Process design traditionally has emphasized simulation. Process integration has, in comparison, received little attention. Yet, unless the most appropriate process structure and operating conditions are selected, major opportunities can be missed. Moreover, raw materials and utilities integration among processes offers significant potential to reduce costs.
Process integration efforts began in the late 1970s, and initially were rooted in energy conservation and, in particular, the design of heat exchanger networks. These efforts led to the development of a variety of tools, the best-known of which are composite curves [1-7]. Such curves give in a single picture the cumulative cooling and heating requirements of a complete process and highlight heat recovery opportunities and targets for external heating and cooling requirements.
The pinch design method was then developed to allow the energy targets to be achieved in practice . Another early application was the grand composite curve , which gives a clear picture of the interface between the process and utility system and allows selection of the most appropriate mix of utilities. The grand composite curve then was extended beyond individual processes to total sites, to produce site composite curves and give a complete picture of the total site utility requirements .
Thousands of successful applications worldwide in a range of industries testify to the value of process integration technology in reducing energy costs and increasing capacity through debottlenecking. In a few cases, the goal has been to cut emissions of combustion gases; this inevitably will become more of a major driving force.
Besides these thermodynamic methods, an alternative approach based on mathematical programming has been developed . It starts with the creation of a “superstructure” that contains all feasible operations and interconnections that are candidates for an optimal design. The problem is formulated mathematically and the superstructure is optimized to reduce the complexity of the design and to set operating conditions.
Both approaches now are well established and available through commercial software.
But what are the recent trends in process integration? Let’;s start by reviewing what is happening in the traditional area of energy integration.
Picking up steam
Any investigation of potential improvements in energy integration should start with a site’;s utility system. This is the only way to establish the true incentives to enhance the energy performance of individual processes. Figure 1 illustrates a typical site setup in which different production units are linked to a common utility system that distributes steam at various pressures. The process units use this steam and, in some cases, generate steam from waste heat and feed it to steam headers. Backpressure steam turbines might generate power from the expansion of steam from high to low pressures. (The steam turbines in Figure 1 also feature some condensing power generation.) Large sites might also employ gas turbines. Power might be imported from centralized power generation or, in some cases, exported from the plant. On such a site, what is the true value of steam saving? The answer turns out to be not so obvious.
Figure 1. Typical site steam system
Optimized conditions, with flow in ton/hr, are shown for a site's existing steam system.
To analyze a utility system, it first is necessary to develop a simulation model . Commercial software is available for this. The model should allow part-load performance of the steam system components and provide a simulation of the complete material and energy balance around the steam system. It should be capable of predicting the fuel, power-generation, water requirements, etc., for any condition of the steam system and must account for the operating constraints around the system. Then, optimization of this model can identify the true value of energy saving at different pressure levels. The steam usage from each header can be gradually reduced and the system re-optimized at each step . This sounds like a lot of work, but it is straightforward given the appropriate software. Figure 2 shows a typical result. Note two key points:
1. There is not a single value for steam at a given level. It depends on how much is being saved, the various costs, equipment performance and the constraints in the utility system.
2. Steam saving is limited by equipment constraints in the utility system. The steps in Figure 2 are created when constraints in the utility system are encountered during optimization.
Figure 2 provides a map and compass for steam saving on a site, but does not show you how to go about making the saving. That requires a detailed analysis of heat-recovery systems, etc.
Figure 2. Marginal steam costs
Constraints in a site's utility system limit steam savings and cause steps in marginal price.
This problem was difficult to address using the thermodynamic methods developed in the 1980s because they essentially considered a retrofit as a pseudo-new design. The pinch approach relates to an ideal new design and not to the arrangement of an existing heat-exchanger network. One of the methods proposed to overcome the drawbacks is the so-called network pinch [14, 15]. It starts by identifying the bottlenecking exchangers within the existing structure. To overcome such bottlenecks, the network pinch requires a modification to the network structure. This could be repurposing of an existing heat exchanger to a new duty, addition of a new exchanger or introduction of a split stream.
