Internal heat integrated distillation columns (iHIDiCs)—New systematic design methodology

chemical engineering research and design 8 7 ( 2 0 0 9 ) 1658–1666

Contents lists available at ScienceDirect

Chemical Engineering Research and Design journal homepage: www.elsevier.com/locate/cherd

Internal heat integrated distillation columns (iHIDiCs)—New systematic design methodology Mamdouh A. Gadalla ∗ Universitat Rovira i Virgili, Departament d’Enginyeria Química, Av. Paisos Catalans 26, 43007 Tarragona, Spain

a b s t r a c t Distillation of close-boiling mixtures, such as propylene–propane and ethyl benzene–styrene systems, is an energy intensive process. Vapor recompression techniques and heat pumping-assisted columns have been adopted for such applications for their high potential of energy savings. In direct vapor recompression columns, the vapors leaving the top of the column are compressed, and in the reboiler of the same column, these vapors are condensed to provide heat for vapor generation. Internal heat integrated distillation columns or iHIDiCs are new developments employing the same concept of vapor recompression. These new column configurations can have significantly lower energy demands than common vapor recompression units. In iHIDiCs, rectifying section is operated at a higher pressure (i.e. higher temperature) than in stripping, and therefore its heat can be used to generate vapor in stripping section. So far, design of these column configurations is performed based on engineering experience, simulation or experimental studies on given cases, including dynamic control simulations. Within previous and most recent research efforts on iHIDiCs, there exist no generalized design methods or systematic approaches for design of these internal integrated distillation columns. The present paper presents a systematic design procedure for iHIDiCs. A design hierarchy for iHIDiCs is developed, which includes two phases of design, thermodynamic and hydraulics. This design procedure is applied using commercial simulation-based design methods. In thermodynamic design, temperature profiles for column sections are used as a design tool to guide designers. On the other hand, hydraulic capacities of stages for heat exchange are analyzed to determine the maximum physical space area available for heat exchange. Hence, feasibility regions for both heat integration and hydraulic design are identified. © 2009 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved. Keywords: Heat integration; Distillation; Process intensification; HIDiC; Temperature profiles; Design feasibility; Hydraulics

1.

Introduction

Distillation is an old separation process and has broadly been used in various chemical and petrochemical industries. High energy consumption of distillation processes is a practical shortcoming. According to various energy estimates, energy consumption of refining and chemical processes is utilized in distillation columns. Besides, distillation of mixtures with low relative volatilities, such as propylene–propane splitting, ethyl benzene–styrene system, is an energy intensive process. The energy inefficiency of distillation has urged the industry’s interest towards adopting the implementation of advanced technologies with higher efficiencies. As a result, thermal

coupling, heat integration, vapor recompression and heat pumps were adopted as new techniques and developments to increase the distillation efficiency (Stupin and Lockhar, 1972; Linnhoff et al., 1983; Freshwater, 1951; Sulzer, 2006). In columns with vapor recompression techniques, the vapors from the top of the distillation column are compressed to a certain pressure (increasing temperature of vapors) and are then condensed in the reboiler of the same column through an indirect contact of the liquid of the column bottom. As result, the condensed vapors provide heat needed for vapor generation at the bottom of the column. Vapor recompression or heat pumping columns have been widely applied for the separation of close-boiling mixtures. Internal heat integrated

∗ Correspondence address: Universitat Rovira i Virgili, Campus Sescelades, Departament d’Enginyeria Química, Av. Païssos Catalans 26, 43007 Tarragona, Spain. Tel.: +34 977 55 8675; fax: +34 977 55 9621. E-mail address: [email protected]. Received 12 September 2008; Received in revised form 9 June 2009; Accepted 13 June 2009 0263-8762/$ – see front matter © 2009 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.cherd.2009.06.005

chemical engineering research and design 8 7 ( 2 0 0 9 ) 1658–1666

distillation columns (iHIDiCs) are further intensifications of vapor recompression principle. These columns combine the advantages of both direct vapor recompression and diabatic operation and can have significantly lower energy demands than common vapor recompression distillation columns or heat pumps (Nakaiwa et al., 1997, 2000; Olujic et al., 2003). The primary concept of iHIDiCs was earlier introduced by Mah et al. (1977) and Fitzmorris and Mah (1980) under the name ‘Secondary Reflux and Vaporization’ (SRV). Seader in his patent (1978/1980) and Glenchur and Govind (1987) advised different column configurations for implementing these internal distillation columns. Thereafter, a shell and tube-type packed column was presented by the patent of Aso et al. (1996/1998). Later, the idea of iHIDiC attracted the interest of many researchers worldwide, such as in Japan and the Netherlands. A group of Japanese researchers such as Nakaiwa et al. (1997, 2000, 2001) and Naito et al. (2000) studied internal heat integrated columns, focusing on theoretical evaluations and pilot plant testing. Their results indicated energy savings of up to 60% with respect to conventional columns. Among the research efforts on iHIDiC designs, few design aspects have been reported. The research rather concentrated on simulation, experimental studies, operational studies, and control aspects. Most recent work by Huang et al. (2008) addressed the effect of feed preheating on the overall energy efficiency of total iHIDiCs. This work reported that the heat integration between the distillate and feed poses additional difficulties to process operation. The work did not include any design suggestions, modeling or simulation aspects. Nevertheless, most simulation studies were performed without any guidelines and not following systematic strategies. Further research efforts were published by Horiuchi et al. (2008), focusing on the energy saving characteristics of the internally heat integrated distillation columns. In this work, iHIDiCs were used for multicomponent petroleum distillation by constructing a pilot plant for the separation of hydrocarbons mixture. The study of the process system was conducted considering a packed column type HIDiC, a concentric double tube packed column. This work did not present any design procedures or refer to how to design an iHIDiC for a given problem, but rather concentrated on the operation results of pilot plant installations. On process design level, simulation and process synthesis aspects, there were no general approaches or methods to deal with new design problems and applications of iHIDiCs; i.e. no systematic method to address such column configurations. Furthermore, design feasibility and hydraulic capacities for viable heat exchange were not defined. In this current work, more attention is directed to the systematic design of iHIDiC. The designs feasibility is introduced and then evaluated with respect to two various aspects, thermal and physical space capacities. A design procedure, simulation and design hierarchy are developed to guide the design of a distillation problem as an internal heat integrated distillation column (iHIDiC). The design procedure is based on rigorous modelsbased commercial softwares, such as AspenHysys and ASPEN Plus (Aspen Technology, 2008), ProII, etc.

