Chemical and Petroleum Engineering, Vol. 35, Nos. 3-4, 1999
INVESTIGATION OF THE PROCESS OF CONTINUOUS P R O D U C T I O N O F S O L I D C A U S T I C S O D A IN " S H A R P " J E T APPARATUS
V. M. Makarov, V. A. Kulikov, and A. B. Vandyshev
UDC 661.322.1:66.093.6
Solid caustic soda is obtained by evaporating water from a 42-45% solution (liquid caustic soda). The most common process is that of periodic evaporation with a constant level of solution in the apparatus (boiler), using combustion or electric heat. The process, which takes place with boiling point of the solution varying as the caustic soda concentration rises to 100%, is called caustic fusion in industrial practice. At the end of the fusion the temperature of the product in the apparatus (boiler) exceeds the caustic soda melting point (320~ by 20-200~ (depending on the pressure in the vapor space of the boiler). The lower temperature corresponds to the vacuum-gauge pressure and the higher temperature to atmospheric pressure. The fusion cycle consists of the following operations: filling the boiler with solution; heating the solution to boiling and evaporating it, keeping a constant level in the boiler by means of periodic fresh solution make-up (59 h); expulsion of the melt from the boiler by compressed air (4--6 h); cooling of the boiler by blowing air at 90-120~ (5-7 h). The fusion cycle lasts for 68-72 h. For the most part of that time the solution evaporates under variable vacuum-gauge pressure depending on the solution temperature (0.79 MPa at the beginning and 0.033 MPa at the end of fusion at the melting point ~300~ The level of the solution in the boiler is maintained automatically while the pressure is regulated manually by the operator, which requires certain skills since an abrupt pressure drop in the apparatus can cause undesirable frothing of the solution. One fusion cycle yields about 6.5 tonnes of caustic soda, which is then cooled and granulated. Periodic fusion has some obvious disadvantages, namely, unproductive loss of time and electric power consumption. A more profound analysis showed that, first, the evaporation productivity is tow. The explanation for this is that heat transfer by natural convection to a boiling solution in a large volume is low. The coefficient c~ of heat transfer from the inner wall of the boiled during fusion by a moderate heat flux (-25 kW/m 2 in the given case) is: roughly 100 W/m2.~ at the beginning of fusion (boiling of a 42% solution) and 67 W/m2.~ at the end. That results in the outer wall of the boiler overheating (to 377427~ and rapid nickel corrosion (especially at the end of the process). The specific evaporated moisture output is about 25 kg or 6.3 kg of caustic soda per hour per 1 m 2 of boiler heating surface area. Second, if the process is fully automated, its algorithm is considerably more complicated than in the case of a continuous process. We have developed a continuous one-stage method of obtaining melts of sodium and potassium nitrates and alkalis from solutions (the method is described below), based on the "sharp" jet principle [1]. A gas (gas-liquid)jet deep below the level of the liquid is said to be "sharp" [2]. In [1] it was proposed that those solutions be sprayed by air (or superheated steam) by means of an injector. The injector was placed below the level of the given melt, which was superheated relative to the melting point. Heat and mass exchange proceeds rapidly in the jet formed at the injector tip: evaporation of water, crystallization and melting of chemical compounds of the sprayed solutions, and superheating of the phases formed. After bubbling through the melt the steam-gas phase is removed and the compounds precipitated from the solutions remain in the melt. The continuity of the process is ensured by continuously adding a certain amount of heat to the melt (in order to maintain the given temperature in accordance with the heat balance of the process) and pouring off part of the melt to maintain the initial level. The advantages of the proposed method are obvious since heat and mass transfer proceeds vigorously in the jet when the media are in direct contact. Moreover, bubbling causes mixing of the melt, which also speeds up the heat transfer to the melt (we call it external) from the heating surface. Institute of Engineering Science, Ural Branch of the Russian Academy of Sciences. Translated from Khimicheskoe i Neftegazovoe Mashinostroenie, No. 3, pp. 13-16, March, 1999. 142
0009-2355/99/0304-0142522.00 9 1999 Kluwer Academic/Plenum Publishers
To terminal block
9
1
~
~
//
i
-J'11Illll
~
_
~
8
7
ol er / -
---~.~
"~
"
So31der
Melt
Fig. 1. Calorimeter for measuring the local coefficients of heat transfer to the melt: 1) copper hemisphere; 2) wire electric heater; 3) mica plates; 4) metal clip; 5) ceramic bead ; 6) copper wire; 7) ceramic straw; 8) thermocouple for measuring the wall temperature; 9) tube; 10) thermocouple in the melt; 11) hot junction of thermocouple.
