Print version ISSN 0104-6632
Braz. J. Chem. Eng. vol.17 n.4-7 São Paulo Dec. 2000
PINCH ANALYSIS OF EVAPORATION SYSTEMS
D.L.Westphalen1and M.R.Wolf Maciel2
1Departamento de Engenharia Química e de Alimentos, Escola de Engenharia Mauá/IMT,
Estrada das Lágrimas 2035, CEP 09580-900, Phone: (55) (11) 741-3060,
Fax: (55) (11) 741-3131, São Caetano do Sul - SP, Brazil
2Faculdade de Engenharia Química, Laboratório de Desenvolvimento de Processos de Separação
(LDPS) / DPQ, UNICAMP, C.P. 6066, CEP 13081-970, Phone: (55) (192) 39-8534,
Fax: (55) (192) 39-4717, Campinas - SP - Brazil,
(Received:November 19, 1999 ; Accepted: April 6, 2000)
Abstract - Evaporation systems are separation processes widely used in chemical industries. Some guidelines can be found in the literature for the process integration of multiple effect evaporators. In the published methodologies some aspects are neglected as boiling point rise, effect of pressure on latent heat of water, sensible heat of liquid streams, heat of mixing, effects configuration and inclusion of accessories. In this work, a new graphical representation for the integration of multiple effect evaporators was developed, using rigorous physical properties. From this representation, an algorithm for optimization of bleed streams was conceived using the concepts of Pinch Analysis. As a case study, a crystal glucose plant was optimized using this new methodology. The optimization of bleed streams showed as result a steam consumption 16% smaller than a similar previous study. From energy and capital costs, it is shown that the integrated evaporator exhibits a total cost 14% smaller than the non-integrated configuration.
Keywords: evaporation, optimization, energy, pinch analysis.
Evaporation systems are separation processes widely used in chemical industries. There is a lack of established methodologies for process design of this unit operation because of the large number of possibilities for number of effects, effects configuration (frontal, reverse or mixed) and inclusion of accessories (mechanical compressors, thermocompressors, heat exchangers, flash cooler, bleed streams, and condensate recovery systems).
Some guidelines can be found in the literature for the process integration of multiple effect evaporators (Kemp, 1986 and MacDonald, 1986). These guidelines are based on a heat cascade representation of the evaporator in a temperature enthalpy diagram, against the grand composite curve of a background process. In the published methodologies some aspects are neglected in this cascade representation of an evaporator, as, boiling point rise, effect of pressure on latent heat of water, sensible heat of liquid streams, heat of mixing, effects configuration and inclusion of accessories.
In this work, it is used a rigorous heat profile of evaporation systems. This new heat profile takes into account all those aspects neglected by other authors cited above. Some examples of industrial evaporation systems were studied and the analysis of the heat profile of these evaporators provides a better understanding of this equipment.
The optimization of evaporators have been traditionally performed as a stand-alone unit operation. This optimization includes basically the optimum number of effects in a multiple-effect configuration and the use of recompression systems. However, an evaporation system is always a part of an overall process and its optimization should take into account integration with the background process. In this work, it was developed a new methodology for process integration of evaporation systems. This tool is based on the rigorous heat profile of a evaporation system and its placement against the grand composite curve of the background process.
Cascade diagrams are frequently used in process integration studies. Townsend and Linnhoff (1983) developed the grand composite curve, where all hot and cold streams of a process are plotted in a temperature enthalpy diagram. In this diagram, all hot streams are shifted in DTmin/2 and all cold streams are shifted in +DTmin/2. This minimum temperature difference DTmin is chosen from economic considerations.