Identifying the most appropriate structural modification is not straightforward, but mathematical programming guided by thermodynamic insights can be used. Commercial software that is now becoming available allows a retrofit to be carried out one step at a time, leaving the designer in control to accept or reject options at each stage.
Reaching a watershed
The developments in the 1980s have extended beyond heat integration. El-Halwagi and Manousiouthakis  created a new area of mass exchange networks from some of the fundamental ideas for the thermodynamic analysis of heat exchanger networks. Later work focused on the particular problem of water networks [17-19]. The design objective in water-using networks is to minimize consumption by maximizing water reuse. Introducing processes for partial treatment of wastewater, termed regeneration, allowing further reuse or recycling of water, can provide further reduction in water use.
Figure 3 illustrates how various water-using processes can be combined to give a composite curve for the entire system, allowing targets to be set for minimal water supply. Extending the methods of energy integration to water integration is, in principle, straightforward. It is possible, although very difficult, to deal with multiple contaminants rather than a single contaminant  and to handle flow-rate constraints that are often encountered with water-using operations. Add to this the desire to introduce cost optimization and forbidden matches, etc., and the simplistic extension of the energy integration approach to water minimization rapidly runs into problems.
Figure 3. Water system integration
Composite curves for water use allows targets to be set for minimal water consumption.
However, formulating the problem using mathematical programming allows all of the complexities of designing a water system to be included . Multiple contaminants and water sources, flow-rate constraints and forbidden matches, as well as water, effluent-treatment and piping costs all can be included in the problem. Commercial software is available for this new area of process integration.
Keeping refineries gassed up
Changing specifications for gasoline and diesel are causing fundamental alterations in petroleum refining throughout the world. When changing the product specifications at a refinery or optimizing or debottlenecking it, the hydrogen supply to a process is often a limiting factor, particularly when making low-sulfur, low-aromatics products. Sources of hydrogen include recovery from fuel gas and tail gas, dehydrogenation processes, heavy-end gasification, steam reforming and outside purchase. The recovery of hydrogen from off gases is analogous to the recovery of process heat. Developing integrated designs that make better use of available resources can provide similar benefits and allow the best overall operation of the refinery. This requires systematic methods of identifying hydrogen needs and available production to optimize site performance and maintain the flexibility to run different crude-oil feedstocks while maintaining maximum equipment utilization.
Graphical targeting can be used to analyze the hydrogen resources available in the refinery and suggest which off-gases should be sent to hydrogen recovery processes . It also can pinpoint targets for required hydrogen production or off-site purchase (Figure 4).
Figure 4. Refinery hydrogen integration
A hydrogen-surplus diagram provides insights on realistic target for consumption of the gas.
Mathematical programming can handle the details of the pipe work, compressors and the capital cost associated with hydrogen recovery processes. Although the methodology is new, it has already seen many successful applications, and commercial software is available. However, to make the most of opportunities from hydrogen network design, process changes for the hydrogen consumers needs to be included in the methodology. For example, exploiting hydrogen network interactions to increase the feed purity to units, such as a hydrocracker, might bring considerable benefits from increased process yields.
Advancing in columns
Distillation system design also is attracting increasing attention. Before we get into this, it is useful to distinguish between simple and complex columns. A simple column has a reboiler and a condenser, and splits a single feed into two products. Although different configurations are possible when using simple columns, the design problem still is a manageable size. However, separations often require more complex arrangements, including thermal coupling, side strippers, side rectifiers, prefractionators and dividing-wall columns (Figure 5). As the number of products increases, the number of possible complex column configurations explodes. In addition to this problem, we need to choose the most appropriate pressure for each separation in a complex arrangement and to consider heat integration simultaneously.
Figure 5. Complex distillation
Opting for more complex arrangements than a series of simple distillation
columns may offer advantages.
