THERMODYNAMIC THERMODYNAMIC PROCESS SIMULATION OF THE DIGESTION UNIT IN IRAN ALUMINA REFINERY Mohammad Farrokh* ACECR, TDI and Sharif branch Box: 13445-686 Tehran Iran and Mälardalens University Akademin för hållbar samhälls- och teknikutveckling 721 23 Västerås Sweden Jafar Mahmoudi Mälardalens University Akademin för hållbar samhälls- och teknikutveckling 721 23 Västerås Sweden ABSTRACT Jajarm bauxite is of the chamosite-diasporic type, one of the hardest types of bauxite for alumina production. The Iran Alumina Refinery uses Jajarm deposits and currently operates at 65-70% of its nameplate capacity. Digestion is one of the most important processes in the refinery and has become the main bottleneck in the production. In this paper we study the effect of selected parameters on the behavior of this important section of the refinery using thermodynamic simulation of the digestion unit. The ASTM tables property method was used for vapor streams and the Elec-NRTL property model was used for the global digestion process. The digestion process was simulated with Aspen plus software. The results show good agreement with experimental data. The relationship between the incoming slurry temperature in the first flash tank and the preheater cold stream (slurry into the furnace) temperature and energy consumption is discussed. Key words: Bauxite, Diaspore, Bayer process, S lurry, Alumina digestion, Simulation, Electrolyte NRTL property model
*
Corresponding author: Phone: Phone: +46 76 5824058 and +98 912 3022771 Fax: +46 21 101370 and +98 21 66012497 E-mail:
[email protected]
INTRODUCTION Jajarm bauxite has a fine grain structure, generally smaller than 10 microns, and is of the chamositediasporic type which is one of the hardest types of bauxite for alumina production. The main aluminous Jajarm mineral is diaspore, which contains 39% of the Al2O3 content of the bauxite. The highly stable lattice of diaspore results in poor digestibility and a low Al2O3 equilibrium solubility during the Bayer process. The hardness of diaspore also causes abrasion problems during the Bayer process. Alumina extraction must therefore be carried out under conditions suitable for processing diaspore, which include high temperature and pressure, and specific caustic concentrations during the digestion stage[1, 2]. Although technology and empirical data of tube digestion are available for gibbsitic and boehmitic bauxite, only a few sporadic studies have been performed on tube digestion of diaspore bauxite. Despite these limitations, considerable amounts of diaspore bauxite are processed in various production plants around the world [3, 4, 5]. As a result of the discovery of the bauxite deposits at Jajarm, the Jajarm alumina plant was funded to refine alumina from locally mined bauxite. The average chemical composition of Jajarm diaspore is shown in Table 1.
Component
Percentage
Al2O3 SiO2 Fe2O3 TiO2 L.O.I Others
47.4 10 22.4 5.6 11.3 Rem
Table 1: Average chemical composition of Jajarm diaspore
Jajarm alumina plant The Jajarm Alumina Plant was designed and built by Technoexport (TEX) from the Czech Republic between 1990 and 1999 and has an annual capacity of 280,000 tons. The plant uses the tubular digestion process. The function of the refinery is to separate the alumina from other minerals and refine it to a purity of 98% Al2O3 in the form of sandy alumina. Table 2 shows the main parameters of the process. The refinery plant is currently limited to operating at
65-70% of its nameplate capacity. This is a result of several problems that have emerged during alumina production in the refinery. Several modifications were therefore made on the production line and in particular the digestion stage. As described above, and because of the complexity of the process, it is very important to determine the effect of changing parameters on the behavior of the process. The present study was performed on a macro simulation of the Jajarm digestion unit using a thermodynamic model to determine the influence of parameters such as pressure, temperature and flow rate on the slurry behavior and on other sections of the digestion unit, in particular the flash tanks and preheaters, with the aim of optimizing the energy consumption of the digestion unit.