The proposed method was investigated on laboratory apparatus with simulated feed of an air-water jet into a sodium nitrate melt [3]. The conditions for stable evaporation of water were determined: superheating of the melt > 20~ and water content in the jet no more than 0.76 kg H20/kg of mixture. An examination of the temperature field in the melt, into which the jet was fed from top downwards, revealed a nonisothermal zone adjacent to the nozzle of the injector. The temperature along the injector axis varies from the temperature of the mixture fed to the temperature of the melt. In the cross sections of the melt (during motion away from the injector axis) the temperature varies from the value on the axis to the temperature of the melt. The temperature is the same in the rest of the melt and is specified in the experiment. The geometric coordinates of points in the nonisothermal zone with a temperature equal to the constant temperature in the melt are called its temperature boundaries (longitudinal L and maximum transverse D). Clearly, internal heat and mass transfer processes occur in that zone. Equations were derived for calculating the dimensions of the temperature boundaries of the zone of internal heat and mass exchange (henceforth, zone) L and D [3]. To keep the solution from boiling and scale from forming on the walls of the apparatus, the surface of the external heat supply should be outside those boundaries. The temperature boundaries of the zone, therefore, are determined by the dimensions of the apparatus. The configuration of the zone boundaries is described by the equation of an ellipsoid of revolution with semiaxes corresponding to L and D and the volume of the zone can be calculated from Vz = 0.525LD 2. The high value obtained for the specific evaporated moisture output (124-426). 103 kg/m3.h indicates a high rate of heat and mass exchange in the zone. An important feature of the proposed method is that the melt is removed from the apparatus by a steam-air jet. The rate of removal was estimated (in order to obtain primary data for the apparatus design) by analyzing steam condensate for sodium nitrate condensate. For various modes of air-water jet delivery, the sodium nitrate content in the condensate was 3-6 g/dm 3. The height of the steam space above the melt, without separation devices, was 1 m and the maximum calculated velocity of the steam-air jet was less than 0.5 m/sec. The dimensions of the part of the apparatus containing the mass of melt are determined, on the one hand, by the dimensions of the temperature boundaries of the zone and, on the other hand, by the area of the external heat-transfer surface outside the zone boundaries. That area depends on the coefficient ~ of heat transfer to the melt. Ramm [4] gives the values of ~ for air-water bubbling systems and systems of SO2(SO 3) plus various solutions (fuming sulfuric acid, sulfuric acid) with a uniform gas distribution in the foaming mode (gas velocity more than 0.4 m/sec). According to those data, o~ = 1350-1850 W/m2.K and does not depend on the physical properties of the liquid. Formulas for calculating the heat-transfer coefficient are given in [5, 6]. I43
(Z L,
W/m 2. K
6000 5000
4000 3000
2000 1000 0
50
100
150
2Q0
250
300
H, mm
Fig. 2. Dependence of the local coefficient o~L of heat transfer to the melt on the distance H between the calorimeter and the bottom of the apparatus as well as on the distance h from the injector axis: 1) h = 0 (along the axis); 2) h = 0.025 m; 3) h = 0.045 m; 4) h = 0.07 m. Experimental conditions: air-wave mixture flow rate 16 kg/h, air flow rate 4 kg/h, diameter of injector nozzle 2.5.10 -3 m, melt temperature 350~ calorimeter heat load 30 kW/m 3.
