WATER GAS FLAME*

N. N. Norkin

I. Introduction

Very little attention has been given in the literature to questions concerning the combustion of water gas; our investigation, carried out with reference to the conditions of industrial methods of burning gaseous fuel, represents a comprehensive treatment of the principal forms of combustion of water gas at constant pressure.

Interest in the study of the water-gas flame is due to the following circumstances: first, to the fact that water gas is a physically homogeneous gaseous mixture consisting chiefly of carbon monoxide, hydrogen, an insignificant amount of methane, and noncombustible impurities of carbon dioxide and water vapor, whereas the thermokinetic properties of the combustible constituents (flammability limits, flame-propagation velocities, etc.) differ sharply; therefore, clarification of the possible influence in the flame of the individual combustible constituents is a matter of current importance, since up to now there have been differing viewpoints on the question of the preferential burning-out of particular constituents present in the mixture. Secondly, there is no doubt that the principal forms of gas combustion—namely, diffusion combustion, combustion of partially prepared combustible mixtures (Bunsen combustion), and catalytic surface (flameless) combustion—are mutually transformable. Therefore a comprehensive investigation of the combustion of water gas is of considerable interest, since it makes it possible to reveal the general properties of water-gas combustion and the characteristic differences inherent in the individual methods of burning. Thirdly, finally, the water-gas flame, as a heat source at one or another temperature level, is used in many cases of production activity, such as in welding burners, in atmospheric burners for flame treatment of parts (for example, at electrode plants), in laboratories, etc. Therefore the quantitative data of the investigation are of direct industrial interest and may be used in the design and regulation of gas-burning devices.

The author carried out the following investigations:

1. The composition of water gas obtained in industrial gas generators during the gasification of semi-anthracite coals. The influence of saturated and superheated steam on the composition and calorific value of the gas was established, as well as the influence of the production regime of gasification.

2. The structure of the water-gas flame. By separating the two-cone flame obtained in partially prepared mixtures, the composition of the intercone gases was studied. In burning combustible mixtures in burners of the Bunsen and Meker types, and in capillary porcelain and platinum tubes, it was estab—

* The results of the investigation were reported by the author at the section on combustion of gaseous fuel of VNITOÉ, Moscow, 24/XI, 1949.

the combustion order of the constituent parts of the gas, the distribution of the oxygen of the primary air among the constituent parts of the gas, and the influence of catalysts.

3. The water-gas flame in partially prepared combustible mixtures. The influence of the composition of combustible mixtures, burner diameter, and outflow regime on the flame length and on the distribution of temperature and heat release in the flame was investigated.

4. The diffusion flame of water gas. The diffusion combustion of water gas was investigated in flat-slot atmospheric burners and in a screened furnace chamber, as well as in an industrial glass-melting tank furnace built at one plant according to the author’s design. The influence of the outflow velocity on the length of the diffusion flat-slot flame was established for equal and unequal gas and air outflow velocities, as were the temperature distribution along the flame length and the heat transfer in a closed furnace chamber.

5. Surface (flameless) combustion of water gas. Along with the development of a general conception of the conditions for the transformation of flame combustion into flameless combustion, owing to deformation of the normal conical flame front under the influence of hydrodynamic and thermokinetic factors, an experimental investigation was carried out of the conditions for the formation and stable development of flameless surface combustion of water gas in a granular bed and a channel burner.

The investigation was carried out in the Laboratory of Thermal Apparatus of the Tomsk Polytechnic Institute of the Order of the Red Banner of Labor. At various times the following scientific and technical staff members of the laboratory were involved in carrying out the work: G. D. Spetsts, A. A. Starikov, N. D. Vylegzhanin, T. Ya. Vaganova.

II. Composition of Water Gas

The fuel used is hard coal from the Aralichevskoye deposit (Kuzbass); the coal is lean, partially semi-anthracitic, and non-caking.

The average composition of the coal is: \(W^{p}=9.0\%\); \(A^{c}=16.0\%\); \(S_{\mathrm{total}}^{o}=0.8\%\); \(C^{r}=88.1\%\); \(H^{r}=4.0\%\); \(N^{r}=1.7\%\); \(O^{r}=5.3\%\); the volatile content averages \(V^{r}=12.0\%\).

The technological regime of gasification is characterized by the following data: the period of air (hot) blast is from 2 to 3 min., the period of steam (cold) blast is from 3 to 4 min. The period of steam blast (gasification) amounts on average to 64% of the time of the complete cycle, and the ratio of the gasification time to the air-blast time averages 1.76. Before charging, the coal is screened through a 25 mm sieve: fines smaller than 25 mm are burned in the furnace of a steam boiler, and coarse coal larger than 25 mm is charged into the gas producer; the coal charged into the gas producer is nonuniform in granulometric composition. The coal is charged in large portions weighing approximately 300–500 kg. The gasification rate averages \(150\ \mathrm{kg}/\mathrm{m}^{2}\mathrm{h}\). The amount of ash and slag (riddlings) averages 18% of the charged coal; the gases of the air blast are not used and are discharged into the atmosphere.

The water gas obtained during the steam-blast period has a temperature of about \(900^{\circ}\mathrm{C}\) at the outlet from the gas producer.

Until November 1947, saturated water vapor at a pressure of 4 atm was used during the steam-blast period.

The average composition of water gas obtained in the interaction of saturated steam was as follows: CO—37.13%; CO₂—4.57%; O₂—0.19%; H₂—52.3%; H₂O—2.3%; N₂—3.21%. The calorific value of water gas of the indicated composition was

\[ Q_n^p = 2470\ \text{kcal}/\text{m}^3. \]

Beginning in November 1947, superheated steam at a pressure of 4 ata and a temperature of 325°C began to be used for steam blowing. As a result of using superheated steam, the composition and calorific value of the water gas changed considerably. Table 1 gives the composition of the water gas obtained when superheated steam was used.

Table 1

Composition and calorific value of water gas obtained with the use of superheated steam

Name		Average, %	Minimum, %	Maximum, %
Chemical composition of dry water gas, in % by volume	CO₂	4.8	3.8	6.4
Chemical composition of dry water gas, in % by volume	O₂	0.4	0.1	0.6
Chemical composition of dry water gas, in % by volume	CO	29.0	24.1	36.0
Chemical composition of dry water gas, in % by volume	H₂	52.2	42.2	57.5
Chemical composition of dry water gas, in % by volume	CH₄	6.2	2.5	8.6
Chemical composition of dry water gas, in % by volume	N₂	7.3	1.3	10.2
Calorific value, kcal/m³		2760	2535	2930

Gas analyses were carried out with a VTI gas analyzer equipped with a platinum capillary for afterburning. The CO₂ content was determined by absorption with alkali, the oxygen O₂ content by an alkaline pyrogallol solution, heavy hydrocarbons CₘHₙ by bromine water, carbon monoxide CO by an ammoniacal solution of cuprous chloride, and the methane CH₄ and hydrogen contents by combustion in a capillary followed by absorption of CO₂ and O₂.

Comparing the composition of water gas obtained from the interaction of fuel with saturated steam and with superheated steam at the same pressure, we conclude that the calorific value of the water gas increases when superheated steam is used, from 2470 kcal/m³ to 2760 kcal/m³ (on average), owing to the increase in the methane CH₄ content in the gas.

The moisture content in the water gas was determined by passing the gas through two consecutively connected U-shaped tubes filled with CaCl₂. From the increase in the weight of these tubes when a certain amount of gas was passed through them, the water-vapor content was determined. On average, the water-vapor content in the gas was 5.2 g/m³, or 6.92 g/kg, which approximately corresponds to a water-gas saturation temperature of 45°C; in reality, however, the temperature of the water gas does not exceed

was higher than 18–20°C; consequently, the moisture in the water gas is present both as vapor and as a finely dispersed mist.

