Small-Sized Tank Furnace for Melting Electric-Bulb Glass, Heated by Water Gas

N. I. Norkin

At one of the plants in Western Siberia, in March 1944, construction was completed of a small-sized tank furnace built according to the author’s design. Beginning in April 1944, after a brief start-up period, the furnace has been in operation and, under the specified temperature regime, has steadily provided the design output without interruptions; in August 1946, after two years of continuous operation, the furnace was stopped for cold repair (replacement of the blocks of the furnace walls and crown), and in September 1946 operation of the furnace was resumed.

In the furnace, a glass melt is produced for manufacturing, by blowing, the following range of glass bulbs:

In addition, part of the melted glass mass is used for making weight glass (drot), from which tubes of diameter 12–13 mm, 15–16.5 mm, and 24–26.5 mm are drawn.

The batch for melting the glass mass is composed according to the following calculation: per 100 parts by weight of sand, 36 parts by weight of soda; 3 parts by weight of sodium sulfate; 24 parts by weight of dolomite.

Glass cullet is added to the batch in an amount of 25–30%. The sand used for melting the glass mass has the following chemical composition:

\[ \mathrm{SiO_2}\ \text{from }96.5\ \text{to }98\% \]

\[ \mathrm{Al_2O_3}\ \text{from }1.62\ \text{to }3.30\% \]

\[ \mathrm{Fe_2O_3}\ \text{from }0.2\ \text{to }0.4\% \]

\[ \mathrm{CaO}\ \text{from }0.4\ \text{to }0.8\% \]

\[ \mathrm{MgO}\ \text{from }0.18\ \text{to }0.35\%. \]

In connection with the article by Prof. D. B. Ginzburg (Glass and Ceramic Industry, Nos. 1–2, 1946), which notes the considerable difficulties of designing small-sized furnaces and the significance of these furnaces in the development of the glass industry, and also bearing in mind that the tank furnace operating according to the author’s design is the smallest in dimensions of all known small-sized furnaces, it seems advisable to elucidate the theoretical considerations underlying the furnace design and the heat-engineering performance indicators of the furnace; it is for this purpose that the present article is being published.

1. Fuel and Heating System

The fuel is water producer gas, whose average chemical composition, excluding moisture, is as follows: CO—36%; CO₂—5%; O₂—0.2%; H₂—53.5%; N₂—3.3%; CH₄—2%.

The composition of the moist gas at a degree of saturation \(\varphi = 1\) for a temperature of 20°C is as follows: CO—35.1%; CO₂—4.87%; O₂—0.19%; H₂—52.3%; N₂—3.21%; H₂O—2.3%; CH₄—1.9%.

The calorific value of the gas, calculated from the above composition and repeatedly determined with a Junkers calorimeter, has the following average value:

\[ Q_{\mathrm{n}}^{\mathrm{p}} = 2578 \ \frac{\text{kcal}}{\text{m}^3}. \]

Water producer gas is produced in Pintsch-system gas generators with a flat grate, with manual slag and ash removal and manual fuel charging. For the production of water gas, Aralichev coal of the semi-anthracite type is used, with an average volatile-matter content of 7 to 9%, and with the following chemical composition on an organic-mass basis: C° = 86.3–89.8%; H° = 3.7–4.1%; O° = 5.3%; N° = 1.4–2% and S° = 0.9%. In the working fuel: \(W_p = 3.5\text{–}12\%\); \(A_p = 14.6\%\). The average calorific value is 6090 kcal/kg.

The use of water producer gas for heating a glass-melting furnace is not dictated by any special requirements imposed on the gas as a fuel. For heating the furnace there is no absolute need to resort to the use of high-calorie water gas as compared with ordinary mixed producer gas, since the production of water gas costs considerably more than the production of ordinary mixed gas, and also requires a greater consumption of coal and steam, which are necessary for obtaining water gas; this is confirmed by the following considerations. If the calorific value of ordinary mixed producer gas is taken as 1250 kcal/m³, and the calorific value of water gas as 2500 kcal/m³, and it is then taken into account that the yield of mixed producer gas is usually 3.0 m³/kg, while the yield of water gas is 1 m³/kg, then the ratio of the coal consumption when heating with water gas to the equivalent coal consumption when heating with mixed gas is expressed by the value

\[ \frac{1250 \cdot 3}{2500 \cdot 1} = 1.5 \]

As regards the pyrometric effects of combustion of water and mixed producer gases, comparison of the calorimetric combustion temperatures under conditions of burning cold gases gives the following relations: for water gas \(t_{\text{cal}} = 2050\text{–}2100^\circ\text{C}\), for mixed gas \(t_{\text{cal}} = 1650\text{–}1700^\circ\text{C}\).