Figure 1: Part of the Jajarm digestion line
THEORETICAL STUDY The main function of the digestion unit is to dissolve the alumina which exit in the bauxite using caustic liquor and to precipitate out undesirable components such as silicates. This is done at high temperature and pressure. Diasporic alumina dissolves in the caustic solution during the Bayer process according to the simplified equation shown below [6,7]. AIOOH diaspore + OH- + H2OÎ Al(OH)4-; OR DIASPORE + 2NAOH Î 2NaAlO2 + 2H2O In order to recover the thermal energy following digestion, the slurry is passed through a series of flash tanks where the excess energy in the superheated liquid is flashed off as steam. There are
frequent disturbances and a large number of interacting processes in the overall digestion process which contribute to there being considerable dead time. The challenge to any alumina refinery is to maximize the production rate (plant flow and yield) and minimize the energy costs per ton of produced alumina. Successful chemical process development and operation requires the ability to accurately predict process conditions. Changes in material and energy consumption can be determined through heat and mass balance calculations. Process simulation is an important tool for evaluating and improving performance in the plant [8]. As the success of process simulation is highly dependent on the modeling method used to model the process, particular attention should be paid to model selection. The most important point when modeling the Bayer process is the strong non-ideality of phase equilibria due to solids and electrolytes. The effect of dissolved ions in the digestion liquor in raising the boiling point of water has a large influence on the process and has to be modeled accurately [9,10,11]. A thermodynamic model that is capable of reliable simulation of the Bayer type solution over a wide range of temperatures, pressures and compositions is therefore of great interest. The Electrolyte NRTL model was originally proposed by Chen et al., for aqueous electrolyte systems and is the most versatile electrolyte property method. It can handle very low and very high concentrations of electrolytes [11, 12, 13]. ASPEN represents the latest generation of production scale chemical process simulation programs. It considers solid components and allows users to write specific unit operation and physical property models. It also contains many useful property models for electrolyte systems and ASTM steam table correlations. The software allows users to mathematically model or simulate a process to predict what happens when different variables (such as heat, pressure, or raw material composition) are changed, thereby making possible increased efficiency and lower costs [10,12,13,14]. A number of studies have been published about the digestion process and its reactions and the effects of additives. However, there are few reports on the energy consumption and optimization of the process. This study considers thermodynamic simulation of the Jajarm digestion unit.
EXPERIMENTAL DATA After grinding and pre-desilication, bauxite slurry enters the digestion line at around 100 bars and about 100oC and passes through high pressure tube preheaters consisting of nine sections where the slurry is heated by exit vapor from the flash tanks. The slurry is further heated in the digestion furnace to 275oC from where it flows through the tube reactor and digestion tanks to achieve sufficient retention time at the required temperature. After digestion, about 73 percent by mass of the Al 2O3 in the bauxite is converted to a soluble form of sodium aluminate, meaning the amount of Al 2O3 in the liquor increases from 87 g/l to 244 g/l.
Parameter
Amount
Al2O3 content in bauxite Silica ratio of bauxite Caustic concentration Na 2Ok αk of digested liquor Digestion temperature Addition of Lime Caustic ratio of pregnant liquor
48.8% 4.61 190 g/l 1.42 275oC 12% 1.49
Table 2: Main parameters of the process The other solid constituents remain insoluble. About 30% of the bauxite therefore remains in suspension as a thin red mud slurry of silicates and oxides of iron and titanium. The average che mical composition of the red mud slurry is shown in Table 4.
Component
mass%
Al2O3
49.1
SiO2
10.7
Fe2O3
20.6
TiO2
5.8
Table 3: Average chemical composition of the bauxite slurry at entry The mud-laden liquor leaving the digestion vessels is flash-cooled by directing the aluminate slurry through a series of flash vessels. The purpose of the flash line is to gradually reduce the pressure and temperature of the aluminate slurry. Pressure is reduced by using orifices of different diameters at
Figure 2: Flow diagram of the Jajarm Refinery digestion process the inlet of each flash tank. Exhaust vapor from the first nine flash tanks is used to heat the bauxite slurry in preheaters before it enters the reactor. The exhaust vapor then enters condensate flash tanks. Aluminate slurry leaves the last flash tank at a temperature between 125 and 130oC and a pressure of around 0.16 MPa and precipitates as alumina hydrate in subsequent units.
Component
mass%
Al2O3 SiO2 Fe2O3 TiO2 CaO
17.9 14.8 27 7.8 15.2
Table 4: Average chemical composition of the red mud slurry
SIMULATION PROCEDURE AND RESULTS The Electrolyte NRTL property model was used for the entire digestion line and ASTM steam tables were used for the vapor phase. The rise in the boiling point of the liquor was also considered. The simulation was run in the Aspen plus software. The composition of the incoming and outgoing slurry and other parameters of the digestion are shown in Tables 2, 3 and 4. The digestion unit was simulated based on data from the Jajarm Refinery. Simulations were first carried out on individual blocks, followed
by sections containing groups of blocks (such as flash tanks), and finally on the entire line. The flow diagram of the simulated digestion process is shown in Figure 2. Using the simulation program, the amount of vapor created in each flash tank (which is used as a hot stream to each preheater) and the preheaters outlet cold stream (bauxite slurry) temperature were calculated and are shown in Table 5.
Flash tank No.
Vapor created (kg/hr)
Preheater No.