The content of hydrogen sulfide, \( \mathrm{H_2S} \), in the water gas was determined by precipitation with acetic-acid cadmium; the cadmium sulfide precipitate was titrated with a 0.1 N iodine solution, followed by back-titration of the excess iodine with hyposulfite in the presence of starch. For this purpose Drechsel bottles were used, connected in series with gas meters. It was found that the hydrogen sulfide content (calculated as sulfur) varies within considerable limits, namely: the minimum sulfur content is \(0.12\ \mathrm{g/m^3}\), the maximum sulfur content is \(1.15\ \mathrm{g/m^3}\), and the most frequently encountered sulfur content is \(0.45\ \mathrm{g/m^3}\). It should be noted that at the plant producing water gas there is no general purification of the gas from hydrogen sulfide, but at some points of consumption local installations have been set up for purifying the gas from hydrogen sulfide by means of bog ore. With proper maintenance of the local purification installations, the hydrogen sulfide content in the purified gas was reduced to \(0.08\ \mathrm{g/m^3}\).

The tar content in the water gas was determined by passing the water gas through two U-shaped glass tubes connected in series and filled with cotton wool. Since, together with the tar, the cotton also absorbs the water vapor present in the gas, after a definite quantity of gas (from 100 to 1000 liters) had been passed through the absorption tubes, the latter were heated for 1.5–2 hours in an electric drying oven at a temperature of 105°C and then weighed. Constancy of the weight of the tubes after heating served as an indication that moisture had been removed from the cotton. The tar content in the gas depends on the point at which the gas samples are taken and on the gasification regime. Thus, in samples taken directly at the outlet from the gas generator, the tar content varies from \(0.1\ \mathrm{g/m^3}\) to \(0.56\ \mathrm{g/m^3}\), depending on the time of gasification. The minimum tar content corresponds to the gasification period before a fresh portion of coal is charged into the gas generator, when the generator contains almost pure coke; the maximum tar content corresponds to the gasification period immediately after a fresh portion of fuel has been charged, when the most intense release of volatiles from the fuel takes place.

In samples taken after gas purification in a scrubber irrigated with water, a tar content of \(0.05\ \mathrm{g/m^3}\) to \(0.19\ \mathrm{g/m^3}\) was found, depending on the time of gasification. In samples taken from the gas pipeline of the plant’s chemical laboratory and in the institute laboratory, i.e., after the gas had passed through a fairly long and branched network of gas pipelines, the tar content in the water gas was more or less constant, independent of the time of gasification; it was on average \(0.04\ \mathrm{g/m^3}\). The tar is a very viscous mass of dark-brown color. Its specific gravity, determined with a pycnometer at 20°C, is \(1.24\ \mathrm{g/cm^3}\). The content of free carbon in the tar averages 11.8%; the moisture content in the tar is about 30%. Fractional distillation established the following group composition of the tar: the first fraction, up to 180°C, amounts to 36%, including: water 32.7%, light oil 3.3%; the second fraction, in the range 230–300°C—heavy oil 2.6%; the third fraction, in the range 300–320°C—the first anthracene fraction 3.7%; and the fourth fraction, in the range 320–345°C—the second anthracene fraction—9.2%; and finally, pitch in the amount of 48.5%. The considerable yield of cyclic compounds (the anthracene fraction) and the large residue of pitch indicate that the tar is secondary and high-temperature.

Although the quantity of tar in the water gas entering the burners is small, averaging \(0.04\ \mathrm{g/m^3}\), purification of the water gas from tar is urgently necessary, since atmospheric burners with small orifice diameters (about 1 mm and less), over a short

over time, they become clogged with tar; the phenomenon of “running away” of the flames is observed, i.e., a sharp shortening of the flame length and, finally, complete extinction of the gas burner. However, no purification of the gas from tar is carried out, apart from preliminary scrubber treatment of the hot gas with water.

III. Structure of the Water-Gas Flame

When partially prepared combustible mixtures containing air in an amount insufficient for complete combustion are burned, a two-cone flame jet is formed, with two combustion fronts. On the surface of the inner cone, that quantity of gas burns which is supplied with the oxygen of the primary air; on the surface of the outer cone, the remaining portion of the gas burns at the expense of the diffusion of oxygen from the surrounding medium (Fig. 1).

The heights of the inner and outer combustion cones depend on the initial composition of the combustible mixture, on the velocity and mode of flow, and on the diameter of the burner orifice.

The purpose of our investigation was to determine the influence of the amount of primary air on the structure of the water-gas flame, which has important practical significance in solving the problem of organizing a flame jet with atmospheric burners.

Figure labels:
Outer cone: combustion at the expense of oxygen diffusing from atmospheric air.
Inner cone: combustion at the expense of the oxygen of the primary air.
Combustible mixture.

Fig. 1. Two-cone flame jet.

The investigation of the structure of the water-gas flame was carried out with separation of the two-cone flame; by chemical analysis, the composition of the water gas, of the gas-air mixture supplied to the burner, and of the intercone gases was determined. The following were investigated:

1. The flame of a Bunsen burner with ejection of primary air from the surrounding atmosphere and with forced supply of gas and air into the burner.
2. The flame of a Meker burner with a platinum gauze.
3. Combustion of a gas-air mixture in porcelain and platinum capillary tubes.

The method of investigation, a description of the experimental apparatus, and the materials obtained in the investigation for each series of experiments have been published in a special article [3].

In the present article, the generalized results and conclusions are given, presented in the form of Table 2 and in Fig. 2.

As an indicator of the relative “combustibility” of any gas, for example hydrogen, one may use the value of the ratio

\[ \gamma = \frac{\text{relative content of hydrogen in the combustible mixture}} {\text{relative content of hydrogen in the combustion products}} = \frac{\beta_{\text{g.m.}}}{\beta_{\text{comb.prod.}}} \tag{1} \]

Table 2

Results of the study of the structure of the water-gas flame

Combustion conditions / chemical composition, percent by volume	Tubular burner, primary-air fraction 0,345, combustible mixture	Tubular burner, primary-air fraction 0,345, combustion products	Tubular burner, primary-air fraction 0,475, combustible mixture	Tubular burner, primary-air fraction 0,475, combustion products	Tubular burner, primary-air fraction 0,580, combustible mixture	Tubular burner, primary-air fraction 0,580, combustion products	Burner with platinum gauze, primary-air fraction 0,580, combustible mixture	Burner with platinum gauze, primary-air fraction 0,580, combustion products	Burner with platinum gauze, primary-air fraction 0,62, combustible mixture	Burner with platinum gauze, primary-air fraction 0,62, combustion products	Burner with platinum gauze, primary-air fraction 0,88, combustible mixture	Burner with platinum gauze, primary-air fraction 0,88, combustion products	Burner with platinum gauze, primary-air fraction 0,96, combustible mixture	Burner with platinum gauze, primary-air fraction 0,96, combustion products	Porcelain tube, primary-air fraction 0,53, combustible mixture	Porcelain tube, primary-air fraction 0,53, combustion products	Platinum tube, primary-air fraction 0,39, combustible mixture	Platinum tube, primary-air fraction 0,39, combustion products
CO₂	2,2	7,4	1,8	7,6	2,0	7,0	2,4	7,7	1,7	10,2	1,2	8,8	1,6	14	2,9	7,2	2,5	8,18
CO	18,8	19,2	14,0	16,6	12,5	10,9	11,5	19	11,4	10,2	8,4	14,8	7,7	1,5	16,2	14,1	12,9	17,35
H₂	27,8	19,8	24,5	19,4	22,2	19,6	20	15,8	20,2	7,5	16	11,7	16,3	1	22,3	19,6	27,3	19,3
CH₄	2,8	3,4	2,4	1,8	1,8	1,2	2,8	2,7	2,5	2,4	2,4	3,2	1,7	0,3	1,1	0,7	3,4	2,47
O₂	10,0	—	11,4	—	12,1	—	12,7	—	12,8	—	14,6	—	14,8	—	11,5	—	10,6	—
N₂	39,3	50,2	45,9	54,6	49,6	61,3	50,6	54,8	51,4	69,7	57,4	61,5	57,9	83,2	46,0	58,5	43,3	52,7
$\beta=\dfrac{\mathrm{H}_2}{\mathrm{CO}+\mathrm{CH}_4}$	1,28	0,88	1,49	1,05	1,53	1,62	1,40	0,73	1,46	0,59	1,47	0,65	1,8	0,55	1,2	1,34	1,66	0,98
$\gamma=\dfrac{\beta_{\mathrm{c.\,mix.}}}{\beta_{\mathrm{comb.\,prod.}}}$	1,45	1,45	1,42	1,42	0,95	0,95	1,91	1,91	2,48	2,48	2,28	2,28	3,28	3,28	0,90	0,90	1,70	1,70
Oxygen distribution during combustion, percent															for CO—35,5; H₂—47,7; CH₄—16,8	for CO—35,5; H₂—47,7; CH₄—16,8	for CO—11,5; H₂—64,0; CH₄—24,5	for CO—11,5; H₂—64,0; CH₄—24,5