However, the pyrometric effect of combustion of mixed gas can be increased by preheating it and the air required for combustion by utilizing the heat of the waste gases.

The comparisons presented for the two types of gas—water and mixed—support the conclusion that there are no special advantages of heating the furnace with water gas over mixed gas; the use of water gas for heating the glass-melting furnace is due to the circumstance that its consumption for the furnace constitutes part of the total consumption of water gas required in this production for the heat treatment of glass and metal parts and for other purposes.

For designing the system for heating the furnace with water gas, it was first necessary to solve the following questions: 1) the preheating of the gas and air required to achieve the necessary pyrometric effect, and 2) the degree of mixing of gas and air in the burner, ensuring normal development of the flame in the furnace.

The first question was solved by means of the following calculations: the mean flame temperature (which can be measured with a suction thermocouple) was taken as 200°C above the maximum temperature of the glass surface; in the present case
\(t_{\text{avg.}}^{\Phi} = 1400 + 200 = 1600^\circ\text{C}\). The pyrometric combustion coefficient in the furnace was taken as the minimum of those usual for glass-melting furnaces (0.65–0.75); in the present case \(\eta_{\text{pyr}} = 0.65\).

Then, taking into account that:
1) \(Q_{\text{H}}^{p} = 2578\ \text{kcal}/\text{m}^3\);
2) the excess-air coefficient \(\alpha = 1.1\);
3) the volume of combustion products is \(3.07\ \text{m}^3/\text{m}^3\);
4) the calorimetric combustion temperature of the gas is 1980°. From the equation

\[ t_{\text{avg.}}^{\Phi} = \eta_{\text{pyr}}\cdot \frac{Q_{\text{H}}^{p} + V_{\text{air}}\, C\, t_{\text{air}}} {V_{\text{c.p.}}\, C_{\text{c.p.}}} \]

we find that the air must be preheated to 800°C in order to ensure the above-mentioned flame temperature.

In solving the second question, concerning the degree of mixing of gas and air in the burner that ensures normal development of the flame in the furnace, the following considerations were taken into account: water gas contains approximately 50% hydrogen, whereas in mixed producer gas

Fig. 1. Small-output tank furnace for melting electric-bulb glass

Principal data

1. Capacity: 2.2 t/day.
2. Water-gas consumption: 130 m³/hour.
3. Air preheating in the regenerators: 800°C.
4. Melting rate: 600 kg/m²·day.
5. Heat intensity: \(19 \cdot 10^3\) kcal/m²·hour.
6. Glass surface area: 3.53 m².

gas from hard coal contains considerably less hydrogen, and usually lies within the range from 12 to 17%. However, it is known that the flame-propagation velocity of hydrogen is the greatest among the flame-propagation velocities of other gases; for example, it is approximately 3.3 times greater than for CO and 8 times greater than for CH₄. In addition, the composition of mixed producer gas contains a significant

the amount of nitrogen is, on average, not less than 50%, whereas water gas contains only 3–5% nitrogen; yet it is known that nitrogen reduces the rate of flame propagation. For the reasons indicated, the rate of flame propagation of water gas is approximately three times greater than the rate of flame propagation of mixed producer gas (under identical temperature conditions). Taking into account these features of water-gas combustion, as compared with the combustion of coal-fired mixed producer gas, and bearing in mind the need to obtain a comparatively long flame, the burner was designed with a low degree of mixing of gas and air, with their separate supply into the furnace—the gas is supplied cold to the burner through the gas pipeline, and the air from the regenerator; mixing takes place only at the outlet into the working space of the furnace. The operating practice of the furnace confirmed the correctness of the conclusion given above, since the temperature distribution along the length of the flame in the furnace is observed to be quite uniform.

Fig. 2. Plan of the tank furnace

The accompanying figure gives a vertical longitudinal section and plan of the furnace and regenerators.

II. Working Space of the Furnace

When designing the dimensions and form of the working space of the furnace from the heat-engineering standpoint, the following principal questions arise:

1) What should be the size of the glass surface in the melting and working parts of the tank, and what should be the ratio between them?