Cold stream temperature (oC)
V06 V07 V08 V09 V10 V11 V12 V13 V14 V15 V16
1899 2146 1567 2681 2352 1992 2713 2427 1595 2170 3195
(9) (8) (7) (6) (5) (4) (3) (2) (1) -
174 166 160 151 144 137 120 115 101 -
Table 5: Preheaters hot stream mass (vapor exiting flash tanks) and cold stream (bauxite slurry) temperature
) C ( e r u t a r e p m e t y r r u l S
) C o ( e r u t a r e p m e t y r r u l S
o
Outgoing preheater cold stream temperature ( oC)
V06 flash tank vapor created (kg/hr)
.
Figure 3: The effect of increasing slurry temperature on the generation of vapor by V06 and the preheater cold stream temperature The effect of incoming slurry temperature on the generation of vapor by V06 flash tank and also on the last preheater cold stream (slurry to furnace) temperature was calculated by the simulation and is shown in Table 6 and Figure 3.
DISCUSSION AND CONCLUSIONS The data shown in Table 5 was computed by the simulation program for the 130 kg/m 3 volume flow rate condition. The data agrees well with the current digestion line conditions and has been accepted by the refinery. Figure 3 shows that the temperature of the incoming slurry into the first flash tank (V06) has a large effect on the efficiency of the preheaters. Decreasing slurry temperature below a specific limit should eliminate flashing and vapor creation. This agrees well with the refinery conditions. Increasing incoming slurry temperature in 2 to 3 oC steps showed that vapor was created by flashing slurry inside the flash tank and this increased according to Figure 3 (right). The last preheater cold stream (outgoing bauxite slurry) temperature is therefore increased by the vapor as shown in Figure 3 (left). There is an upper limit for incoming slurry temperature into the flash tank due to heat loss in the digestion line. 260 oC is therefore an acceptable temperature and is close to the operating condition
of refinery. The Jajarm Alumina Refinery digestion line was simulated using the Elec-NRTL property model and Aspen simulating tool. The simulation showed good agreement with the real conditions and experimental data. This simulation program can be used to optimize energy consumption and operation of the digestion line.
slurry temperature(oC)
Vapor created(kg/hr)
Cold stream (Preheater)(oC)
250
0
178
252
306
180
255
1178
182
257
1094
182
260
1900
185
262
2446
186
265
3283
189
267
3853
191
270
4731
194
Table 6: The effect of slurry temperature on vapor generated in V06 flash tank and the last preheater cold stream temperature
ACKNOWLEDGMENTS The authors would like to acknowledge the Iran Alumina Company and their managers and engineers for their kind help in providing essential data for this research. We would also like to thank Miss Zeinab Farrokh for her assistance in preparing this paper.
REFERENCES [1] Shadloo M.T, Ramezani M. Elburz bauxite mining treatment in a high temperature tube digestion plant . Travaux 2000;31(27):47-62. [2] Technical Description of Heaters11-H-1ABCD. TR:226 Rev.1 Techno Export Erection Documentation [3] Taxiarchou M., Paspaliaris J, Papanastassiou D. Possibilities for improved digestion of diasporic bauxites for alumina production . TRAVAUX de ICSOBA 1997;28(26):235-245. [4] Papanastassiou D.J. Research directions for improved digestion of Greek diasporic bauxites. TRAVAUX de ICSOBA 1999;30(26):137-145. [5] Wargalla G., Brandt W. Processing of diasporic bauxites. Light Metals 1981:83-100. [6] Panias D., Asimidis P, Paspaliaris I. Solubility of boehmite in concentrated sodium hydroxide
solutions: model development and assessment . Hydrometallurgy 2001;59:15-29. [7] Mock B., Evans L. B. and Chen C.C, Thermodynamic representation of phase equilibria of mixed-solvent electrolyte systems , AICHE J. 1986;32(10):1655-1664 . [8] Handbook of Chemistry and Physics, 56th Edition, CRC Press, 1975, p. E-1 [9] Dewey John L. Boiling point rise of Bayer plant liquors. Light Metals 1981:185-197. [10] Zigao L., Shu Lu, Experimental data on the Bayer digest process for Pingguo bauxite. Light Metals 1993:21-25. [11] Konigsberger E., Eriksson G., May P. M., Hefter G. Comprehensive model of synthetic Bayer liquors, Part 1, Overview. Ind. Eng. Chem. Res. 2005;44:5805-5814. [12] Kumagoe Y., Simulation of the Bayer digestion process with solids and electrolytes. Light Metals 1990:69-72 . [13] Chen C.C, Evans L.B, A local compositions model for the excess Gibbs energy of aqueous electrolyte system AICHE. 1986;32(3):444 . [14] Langa J.M, Russell T.G, O’Neill G.A, Gacka P. Aspen modeling of the Bayer process, Process Control and Computer Technology 1991:169178