As is evident from Table 2, for the tubular burner, when the proportion of primary air in the combustible mixture is increased from \(\alpha_1 = 0.345\) to \(\alpha_3 = 0.580\), the value of

\[ \gamma \text{ decreases: } \gamma_1=\frac{1.28}{0.88}=1.45;\quad \gamma_2=\frac{1.49}{1.05}=1.42 \text{ and } \gamma_3=\frac{1.53}{1.62}=0.95. \]

It follows from this that when the amount of primary air in the combustible mixture is less than \(\frac{1}{2}\) of the theoretically required amount, hydrogen “burns out” preferentially in comparison with CO and CH\(_4\); but when the proportion of primary air is greater than \(\frac{1}{2}\), the preferential “combustibility” of hydrogen is not preserved and gives way to CO and CH\(_4\). If the combustible mixture is burned on a platinum gauze, then, as the proportion of primary air increases, the relative “combustibility” of hydrogen increases.

Comparing the “combustibility” indices of hydrogen when combustible mixtures of approximately the same composition are burned in a porcelain tube and in a tubular burner without a platinum gauze, we come to the conclusion that the “combustibility” of hydrogen in both cases is almost the same and, consequently, the influence of porcelain as a catalyst on selective combustion is imperceptible. Comparing the “combustibility” indices when combustible mixtures of approximately the same composition \((\alpha_1 = 0.345\) and \(\alpha_2 = 0.39)\) are burned in a platinum tube and in an ordinary tubular burner, we come to the conclusion that the relative “combustibility” of hydrogen is increased in the presence of platinum even at a small proportion of primary air.

Fig. 2. Relative combustion intensity of H\(_2\) and CO.

In the figure: vertical axis—\(\zeta = \dfrac{\left(\frac{H_2}{CO}\right)\text{ in the combustible mixture}}{\left(\frac{H_2}{CO}\right)\text{ in the combustion products}}\); curve 1—“without catalyst”; curve 2—“in the presence of a catalyst”; horizontal axis—\(\alpha\), proportion of primary air in the combustible mixture.

Our study also shows that, in combustible mixtures containing H\(_2\), CO, and CH\(_4\), the combustion of the constituents proceeds not selectively in succession, but simultaneously, though with different intensity. If combustion proceeded selectively in succession, then, upon combustion of mixtures containing primary air in an amount greater than \(\frac{1}{2}\) of that theoretically required, hydrogen should have been absent from the combustion products; and upon combustion of mixtures with a smaller proportion of primary air, CO and CH\(_4\) should have remained unchanged, since

the amount of air required for complete combustion of \( \mathrm{H_2} \), corresponds to \( \alpha = \dfrac{1}{2} \).

As an index of the intensity of combustion of \( \mathrm{H_2} \) in comparison with CO, the following ratio may be used:

\[ \frac{ \left(\dfrac{\mathrm{H_2}}{\mathrm{CO}}\right)\ \text{in the combustible mixture} }{ \left(\dfrac{\mathrm{H_2}}{\mathrm{CO}}\right)\ \text{in the combustion products} } = \delta \tag{2} \]

In Fig. 2 the curves \(\delta = f(\alpha)\) are shown on the basis of the results presented in Table 2. Curve 1 shows the change of \(\delta = f(\alpha)\) in the absence of a platinum catalyst, and curve 2 shows the change of \(\delta = f(\alpha)\) in the presence of a platinum catalyst.

From the character of curve 1 we see that in combustible mixtures of composition \(\alpha < \dfrac{1}{2}\), the combustion intensity of \( \mathrm{H_2} \) is approximately 1.5 times greater than the combustion intensity of CO; with an increase in \(\alpha\), the combustion intensities of \( \mathrm{H_2} \) and CO are almost identical (\(\delta = 0.98\)); consequently, the “chemical affinity” of oxygen for \( \mathrm{H_2} \) and CO under the conditions of a stationary flame front is almost the same. Confirmation of this may be provided by comparing the values of the free energy of formation of \( \mathrm{CO_2} \) from CO and \( \mathrm{H_2O} \) from \( \mathrm{H_2} \): \((\Delta F)_{\mathrm{CO_2}} = -61400\ \text{kcal/mole}\) and \((\Delta F)_{\mathrm{H_2O}} = -54507\ \text{kcal/mole}\).

Curve 2 shows that, in the presence of a platinum catalyst, the combustion intensity of \( \mathrm{H_2} \) exceeds that for CO by 1.9–3.3 times as the proportion of primary air in the combustible mixture increases.

IV. Flame of Water Gas in Partially Prepared Combustible Mixtures (Atmospheric Burners)

The combustion of gaseous fuel in burners with partially prepared combustible mixtures is quite widespread in various fields of technology. For example, in burners used in the flame treatment of parts and in the assembly of electric lamps, in soldering and welding burners, in injection gas burners, in household gas apparatus, in laboratory practice, etc. In all these cases, a prepared combustible mixture consisting of fuel gas and air is supplied to the burner, or is formed with its aid to one degree or another. Usually, the combustible mixture contains air in a smaller amount than is required for complete combustion of the gas according to stoichiometric calculations of the combustion process, i.e., in combustible mixtures the excess-air coefficient is less than 1. The air in the combustible mixture is primary. The amount of air lacking for complete combustion either diffuses from the surrounding atmosphere, as occurs in atmospheric burners, or is additionally supplied as secondary air to the gas-combustion chamber.

At the surface of the front, more precisely at a certain very small depth of the flame front (fractions of a mm), the chemical act of combustion takes place; in the inner front the gas burns with the primary air, and the amount of gas burned depends on the amount of primary air. In the outer flame front, the unburned gas burns out with air diffusing from the surrounding medium.

In the present work, the many-sided question of the influence on the length of the inner and outer cones of the flame of the burner diameter, the efflux velocity, the flow regime, and the composition of the combustible mixtures has been investigated. Simultane-

namely, an attempt was made to establish certain ideas about the normal (fundamental) propagation velocity of the water-gas flame by measuring the height (length) of the cone and making subsequent calculations; in doing so it was assumed that the surface of the flame front has a regular conical shape.

In experiments on the study of flame length and temperature distribution, the quantities of gas and air were determined with the aid of gasometers (Fig. 3). Both gasometers were of identical design, each with a working capacity of 17 liters. Each of them is arranged as follows: the lower vessel “A” is hermetic; it can be completely isolated from the atmosphere. The diameter of the vessel is 280 mm, the height 500 mm. At the bottom it has a pipe “d,” closed tightly by a lid with a gasket, which serves for draining water from the vessel. In the upper part of the vessel there is a small cock “C” for discharging the gas or air collected in it. The vessel is provided with a water-gauge glass “n” with a scale and an outlet “K” for connecting a manomete-

Fig. 3.

On the lid of vessel “A,” on four legs (tubes), the upper pressure vessel was fastened. The diameter of this vessel was the same as that of the lower one; its height was 870 mm.

The lower vessel communicates with the upper one by means of cock “b,” installed in the central tube, and cock “a,” installed in one of the side tubes, the lower end of which is lowered almost to the very bottom of the lower vessel.