Kemp (1986) proposes a graphic representation of multiple-effect evaporators in a shifted temperature enthalpy diagram. In this representation, sensible heat and changes in latent heat of water are neglected. In this way, an evaporation effect can be represented by a rectangle as shown in Figure 1a. In this Figure, the upper line of the rectangle represents the heat duty received by the effect for water vaporization. The lower line represents the heat duty available if the vaporized water in the effect is condensed and used as heating medium in another equipment. The vapor leaving the effect is in the same temperature of the liquid boiling inside the effect. However, since in this diagram it is used the shifted temperature scale, the distance between these two lines is exactly the minimum temperature difference. In Figure 1b, it can be seen an evaporator with two effects, where the shifted temperature of the vapor from the first effect is the same shifted temperature of the boiling liquid in the second effect. In Figure 1c, it is plotted an evaporator with three effects, operating with the same temperature difference. This temperature enthalpy diagram emphasizes how heat is cascaded through the effects.
MacDonald (1986) and Kemp (1986) explain that energy integration of an evaporator with a background process can be obtained when the "boxes" representing evaporation effects are accommodated in the grand composite curve of the background process.
Smith and Linnhoff (1988) established some general principles for the optimization of separation processes in the context of a global process. From the heat profile described above, these authors explain how the operation pressures of the effects affect the heat integration of the equipment. From the same principles, Smith and Jones (1990) developed an algorithm for the heat integration of evaporation systems based on temperature enthalpy diagrams. Again, some aspects as sensible heat of liquid streams, changes in latent heat of water, use of vapor recompression and effect configuration were not studied by the authors. Also, these authors assume that the heat transfer coefficient has the same value in all effects, and this assumption leads to a conclusion that the minimum capital cost of the equipment will be obtained if all effects operate using the same temperature difference. However, it is observed that heat transfer coefficients are strongly influenced by temperature and solids concentration (Pacheco et al., 1999) and therefore their conclusions are not suitable for real cases. Indeed, the algorithm proposed by Smith and Jones (1990) is based on shifting effects pressure, however, in some cases as food processing, shifting pressures is limited because of restrictions as undesirable reactions.
It can be concluded that a rigorous heat profile of an evaporation system should be developed and it is shown in this work how this new heat profile provides a better understanding of the heat interactions of an evaporator. Also, a new algorithm for process integration of evaporation systems is presented and it is discussed that this new algorithm does not rely on simplifications that could result in unreliable process configurations.
HEAT PROFILES OF EVAPORATION SYSTEMS
The rigorous representation of a simple evaporation effect can be plotted in two basic manners, depending on feed stream temperature. In Figure 2a, it can be seen an effect where feed stream temperature is lower than the boiling temperature of the fluid inside the effect. In this Figure, lines representing steam and cooling water are also plotted. As steam only changes latent heat, its temperature is constant and therefore it is drawn as an horizontal line. At the other side, cooling water changes sensible heat and it is drawn as an inclined line. The evaporation effect is drawn as a trapezoid. The upper line can be understood as the heat duty received by the effect and it is plotted at the boiling temperature of the liquid shifted +DTmin/2. The bottom line of the trapezoid corresponds to the heat duty available from the vaporized water in the effect. In this case, all this heat duty is received by cooling water at the condenser. The shifted temperature used in this bottom line is taken from pure water boiling temperature at the effects pressure. When the boiling point elevation of the solution is considered, water leaves the effect as superheated vapor. However, when this vapor is used as a heating medium, it condenses at the pure water saturation temperature, and this value has to be taken into account for heat transfer driving forces purposes.
When the boiling point elevation is not considered, the distance of the top and bottom lines of the trapezoid is exactly the minimum temperature difference. When this aspect is considered, the distance between these lines is greater than the minimum temperature difference. The difference in length of these two lines in the enthalpy axis is equal to the sensible heat duty necessary to increase feed stream temperature until the effect temperature. It is concluded that part of steam is used to increase feed stream temperature and this amount of heat can not be cascaded to other effects by means of vaporized water.
In Figure 2b, feed stream temperature is greater than effect temperature. In this case, sensible heat of the liquid stream assists the evaporation and as a result the amount of available energy for heating purposes is greater than the consumed energy. This aspect can be seen in Figure 2b, where the bottom line of the trapezoid is greater than the top one.