2.

Simulation background of iHIDiCs

A schematic diagram for an internal heat integrated distillation column is illustrated in Fig. 1. As seen, iHIDiC configuration comprises two separate distillation columns,

1659

Fig. 1 – Schematic representation of an iHIDiC.

stripping and rectifying columns. There is a pressure difference between the two columns; the overhead vapor of the stripping column is compressed and then enters at the bottom of the rectifying column. The rectifying column therefore operates at a higher pressure and hence at a higher temperature. The liquid from the bottom of the rectifying column is fed into the top of the stripping column, as reflux stream. The pressure of the recycled liquid stream from the rectifying column is equalized with the pressure of the stripping column through a throttling valve. The vapor leaving the top of the rectifying column is the light product, while the heavy product is the bottom stream of the stripping column. The two columns are configured in a particular way so that the energy of the hot rectifying column can be used to heat the stripping column (Glenchur and Govind, 1987; Seader, 1978/1980). The amount of heat transfer between the two columns can vary, and correspondingly the reboiler duty will change. The schematic representation of Fig. 1 explains the manner through which internal HIDiCs are simulated by using commercial simulators (Aspen Technology, 2008). Therefore, the simulation is performed using two separate distillation columns, one for stripping column and another for rectifying column. However, the implementation of internal HIDiCs is executed such that the two columns of stripping and rectifying are configured in various forms, e.g. concentric configuration with the rectifying column placed inside the stripping shell, or as adjacent-two columns (Gadalla et al., 2007). In simulation of internal HIDiCs, heat is transferred from hot rectifying column to stripping column in order to heat the liquid on a stage-basis. Initially, when no heat is exchanged, the reboiler duty will have a large value and is comparable with conventional duty. In this case, the reboiler provides all heat required for vapor generation. Heat is transferred on each column tray through an indirect contact of the rectifying hot vapor and the stripping cold liquid streams. This implies that a continuous condensation of the vapor phase occurs along the rectifying column and a continuous evaporation, i.e. vapor generation takes place in the stripping column. This heat integration will result in a reduction in both the reboiler and condenser duties. The heat transfer is achieved in an external medium (device), such as heat panels (Gadalla et al., 2007). Heat panels are placed either on the rectifying side or stripping side of the column trays, depending on the larger space area. Hot vapors of rectifying column enter panels, while cold liquids of stripping column flow across the outer surface of the panels. As a result of this integration, the energy requirement

1660

chemical engineering research and design 8 7 ( 2 0 0 9 ) 1658–1666

Fig. 2 – A hierarchy for iHIDiC design. in the reboiler is reduced as part of the vapor is generated by heat exchange. The more the heat exchanged, the less the energy consumed in the reboiler. Design of iHIDiCs can be partial, when the reboiler energy consumption is decreased to a lower value, or ideal, when the reboiler duty is reduced to zero. For an ideal iHIDiC, the reboiler unit is not needed since all the energy required for vapor generation is provided by the rectifying section; however, for startup necessities, a reboiler unit may be available. As mentioned above, design of internal HIDiCs is typically performed using process simulations (e.g. HYSYS). The design task includes the calculations of the compressor, condenser and reboiler duties, stage heat transfer rates, and the heat transfer area required per stage. The heat transfer area is the area of heat panels that is placed inside the column on stages to achieve the heat exchange between the hot vapor and the cold liquid of the rectifying and stripping stages, respectively.

3.