In order to maintain the gas and air pressure constant during each experiment, an auxiliary indicating device was made, consisting of a cylindrical float “g” and a light vertical rod “m” with two pointers.

With cock “a” (and “e”) open during the experiment, the inflow of water into the upper pressure vessel was regulated in such a way that the difference between the water levels in the upper and lower vessels remained constant; both pointers had to coincide with the water menisci in the two water-gauge glasses (upper and lower). The pressure was thus maintained constant, which could also be judged from the readings of the manometers.

Before the beginning of an experiment, both gasometers were filled: one with gas, the other with air; first the cocks “a” and “c” of the gas gasometer were opened. The gas passed through the nozzle of the glass injector “l” of the mixer to the burner “f” and was ignited; then the corresponding cocks of the air gasometer were opened.

Regulation of the quantities of gas and air flowing out was carried out by cocks or clamps on the rubber tubes at the mixer.

After a flame of the desired kind had been obtained, the experiment was begun. At the same time, the water levels in the water-gauge glasses of both gasometers were noted, and two stopwatches were started, for determining and checking the duration of the experiment. Then measurements were made of the temperature of the flame plume and of its height. The temperature was measured by means of a Pt—PtRh thermocouple at several points of the plume: at the top of the inner cone, at three points located at equal distances from one another along the height of the lateral surface of the inner cone, and at three points along the height of the outer cone (along its axis).

Fig. 4

The length of the flame plume was measured by means of a special device consisting of two mirrors 1—1, arranged at an angle of 45° to each other, placed in a metal chamber 2 with darkened walls (Fig. 4).

Thanks to the darkening, the image of the flame appeared clearer; in addition, the walls of the chamber, protecting the flame from the influence of air fluctuations, helped make more accurate readings.

A metal scale was attached to the burner (3). In a mirror placed opposite the chamber opening, the image of the scale and, against its background, the flame could be observed. Owing to this, it was easy to take readings without taking into account the distortion of the image dimensions.

The experiments were carried out with burners of various diameters at different outflow velocities and mixture compositions. The amount of combustible mixture flowing out was determined from the volumes of gas and air issuing from the gas meters during the time of the experiment, as measured by a stopwatch.

The outflow velocity of the mixture from the burner mouth, the coefficient of excess air, and the flame-propagation velocity were then determined by calculation.

The composition of the combustible mixtures varied from 15.8% to 73.2% water gas in the combustible mixture, which corresponds to an excess-air coefficient from \(\alpha = 2.13\) to \(\alpha = 0.14\). The outflow velocities of the combustible mixtures varied from \(1.58\ \text{m/sec}\) to \(123.3\ \text{m/sec}\), and the Reynolds numbers from 738 to 9250. The combustion of the combustible mixtures was carried out by means of tubular burners of diameter \(0.8\)—\(1.25\)—\(2.9\)—\(3.7\)—\(5.0\)—\(8.0\ \text{mm}\) without an igniting pilot belt, and of diameter \(1.18\)—\(1.54\ \text{mm}\) with an igniting pilot belt. Such a burner is shown in Fig. 5.

The amount of combustible mixture passing through the pilot belt is 12% of the total combustible mixture. With burners without a pilot belt, 57 experiments were carried out, and with burners with a pilot belt, 29 experiments.

[In Fig. 5: 1. copper tube; 2–4. nozzle; 3. iron burner; 5. side openings.]

Fig. 5. Burner with ignition

On the basis of data on the measurement of the height of the inner cone of the flame, according to the well-known dynamic method of V. A. Mikhelson [1], the flame-propagation velocities were calculated; depending on the composition of the combustible mixtures and on the velocity and regime of outflow, the magnitude of the flame-propagation velocity varied from \(0.19\ \text{m/sec}\) for combustible mixtures of the upper limiting composition at \(\alpha = 0.14\) to \(2.27\ \text{m/sec}\) for combustible mixtures of composition \(\alpha = 0.98\).

In Fig. 6, straight lines of the form

\[ h = f\!\left(\frac{w}{u}\right), \tag{3} \]

are plotted, showing the dependence between the height of the inner cone \(h\), the outflow velocity \(w\), and the flame-propagation velocity \(u\), for 4 burners of diameter \(1.18\)—\(1.54\)—\(2.9\)—\(3.7\ \text{mm}\). For a burner of any given diameter, the height of the inner cone \(h\) increases proportionally to the ratio \(\frac{w}{u}\), and the coefficient of proportionality is equal to the radius of the burner \(r\), i.e.

\[ h = r \frac{w}{u}. \tag{4} \]

The equations of the straight lines correspond to the well-known equation of V. A. Mikhelson

\[ h = r \sqrt{\frac{w^2}{u^2} - 1}, \tag{5} \]

in which, owing to the large values of \(\left(\frac{w}{u}\right)^2\), the \(-1\) has been discarded. In our experiments the quantity \(\left(\frac{w}{u}\right)^2\) varied from 360 to 4900.

In Fig. 7, curves of the form

\[ h = f(a). \tag{6} \]

are plotted for the 4 above-mentioned burners, showing the change in the height of the inner cone as a function of the composition of the combustible mixtures. These curves do not have smooth transitions between the individual experimental points and intersect one another.

Fig. 6. Change in the length of the inner cone with changes in the outflow velocity “\(w\)” and the flame-propagation velocity “\(u\)” for burners of different diameter.

Such an indeterminate character of the curves indicates that the height of the inner cone depends simultaneously on \(a\), \(w\), \(Re\), and \(d\). To establish a more definite relationship between the height of the inner cone and the indicated factors, in the coordinate system \(\frac{h}{w} m\) and \(a\), curves of the form

\[ \frac{h}{w} m = f(a)_d . \tag{7} \]

are plotted in Fig. 8 from the experimental data.

In this equation \(m = \frac{Re}{2320}\), and for a laminar outflow regime \(m = 1\); for the transitional regime, when \(5000 > Re > 2320\), the quantity

Fig. 7. Change in the height of the inner flame cone as a function of the composition of the combustible mixture, velocity, flow regime, and burner diameter.

\[ h=f(\alpha,w,Re,d) \]

Vertical axis: Height of the inner cone

Horizontal axis: \(\%\)

Curve labels visible in the figure:

• ① \(d=1.18\ \mathrm{mm}\), \(Re>5000\)
• ② \(d=1.54\ \mathrm{mm}\), \(5000>Re>2320\)
• ③ \(d=2.9\ \mathrm{mm}\), \(5000>Re>2320\)
• ④ \(d=3.7\ \mathrm{mm}\), \(Re<2320\)

Additional visible annotations:

\[ 1<\alpha<4 \]

\[ \alpha<1 \]

Fig. 8. Generalized curves of the variation of the height of the inner cone.

\[ m=\frac{Re}{2320} \]
and for the stable turbulent regime, when \(Re>5000\), the quantity

\[ m=\frac{Re^{0.5}}{2320}. \]

It follows from this that in the laminar regime \(h\) does not depend on the numerical value of \(Re\); in the transitional regime \(h\) is inversely proportional to

\[ \frac{Re}{2320}, \]

and in the stable turbulent regime \(h\) is inversely proportional to

\[ \frac{Re^{0.5}}{2320} \]

(when plotting the curve for the turbulent regime, for reasons of scale the quantity \(m\) is increased by a factor of 10).

For any outflow regime \(h\) is directly proportional to \(w\) and \(d\). The smallest height of the inner cone is obtained for combustible mixtures of composition from \(a=0.9\) to \(a=0.7\), all other conditions being equal. When \(a>1\) is increased and when \(a<0.7\) is decreased, the height of the inner cone increases.

On the basis of the experimental material and subsequent graphical-analytical processing, the following generalized mathematical relationships have been established, making it possible to calculate the height of the inner cone for combustible mixtures of one and the same composition when \(d,w\), and \(Re\) are varied, using for this purpose experimental data obtained for some regime, namely:

for the case when the outflow regime is

\[ \text{laminar}\ \underset{\leftarrow}{\rightarrow}\ \text{transitional} \]

\[ \frac{h}{wd}\,Re=\mathrm{const}, \tag{8} \]

for the case when the outflow regime is

\[ \text{laminar}\ \underset{\leftarrow}{\rightarrow}\ \text{turbulent} \]

\[ \left[\frac{hRe^{0.5}}{wd^{1.65}}\right]_{\text{turb.}} =35.6\left(\frac{h}{wd^{0.65}}\right)_{\text{lam.}} \tag{9} \]

We illustrate the foregoing by examples.

