Acta Chini. Slov. 2001, 48, 565-573. 565 MULTIPLE-EFFECT EVAPORATION IN ALUMINOSILICATE TECHNOLOGY* Dušan Klinar Faculty of Chemistry and Chemical Engineering, University of Maribor, Smetanova 17, 2000 Maribor, Slovenia Janvit Golob Faculty of Chemistry and Chemical Technology, University of Ljubljana, Aškerčeva 5, 1000 Ljubljana Slovenia This paper is dedicated to Professor Roman Modic at his 90th birthday Received 30-05-2001 Abstract The article presents an engineering approach to the development of an alternative production with the existing equipment. Zeolite production is becoming an alternative to the old "Bayer" process for aluminum hydrate production. By applying appropriate know-how old equipment was used for an economically as well as ecologically acceptable production. In this equipment adaptation among other production steps the multistage evaporation battery was the most critical part of the process. We applied semi empirical approach on the basis of Z. Rant ideas to find heat transfer coefficients and to determine which equipment elements are more relevant for efficient performance and need serious maintenance. Introduction In late 80's and early 90's aluminosilicate - zeolite applications and productions spread rapidly worldwide, mainly for environmental reasons. The environmental reasons affect both the production and the application side. Many old extraction (leaching) processes had to be abandoned and one of the current alternatives is a restructuring process for synthesizing of zeolite from silicate and aluminum compounds in an alkaline medium. In the case discussed here the equipment of an over 50 years old »Bayer« aluminum hydrate production process was adapted to achieve economical zeolite production. It was mainly the economical reason that directed this decision. Relatively low and still falling world prices of zeolite disabled many new investments and old processes can now offer economical production. Fig. 1 presents the basic process principles of zeolite production. From the economic and environmental points of view, a crucial part of the process is concentration of dilute NaOH (cca 6%) solution by evaporation. Disposal of 50 m of the solution per hour is ecologically impossible, and D. Klinar, J. Golob: Multiple-effect evaporation in aluminosilicate technology 566 Acta Chini. Slov. 2001, 48, 565-573. recycling can not be economically efficient. The solution to this is a multistage evaporation process. The purpose is how to test, evaluate and set the parameters of old equipment for the new product. Here we present a concept and the principles of solution, which enables economical operation of the equipment in question. WATER GLASS SOL. AL HYDRATE NaOH il HIGH PRESSURE REACTOR CD hk CD DISSOLUTION PROCESS CONCENTRATED LIQUOR = 15 % NaOH Pti CD ZEOLITE SUSPENSION DILUTE SOLUTION = 8 % NaOH MULTISTAGE EVAPORATION T D R Y E R WASTE STEAM Y DRY ZEOLITE Figure 1. Principal process flow chart of zeolite production Theoretical principles and concepts The basic principles and equations can be found elswere, " but here we limit the discussion to the critical relations and parameters necessary to understand the basic concept. Thermodinamic potential of the process is defined with the condensation temperature of fresh steam in the first stage and the boiling temperature of the solution in o the last stage, while the boiling-point elevation (ebullioscopic) - BPE of the increasingly concentrated solution lowers the available temperature difference. The BPE of the boiling temperature, due to the more concentrated solution in the boiling chamber, can be expressed as: D. Klinar, J. Golob: Multiple-effect evaporation in aluminosilicate technology Acta Chini. Slov. 2001, 48, 565-573. 567 D* rpO.2 * M e m{ *M2 AH, Number 1 indicates the solvent (water) and 2 the solute, m is mass, M denotes molecular weight and AH the latent heat of evaporation, Tb is the boiling temperature. From equations 1, the relationship can be expressed: ATe=f(C2,Pahs) (2) The relationship is used to understand the basic relations between the lost BPE, the pressure and the concentration of the solution boiled. We found BPE data by the help of Duhring's rule, which states that the boiling point of a given solution is a linear function of the boiling point of water at the same pressure. The rule is applicable mainly to the strong inorganic solutions, what is also the situation under consideration. Besides the thermodynamical potential, heat transfer is the most important factor influencing the capacity of the evaporation system. In the discussed case, heat transfer rate is a limiting process for the whole system. Heat flows from the warmer to the colder fluid separated by a solid wall. In this process convection and conduction play the mayor role and thus have become the object of this study. Typical situation is presented on Fig.2. While mayor part of the equipment and process parameters are not known a semi-empirical approach to model the process is applied. The model should enable to analyze the existing equipment configuration, to define the process limitations and to optimize the process parameters. The approach is based on a