A hierarchy for design of iHIDiCs

Prior to any design calculations of iHIDiCs, conventional distillation calculations are easily performed, besides in other cases, heat pump data may be available. The design process of iHIDiCs therefore may start from this basic step. Based on available conventional calculations, a design hierarchy is proposed in Fig. 2 for internal heat integrated distillation columns. The hierarchy is to guide the designer to perform a systematic design and to provide valuable insights for optimum results. The design process starts by simulating a conventional column (or heat pump configuration), if not available, for the given design problem. Required data (given data) for simulation design are typically: (1) feed flow rate and conditions, (2) components composition, (3) product and separation requirements, and (4) column pressure and pressure drops. Simulated design parameters (output parameters) are: (1) number of stages in each column section, (2) reboiler duty, (3) condenser duty, and (4) product compositions. On the other hand, for separation systems where heat pumps are adopted, the column configuration is simulated to calculate, in addition to the above parameters, the compressor electricity consumption and working pressure ratio. The design of iHIDiCs follows up according to the hierarchy to first a design of basic iHIDiC which has no heat exchange or integration. Then a complete iHIDiC is designed based on the design obtained in the previous step. This complete iHIDiC is simulated by increasing the level of heat transfer between the individual columns step by step until the reboiler duty is reduced to minimum (partial iHIDiC) or zero (ideal iHIDiC). The two designs are different only in the heat exchanged between the two columns, rectifying and stripping. The design step of complete iHIDiC is

performed through a two-step approach, first according to the thermodynamic capabilities of the design and then based on the hydraulic capacity of the stages. The final design is completed by an optimization procedure and some improvement modifications. However, the optimization procedure of the design will not be taken into account in the current work. Improvements may include addition of stages, flash drum, gas turbine for more energy and power efficiency. Certainly when heat is exchanged between the rectifying and stripping columns, liquid and vapor flows along the column stages will change. It must be noted that the integration of the two columns does not necessarily imply heat exchange on all trays. If the two columns are asymmetric, this means that not all stages of the two columns will be integrated. Part of the two columns will perform as conventional column, i.e. vapor and liquid flows are almost constant throughout this section. Typically when heat is integrated, vapor flow increases in stripping column due to an internal evaporation in the heat integrated part, while in the corresponding integrated part of rectifying, vapor flow decreases due to an internal condensation. On the other hand, in nonintegrated part on both sides of the columns, vapor and liquid flows are almost unchanged.

3.1.

An iHIDiC basic design

When iHIDiC is simulated, considering the column configuration of Fig. 1 with no heat transfer between the two columns, this simulation design is called basic design. This basic design will be considered a first step in designing a full iHIDiC. The design is expected to show maximum reboiler duty which is comparable with conventional designs. Therefore, the heat transfer per stage Qstage is equal to 0. The profile of vapor and liquid flows for this design throughout the column is constant, which is similar to conventional columns. The principal question to be addressed when designing iHIDiCs is where to start the design process for the two designs, basic iHIDiC and complete iHIDiC, given the basic data for separation. In this context, several basic assumptions need to be taken into account for both design cases, including: (1) feed location in the iHIDiC configuration, (2) number of stages in each column, and (3) pressure before and after the compressor (pressure ratio). The conventional column design which is the only available data for iHIDiC is a key for starting the design. The column is split around the feed entrance into two separate columns for iHIDiC, rectifying and stripping. The relative feed location is kept unchanged, i.e. the feed enters the stripping column at the top stage. So, the upper section of the conventional column will be rectifying column, while the lower section (stripping) will be the stripping column with the column feed

chemical engineering research and design 8 7 ( 2 0 0 9 ) 1658–1666

entering at the top. Consequently, the numbers of stages for both iHIDiCs columns are identified from the upper and lower sections of the conventional column design. This simplified assumption seems reasonable since the separation performed in HIDiC two columns is the same as achieved in the conventional two sections. The pressure in rectifying column of iHIDic is assumed to be as high as the pressure at the bottom of the conventional column. On the other hand, the pressure of the stripping column is taken to be the same for the conventional top section. Therefore, a reasonable pressure ratio is obtained from the pressure at the bottom of the rectifying column and the pressure of the top of the stripping column. This assumption will allow a logical temperature difference for heat integration. Pressure ratio needs to meet some economic constraints since it leads to a high operational cost of electricity; so it can undergo through an optimization procedure in late design stage. The pressure distribution suggested is an initial guess for the pressure to be used in both columns of stripping and rectifying columns. However, the pressure for a final design and hence the pressure ratio can be adjusted through an optimization, and thus the pressure distribution will be re-estimated. A reasonable pressure drop per stage is assumed for both columns. By considering the same pressure drop per stage as for the conventional column as an approximation, the pressure profile across the two columns (stripping and rectifying) can then be estimated. Certainly these underlying assumptions of the distribution of stages and pressure will have implications on the total cost of the total design. Additionally, the pressure drop assumed will affect the relative volatility (i.e. separation) and hence to fulfill the same separation, reflux and utilities will change. After obtaining the basic design, optimization can be applied in order to reach an optimum number of stages in each column and best pressure ratio values, by considering all cost components of column capitals and variable costs of cooling water, electricity and heating steam. The column diameter will be also calculated according to the flooding and pressure drop encountered in each column stage. The diameter obviously will have an influence on capital costs. When a heat pump design is available for the separation problem, the basic assumptions will be different, and are rather close to design of iHIDiCs. This implies that rectifying and stripping columns will be equivalent to the top and bottom sections of heat pump column design. The pressure difference between the top and bottom of heat pump column, i.e. pressure ratio will be used directly for both columns, i.e. rectifying and stripping. Further optimization will result in optimum values of number of stages and pressure distribution, taking into account all implications of capital costs and utility costs (water, electricity, steam).

3.2.