Determine the height of the inner flame cone \(h\) and the flame propagation velocity \(u\) in a combustible mixture of water gas–air of composition \(41.7\%\) gas and \(58.3\%\) primary air \((a=0.55)\), when issuing from a burner of diameter \(1.18\ \mathrm{mm}\) at a velocity of \(89.2\ \mathrm{m/sec}\).

Also find how \(h\) changes for burners of diameter \(1.54\ \mathrm{mm}\), \(2.9\ \mathrm{mm}\), and \(3.7\ \mathrm{mm}\), if the composition of the combustible mixtures is the same in all cases, but the velocity and outflow regime change.

1. For a burner \(d_1=1.18\ \mathrm{mm}\), with \(w_1=89.2\ \mathrm{m/sec}\), the number \(Re_1=6850\). From curve 1 in Fig. 8, along the ordinate axis for \(a=0.55\), we find

\[ \frac{h_1}{w_1}m_1=0.1. \]

Since the outflow regime is turbulent,

\[ m_1=\frac{\sqrt{Re_1}}{2320}=0.0356. \]

Therefore, the height of the inner cone for the burner \(d = 1.18\) mm:

\[ h_1 = 0.1 \frac{w_1}{10 m_1} = 0.1 \frac{89.2}{0.356} = 25 \text{ mm}. \]

Check: according to curve 1 in Fig. 7, \(h_1 = 25\) mm.
The flame-propagation velocity \(u\) is found from equation (4) or from the curve in Fig. 6. For the burner \(d_1 = 1.18\) mm, with \(h_1 = 25\) mm,

\[ \frac{w_1}{u_1} = 42.2, \]

whence

\[ u_1 = \frac{w_1}{42.2} = \frac{89.2}{42.2} = 2.12 \text{ m/sec}. \]

1. For the burner \(d_2 = 1.54\) mm, we set the outflow velocity \(w_2 = 48.8\) m/sec; then \(Re_2 = 4500\). Since the outflow regime is transitional, in calculating \(h_2\) from the experimental data obtained for the turbulent regime, equation (8) should be used; then

\[ h_2 = h_1 \frac{w_2 d_2 Re_1}{w_1 d_1 Re_2} = 25 \frac{48.8 \cdot 1.54 \cdot 6850}{89.2 \cdot 1.18 \cdot 4500} = 27 \text{ mm}. \]

The flame-propagation velocity \(u_2\) is found from equation (4) or from curve 2 in Fig. 6:

\[ \frac{w_2}{u_2} = 32.4; \quad u_2 = \frac{w_2}{32.4} = \frac{48.8}{32.4} = 1.5 \text{ m/sec}. \]

Check: according to curve 2 in Fig. 8,

\[ \frac{h_2}{w_2} m_2 = 0.94, \]

but by calculation

\[ \frac{h_2}{w_2} m_2 = \frac{27}{48.8} \cdot \frac{4500}{2320} = 1.1. \]

According to curve (2) in Fig. 7, the value \(h_2 = 25\); the discrepancy between the values of \(\frac{h_2}{w_2} m_2\) and \(h_2\), as a result of the inaccuracy of the experiments, is

\[ \frac{27 - 25}{25} \cdot 100 = 8\%. \]

1. For the burner \(d_3 = 2.9\) mm, we set the outflow velocity at \(22.4\) m/sec; then \(Re_3 = 4100\). Since the outflow regime is transitional, in calculating \(h_3\) from the experimental data obtained for the turbulent regime, equation (8) should be used; then

\[ h_3 = h_1 \frac{w_3 d_3 Re_1}{w_1 d_1 Re_3} = 25 \frac{22.4 \cdot 2.9 \cdot 6850}{89.2 \cdot 1.18 \cdot 4100} = 25.8 \text{ mm}. \]

The flame-propagation velocity according to curve (3) in Fig. 6 is

\[ \frac{w_3}{u_3} = 17.2; \quad u_3 = \frac{w_3}{17.2} = \frac{22.4}{17.2} = 1.13 \text{ m/sec}. \]

Check: from curve (3) in Fig. 7, \(h_3 = 25\) mm, and from curve (3) in Fig. 8

\[ \frac{h_3}{w_3} m_3 = 1.99; \]

by calculation

\[ \frac{h_3}{w_3} m_3 = \frac{25.8 \cdot 4100}{22.4 \cdot 2320} = 2.01, \]

the discrepancy is insignificant.

1. For a burner with \(d_4 = 3.7\) mm, we set the outflow velocity at \(11.2\) m/sec; then \(Re_4 = 2240\). Since the outflow regime is laminar, in calculating \(h_4\) from the experimental data obtained for the turbulent regime, equation (9) should be used; then

\[ h_4 = h_1 \frac{Re^{0.5} w_4}{35.6 w_1} \left(\frac{d_4}{d_1}\right)^{0.65} \simeq 25 \frac{\sqrt{6850 \cdot 11.2}}{35.6 \cdot 89.2} \left(\frac{3.7}{1.18}\right)^{0.65} \simeq 15.2 \text{ mm}. \]

The flame-propagation velocity from curve (4) in Fig. 6 is

\[ \frac{w_4}{u_4} = 8.1; \qquad u_4 = \frac{11.2}{8.1} = 1.37 \text{ m/sec}. \]

Check: from curve (4) in Fig. 7,

\[ h_4 = 15.0 \text{ mm} \]

and from curve (4) in Fig. 8

\[ \frac{h_4}{w_4} m_4 = 13.1. \]

By calculation

\[ \frac{h_4}{w_4} m_4 = \frac{15.2}{11.2} = 13.5, \]

the discrepancy is insignificant.

An essentially important characteristic of the flame cone of partially prepared combustible mixtures is the relation between the height of the inner cone \(h\) and the total flame length \(L\). Figure 9 shows curves of the form

\[ \frac{h}{L} = \varphi(a) \tag{11} \]

for burners of diameter \(1.18\) mm—\(1.54\) mm—\(2.9\) mm—\(3.7\) mm. As is evident from the figure, the curves form two groups: one group—curves (1) and (2), for burners of diameter 1.18 and 1.54 mm, and the second group—curves 2 and 4, for burners of diameter 2.9 and 3.7 mm. The first group of curves is for a turbulent, or nearly turbulent, outflow regime; the second group is for a laminar regime. Both groups of curves have a characteristic bend with a minimum value of \(\frac{h}{L}\) at \(a \sim 0.5\). Let us follow the change in the ratio \(\frac{h}{L}\) for burners of diameter 1.18 mm and 1.54 mm. At small values of \(a\), the flame length is greatest and the height of the inner cone amounts to about half the flame length; as \(a\) increases, the flame length \(L\) decreases, but the height of the inner cone decreases still more rapidly, reaching at \(a \sim 0.5\) the minimum value of \(\frac{h}{L}\); with a further increase of \(a\), a further reduction occurs

\(L\) and \(h\), but \(h\) decreases more slowly than \(L\), as a result of which the ratio \(\frac{h}{L}\) increases. When \(\alpha\) approaches 1, the inner and outer cones merge. The value of \(\frac{h}{L}\) changes similarly for burners of diameter 2.9 mm and 3.7 mm. Thus it may be concluded that, under a turbulent outflow regime in combustible water-gas–air mixtures, the total

Fig. 9. Relationship between the height of the inner cone \(h\) mm, the length of the flame \(L\) mm, the composition of combustible mixtures \(\alpha\), and the diameter of the burners.

flame length is 2–2.5 times greater than the height of the inner cone, whereas under a laminar outflow regime the total flame length is 5.5–6.5 times greater than the height of the inner cone.