method of Z. Rant, where it is necessary to establish the relationship AT = f (q) for a used range of heat flow - q. With this result the temperature gradient from a condensing steam side to the boiling solution side is found. By the help of equations the heat transfer from the surface to the boiling solution and the heat transfer from the steam to surface in film condensating regime the limiting heat transfer side and the corresponding heat flux is determined. On the basis of heat flux and total available temperature difference, heat transfer through the tube wall and the scale (usually on the boiling side) can be determined. From these results it is possible to calculate the total heat transfer coefficient in the equipment or, backwards from known or estimated heat transfer, it is possible to establish the scaling thickness on D. Klinar, J. Golob: Multiple-effect evaporation in aluminosilicate technology 568 Acta Chini. Slov. 2001, 48, 565-573. the equipment. Such an approach also implies simultaneous calculation of heat and mass balance, because it is very difficult to get all necessary data for separate calculations. Close attention has to be paid to the scaling and fouling processes because some silicon dioxide alkali products are present in NaOH feed exceeding trace quantities. The effects of scaling and fouling can be monitored through the deviation from initial process parameters and through different model scenarios. dw = wall thickness ds.= scale thickness N—> steam side Tsat Tdiff boiling solution side Tb condensate Figure 2. Heat transfer profile Calculation procedure To calculate the model we used MathCAD program. Calculations were preceded step-by-step, starting with an approximation. In the first step, concentration distribution among evaporators was calculated to find out the evaporated steam flow in each stage. The feed of dilute NaOH was 52 t/h and it contained 8,3% of NaOH, the final concentration should have been 14,5 % of NaOH. It was necessary to evaporate 22.3 t of water per hour in all stages. Heat transfer areas of all four stages were the same (Fig 3). The calculation began with an equal distribution of evaporated water from each stage but the final results gave: 1.00, 0.74, 0.31 and 0.80 in the last fourth stage. From these results concentration distribution was calculated: 12,14 %, 14,5% at second (exit) stage, 8,7% at the third (input) stage and 9,95% at the fourth stage were the results. D. Klinar, J. Golob: Multiple-effect evaporation in aluminosilicate technology Acta Chini. Slov. 2001, 48, 565-573. 569 In the second step, the thermodynamic situation in the evaporation process was evaluated. With the help of Duhring's rule: Tbsol = f (Tbwater, Csol), reverse Antoine's equation for steam condensation temperature, the latent heat of water evaporation at different temperatures and enthalpy of steam at different temperatures enabled us to calculate pressure distribution among stages. The pressures should have been: 2.3, 1.6, 0.9 and 0.2 bar respectively. At this step, all parts of the lost AT can be calculated. The first one was BPE, then hydrostatic pressure (the pipes are 2.4 m high), and transport AT is estimated to be approx. 1 deg on each stage. Finally the temperature differences lost in each stage were computed: 6.6, 8.4, 6.8 and 14.8 degrees. The whole AT available (netto) for all stages was 42.3 degrees and represented the thermodynamic potential for heat transport. Figure 3. Flowchart of four-stage evaporation process In the third part, the heat transfer was estimated. As we mentioned with regard to the principle published by Z. Rant AT had to be established as a function of heat flux. According to presented process the computation scheme started with heat flux value 5.4 D. Klinar, J. Golob: Multiple-effect evaporation in aluminosilicate technology 570 Acta Chini. Slov. 2001, 48, 565-573. GJ/h up to 160 GJ/h. From the heat flux we calculated the corresponding water condensate flow and then the Reynolds number for a different heat flux. With the known Re heat transfer coefficient a from the steam to the heat exchanger tube wall, following was established (film condensing regime): ' ac :1,47-Re~a 'cond 2 fécond g Iconti J (3) In the eq. (3) A is heat conductivity, p is specific gravity, rj viscosity and cond denotes (water) condensate. From the heat transfer coefficient a and known specific heat flux through the tube surface, the corresponding temperature differences on the steam side were calculated. This was the first set of data for the construction of a graph AT = f (q). As heat flowed from the surface of the tube wall through the metal wall (Fig. 2), the temperature drop on a base of heat conductivity was calculated. We got a second set of data AT = f (q). Almost the same situation was on the scale layer in the heat exchanger tubes. This represented the third set of data AT = f (q). Finally, it was necessary to calculate AT for the convection heat transfer to the solution. For pure water the heat transfer coefficient a was calculated using an equation by Kruschlin: aw = 1.95 * q ' * p ' and corrected with factor (J) for solution at the same conditions. The correction factor