a detailed insight into the complicated turbulent flame field. It appears that the eddies are generally elliptical. In hexane and methanol pool flames, the lengths and widths of the largest eddies increase with increasing pool diameter. The dimensions of the smallest eddies, which are almost spherical and in which the transfer of kinetic energy into heat occurs, remain constant. Using equidensities obtained from a series of high-speed frames the dissipation of a single eddy can be visualized in detail. Presently, we are using a high-speed camera to make a film of the equidensities obtained from the pool flames. In this way, the migration velocity, the dynamics of the geometrical dimensions, and the lifetime of the eddies, including their fluctuation quantities, can be determined.

Thermal radiation of di-tert-butyl peroxide pool fires-Experimental investigation and CFD simulation.

Instantaneous and time averaged flame temperatures T , surface emissive power SEP and time averaged irradiances E of di-tert-butyl peroxide (DTBP) pool fires with d=1.12 and 3.4m are investigated experimentally and by CFD simulation. Predicted centerline temperature profiles for d=1.12m are in good agreement with the experimental emission temperature profiles for x/d>0.9. For d=3.4m the CFD predicted maximum centerline temperature at x/d=1.4 is 1440 K whereas the emission temperature experimentally determined from thermograms at x/d approximately 1.3 is 1560 K. The predicted surface emissive power for d=1.12m is 115 kW/m(2) in comparison to the measured surface emissive power of 130 kW/m(2) whereas for d=3.4m these values are 180 and 250 kW/m(2). The predicted distance dependent irradiances agree well with the measured irradiances.

Simultaneous observation of organized density structures and the visible field in pool fires

Organized density structures and the visible field of pool fires were observed simultaneously with a real-time holographic interferometric method. The pool diameters d varied between 1≦d≦10 cm and the fuels studied were methane, LNG, n-butane, n-pentane, n-hexane, n-heptane, diesel/gasoline, liquid paraffin; methanol, ethanol, propanol, cyclohexanol, glycerin; acetaldehyde, acetone, diethyl ether; acetic acid, methyl acetate; benzene. The organized structures depend on the height x above the pool rim, the pool diameter, the fuel supply rate, the equivalence ratio and the fuel type. These structures show mono-, quasi- and nonperiodic behaviour. In the height region Δx1=0 to 8 cm above the fuel surface of an n-hexane pool fire, d=4.6cm, six monoperiodic subprocesses exist with one independent frequency of 3 Hz and two harmonics of 9 and 12 Hz. The independent frequencies of the first quasiperiodic process are 12, 39 and 41 Hz in the height region Δx2=8 to 16 cm. The second quasiperiodic process with the two independent frequencies of 107 and 78 Hz occurs in the height region Δx3=16 to 24 cm. A third quasiperiodic process with at least the two independent frequencies of 177 Hz and 192 Hz exists in the height region Δx4=24 to 32 cm. For the mean frequencies f of the monoperiodic subprocesses the correlation f ¯ ( d ) = b ¯ D B ( d ) d − 0.5 ≈ 1.83 d − 0.63 , (2.0≦d≦2500 cm) is derived, based on the buoyant acceleration bDB of density parcels. A density structure forming concentric ring-shaped interference fringes is defined as a density parcel. With a simplified momentum equation the axial convection velocity of fire parcels, which is up to three times faster than that of the density parcels, was calculated. Instantaneous mass densities and temperatures were calculated by the Abel-inversion, including the concentration profiles of 15 stable fire gas species. From the frequency distributions of the geometric dimensions of density parcels macro-and microscales, which agree with the scales obtained from power spectra, were determined.

Determining the effect of species composition on temperature fields of tank flames using real-time holographic interferometry

Interference fringe fields and the visible flame field of a 50 mm diameter n-hexane tank flame were simultaneously measured using a real-time holographic interferometer with special image optics. An inhouse developed image processing method was applied to the holographic images to calculate the interference fringe order profiles. The effect of species composition on temperature profiles was studied by considering three different cases: using the measured species profiles, using an overall reaction mechanism based on stoichiometric combustion, and by assuming that the flame consists of hot air. The results show that species composition has the largest effect on temperature fields in regions near the flame axis at lower axial distances. In the region of the plume zone, the flame consists primarily of hot air due to the increase in total entrained air.