Experiments on the combustion of gas–oxygen mixtures have established that the relative height of the inner cone is sharply reduced in comparison with gas–air mixtures; thus, for gas–oxygen combustible mixtures of composition \(\alpha \sim 0.5\), the value of \(\frac{h}{L} \sim 0.1\) under a turbulent outflow regime.

This is explained by the fact that in gas–oxygen mixtures the flame-propagation velocity is considerably greater than in gas–air mixtures, as a result of which \(h\) decreases substantially.

In the combustion of air mixtures of natural gas containing more than 87% \(CH_4\), the value of \(\frac{h}{L}\) increases greatly and is approximately 0.8–0.9; owing to the low flame-propagation velocity of methane, the height of the inner cone \(h\) increases greatly.

Pyrometric characteristics of the flame in partially prepared combustible mixtures

Heat density of combustion. The heat density of combustion is the amount of heat released during combustion per unit time per unit surface area of the inner cone, i.e.

\[ q\ \frac{\text{kcal}}{\text{m}^2\text{ sec}}. \]

In Fig. 10 a curve of the form

\[ \frac{q}{u}=\psi(\alpha), \tag{12} \]

is shown, indicating the dependence between the heat density of combustion \(q\), the flame-propagation velocity \(u\), and the composition of the combustible mixtures \(\alpha\) for burners of diameter \(1.18\ \mathrm{mm}\)—\(1.54\ \mathrm{mm}\)—\(2.9\ \mathrm{mm}\)—\(3.7\ \mathrm{mm}\). In studying the flame of these burners, the quantity \(q\) varied from \(59\ \dfrac{\mathrm{kcal}}{\mathrm{m^2\,sec}}\) to \(1515\ \dfrac{\mathrm{kcal}}{\mathrm{m^2\,sec}}\). However, the ratio \(\dfrac{q}{u}\) does not depend on the burner diameter or on the outflow regime, but depends only on the composition of the combustible mixtures. Indeed, according to the definition

\[ q=\frac{V'_{\mathrm{gas}} Q^p_n}{F}, \tag{13} \]

where \(V'_{\mathrm{gas}}\ \mathrm{m^3/sec}\) is the amount of gas burned on the surface of the inner cone, \(F\ \mathrm{m^2}\) is the surface of the inner cone, and \(Q^p_n\) is the calorific value of the gas.

In combustible mixtures with a deficiency of primary air, the amount of gas burned on the surface of the inner cone is determined by the amount of primary air, i.e.,

\[ V'_{\mathrm{gas}}=\frac{V_{\mathrm{air}}}{V_{\mathrm{theor}}} =\frac{\alpha V_{\mathrm{theor}} V_{\mathrm{gas}}}{V_{\mathrm{theor}}} =\alpha V_{\mathrm{gas}} . \tag{14} \]

Here \(V_{\mathrm{air}}\ \mathrm{m^3/sec}\) is the amount of primary air; \(V_{\mathrm{theor}}\ \dfrac{\mathrm{m^3}}{\mathrm{m^3}}\) is the theoretically required amount of air; \(V_{\mathrm{gas}}\ \mathrm{m^3/sec}\) is the amount of gas in the combustible mixture. But in accordance with the method of determination

\[ u=\frac{V_{\mathrm{mix}}}{F} =\frac{V_{\mathrm{gas}}+V_{\mathrm{air}}}{F} =\frac{V_{\mathrm{gas}}+\alpha V_{\mathrm{theor}}V_{\mathrm{gas}}}{F} =\frac{V_{\mathrm{gas}}(1+\alpha V_{\mathrm{theor}})}{F}, \]

where \(V_{\mathrm{mix}}\ \mathrm{m^3/sec}\) is the amount of combustible mixture. Consequently,

\[ \frac{q}{u} =\frac{\alpha V_{\mathrm{gas}} Q^p_n}{V_{\mathrm{gas}}(1+\alpha V_{\mathrm{theor}})} =\frac{\alpha Q^p_n}{1+\alpha V_{\mathrm{theor}}} =\frac{Q^p_n}{\dfrac{1}{\alpha}+V_{\mathrm{theor}}}. \tag{15} \]

In combustible mixtures with an excess of primary air, a single-cone flame is formed, and on its surface all the gas supplied to the burner burns; consequently,

\[ \frac{q}{u} =\frac{V_{\mathrm{gas}} Q^p_n}{V_{\mathrm{mix}}} =\frac{V_{\mathrm{gas}} Q^p_n}{V_{\mathrm{gas}}\,1+\alpha V_{\mathrm{theor}}} =\frac{Q^p_n}{1+\alpha V_{\mathrm{theor}}}. \tag{16} \]

Thus, for combustible mixtures of both types, according to equations (15) and (16), the quantity \(\dfrac{q}{u}\) does not depend on the burner diameter or on the outflow regime, but is a function of \(\alpha\).

Temperature of the flame. The temperature of the surface of the inner and outer cones of the flame was measured at several points with Pt—PtRh thermocouples. The measurement results were processed graphically in the form of curves

\[ \frac{t}{q/u}=f(\alpha). \tag{17} \]

As is seen from Fig. 11, both in the region of small values of $\alpha$ (from 0.14 to 0.6) and in the region of large values ($\alpha > 1.2$), the flame temperature depends on the value $\dfrac{q}{u}$ and on the burner diameter. But in the range of values of $\alpha$ from 0.6 to 1.2, i.e. the most important in practice, the burner diameter has little effect on the flame temperature; in this range of values of $\alpha$ the flame temperature depends on the value $\dfrac{q}{u}$, i.e., ultimately, on the composition of the combustible mixtures. Mathematically this dependence is expressed, as is seen from Fig. 11, by the empirical equation

\[ \frac{t}{q/u} = 1.75 \div 1.9 \]

or, approximately on average,

\[ \frac{t}{q/u} = 1.82, \]

i.e.

\[ t = 1.82\,\frac{q}{u}. \tag{18} \]

The temperature on the surface of the outer cone of the flame is lower than the temperature on the surface of the inner cone by 50–80°C; the smaller $\alpha$, the greater this difference; at $\alpha \sim 1$ the temperature difference is zero.

In addition to the examples given earlier, which illustrated the possibility of calculating the flame length from the relations derived by us, let us determine the pyrometric characteristics of the flame.

For the conditions of example 1, i.e. for a burner with $d_1 = 1.18 \text{ mm}$.

According to Fig. 10, for $\alpha_1 = 0.55$

\[ \frac{q_1}{u_1} = 650 \quad \text{and} \quad q_1 = 650\,u_1 = 650 \cdot 2.12 = 1360\ \frac{\text{kcal}}{\text{m sec}} . \]

According to Fig. 11, for $\alpha_1 = 0.55$

\[ \frac{t_1}{q_1/u_1} = 2.24 \quad \text{and} \quad t_1 = 2.24\,\frac{q_1}{u_1} = 2.24 \cdot 650 = 1460^\circ\text{C}. \]

According to Fig. 9, for $\alpha_1 = 0.55$

\[ \frac{h_1}{L_1} = 0.425; \quad L_1 = \frac{h_1}{0.425} = \frac{25}{0.425} = 59 \text{ mm}. \]

For the conditions of example 2, i.e. for a burner with $d_2 = 1.54 \text{ mm}$, according to Figs. 10, 11, 9, for $\alpha_2 = 0.57$

\[ \frac{q_2}{u_2} = 650; \quad q_2 = 650\,u_2 = 650 \cdot 1.5 = 975\ \frac{\text{kcal}}{\text{m}^2\text{ sec}}; \]

\[ \frac{t_2}{q_2/u_2} = 2; \quad t_2 = 2\,\frac{q_2}{u_2} = 2 \cdot 650 = 1300^\circ\text{C}; \]

\[ \frac{h_2}{L_2} = 0.425; \quad L_2 = \frac{h_2}{0.425} = \frac{27}{0.425} = 64 \text{ mm}. \]

Vertical axis:

\[ \frac{t}{q,\,U} \]

Horizontal axis:

\[ \alpha \]

Coefficient of excess air.