2) What length should the path be for the free development of the flame?

3) How should the heat and temperature regimes be carried out in the working part of the furnace?

Since the furnace being designed, in comparison with those usually in operation, differs not only in its heating system (water gas) but also in its very small dimensions (tank surface area 5.58 m²), it is quite natural that the solution of the questions indicated above acquired all the greater importance, for it was impossible to borrow data from the production practice of other small-size furnaces. In solving the first question concerning the surface of the glass surface, the following mutually related factors were taken into account: a) the specific output of glass melt from 1 m² of glass surface in the melting and working parts of the furnace; b) the relationship between daily output, the surface of the glass mirror, and the volume of glass melt in the furnace basin; c) the relationship between heat intensity and the surface of the glass mirror under conditions of heat transfer from the flame of water gas. This last factor is of essential importance, since the difference in the flame-propagation rates of water gas and mixed producer gas, and the difference in the composition of the combustion products of these gases, entail a considerable difference in the degrees of blackness of the flames of both gases; the degree of blackness of the flame of mixed—

producer gas does not exceed 0.15, while the degree of blackness of a water-gas flame is about 0.25 of absolutely black radiation. By varying, by calculation, the ratios of the factors cited above, the following optimal indices were ultimately obtained: a) the specific output of glass melt is 600 kg per 1 m² per day over the area of the melting section, and 400 kg per 1 m² per day over the total area of the furnace hearth; b) the daily capacity of the furnace is 17.5% of the output of the total volume of glass melt in the tank; c) the thermal intensity of 1 m² of the heated (melting) part of the furnace is expressed by the value 119,000 kcal/m²·h, which corresponds to a specific heat consumption, in terms of gas, of 4750 kcal per 1 kg of melted glass.

The consumption of water gas in operating practice is monitored by means of a sharp-edged orifice, 63 mm in diameter, installed in a gas line 105 mm in diameter. The calculation formula for determining the gas consumption at a diameter ratio of \(\frac{63}{105}\) was derived as follows:

\[ V = 39\sqrt{h}\ \frac{\mathrm{m}^{3}}{\mathrm{h}} \]

Here \(h\) is the pressure drop across the orifice used to create the velocity of gas motion. Usually \(h\) varies within the range 20–26 mm water column, which corresponds to a gas consumption from 175 m³/h to 198 m³/h. According to the design, the gas consumption was set at 175 m³/h.

Under these conditions the flame length proved to be 4.9 m, and the distance between the axes of the flame in the forward and reverse directions was 1100 mm, with the total furnace width being 1600 mm.

As for the third question—that of the heat and temperature regime of the working-out section of the furnace—in order to solve it it was necessary to find the relationship between heat transfer to the surrounding medium through the enclosing walls of the working-out section of the furnace and the drop in heat content of the glass melt continuously entering from the melting section into the working-out section. It was found that the drop in heat content of the glass melt from 1400°C (in the melting section) to 1150°C (in the working-out section) does not fully compensate for the heat losses to the surrounding medium, as a result of which the temperature of the glass melt in the working-out section would inevitably fall below the specified normal level; therefore, to avoid such a decrease, the amount of heat lacking for the working-out section is supplied by radiation from the melting section by means of a lattice thermal screen made of refractory bricks, located above the throat. The cross-sectional area of the openings of the lattice screen is calculated according to the formulas for a thermal diaphragm.

The tank bottom is at the same level in both the melting and working-out sections of the furnace; the tank depth is 900 mm; the melting section of the furnace is separated from the working-out section by a throat measuring 600 × 300 × 1000 mm. During operation, no freezing of glass in the throat was observed.

III. Air Regenerators

The principal quantities determining the normal operation of the regenerators and of the entire heating system are:

1) the thermal capacity of the checkerwork, providing the specified preheating of the air and sufficient utilization of the heat of the gases leaving the furnace, and
2) the height of the regenerators, providing such a geometric head of hot air as would be sufficient to create a specified

…of the specified velocity of air discharge from the burner into the working space of the furnace.