Simulation of iHIDiC basic design

The basic design of HIDiC is simulated using, as mentioned above, a commercial simulator. In this work, HYSYS (version 2006.5) is used to perform the process simulation. Chemical components included in the system need to be specified, followed by selecting a physical property model, such as Peng Robinson (Aspen Technology, 2008), for the calculations of the physical and thermodynamics properties. After specifying the components, together with the property model, the column configuration is simulated as shown in Fig. 1. First, stripping column is simulated for the feed conditions and flow rate. Number of stages is specified and reflux ratio is calculated

1661

Fig. 3 – Basic iHIDiC for methanol–water separation (no heat exchange) (numbers inside columns refer to number of stages). for the required separation (product composition). Then the vapors from the top are compressed to enter the rectifying column. Similarly, rectifying column is simulated to achieve the product specifications. Required stage diameters for separation are then calculated based on flooding conditions given by Fair’s correlations (Kister, 1992) and liquid and vapor flow rates inside the column (Aspen Technology, 2008). Fig. 3 shows a basic iHIDiC designed for separating methanol–water system. Design assumptions are made in accordance with the above discussion, and based on a conventional design and available data (confidential). The column is to separate a nearly pure methanol as top product and almost pure water in the bottom product. Separation data and column specifications are shown in Fig. 3. In addition, the feed to this column is at atmospheric conditions: physical and thermodynamic properties are calculated using Peng Robinson model (HYSYS). Design assumptions of Fig. 3 are: (1) Number of stages in the stripping column is 30, whereas the rectifying column contains 65 stages. Feed enters at the first stage from top of stripping column. (2) The pressures for the striping column are 1.2 and 1.6 bar at column top and bottom, respectively. (3) For rectifying column, the bottom pressure is 2.6 bar, while the top pressure is 1.8 bar. (4) The pressure drop per stage is kept unchanged for the basic design as for the conventional base case. (5) The working pressure ratio is calculated to be 2.1. Opting for the above assumption, a basic configuration design of iHIDiC can be simulated to meet the separation requirement of the design and obtain the required products. From Fig. 3, it may be noted that the simulated values of reboiler duty (41 MW) and condenser duty (39 MW) of basic iHIDiC are close to those of conventional column design (40 MW for condenser, 45 MW for reboiler). Also the simulated diameter is 3.8 m for both columns, compared with 4 m for the conventional column. The basic HIDiC remains an intermediate stage between conventional column and complete HIDiC. It represents a fundamental configuration that leads eventually to an optimum full HIDiC.

3.3.

An iHIDiC complete design

After obtaining the basic design of internal HIDiC, a complete design is simulated to achieve both conditions, ideal and partial designs. The simulation of internal HIDiCs proceeds

1662

chemical engineering research and design 8 7 ( 2 0 0 9 ) 1658–1666

via integrating (exchanging) heat between the rectifying and stripping columns. In other words, on tray-basis the heat contents of the vapors of rectifying stages are exchanged in an indirect contact in heat panels to vaporize the liquid flowing downwards across stages of stripping column. It is assumed that the two columns are configured as concentric with rectifying column in the middle. So panels can be placed on trays in the space area available (active areas) such that they are exposed to both vapors of rectifying and liquids of stripping. The location of panels can take place in rectifying side or stripping side, subject to availability of physical space (Gadalla et al., 2007). In simulation terms, this step is performed via energy streams withdrawn from the rectifying column at a given stage and added to the stripping column at the corresponding stage. The value of the energy amounts exchanged is increased gradually until the reboiler duty is reduced to a minimum value, keeping the product specifications for the base case fixed. If the ultimate value of reboiler duty reaches zero, the column design is called ideal, otherwise the design is partial with reboiler duty above zero. As a result of this energy exchange, the vapor flow in rectifying column is expected to decrease because of condensation. Consequently, the vapor flow inside stripping column in turn increases due to an external vaporization. As stated earlier not all stages have to be integrated. In this case, vapor and liquid flow profiles will change in integrated stages from non-integrated stages, i.e. constant profiles in non-integrated parts. The condensation of vapors in rectifying column will lead to a lower duty in the external condenser. In the same way, the external reboiler duty of the stripping column will decrease due to the vaporization taken place by heat integration with rectifying column. The simulation of complete HIDiCs seems timeconsuming; however it is useful to overcome convergence problems of commercial simulators by decreasing the incremental increase in the energy transfer value. After completing a full design of internal HIDiCs, the calculations of heat transfer areas that are required on each stage for energy integration are performed, according to the following relationships: A=

Qstage U × Tstage

Tstage = TR − TS

(1) (2)

where, A is the heat transfer area of panel, Qstage is the amount of energy exchanged per stage, U is the overall heat transfer coefficient, and Tstage is the temperature difference between the rectifying stage temperature (TR ) and the stripping stage temperature (TS ). These two temperatures can be extracted from simulation results. Note that the temperature difference between stages indicates the driving force that allows heat integration. The overall heat transfer coefficient can be assumed with a reasonable value (≈1 kW/m2 ◦ C), or is replaced by experimental data.

3.3.1.