Legend:

1. \(d = 1.18\ \mathrm{mm}\)
2. \(d = 1.54\ \mathrm{mm}\)
3. \(d = 2.9\ \mathrm{mm}\)
4. \(d = 3.7\ \mathrm{mm}\)

Fig. 11. Dependence between the flame temperature \(t^\circ\mathrm{C}\), the heat intensity of combustion \(q\ \frac{\mathrm{kcal}}{\mathrm{m}^2\cdot\mathrm{sec}}\), the flame-propagation velocity \(U\ \frac{\mathrm{m}}{\mathrm{sec}}\), the composition of the combustible mixture \(\alpha\), and the burner diameter.

For the conditions of example 3, i.e., for burner \(d_3 = 2.9\) mm,

\[ \frac{q_1}{u_3} = 630, q_3 = 630 u_3 = 630 \cdot 1.3 = 820 \ \frac{\text{kcal}}{m^2 \text{ sec}}; \]

\[ \frac{t_3}{q_3/u_3} = 2.11; t_3 = 630 \cdot 2.11 = 1320^\circ\text{C}; \]

\[ \frac{h_3}{L_3} = 0.31; L_3 = \frac{h_3}{0.31} = \frac{25}{0.31} = 80 \ \text{mm}. \]

For the conditions of example 4, i.e., for burner \(d_4 = 3.7\) mm,

\[ \frac{q_4}{u_4} = 680; q_4 = 680 \cdot 1.41 = 960 \ \frac{\text{kcal}}{m^2 \text{ sec}}; \]

\[ \frac{t_4}{q_4/u_4} = 1.90; t_4 = 680 \cdot 1.90 = 1300^\circ\text{C}; \]

\[ \frac{h_4}{L_4} = 0.26; \quad L_4 = \frac{h_4}{0.26} = \frac{15.5}{0.26} = 56 \ \text{mm}. \]

The experimental relationships established make it possible to assume and to hope that the results obtained and the generalized relationships can be used in the design and adjustment of gas burners for flame combustion.

As indicated above, the water-gas flame of partially prepared combustible mixtures is used for flame treatment in the production process of incandescent electric lamps at some plants. To study the mode of flame adjustment, we carried out an investigation under production conditions at the Tomsk Electric-Lamp Plant. As the object of investigation, an eight-station semiautomatic machine for assembling lamp stems was taken. Fig. 12 shows a photograph of the semiautomatic machine.

The processes of treating the stem at the individual stations consist of the following: at the first station, in the soft flame of burners installed on both sides, the glass parts of the stem are heated while the tongs rotate; at the second station the glass is heated to red heat by the flame of burners likewise located on both sides; at the third station the blade, the rod, and the stem tube are softened, also under the effect of two-sided heating. During the transition of the tongs from the third station to the fourth, the blade of the stem is stamped. At the fourth station compressed preheated air is blown through the stem tube; the jet of air ruptures the softened glass, forming an opening through which air will subsequently be pumped out of the vacuum lamp. At the fifth station the pierced opening is fire-polished and the neck and blade of the stem are heated through; at the sixth, partial cooling of the stem takes place in the atmosphere of a weak flame; at the seventh, the finished stem is removed from the tongs, and at the eighth station the next loading of semifinished products is carried out. A detailed description of the processes of assembling electric lamps is set forth in the book by L. G. Ulmishek [7].

The semiautomatic machine tested by us is equipped with injection gas burners similar to those described above. The burners are assembled in groups of three, forming crossing flames. The general appearance of burners with crossing flames is shown in Fig. 13.

In the investigations, measurements were made of the following quantities: the temperature at the apex of the inner cone of the flame, the temperature

Fig. 12. Eight-position semiautomatic machine for assembling electric-lamp legs.

Fig. 13

at the point where the flames intersect, the length of the inner cone, the analysis of the combustible mixture, and the pressure of the gas, air, and combustible mixture.

The measurement scheme is shown in Fig. 14.

Table 3 gives summary results of the investigation for a burner with a diameter of 1.2 mm.

Let us compare the results of the investigation obtained under production conditions with the same results obtained under laboratory conditions for a burner with a diameter of 1.18 mm. The flame temperature under production conditions is considerably lower than under laboratory conditions for combustible mixtures of the same composition. Thus, for a burner with a diameter of 1.18 mm, tested under laboratory conditions, with a combustible-mixture composition of 35–37% gas and 65–63% air \((\alpha = 0.37 \div 0.67)\), the flame temperature is approximately 1500°C, whereas for a burner with a diameter of 1.2 mm, tested under production conditions with approximately the same combustible-mixture composition, the flame temperature does not exceed 1260°C. Such a sharp difference in temperature is due to the great difference in the velocities of outflow of the combustible mixtures: in the laboratory burner the outflow velocity is approximately 80 m/sec, while in the production burner the outflow velocity varied within the range 25–43 m/sec. The influence of the reduced outflow velocity under production conditions should also explain the smaller height of the inner cone of the flame: for the burner under production conditions the height of the inner cone varied from 12.6 mm to 22.8 mm, whereas for the laboratory burner it was 24 mm.

Fig. 14. Measurement scheme on the bench semiautomatic machine.

Labels in the figure:

• water
• gas
• To the manometer
• Point IV was also used for taking a mixture sample for analysis.
• tube for connection to the manometers
• For the measurements, the position was selected after which stamping is performed
• Only one position selected for measurements is shown in the diagram

Table 3

Test data for an injection burner of diameter 1.2 mm on the assembly semiautomatic machine

No.	Pressure of combustible mixture before the burner, mm w.c.	Outflow velocity, mm/sec	Combustible mixture composition: gas, %	Combustible mixture composition: air, %	$\alpha$	Height of internal cone, mm	Temperature, °C: at the tip of the internal cone	Temperature, °C: at the place where the flames cross
1	141	34,0	62,0	38,0	0,23	14,8	1130	1080
2	164	36,8	62,0	38,0	0,23	20,0	1130	1100
3	150	35,0	51,2	48,8	0,36	18,3	1065	1090
4	157	36,0	51,2	48,8	0,36	21,0	1155	1130
5	153	35,2	46,1	53,9	0,45	23,4	1130	1105
6	159	36,0	40,5	59,5	0,56	20,0	1175	1130
7	159	35,9	38,0	62,0	0,62	18,9	1040	1130
8	142	34,0	34,2	65,8	0,73	22,8	1058	1065
9	206	41,0	34,2	65,8	0,73	16,8	1145	1114
10	90	25,0	33,3	66,7	0,76	12,6	1100	790
11	95	27,8	33,3	66,7	0,76	12,8	1100	800
12	107	29,6	33,3	66,7	0,76	13,0	1140	940
13	124	32,0	33,3	66,7	0,76	14,2	1220	1000
14	144	35,0	33,3	66,7	0,76	15,2	1210	1080
15	226	43,0	33,3	66,7	0,76	18,8	1260	1170
16	264	46,5	32,6	67,4	0,79	19,8	1260	1170
17	187	39,0	31,7	68,3	0,82	16,2	1200	1170
18	174	38,0	31,7	68,3	0,82	15,7	1205	1070

Then, also under production conditions, a burner of diameter 1.5 mm was tested on a semiautomatic machine for assembling blades. The results of the study are presented in Table 4.