The melt system under consideration is characterized by a substantially nonuniform distribution of bubbling phases over the cross section of the apparatus. Clearly, in that case oc, which depends on the hydrodynamic conditions near the heat-transfer surface, varies along that surface. Data on the local heat-transfer coefficient c~L, however, are not available in the literature. The necessary experimental data were obtained and tests of the apparatus described in [3] were carried out as follows. A vertical pipe (length 0.35 m, diameter 0.01 m) was welded into the bottom of the melting pot for draining off the production melt and four thermocouples were caulked into the casing wall of one electric heater. Three of the thermocouples were put on the generatrix of the cylindrical casing, turned toward the jet (one at the middle and one 50 mm from each end of the generatrix). The fourth was placed at the middle of the opposite generatrix of the casing. The local values eL were obtained by placing a small calorimeter (instead of a moveable thermocouple) at various places in the melt. The calorimeter (Fig. 1) was a 0.015-m hollow sphere consisting of two hemispheres 1. Inside the sphere was an electric heater made of nichrome wire with a resistance of 1.7 ft. The turns of the coil were insulated from the sphere wall and from each other by mica plates 3. One end of the coil was caulked to the sphere wall and the other was connected to the insulated copper wire by a metal clip. The wire was insulated by a ceramic bead 5 and straw 7. The junction of a thermocouple, made of thermocouple cable 0.0015 m in diameter, was caulked to the inner wall of the sphere. Through the tube 9 soldered to the sphere, the cable and the copper wire from the coil were led out to the terminal block mounted on the device for moving the calorimeter. The thermocouple 10 was mounted on the tube and its junction was attached to the outside of the tube, 0.003 m from the surface of the sphere; the cold ends were led out to the terminal block. With an ac voltage of 6-7 V applied to the calorimeter heater, the specific heat load on the calorimeter surface was 30-40 kW/m 2. Moving the calorimeter to various points in the melt with given coordinates, we recorded the value of ~L for the given point of the surface of heat transfer to the melt, assuming that because of its small size the calorimeter does not affect hydrodynamic conditions of the motion of the melt at the point where otL is measured. The proposed method and calorimeter can be useful for studying the local heat transfer in various liquid media with nonuniform hydrodynamics near the heat-transfer surface. Experiments on obtaining caustic soda were carried out on laboratory apparatus so modernized [3]. Into a caustic soda melt 45% caustic soda solution was supplied (at a rate of 12 kg/h) by spraying with air flowing at 4 kg/h. The temperature of the melt was 350~ in one case and 435~ in the other. The solution and air flow rates were chosen so that the velocity of the vapor-gas phase calculated for the area of the free cross section of the apparatus was about 0.5 m/sec. The process ran stably as melt was poured off into a container where it crystallized. Each experiment lasted 6-8 h. The moisture content of the solid was 144
,
•
Steamfrommains(dry,saturated)0.8MPa olution(42%NaOH)262kg/h,tmax =113.6"C
t = 450"C,195kgsteam/h+ p0~59kg NaOH/h(removed)
. • t = 150,*C,223kgstearn/h+0.59NaOHtn
" ~
1. E
'
Condensate283'k g / h , . ~
6
-.-
t = 1000C
-
73.71 kg steam/h, ..1 t = tmax+154oC Solution(47%NaOH) 234kglh, tb= 138~
~
SteamtOatmOsphere k gtO / 0h~' ' = 195
'
'
5
Melt(109.41NaOH/h + 1.5kgH2Olh),t = 4500C
268kg/h,t = 20~
Fig. 3. Apparatus-technology flowchart for continuous production of solid reactive caustic soda: I) melter; 2) heat exchanger; 3) heater for initial solution; 4) tank with the initial solution; 5, 6) controlled-volume pumps for the solution.
determined; it was 1.35% when the experiment was performed at 435~ and 2.3% at 350~
The zone dimensions L = 0.105 m
and D = 0.03 m, measured from the temperature boundaries, agreed with the calculated values. The specific evaporated moisture output of the zone was 132.103 kg/m3.h; the output of the laboratory apparatus was 130 kg of melt/h (per m 3 of apparatus volume) at a heat load of 30 36.8 kW/m 2, which was almost five times that for periodic melting. During the experiments the values obtained for a calorimeter placed near the surface of the casing of the main electric heater were taken for the local heat-transfer coefficient
O~L
and the values obtained for the main electric heater were taken
for the average ~ave" In determining O~ave we used the root-mean-square value of the readings of four thermocouples measuring the wall temperature of that electric heater. The dependence of cxL on the location of the calorimeter in the melt (the calorimeter was beyond the zone boundaries) was thus obtained. Analysis of the experimental data (Fig. 2) allows us to make the following conclusions. The external heat transfer is nonuniform over the height of the apparatus: the rate is highest at the zone boundaries. If we consider the surface of heat transfer to the melt in the form of vertical tubes arranged in that melt (as in the given apparatus) or in the form of the side walls of the apparatus, the coefficient ccL decreases with their distance from that boundary. The coefficient
(~L
the melt,
also varies over the height of such a surface. In the bottom part of the apparatus, where a gas-liquid jet penetrates into