In a particular case, feed stream temperature could be exactly equal to effect temperature. In this situation, only latent heat would be involved and top and bottom lines of the trapezoid would have the same length.
Figure 3 shows three evaporation systems with three effects, at the following configurations: frontal, reverse and mixed. For all these situations, it was assumed that the feed stream temperature was lower than the corresponding effect temperature. Also, temperature differences between all effects were the same and equal to the specified minimum temperature difference.
In the frontal configuration (Figure 3a), liquid from any effect always follows to a lower pressure effect, and consequently, to a lower temperature effect. Sensible heat from liquid streams assists evaporation and as a result the amount of evaporated water increases through effects.
At reverse configuration (Figure 3b), liquid from any effect always follows to a greater pressure effect, and consequently, to a greater temperature effect. In this situation, heat is consumed in all effects in order to increase liquid temperature (sensible heat) and therefore the amount of evaporated water decreases through effects.
At mixed configuration (Figure 3c) there is no fixed rule for liquid flow pattern. The profile of evaporated water will be defined as a result of the liquid flow pattern.
As a practical rule, the most efficient evaporation system, from an energetic point of view, will be that one where the smallest amount of evaporated water is sent to the condenser. It can be seen from the heat profiles shown above that the profile of evaporated water is dependent of equipment configuration. From this principle, it can be concluded that reverse configuration shows the lower energy consumption. This is not a general rule, because the influence of feed stream temperature has to be included in a thorough analysis. However, it has to be stressed that sensible heat terms of liquid streams can not be neglected in process integration studies of evaporation systems.
Flash condensate system
Commonly, condensate streams are returned to boiler, or discarded to a effluent system. As the heating medium provided in each effect is used in a different pressure level, it is possible to send the condensate stream from an effect to be mixed with the heating medium of next effect. This technique allows the reduction of energy consumption and some studies have concluded that savings of 5% can be reached in steam consumption (Westphalen, 1999). Figure 4 shows the representation of an evaporator using and not using flash condensate system. When a condensate stream is not flashed on next effect, it can be seen in this figure that not all the available energy from one effect is cascaded to next effect.
This technique allows the integration of the evaporator with other process streams. Vapor from an effect is divided and one part flows to next effect and the other part is sent as heating medium to a process stream external to the evaporator. Vapor bleeding implies that not all heat is cascaded through the effects, as can be see on Figure 5.
Vapor leaving the last effect can not be used as heat source because of its low pressure level. Therefore, one way to reutilize this vapor is to recompress it. This can be performed with mechanical compressors or thermocompressors. In both ways, the amount of energy provided by recompression to the first effect is greater than the amount of heat that is not sent to condenser, as can be seen in Figure 6. This is due to the release of shaft work as heat (compressor) or the addition of the jet steam enthalpy (thermocompressor).
PROCESS INTEGRATION OF EVAPORATION SYSTEMS
In this work, it is proposed a graphical representation for process integration of evaporation systems. This representation consists of the grand composite curve of the background process and the heat profile of the evaporator mirrored around the temperature axis. This graphical representation is called evaporators placement diagram (EPD).
In Figure 7, it can be seen the EPD of a system with three effects and frontal feed in two different situations. In Figure 7a, the minimum temperature difference of the background process and temperature difference between effects are the same. In Figure 7b, these temperature differences are not the same, and it can be observed that the temperature difference between effects #1 and #2 is lower than the background process minimum temperature difference, while for effects #2 and #3, temperature difference between effects is greater than the process minimum temperature difference. When a multiple-effect evaporator is designed following the same area criteria, temperature differences between effects are not the same leading to a situation exemplified in Figure 7b.
Process integration between an evaporation system and a background process can be performed by means of bleed streams. Figure 8a shows a triple effect evaporator where bleed streams are extracted from effects #1 and #2. The cascade philosophy of this diagram emphasizes that part of the heat provided by effects #1 and effect #2 are cascaded to the next effects and part are cascaded to the background process.