Fig. 4 – Constant heat transfer rate design for iHIDiC (stages w.r.t. column top). requires different panel areas to be installed on each stage. This design may lead to unfeasible requirements of heat transfer areas on some stages if the required area cannot be fulfilled by the existing stage capacities. On the other hand, for the other design, the same amount of heat transfer areas is to be installed on each stage and correspondingly the heat transfer rate will vary throughout the stages as function of the temperature differences. This last design is more practical since it exploits the variation in temperature driving force and hence it ascertains the design feasibility. This is because the installed heat transfer areas will take into account the maximum tray capacity. Fig. 4 demonstrates a HIDiC designed for methanol–water separation (data of Fig. 3), considering a constant heat transfer scenario. As can be seen, heat transfer area (A) increases rapidly with stages from top to bottom locations (N). For this design, the heat transfer rate per stage is 1.1 MW. In contrast, the design with constant heat transfer area per stage of 170 m2 is shown in Fig. 5. The two designs are identical in the total amount of heat integration and separation specifications; however, they differ in the heat transfer area distributions. Further differences between these designs may include vapor and liquid flows, external energy consumption, column diameters, etc. To decide on the preference of these designs, two important criteria must be taken into account: (1) feasibility of design, i.e. if the design is practically feasible and (2) optimization. The feasibility will take into account whether the heat transfer areas can be accommodated by the existing column diameters or not. On the other hand, optimization will include all cost variations of the two cases caused by the change in vapor and liquid flows (i.e. diameter), heat transfer areas, reboiler and condenser duties, and electricity. A comparison of the two designs is shown in Table 1. The HIDiC design with constant duty per stage assumes a rate of 1.1 MW, while the other design installs a 172 m2 per stage. The last

Design scenarios of iHIDiCs

For achieving a full iHIDiC design, there exist two scenarios for transferring heat from rectifying column to stripping column: (i) constant energy transfer and (ii) constant heat transfer area. In the first scenario, the amount of heat transfer is fixed throughout the column stages, i.e. Qstage is a fixed value. Therefore, the heat transfer area of heat panels will vary from one stage to another, in accordance with the temperature driving force on each stage. As a result, this design

Fig. 5 – Constant area design for iHIDiC (stages w.r.t. column top).

1663

chemical engineering research and design 8 7 ( 2 0 0 9 ) 1658–1666

Table 1 – Comparison of design scenarios of iHIDiC. Conventional Reboiler duty (MW) Condenser duty (MW) Compressor load (MW) Total heat transfer areas (m2 )

45 40

decision for comparison should be cost-oriented, and preferably is resulting from optimization and feasibility criteria. As seen in the above discuss, the iHIDiC is simulated, then the calculation of heat transfer areas is performed in a twostep procedure. However, a simultaneous simulation tool has been developed within HYSYS simulator (Aspen Technology, 2008) and will be considered in future publications. This tool allows performing the calculation of heat transfer at the same time the simulation is executed. In order to achieve this, a set of equations is built inside the simulator. Calculations are carried out using the online values of temperatures. Note that this new tool can handle the two scenarios of designs. Given that the principal aspect (benefits) of iHIDiCs is the energy transfer from the rectifying column to stripping one, it is an obvious necessity to consider this key parameter in more detail in design. For this respect, the design task of iHIDiCs in the hierarchy presented in Fig. 2 entails two important stages: (a) thermodynamic design and (b) hydraulic design. This implies that internal integration is designed according to two fundamental considerations, firstly from the point of view of energy exchange, and then according to the capacity of trays to satisfy exchange requirements. It must be noted that these aspects of thermodynamic and hydraulic designs have not been accounted for in previous works.

3.4.

Thermodynamic design of iHIDiCs

As mentioned earlier, heat is exchanged from rectifying stages to stripping stages. Certainly, there must be enough heat (quality) to perform this exchange; therefore rectifying stages need to be hotter than the corresponding stripping stages. For this reason, stage temperature profiles of basic iHIDiCs are found to be key parameter for such a measure leading to heat integration. Temperature profiles are obtained from simulation results of basic iHIDiC using HYSYS (Aspen Technology, 2008) by extracting the temperature of each stage for both columns, i.e. rectifying and stripping. Hence, two temperature profiles are obtained; hot profile for rectifying column and cold profile for stripping column. Then, these profiles can be plotted against the stages number for both columns. Temperatures for both columns are plotted on Y-axis, whereas stage number is represented on X-axis. For a possible heat transfer between the two columns, the hot profile of the rectifying column must be above (hotter) the cold temperature profile. This means that there should be positive temperature driving forces between the rectifying and stripping columns. The positive temperature difference is obtained by a specific pressure ratio; this leads to the definition of a minimum pressure ratio. This minimum value is defined as the lowest value of pressure difference that ascertains a positive driving force between the rectifying and stripping columns. This value is crucial for good design and can be obtained thus by plotting the temperature profiles for basic iHIDiCs. Fig. 6 shows a typical stage temperature profile of iHIDiC for separating benzene–toluene mixture