O~L =
2300-3300 W/m2.K is the smallest value. With distance from the bottom of the apparatus cxL increases, reaching
4500-5500 W/m2-K. The quantity ~xL stabilizes at a height of roughly 2L: the largest value c~L = 5500 w/mZ.K corresponds to the axis of the apparatus and the distance from it to the zone boundary. In experiments at an average distance of 0.07 m between the surface of the casing of the main electric heater and the zone of internal mass transfer, O~ave = 3860 W/m2-K. The calculation (Xave -- 3530 W/m2.K, obtained with a calorimeter, differs 8% from the ~ave measured on the surface of the main heater, which is within the limits of measurement accuracy of ~. That attests to the reliability of the method of determining a L by using a small calorimeter. When the external heat transfer during the production of caustic soda was studied by the proposed method, the heat-transfer coefficient was found to be 30-50 times that for the boiler wall under periodic melting. That means that despite the higher average process temperature (350-450~
the apparatus wall temperature is substantially lower than during peri-
odic melting. The experiments revealed that the continuous method of melting reactive caustic soda has the following advantages over periodic methods: fast process (five times more product per sq. m. of apparatus volume); simple control and automation of the process, according to one parameter (melt temperature), the same for the entire volume of technological space; slower corrosion of the material because the apparatus wall temperature is considerably lower than its maximum value under periodic melting. The experimental data obtained are necessary for designing melters with a "sharp" jet. An apparatus-technology flowchart for continuous production of reactive caustic soda is proposed on the basis of the results of the tests (Fig. 3). The setup is assumed to have an output of 110 kg of caustic soda melt per hour, which is equivalent 145
to a nominal continuous output of melt in a periodic process. The process temperature was 450~ which ensures that the product has a moisture content of less than 1.35%. To prevent the product from carbonizing, dry saturated steam at a pressure of 0.8-1 MPa is chosen as the spraying agent. The technological flowsheet (Fig. 3) shows the calculated material flow parameters (composition, flow rate, temperature). The initial solution is sent by the pump 5 to the condenser/heat-exchanger 3, where it is heated to 113~ by the heat of the liquor vapor from the melter. In the apparatus 2 in direct contact with the superheated steam from the melter, the solution is heated further and part of the water evaporates. As a result, the concentration of the NaOH solution rises to 47%, and the temperature to 138~ The hot 47% NaOH solution is sent by the pump 6 through the injector into the melter. Spraying steam, superheated to 360~ in the apparatus 2 by the superheat of the liquor vapor from the melter, is fed into the injector. The liquor vapor at 450~ gives up part of the heat to the solution in apparatuses 2 and 3, and at a temperature ~ 100~ is discharged into the atmosphere (or can be condensed by supplying cooling water to the condenser). Hot melt from the melter is continuously poured off and transported by pipe for granulation. The melter is a cylindrical apparatus of diameter 0.75 m and height 2.2 m with two injectors in the bottom part and a lid on a flange connector. The melt vat 1.2 m high (volume 0.44 m 3) is heated from the outside by a 100-kW electric heater that delivers 37 kW/m 2. The melter can be made of two-ply steel (St3-nickel), with a 1.5-mm nickel cladding. For a nickel corrosion rate of 0.135 g/m2.h under such conditions [7], the melter lifetime is roughly 10 yr (design wall temperature 467~ The main design technological indicators of the continuous process under consideration are: specific electricity consumption 1.3-1.4 kW/kg of melt and specific steam consumption 670 kg/tonne. The use of a continuous-process facility with the main apparatus in the form of a melter with a "sharp" jet saves nickel; the process is controlled and can be fully automated; the equipment has smaller temperature drops during operation and thus has a longer lifetime; and the output is higher without enlarging the production area or increasing the personnel. The proposed technological scheme for continuous production of a caustic soda melt can be applied., e.g., to the preparation of electrolyte in the production of metallic sodium.
REFERENCES
. .
.
5.
.
7.
146
Inventor's Certificate 355101 SSSR,"Method of producing melts of nitrates and bases of alkali metals," ByulL Izobret., No. 8 (1989). E. D. Khmelevskaya and V. F. Chukhanov, "Investigation of the hydrodynamics and mass exchange of a "sharp" gas jet with a liquid," Dokl. Akad. Nauk SSSR, 168, No. 6, 1307 (1966). V. A. Kulikov, V. I. Davydov, and L. L. Murav'ev, "Zone of heat and mass exchange of air-water jet in a sodium nitrate melt," Inzh.-Fiz. Zh., 23, No. 3,435 (1972). V. M. Ramm, Absorption of Gases [in Russian], Khimiya, Moscow (1966). V. N. Sokolov and A. D. Salamakhin, "Heat exchange between a gas-liquid mixture and heat-exchange elements,"
lnzh.-Fiz. Zh., 25, No. 11, 2570 (1962). V. V. Konsetov, Int. J. Heat Mass Transfer, 9, 1103 (1966). E. I. Kogan, Research on Corrosion Protection of Metals in the Production of Calcined Soda and Coproducts [in Russian], NIOKhIM, Kharkov (1972).