For heating the air to 800°C, the total heating surface of the checkerwork in the air chamber is 81 m². The checkerwork is of the Siemens type, with cell dimensions of 65·90·115 mm and with a specific heating surface of 15 m²/m³. The heat-transfer coefficient in the checkerwork is 3.17 kcal/m²·deg·period. Per 1 m² of glass surface in the melting section there are 22 m² of checkerwork heating surface, with a heat load of 6100 kcal/m²·h. The geometric head produced by the regenerator is 2.78 mm w.g.; as the calculations have shown, this value is quite sufficient both for overcoming the resistances along the path of the air movement and for creating a burner discharge velocity equal to 6 m/s. The resistances along the path of the gases leaving the furnace are overcome by a low-pressure screw fan, producing a draft of 9–10 mm w.g. at a waste-gas temperature of about 150–160°C. Such a hydraulic regime is stable during operation of the furnace.

Summary of the principal design and operating indices of the furnace

Type of furnace—a continuous tank furnace with a flow and a horseshoe-shaped flame. Dimensions of the tank:

total length—4.4 m;
length of the melting section—2.3 m;
length—the radius of the working-out section—1.1 m;
width of the melting section—1.6 m;
depth of the melting section—900 mm;
depth of the working-out section—900 mm;
glass-surface area of the melting section—3.68 m²;
glass-surface area of the working-out section—1.9 m²;
total glass-surface area—5.58 m²;

\[ \text{ratio}=\frac{\text{melting section}}{\text{total area}}=\frac{3.68}{5.58}=0.66; \]

total capacity of the tank—5.03 m³;
capacity of the melting section—3.31 m³;
capacity of the working-out section—1.72 m³.

\[ \text{ratio}=\frac{\text{daily output}}{\text{weight capacity of the furnace}}=0.175. \]

Dimensions of the checkerworks (air): length—2 m, height—3 m, width—0.9 m, volume—5.4 m³, checkerwork heating surface—81 m²;

\[ \text{ratio}=\frac{\text{heating surface}}{\text{total glass-surface area}}=\frac{81}{5.58}=14.5\ \text{m}^2/\text{m}^3; \]

\[ \text{ratio}=\frac{\text{heating surface}}{\text{glass-surface area of the melting section}}=\frac{81}{3.68}=22\ \text{m}^2/\text{m}^2; \]

Note: After 2 years of continuous operation, during a cold repair, at the suggestion of engineer F. P. Gurskii, with the consent of the designer, the length of the checkerworks was reduced to 1.5 m (instead of 2). End burner: discharge width—500 mm; discharge height in the arch opening—450 mm, discharge cross-section 0.2 m²,

\[ \text{ratio}=\frac{\text{discharge cross-section}}{\text{glass-surface area of the melting section}}=0.0545 \]

Furnace output: 2.2 t/day.
Glass-melt withdrawal: per area of the glass surface of the melting section, 600 kg/m²·day; per total glass-surface area, 400 kg/m²·day.

Fuel type: water gas from Aral semi-anthracite coals.
Fuel consumption (average): 190 m³/hour.
Specific consumption of the fuel’s potential heat: for glass melt—4750 kcal/kg; per area of the hearth surface of the heated section—119000 kcal/m²·h.

External surface of the furnace walls: 30 m².
External surface of the furnace crown: 13.5 m².
External surface of the furnace bottom: 12.2 m².
External surface of the checker chamber: 28 m².

Temperature conditions: melting section, 1400–1450°C; working-out section, 1150–1200°C; air preheating, 800°C; waste gases, 150–160°C.

Composition of the outgoing combustion products:

\[ CO_2 = 13.6\%;\quad H_2O = 17.3\%;\quad O_2 = 3.1\%;\quad N_2 = 66\%. \]

Draft before the fan: 9 mm water column.

Batch-charging regime: until the furnace was stopped for repairs, during 2 years of continuous operation, the batch was charged every 2½ hours in an amount of 280–300 kg. After the repair, in order to increase the intensity of furnace operation, on the proposal of engineer F. P. Gursky, with the consent of the author of the design, charging every 1 hour in an amount of 130 kg was introduced.

Conclusion

The article sets forth the basic principles adopted by the author in designing a small-sized continuous tank glass-melting furnace fired with water gas, as well as the principal indicators of the furnace’s construction and operation.

The continuous operation of the furnace for 4 years at one of the plants in Western Siberia showed good results, in agreement with the design data.

The construction of the furnace and its subsequent operation were carried out under the supervision of engineer I. G. Litvinenko.

The repair of the furnace after 2 years of operation was carried out under the direction of engineer F. P. Gursky.