Flat flame experiments

Instrumental methods have become more sophisticated to face these challenges. In particular, Westmoreland and Cool have developed a flame-sampling mass spectrometer that has provided several revelations in terms of relevant molecular intermediates in combustion. " Their setup couples a laminar flat-flame burner to a mass spectrometer. This burner can be moved along the axis of the molecular beam to obtain spatial and temporal profiles of common flame intermediates. By using a highly tunable synchrotron radiation source, isomeric information on selected mass peaks can be obtained. This experiment represents a huge step forward in the utility of MS in combustion studies lack of isomer characterization had previously prevented a full accounting of the reaction species and pathways. 

Among other new methods, tunable laser absorption spectroscopy using infrared diode lasers offers prospects for improved accuracy and specificity in concentration measurements, when a line-of-sight technique is appropriate. The present paper discusses diode laser techniques as applied to a flat flame burner and to a room temperature absorption cell. The cell experiments are used to determine the absorption band strength which is needed to properly interpret high temperature experiments. Preliminary results are reported for CO concentration measurements in a flame, the fundamental band strength of CO at STP, collision halfwidths of CO under flame conditions, and the temperature dependence of CO and NO collision halfwidths in combustion gases. 

Experiments are currently in progress to measure CO and NO concentrations in a flat flame burner by diode laser spectroscopy. Comparative measurements are also being made using microprobe sampling with subsequent analysis by non-dispersive infrared and chemiluminescent techniques. Some preliminary laser absorption results for CO are reported here initial results for NO have been published separately (4). Also reported are initial data for collision halfwidths in combustion gases. 

Carlsson [97] used several different CFD models to simulate flame spread on wood both on a flat wall and in a room/corner experiment. Figure 20.7 gives a comparison of the measured HRR and... 

Stream, is equivalent by symmetry to a one-stream problem in which one of the reactant streams is directed normally onto an infinite adiabatic flat plate at which the no-slip condition is replaced by a zero-stress condition a number of theoretical analyses of this problem have been published (for example, [101], [102], [103], [107], [112], [118]), and nonadiabatic problems also have been studied [112], [118]. The results of these analyses are qualitatively similar to those just discussed, but the multiple-valued dependence of upon k is now found to occur for J > 4 in the symmetric problem by removing the possibility of thermal adjustments occurring in the product stream, the introduction of symmetry strengthens the dependence of the flame temperature on the Lewis number and enables abrupt extinctions to be encountered for many real reactant mixtures that have Lej > 1. There are clear experimental confirmations of this qualitative prediction [114], [120]. As K is increased for reactants with Lcj > 1 to a sufficient extent, the two flames move closer together but experience abrupt extinction at a critical value of k before they reach the stagnation point for reactants with Lcj < 1, the two flames tend to merge at the plane of symmetry prior to abrupt extinction. 

Experiment 4.—A piece of glass rod about 25 centimeters long and 5 millimeters in diameter is needed. Heat the rod in the middle in the ordinary — not flat — Bunsen flame, and when soft draw it out slightly into the shape shown in Fig. 12. Cut it-into two rods by making a slight scratch at the desired point of the narrow... 

The numerical model for n-butane oxidation, by Pitz et al. [228], was used also by Carlier et al. [21] to simulate experimental studies of the two-stage combustion of n-butane at 0.18 MPa on a flat-flame burner and, following this validation, to simulate the ignition delays of n-butane in a rapid compression machine. The numerical studies of the burner experiments were extended by Corre et al. [233]. For simulations of the behaviour on a flat-flame burner the chemical model was computed in an isothermal mode, the experimental one-dimensional temperature profile being introduced as an input parameter. Among the important aims of the tests by Corre et al. [233] was the rationalization of the predicted extent of n-butane consumption throughout the development of the first (cool-flame) and second stages of combustion, with that observed experimentally. The experimental study by Minetti et al. [22, 116] included the detection and measurement of RO2 and HO2 radicals by esr, the one-dimensional spatial profiles of which were simulated by Corre et al. [233],... 

In flat flame laboratory experiments, Shaw and Thomas (28) have shown that addition of fuel nitrogen compounds such as pyridines, amines, and cyanides to CO flames increases the NOa emissions. Values to 1300 ppm NOa were measured with and without molecular nitrogen during low temperature CO combustion (860-1145°K). Both CO-02-argon and CO-O2-N2 mixtures were studied with maximum conversions of about 40-50% of the fuel nitrogen to NOar. 

Potassium acetate (Kalium aceticum pur. Ph. G. III.), differing from sodium acetate (compare page 127). crystallises without water of crystallisation. Nevertheless. for this experiment it must be heated to fusion over a free flame in an iron or nickel dish. The melted salt is poured into a shallow, flat iron or copper dish, in a thin layer. While slill warm it is pulverised as finely as possible, and must be at once transferred 10 a bottle which is to be kept tightly closed. For this expert ment 200 grammes of the salt are fused. 

Low Pressure Flame Apparatus. All experiments were performed on flat, premixed low pressure, 2.7 kPa (20 Torr) C2H2/O2 flames stabilized on a water cooled burner. This 8.6 cm diam burner was constructed of approximately 900, 0.12 cm i.d. stainless steel tubes microbrazed into two stainless steel perforated plates to form a water jacket around the tubes. Gases were metered using calibrated critical flow orifices. The burner was installed in a low pressure vessel pumped by a 140 L s" (300 CFM) mechanical vacuum pump. Unburned gas velocities (298 K and 2.7 kPa) in all cases were 50 cm s . Equivalence ratios from = 1.5 to 4.0 were studied with most emphasis on a sooting ( ) = 3.0 flame. Visible soot emission became apparent at a soot threshold of <() = 2.4 to... 

Combustion occurs with a large number of intermediate steps and even simple processes, such as the ones listed in Table 10.1, occur through dozens of coupled elementary reactions. With computer simulations it is possible to describe the interaction between the reactions, and concentration profiles can be calculated. In order to perform the computer calculations it is necessary to know the rate constants for the individual elementary reactions. Comparisons between theory and experiments are best made for a flat, premixed flame, which in its central part can be considered to have only onedimensional (vertical) variation, allowing computer calculations to be performed comparatively easily. The most important reactions are included in the computer description. In Fig. 10.1 experimental and theoretically calculated concentration curves are given for the case of low-pressure ethane/ oxygen combustion. As examples of important elementary processes we give the reactions... 

---