Bleed streams are always represented by "steps" located at the right side of the evaporator heat profile, and these steps overlap the process grand composite curve. If vapor recompression is employed, it is represented at the left side of the evaporator heat profile, suggesting that this technique does not affect the integration of the equipment with the global process, as exemplified in Figure 8b.
ALGORITHM FOR OPTIMIZATION OF BLEED STREAMS
In this work, it was developed an algorithm for optimization of bleed streams based on Pinch Analysis principles. The algorithm, was implemented in an evaporation systems simulator (Westphalen and Wolf Maciel, 2000) and can be described by the following steps:
Step 1) From hot and cold streams of the background process, and a specified process minimum temperature difference, calculate the hot and cold utility targets and locate pinch temperature, using the "Problem Table" algorithm described by Linnhoff and Flower (1978), Linnhoff et al. (1982) and Douglas (1988).
Step 2) From the evaporation system previously described by the user (number of effects, liquid flow pattern, accessories, feed stream information, product concentration and operating pressures of all effects), identify the effect operating with the smallest pressure, that is, connected to the condenser.
Step 3) Verify if there is a bleed stream connected to the effect analyzed. Following a philosophy where the process engineering is responsible for the main decisions during the optimization, only those effects where the user has explicit added a bleed stream will be analyzed. So, if the effect in analysis does not have a bleed stream, go to the next one. If the effect is the last one and does not have a bleed stream, stop.
Step 4) Calculate temperature of the effect and shift this value by DTmin/2, because for the background process, a bleed stream must be considered as a hot stream.
Step 5) Compare the effect shifted temperature and process pinch temperature. Following the basic pinch principles, hot utilities can not be used in a temperature level lower than the process pinch temperature. So if the effect shifted temperature is lower than the process pinch temperature, zero is assigned to the bleed stream flow rate and the algorithm search for the next effect and goes back to step 3. Otherwise, the algorithm goes to the next step.
Step 6) Calculate the heat flow from the grand composite curve for the effect shifted temperature.
Step 7) Verify if this heat flow is located inside an envelope in the grand composite curve. Figure 9 shows two cases for heat integration of an effect. In Figure 9a, all heating source for the effect comes from steam or from a previous effect. In Figure 9b, part of the heat consumed by the effect comes from the process and the rest comes from steam or from a previous effect. In this case, no additional hot utility would be saved because an envelope in the grand composite curve means the possibility of process to process heat transfer. Indeed, the integration of an effect inside an envelope would lead to the use of smaller temperature differences, and as result greater heat transfer areas. So, it is important to avoid this kind of heat integration. This verification is done comparing the heat flow calculated in step 6 with the hot utility target. If this heat flow is greater than the hot utility target, the heat duty of hot utility target is assigned to the bleed heat duty of the analyzed effect. Otherwise, the heat duty calculated in step 6 is directly assigned as the bleed heat duty.
Step 8) Calculate the bleed stream flow rate as the bleed heat duty divided by the latent heat of water at effect pressure.
Step 9) Verify if the bleed stream flow rate is greater than the maximum bleed stream flow rate specified by the user. When heat duties existing in a evaporator are too small when compared to process heat duties, the evaporator placement diagram depicts a situation as shown in Figure 10. In this Figure, it can be observed that there is no overlap between effects #2 and #3. Mathematically, this corresponds to a negative flow of vapor from effect #2 to effect #3, that is, an infeasible process. This kind of situation can not be anticipated, and in order to avoid it, the user has to specify a maximum bleed stream flow rate. In case the calculated flow is greater than the specified value, this specified value is assigned to the effect bleed stream flow rate.
Step 10) Search for the next effect in the multiple-effect structure and goes back to step 3.
The procedure is repeated until all effects, from the lowest to greatest pressure, are analyzed.