HIDiC with constant energy rate per stage 19.5 17.9 3.5 9410

HIDiC with constant heat transfer area per stage 19.9 17.9 3.5 4140

with a flow of 8.5 t/h available at 40 ◦ C and 1 atm. The mixture contains benzene and toluene with equimolar composition. The iHIDiC design contains 10 stages for each column, and is to separate benzene and toluene products with molar purities of 99.7% and 99.5%, respectively. The pressure ratio for these profiles is 2:1 (rectifying pressure/striping pressure). The stage profiles were obtained for a basic iHIDiC designed for this system. External reboiler and condenser duties are found to be 1347 kW and 1492 kW, respectively, while the compressor load is 170.3 kW. The Peng Robinson property model was used for this simulation. It can be seen from the plot that the rectifying temperature profile is hotter than the corresponding cold temperature profile with a temperate difference varying along the column stages. There is a positive temperature difference throughout all columns’ stages. This indicates that energy can be exchanged from rectifying column to stripping column along all stages. It is clear that some stages (top and bottom sections) show large driving forces, whereas others have smaller temperature differences. Referring to the previous scenarios of design, when duty per stage is fixed the required heat transfer areas will be large on some stages with small temperature driving forces (middle section of Fig. 6). This design may represent a practical difficulty in the case that the column diameter is incapable of providing the required design area of heat transfer. Thus, the variation of the temperature differences of both profiles should correspond to the distribution of the heat transfer between the columns. In other words, the amount of heat that can be transferred on stages should vary according to the temperature difference. Therefore, more heat can be transferred on those stages of large driving forces and conversely stages with smaller temperature difference preferably transfer less heat. The area requirement of this design justifies the recommendation of transferring more heat on those stages with large driving force. The larger driving force, the less area required. Indeed, this leads to a minimum capital cost and is accompanied by less external energy consumptions. On the other side, if the same heat is to be transferred on stages with

Fig. 6 – Stage temperature profiles for basic iHIDiC.

1664

chemical engineering research and design 8 7 ( 2 0 0 9 ) 1658–1666

small temperature differences, the area requirements will be substantial. Thus, the suggestion in this case is to transfer less heat, but this will incur more external energy consumptions. In conclusion, stages with minimum temperature differences are limiting stages and may define the key bottleneck for heat integration. It must be noted that the shape of the temperature profiles is a function of the system to be separated. The shape of profiles can be divided into three groups: (1) parallel profiles where temperature difference is almost fixed, (2) variable profiles with large difference on the terminals and minimum in the middle, and (3) decreasing or increasing profiles. Each profile category has its implications on both duty transfer and heat transfer areas. The first group will have nearly constant heat transfer areas across all column stages to transfer the same amount of energy per stage. Benzene–toluene mixture is an example for the second group profiles, where energy exchange is a maximum at top and bottom sections, while middle section is characterized by minimum energy transfer. Group 3 will show area and duty requirements as those given in Figs. 4 and 5 for methanol–water system. For such profiles, the energy transfer and transfer areas are large at one end of the profile and small at the other end.

3.4.1.

Pinched and limiting stages of iHIDiCs

From Fig. 6, the temperature difference is shown to be at its minimum value at stage number 6 (from top). This difference is called the minimum temperature difference for iHIDiC design (Tmin ). This value is a trade-off parameter with its impacts on capital-energy costs; it needs to be determined through an optimization. When HIDiC is designed for a smaller temperature difference, more heat can be exchanged, and thus the external utility consumptions are reduced. In addition, larger heat transfer areas will be required. On the other hand, larger values of Tmin will imply less heat to be exchanged, more utility consumption and less capital cost of transfer areas. Again, it is clear that heat transfer is possible for this system throughout all column stages. The temperature driving force is a maximum at top and bottom locations of the column, and is reduced, moving towards the middle of the column. Note that this temperature profile was obtained for a basic iHIDiC, i.e. no heat transfer between columns. Thus, when heat is transferred on stage-basis, vapor and liquid flows will change across each stage and hence temperatures will change. Therefore, temperature driving forces will decrease for column stages, i.e. temperature profiles will move against each other. Repetitively, heat can be exchanged until the two profiles meet at a certain stage (or stages). Fig. 7 shows temperature driving force for basic iHIDiC compared with iHIDiC with heat transfer per stage of 90 kW. On the other hand, Fig. 8 represents the vapor flows for these designs, compared with conventional column for benzene–toluene separation. It is obvious for iHIDiC with energy exchange that vapor flows decreases in rectifying column because of condensation, while in stripping column these flows increase due to vaporization. Back to Fig. 7, the temperature driving force decreases with heat integration (exchange). The driving force is minimum and equal to zero at one specific stage (stage #6). This stage is called pinched stage and indicates the condition of maximum heat recovery for this particular stage. This means that the maximum energy exchange on this stage is 90 kW and heat cannot be integrated further. Stages with temperature difference around this pinched stage are called limiting stages. These stages are then responsible for impractical heat trans-

Fig. 7 – Temperature driving force for iHIDiC with zero duty and 90 kW per stage. fer when their temperature differences are smaller than a minimum specified value (Tmin ). Therefore, maximum heat integration on a stage can be calculated knowing its temperature driving force. Pinched and limiting stages can be predicted ahead of design by moving the temperature profiles towards each other till their touch point (or points). Pointing to Fig. 6, if the profiles are moved vertically against each other, stage #6 is pinched; this agrees with the conclusion drawn from Fig. 7. In conclusion, temperature profiles can identify stages with limiting temperature difference for heat exchange, also the maximum heat recovery between the two columns. One more advantage of such profiles is that a minimum pressure ratio (or rectifying pressure) can also be identified. This pressure ratio is the smallest value that allows a possible heat transfer from rectifying to stripping stages (or positive driving force). For a given system, the hot profile is moved vertically away from the cold profile until a positive value of T is observed. Therefore, this minimum difference corresponds to a minimum working rectifying pressure that leads to a feasible heat recovery. So the design of complete iHIDiCs can proceed by: (1) defining pinched and limiting stages, (2) identifying minimum pressure ratio, and (3) calculating maximum heat recovery for each stage. Moreover, heat integration schemes can also be obtained for a given pressure ratio by performing more heat recovery on stages with higher temperature differences. This step may result in several feasible integrated schemes with regard to the same problem. In other words, one integration