Mass Burning Rates of Di-tert-butyl Peroxide Pool Fires—Experimental Study and Modeling

Data and predictions for the mass burning rates of di-tert-butyl peroxide (DTBP) pool fires (0.003 m < pool diameter < 3 m) are presented. The mass burning rates of DTBP fires are up to five times higher and are less dependent on pool diameter compared to hydrocarbon pool fires caused by an additional heat release rate due to exothermic decomposition reaction in the liquid phase. This heat release rate is calculated using a first-order reaction kinetic obtained from microcalorimetric measurements. A new model is derived considering the heat release rate due to the decomposition reaction, which is shown to be 40% of the heat release rate radiated to the pool surface. With the presented model, which also includes physical quantities, especially the limiting fuel concentration for upward flame propagation, it is possible to predict the mass burning rates of large DTBP pool fires. The predicted values are in very good agreement with the experiments.

Modelling of the thermal radiation of pool fires

Since the thermal radiation of large-scale fires can be a hazard to surrounding buildings, there is a need for suitable models to describe the thermal radiation of these fires. With the presented radiation model OSRAMO II, which is experimentally validated, it is possible to calculate the mean surface emissive power of pool fires with pool diameters 8 m ≤ d ≤ 25 m. Direct numerical simulations of a small-scale ethylene pool fire show that the periodical rise of vortex structures like soot parcels and hot spots has a major influence on the temporal variation of flame temperature, soot volume fraction and volume emissive power. Furthermore, the modelling of a large-scale kerosene pool fire shows a time depending temperature field due to the periodical rise of vortices.

Static and dynamic radiance structures in pool fires

Digital image analysis of pool fire photograms and flame radiance measurements, time-averaged and instantaneous, was applied in observing static and dynamic radiance structures in the visible and infrared spectral ranges. Small-view-angle pyroelectric radiometer measurements as well as measurements of instantaneous radiances with Si- and Ge-photodiodes and of flame temperatures, concentrations of stable species and gas flow velocities were carried out. Time-averaged fields of radiances L ¯ λ ( y , x ) are defined as static equidensitometric structures consisting of symmetric W- or M-shaped core structures and elliptical plume structures. The fuel type and the pool diameter change only the geometric part of the core and plume structures. The structures L ¯ λ ( y , x ) in the visible spectrum represent in good approximation the spatial distribution of the radiances L ¯ r a d ( y , x ) in the infrared spectrum. Volumetric emission coefficient Ψ ¯ λ ( r , x ) are determined by Abel-inversion from L ¯ λ ( y , x ) . A simplified model, based on the assumptions of the optical thin limit and inhomogeneities in temperatures and species concentrations is presented to calculate Ψ ¯ λ ( r , x ) . The dynamic radiance structures are totally unsymmetric and show monoperiodic and nonperiodic properties. Formation frequencies for fire parcels, fire mushrooms, soot parcels and hot spots as well as oscillation frequencies for visible fire shapes are determined. Spatial power spectra are calculated from instantaneous radiance profiles by Fourier analysis. The macroscales determined depend strongly but the microscales only weakly on the height x above the fuel surface. It seems that the microscales reach a lower limit of 5 mm.

Measurement and Prediction of the Inert Gas Influence on Explosion Limits for Ethylene/Nitrogen/Air and Ethylene/Carbon‐Dioxide/Air Mixtures at Elevated Pressures

To determine the explosion limits of combustible gas/inert-gas/air mixtures at elevated initial pressures and/or temperatures, a high experimental effort is necessary. No standard test method exists and values are rarely to find in the literature. In order to reduce the large number of experiments, models suitable for simulating the inert gas influence on explosion limits at elevated pressure have been developed. In this work, an experimental method to determine the explosion limits at elevated pressures and the simulation of the inert gas influence of nitrogen and carbon dioxide on the explosion limits of C 2 H 4 /air mixtures are presented.

Simulation of a reaction network in a semibatch reactor

With regard to process safety of technological processes, chemical reactions were assumed to be single reactions. After a typical runaway incident (IC-MESA, Seveso/Italy, 1976), the technical trichlorophenol synthesis was investigated thoroughly. Using analytical, reaction-kinetic and thermodynamic data obtained in cooperating with Boehringer Ingelheim, the authors have established a kinetic model for the TCP synthesis to illustrate simulations of the normal-operating semibatch reactor as well as of other selected incidents.

Simulation of the Base-Catalyzed Ethoxylation of Methanol in the Semibatch Reactor

A model for base-catalyzed ethoxylation of methanol in the semibatch reactor was established. The applied method for the parameter detn. was described. Simulations were carried out to find optimized operating conditions using the limiting operating conditions for max. possible selectivity (output). Substantial deviations between the simplified single reaction model (SRM) and the reaction network model (RNM) were shown to occur in the high sensitivity range (jacket temp. TM 85 Deg, SRM gave lower calcd. max. pressure and temp. than those predicted by RNM. These deviations were classified as dangerous from the point of safety engineering.