CASE STUDY: CRYSTAL GLUCOSE PLANT
The proposed algorithm for optimization of bleed streams will be illustrated in a case study: the production of crystal glucose plant, using data presented by Klemes et al. (1998).
The industrial process for production of glucose consists basically of the hydrolysis of corn starch, followed by product purification (Shenk, 1989). Stream data for this process were adapted from data presented by Klemes et al. (1998) and are shown in Table 1. Evaporator data is not included in Table 1, as in the original paper, because the integration of this equipment will be performed using the algorithm developed in this work.
Figures 11a and 11b show the composite curve and grand composite curve, respectively, of the background process using data from Table 1 and a minimum temperature difference of 8°C, as published by Klemes et al. (1998). Hot and cold utility targets and pinch position were calculated using the problem table method, resulting in 2718 kW, 634 kW and 56°C, respectively.
In this process, there is a triple effect evaporator with frontal feed configuration. This equipment was simulated with an evaporation systems simulator (Westphalen, 1999) using data presented in Table 2. A steam consumption of 1781 kg/h was calculated.
In this process, two steam levels are available: 110°C (144 kPa, low pressure level) and 150°C (477 kPa, medium pressure level). From the grand composite curve, it can be concluded that the hot utility target can be divided in these two steam levels: 2399 kW and 319 kW, or 3883 kg/h and 550 kg/h for the levels 110°C and 150°C, respectively.
The algorithm developed in this work was used to optimize the integration of the evaporator and the background process, using 3000 kg/h as the maximum bleed stream flow rate. The evaporator placement diagram is shown in Figure 12, and Table 3 summarizes the results.
The integration of the evaporator and the global process allows a saving of 3435 kg/h of steam at low pressure level. This value is obtained from the sum of all bleed flow rates. For the non integrated process, the total low pressure steam consumption is 5664 kg/h: 3883 kg/h from process streams and 1781 kg/h from the evaporator. For the optimized process, total low pressure steam consumption is 4343 kg/h: 448 kg/h from process streams and 3895 kg/h from the evaporator. This represents savings of 23% of this hot utility. Considering the total low pressure level steam consumption and also the medium pressure level steam consumption, it can be calculated a total hot utility target of 2894 kW for the integrated process. Klemes et al. (1998) using traditional Pinch Analysis tools to the same process reached a total hot utility target of 3435 kW. The value obtained in this work is 16% smaller than the value obtained by those authors.
Process industries are not exclusively concerned with energy savings. Capital costs have also to be analyzed together with these energy savings. The trade-off between energy costs and capital costs has to be computed for a decision concerning process modifications. A economic study was developed using data presented in Table 4.
Tables 5 and 6 present the results for the economic analysis of the crystal glucose plant, without the integration of the evaporator and integrating the evaporator, respectively. The cost of the heat exchanger network was performed using the methodology developed by Ahmad and Smith (1989), using 0.4 kW/m2°C for the heat transfer coefficient of all process streams. Global heat transfer coefficient for the evaporator effects were calculated using the Baloh equation (Radovic, et al., 1979).
Comparing the figures shown on Tables 5 and 6, it can be noticed that the integration of the evaporator results in a reduction of energy costs. This fact is a direct consequence of the global reduction in steam consumption as already stated. It can be also noticed that the integrated process shows a smaller capital cost for the evaporator. As the initial and final concentrations of liquid streams of the evaporator are the same in both economic evaluations, total evaporated water is the same. However, the use of bleed streams cause a different pattern of how this water is evaporated. It can be observed from Table 7 that when bleed streams are employed, a large amount of water is vaporized from the first effect and consequently this is the effect with the largest heat duty. In a frontal feed configuration, the first effect boils a liquid with the largest temperature and the smallest solid concentration. Due to these two factors, the global heat transfer coefficient of this effect is the largest of the multiple-effect equipment. In conclusion, the redistribution of heat duties and the influence of global heat transfer coefficients result in an integrated evaporator with a smaller capital cost. Adding these two factors, the comparison of the results presented on Tables 5 and 6 shows that the integrated process exhibits the smallest total cost for the global process.