Fig. 8 – Vapor flows for iHIDiC designs compared with conventional design (rectifying stages from 1 to 10, stripping stages from 11 to 20).

chemical engineering research and design 8 7 ( 2 0 0 9 ) 1658–1666

1665

scheme could be integrating only the top and bottom stages of the two columns since they have the largest driving forces. Another scheme may integrate all stages except the limiting and pinched stages. Each scheme differs in the overall external energy consumption (reboiler/condenser), areas of heat transfer, compressor load, and capital costs of equipments. Therefore an optimization is necessary to screen all these feasible options. During optimization, feasibility of design schemes can be assessed and guaranteed using temperature profiles.

3.5.

Hydraulic design of iHIDiCs

After reaching a reasonable design with respect to heat integration, the subsequent step is to design the heat transfer medium. Traditional design of such a step only refers to the calculations of the heat transfer areas (panels) required to perform the heat integration. Heat transfer area required for an energy exchange on a particular stage can be calculated from the relationships presented previously (Eqs. (1) and (2)). The availability of enough heat to be integrated does not necessarily imply that the design is feasible in overall. As an example, it is likely that a designed column provides the necessary heat from rectifying to stripping but the calculated stage diameters are not able to provide space area (to place panels). Therefore, the overall design is not attainable. In conclusion, there must be a distinction between two feasibility factors in HIDiC design, firstly the thermodynamic feasibility and secondly the hydraulic feasibility. The design is thermodynamically feasible, as seen in Section 3.4, when there is sufficient heat or positive temperature driving force. From another side, the presence of enough space and large column diameters will lead to reasonable designs with respect to hydraulic design. Therefore, calculated heat transfer areas must be compared with the real capacity of stages. To assess hydraulic feasibility, the maximum available area on each tray needs to be quantified, and the available area should allow the placement of heat panels to perform the required energy exchange. Providing the configuration of rectifying with stripping columns, e.g. concentric layout (Gadalla et al., 2007), the maximum physical space areas on stages can be calculated. Within these space areas, heat transfer devices can then be placed to allow heat exchange. The space calculation is simple and only requires: (1) hydraulic column diameter, (2) dimensions of the heat transfer device, and (3) layout of heat panels with column stages. The hydraulic diameter is determined by process simulation, which is calculated from flooding limits given by Fair’s correlations (Kister, 1992). The diameters are calculated for each stage of rectifying and stripping columns. Dimensions of heat panels (devices) are length, thickness, and width, of panels. Heat panels are normally placed repeatedly in vertical layout on the space available by the trays. Then, the geometry of the column stages together with placing heat panels is analyzed. The analysis results in estimating all the physical space areas that are available to place heat panels. Also, the number of heat panels that can be handled in these physical spaces is calculated, given the panel dimensions and assuming a reasonable free allowance between each pair of panels. Therefore, on each column stage, the maximum available area is determined. If these areas are plotted against the stage number, a profile for maximum hydraulic area available is obtained. This profile represents the maximum area possible for energy exchange. Details of such models and calculations are given by Gadalla et al. (2007).

Fig. 9 – Stage feasibility profile for iHIDiC. The maximum area is a limiting key for hydraulic design and feasibility. For that reason, the required heat transfer area for heat integration needs to be smaller than the maximum available one (hydraulic area). If the required heat transfer areas (design areas) are superimposed on the hydraulic profile, a feasibility profile is obtained (Fig. 9). As a result, a hydraulic feasibility indicator can be determined to decide whether the energy can be exchanged, satisfying the stage hydraulic design or not. This indicator is therefore calculated for each stage, and is equal to the ratio between the required areas to the maximum available areas. For a feasible hydraulic design, the indicator must be smaller than unity. The calculations of Fig. 9 are made for iHIDiC design (10 stages for each column) for an equimolar benzene–toluene mixture (8.5 t/h, 40 ◦ C, 1 atm.). The reboiler duty of this design is zero, i.e. ideal iHIDiC. The pressure ratio between the two columns is 2.5. The separation specifications (by mole) for top and bottom products are 99.7% benzene and 99.5% toluene, respectively. The design heat transfer rate to achieve ideal iHIDiC is 287.2 kW/stage. From the profile, feasibility regions can be identified, where energy integration can be processed due to the presence of enough space areas. Conversely, stages which cannot allow the energy exchange inside columns are determined. This is because these stages have no enough spaces to place panels for energy exchange. As seen in the figure, stages #4–7 cannot provide the required heat transfer areas (design). Therefore energy rates have to be decreased on these stages. On the other hand, all other stages can satisfy the designed heat transfer rates. This finding agrees with the observation of Fig. 6, which indicated that the middle stages of the column have limiting temperature driving forces. In summary, the feasibility profile identifies regions of hydraulic feasibility in addition to thermodynamic feasibility. Note that the feasibility profile is not only a measure for evaluating designs but it can also determine the maximum allowable heat transfer rates for stages with indicator greater than 1. This maximum energy rate is calculated by multiplying the maximum available area for a given stage by the temperature difference and the overall heat transfer coefficient.