Table 4

Test indices of an injection burner of diameter 1.5 mm
on an assembly semiautomatic machine

No.	Pressure of the combustible mixture before the burner, mm water column	Efflux velocity, m/s	Combustible-mixture composition: gas, %	Combustible-mixture composition: air, %	\(a\)	Height of the inner cone, mm	Temperature, °C: at the apex of the inner cone	Temperature, °C: at the place where the flames cross
1	470	62.0	65.0	35.0	0.2	48.1	1160	1275
2	462	61.5	54.0	46.00	0.33	47.8	1100	1275
3	452	61.0	52.0	48.0	0.35	34.9	1170	1240
4	620	71.0	52.0	48.0	0.35	47.6	1240	1290
5	452	61.0	50.5	49.5	0.37	34.2	1220	1280
6	600	70.0	50.5	49.5	0.37	46.5	1255	1305
7	600	70.0	48.0	52.0	0.41	41.2	1265	1305
8	432	59.0	47.6	52.4	0.42	39.1	1160	1300
9	428	58.8	44.3	55.7	0.47	29.0	1220	1275
10	590	69.3	41.6	58.4	0.53	39.0	1285	1310
11	580	69.2	40.0	60.0	0.57	33.0	1290	1310

Comparing the results of the study obtained under production conditions with those obtained under laboratory conditions for a burner of diameter 1.54 mm, we find that in both cases the temperatures in the flame cone were obtained of the same order for combustible mixtures of approximately the same composition, but the height of the inner cone of the flame for the burner under production conditions is somewhat greater than for the burner under laboratory conditions. The approximate coincidence of the temperature in both cases and the small discrepancy in the heights of the inner cone are explained by the fact that the efflux velocities of the combustible mixture in production burners have a somewhat greater value.

V. Surface flameless combustion of water gas

One form of gas combustion is surface catalytic flameless combustion; this combustion is characterized by the following: 1) before entering the burner, the gas and air are first thoroughly mixed to the state of a homogeneous mixture; 2) the combustible mixture is burned on the surface of the refractory material of the burner, which has a high temperature and exerts a catalytic effect on the combustion process.

The external signs of surface combustion are the complete or almost complete absence of a visible flame front. The flame front coincides with the surface of catalytically active refractory materials or is located in the pores of the granular bed. Owing to the high thermal density of combustion, the surface of the burner or of the bed has intense thermal radiation.

The conditions for the mutual transformation of different forms of gas combustion were formulated by the author [2].

In surface combustion the flame front is significantly deformed under the influence of hydrodynamic, thermal, catalytic, and other phenomena. The causes affecting the deformation of the normal conical flame front are as follows: a) as a result of an increase in the efflux velocity of the combustible mixture, the parabolic epure of the distribution of velocities over the cross section

the flow is straightened, as a result of which the conical flame front becomes flat; b) the surface of the flame front, in the transition from torch combustion to flameless combustion, decreases, as a result of which the heat release during combustion is concentrated on a smaller surface; c) in surface combustion the rate of flame propagation increases by tens and hundreds of times in comparison with the same rates in torch combustion.

The investigation of surface flameless combustion of water gas was carried out by the following methods: 1) combustion in a quartz tube filled with refractory grains; 2) combustion in an experimental furnace with a granular bed; and 3) combustion in a channel burner. Experiments on the combustion of water gas in a quartz tube and in an experimental furnace with a granular bed are described in a published work by the author [2], and are also partly given in the monograph of Prof. M. B. Ravich [6]. As for the combustion of water gas in a channel burner, this is reported in the present work.

The channel burner—a refractory block—is made of high-alumina refractory fireclay material. The dimensions of the block are: outside diameter 195 mm, thickness 100 mm. The block has 48 openings with diameters of 3/15 mm. A drawing of the burner is given in Fig. 15.

The velocity of passage of the combustible mixture through the narrow openings of diameter 3 mm was 20 m/sec, and through the wide openings of diameter 15 mm—0.8 m/sec. The gas flow rate in the burner was 7 m³/hour and that of air 18 m³/hour; the composition of the combustible mixture corresponds to an excess-air coefficient \(\alpha = 1.02\).

The pressure in the mixer was 50 mm water column. Under this hydraulic regime, no flashback of the flame into the supply chamber or into the mixer was observed. The heating on the surface and inside the block was uniform; the temperature, measured with an optical pyrometer, was about 1500°C.

Fig. 15. Channel burner for flameless combustion.

The surface heat intensity of combustion, i.e., the amount of heat released referred to the radiating surface of the burner, averaged 650,000 kcal/m²·hour.

The volumetric heat intensity of combustion, i.e., the amount of heat released referred to the volume of the channels in the block, amounted to \(37 \cdot 10^6\) kcal/m³·hour.

VI. Conclusion

The author has carried out a comprehensive investigation of the combustion of water gas. This investigation confirms the author’s proposition concerning the mutual transformability of the principal typical forms of gas combustion and reveals the conditions, specific for each form, for the stable development of combustion.

The investigation of the combustion of water gas was carried out under laboratory conditions.

conditions, in an environment approaching production conditions, and in industrial heat installations.

Thus, under laboratory conditions a study was carried out of the structure of a water-gas flame by separating a two-cone flame; the effect of platinum as a catalyst on the combustion process; the effect of the composition of combustible mixtures, burner diameter, and outflow regime on the length of the flame and on the distribution of temperature and heat intensity in the flame; the diffusion combustion of water gas in flat-slot atmospheric burners; and the surface flameless combustion of water gas and the effect of the granulometric composition of the granular bed on stability and heat intensity.

The study of diffusion combustion in a shielded furnace chamber was carried out in an environment which, in scale and conditions, approached production practice. A detailed description of the study of diffusion combustion of water gas is given in the author’s article [5]. The operation and construction of an industrial tank glass-melting furnace, built according to the author’s design at one state Union plant producing electric incandescent lamps, is described by the author in article [4].

The object of the study was water gas produced in the gasification of lean hard coals.

The most important results of the comprehensive study of the combustion of water gas are the following.

1. In studying the structure of a water-gas flame by the method of two-cone separate combustion in Bunsen and Meker burners, with a platinum gauze, and also in porcelain and platinum tubes, the sequence of burnout of the individual constituents of water gas was established, as well as the effect of the proportion of primary air and of platinum on the sequence of burnout.

2. In studying the flame of water gas in partially prepared combustible mixtures with the aid of burners of different diameter and construction, the relationship was established between the flame length, outflow velocity, character of motion, burner diameter, composition of the combustible mixtures, and temperature distribution in the flame. There is every reason to believe that the factual material obtained and the mathematical relationships can be used as the basis for designing atmospheric burners and for regulating the firing regime in industrial installations during operation.

3. In studying the diffusion flame of water gas in a flat-slot atmospheric burner and in a shielded furnace chamber, it was shown that, despite a number of difficulties arising in the practical use of the mathematical apparatus of the physical theory of diffusion, and despite the inevitability of introducing certain simplifying assumptions, a preliminary calculation of a diffusion flame is nevertheless possible and necessary, since the quantities actually observed in experiments are fairly close to the calculated ones. In addition, a method is proposed for calculating a diffusion flame at unequal outflow velocities of gas and air with the aid of the equivalent thickness of the diffusion layer.

4. According to the author’s design, a tank glass-melting furnace using water gas with air preheating in regenerators was built. This furnace has been in operation at an electric-lamp plant since 1944. A distinctive feature of the author’s furnace is its small overall dimensions; successful design of small tank furnaces, as is known, is difficult, and the author had no experimental data at his disposal.

5. In studying the surface combustion of water gas in a granular bed and in a channel burner, the conditions for the stability of this process were determined, the influence of the diameter and specific gravity of the grains composing the bed was established, and a pyrometric combustion coefficient of 0.85 was achieved. The results of this

research can serve as the basis for designing high-temperature muffle (crucible) furnaces operating on a cold combustible mixture.

References Cited

1. W. A. Michelson. Collected Works, vol. 1, 1930.
2. N. N. Norkin. Proceedings of the Tomsk Polytechnic Institute, vol. 64, 1948.
3. N. N. Norkin and G. D. Spet︠s︡i. Proceedings of the Tomsk Polytechnic Institute, vol. 66, no. 3, 1948.
4. N. N. Norkin. Proceedings of the Tomsk Polytechnic Institute, vol. 64, 1948.
5. N. N. Norkin. Proceedings of the Tomsk Polytechnic Institute, vol. 70, 1950.
6. M. B. Ravich. Surface Flameless Combustion. Publishing House of the Academy of Sciences of the USSR, 1949.
7. L. G. Ulmishek. Production of Electric Incandescent Lamps, 1949.