A new algorithm for integration of evaporation systems with a global process was developed, using Pinch Analysis principles. This new tool was implemented in a computer software and exemplified in a case study. It can be concluded from this work that the integration of evaporation systems and a background process is economically viable. It has to be stressed that the optimization method developed in this work does not make use of simplifications as neglecting boiling point rise of solutions, sensible heat of liquid streams and differences on latent heat of water. This algorithm can be used in any kind of evaporator configuration, even when some complex accessories are used as vapor recompression, flash coolers, flash condensate systems, and others. The existing algorithm is being expanded for the analysis of effects with temperatures lower than the process pinch temperature. Also, a new methodology for retrofit of existing evaporators is under development.
|A||Heat transfer area (m2)|
|Cef||Cost of evaporation effects ($)|
|Chx||Cost of heat exchangers ($)|
|DTmin||Minimum temperature difference (°C)|
Ahmad, S., and Smith, R., Targets and Design for Minimum Number of Shells in Heat Exchanger Networks, Chem. Eng. Res. Des., 67, September (1989). [ Links ]
Douglas, J. M., Conceptual Design of Chemical Processes. McGraw-Hill, Inc, New York (1988). [ Links ]
Kemp, I. C., Analysis of separation systems by process integration, J. Separ. Proc. Technol., 7, No. 9 (1986). [ Links ]
Klemes, J., Kimenov, G., Nenov, N. e Nedanova, A., Application of process integration in food industry, Proceedings of 13th International Congress of Chemical and Process Engineering CHISA 98, Prag, Czech Republic (1998). [ Links ]
Linnhoff, B., Flower, J. R., Synthesis of Heat Exchanger Networks: I. Systematic Generation of Energy Optimal Netwoks. AIChE J., 24, No. 4, 633 (1978). [ Links ]
Linnhoff, B., Towsend, D. W., Boland, D., Hewitt, G. F., Thomas, B. E. A., Guy, A. R., Marsland, R. H., A User Guide on Process Integration for the Efficient Use of Energy. The Institution of Chemical Engineers, Rugby (1982). [ Links ]
MacDonald, E., Process integration gives new insights on evaporators, Process Engineering, November, 25 (1986). [ Links ]
Pacheco, C. R. F., Cézar, C. A. and Song, T. W., Effect of the solute concentration on the performance of evaporators, Chem. Eng. and Proc., 38, 109 (1999). [ Links ]
Radovic, L. R., Tasic, A. Z., Grozdanic, D. K., Djorjevic, B. D., and Valent, V. J., Computer desing and analysis of operation of a multiple-effect evaporator system in the sugar industry, Ind. Eng. Chem. Process Des. Dev., 18, No. 2, 318 (1979). [ Links ]
Shenck, F. W., Glucose and Glucose-Containing Syrups, Ullmanns encyclopedia of industrial chemistry, VCH Verlagsgesellschaft mbH, Weinheim (1989). [ Links ]
Smith, R. and Linnhoff, B., The design of separators in the context of overall processes, Chem. Eng. Res. Des., 66, May, 195 (1988). [ Links ]
Smith, R. and Jones, P. S., The optimal design of integrated evaporation systems, Heat recovery systems & CHP, 10, No. 4, 341 (1990). [ Links ]
Townsend, D. W. and Linnhoff, B., Heat and power networks in process design: I. Criteria for placement of heat engines and heat pumps in process networks, AlChE Journal, 29, No. 5, pp.742-748 (1983). [ Links ]
Westphalen, D. L., Modelagem, Simulação e Otimização de Sistemas de Evaporação, D.Sc. Thesis, Campinas State University (1999). [ Links ]
Westphalen, D. L. and Wolf Maciel, M. R., Desenvolvimento de um simulador estático e dinâmico de sistemas de evaporação, submitted to 13o. Congresso Brasileiro de Engenharia Química (2000). [ Links ]