4.

Concluding remarks

A systematic design hierarchy has been proposed for iHIDiCs, including thermodynamic and hydraulic approaches. Starting from a conventional design or heat pump, a full iHIDiC design is realized by performing basic design assumptions to conventional data. A basic design of iHIDiC is first simulated, and

1666

chemical engineering research and design 8 7 ( 2 0 0 9 ) 1658–1666

then a complete iHIDiC is obtained, ideal or partial. Temperature profiles are keys for heat integration, while hydraulic calculations are necessary to quantify the ability of a column design to place heat panels. Following the two design criteria, together with the design hierarchy, a feasible internal integrated column design is easily obtained. The systematic design methodology can be applied to any separation systems, providing a basis for further optimization and improvement studies. Designs obtained by the new design hierarchy guarantee feasibilities with respect to thermodynamics and physical space availabilities.

Acknowledgements The author is grateful to the Program Office on Economy, Ecology and Technology for financial support and to the partners ABB-Lummus, AKZO-Nobel, BP, DSM, ECN, SHELL GS and Sulzer Chemtech. Special thanks as well are given to Professor Zarko Olujic and Professor Peter Jansens from Delft University of Technology (The Netherlands) for their valuable support, precious discussion and great advices related to this work. I would also appreciate the help of Dr. Aris de Rijke from the department of Process and Energy at Delft University of Technology.

References Aso, K., Matsuo, H., Noda, H., Takada, T. and Kobayashi, N., 1996/1998, Heat integrated distillation column, US Patent, 5, 783, 047. Aspen Technology, Inc., 2008, AspenTech Version 2008, 20.0.0.6728. Fitzmorris, R.E. and Mah, R.S.H., 1980, Improving distillation column design using thermodynamic availability analysis. AIChE Journal, 26(2): 265–273. Freshwater, D.C., 1951, Thermal economy in distillation. Transactions IChemE, 29: 149–160. Gadalla, M., Jimenez, L., Olujic, Z. and Jansens, P.J., 2007, A thermo-hydraulic approach to conceptual design of an internally heat-integrated distillation column (i-HIDiC). Computers and Chemical Engineering, 31: 1346–1354.

Glenchur, Th. and Govind, R., 1987, Study on a continuous heat integrated distillation column. Separation Science and Technology, 22: 2323–2328. Horiuchi, K., Yanagimoto, K., Kataoka, K., Nakaiwa, M., Iwakabe, K. and Matsuda, K., 2008, Energy saving characteristics of the internally heat integrated distillation column (HIDiC) pilot plant for multicomponent petroleum distillation. Journal of Chemical Engineering of Japan, 41: 771–778. Huang, K., Shan, L., Zhu, Q. and Qian, J., 2008, A totally heat-integrated distillation column (THIDiC)—the effect of feed preheating by distillate. Applied Thermal Engineering, 28: 856–864. Kister, H.Z., (1992). Distillation Design. (McGraw-Hill, New York). Linnhoff, B., Dunford, H. and Smith, R., 1983, Heat integration of distillation columns into overall processes. Chemical Engineering Science, 38(8): 1175–1188. Mah, R.S.H., Nicholas, J.J. and Wodnik, R.B., 1977, Distillation with secondary reflux and vaporization: a comparative evaluation. AIChE Journal, 23: 651–658. Naito, K., Nakaiwa, M., Huang, K., Endo, A., Aso, T., Nakanishi, T., Nakamura, T., Noda, H. and Takamatsu, T., 2000, Operation of bench-scale HIDiC: an experimental study. Computers and Chemical Engineering, 24: 495–499. Nakaiwa, M., Huang, K., Naito, K., Endo, A., Owa, M., Akiya, T., Nakane, T. and Takamatsu, T., 2000, A new configuration of ideal heat integrated distillation columns (HIDiC). Computers and Chemical Engineering, 24: 239–245. Nakaiwa, M., Huang, K., Naito, K., Endo, A., Akya, T., Nakane, T. and Takamatsu, T., 2001, Parameter analysis and optimization of ideal heat integrated distillation columns. Computers and Chemical Engineering, 25: 737–744. Nakaiwa, M., Huang, K., Owa, M., Akiya, T., Nakane, T., Sato, M. and Takamatsu, T., 1997, Energy savings in heat-integrated distillation columns. Energy, 22: 621–625. Olujic, Z., Fakhri, F., de Rijke, A., de Graauw, J. and Jansens, P.J., 2003, Internal heat integration—the key to an energy-conserving distillation column. Journal of Chemical Technology and Biotechnology, 78: 241–248. Seader, J.D., 1978/1980, Continuous distillation apparatus and method, US Patent, 4, 234, 391. Stupin, W.J. and Lockhar, F.J., 1972, Thermally coupled distillation—a case history. Chemical Engineering Progress, 68(10): 71–72. Sulzer Chemtech, 2006, Distillation and heat pump technology, Brochure 22.47.06.40-V.91-100.