Whereas noteworthy advances have been achieved in the development of process integration design methods for the synthesis of heat-integrated distillation systems, much work remains before the full potential can be exploited commercially. Methods now can synthesize complex column systems, but not those that are heat-integrated. Commercial software can handle certain aspects of the problem. Some progress has been made for a particular case of heat-integrated refinery distillation, but further development work is required .
One of the most significant developments in distillation in recent years has been the application of partition, or dividing wall, columns [22-25] (Figure 6). Such a column can yield three essentially pure products with efficiency greater than that possible with conventional designs. These columns allow energy reductions on the order of 30%, together with capital cost reductions, when compared with a series of simple distillation columns. Although the basic idea has been around since 1947, only in recent years has the arrangement been exploited commercially [22, 24, 25]. Commercial software is now available that can simulate such columns, but software that identifies appropriate use in an integrated context isn’;t.
Figure 6. Dividing-wall column
This column provides three essentially pure products, as well as energy
and capital cost savings.
So far, we have considered non-azeotropic systems. With azeotropes, a new set of problems emerges. Graphical techniques have been developed to help designers visualize the difficulties associated with azeotropic and extractive distillation . Unfortunately, these tools are restricted largely to ternary systems. With the use of elaborate graphics, the principles can even be extended to quaternary systems. However, such simple graphical tools cannot handle the most complex cases. Even though much progress has been made in recent years toward a better understanding of azeotropic and extractive distillation, much work still needs to be done. Fundamental knowledge of heterogeneous systems needs to be improved before the potential for simple phase separations can be fully exploited.
Reactive distillation is even more complex . Again, graphical tools can provide insights about simple systems. However, much work remains to be done to fully understand how to synthesize reactive distillation systems for the most complex cases. The development of better simulation tools for this problem has been welcome, but process integration of these systems is at an early stage.
Tackling reactions and beyond
Obstacles certainly need to be overcome in established areas of process integration — and new areas pose enormous challenges. Little work has taken place in new areas to develop methods that can be exploited commercially.
For instance, process integration of reactions systems is still in its infancy. Reactor design generally considers the reactor to be in isolation even though, in most cases, the design interacts closely with the rest of the process. Taking this a step further, catalyst design often occurs earlier and without too much consideration of reactor design. Yet, the design of the catalyst, reactor and separation system should be carried out simultaneously and in an integrated fashion. This is particularly important when the separation system is extremely complex and costly, as is the case with many azeotropic systems. Biochemical reaction systems present similar and, in many ways, more daunting challenges due to the difficulties in modeling such systems at a level of sophistication appropriate for process integration methods.
The process integration of separation systems other than distillation has hardly been considered. Membranes, crystallization and solids-processing operations in general all require further development.
In addition, more attention needs to be paid to batch processes. Whereas heat integration of batch processes usually is not worthwhile, there still is significant potential for improvement through better process integration. The time element in batch processes provides flexibility in design and operation. However, this flexibility is often taken as an excuse for not doing a thorough and optimized design. In these applications, data are often the biggest problem. How can we design and optimize such a process if we don’;t have basic information such as reaction kinetic data? Such problems require experimental investigation (and the design of the experiments) to be conducted in parallel with the design of the process and its integration. The experimental design and the process design must be integrated in a coordinated activity.
The trend in process integration has been moving away from using graphical systems and moving toward employing automated optimization techniques. Graphical techniques still have a role to play, but are limited. We need the power of mathematics combined with graphical displays for better understanding. Lack of software for the newer areas of process integration is a bottleneck. There is plenty of commercial software for energy integration. More recently, software has come on the market for water- and hydrogen-system integration and the design of distillation schemes.
However, software on its own is not enough. To take the fullest advantage of process integration, engineers need to be educated about newer techniques. Although university curricula are already crowded, this education must start at school and continue through professional training. Otherwise, process engineers will miss significant opportunities to improve designs.
Dr. Robin Smith is a professor in the School of Chemical Engineering and Analytical Science at the University of Manchester. A Chartered Engineer, Fellow of the Institution of Chemical Engineers and Fellow of the Royal Academy of Engineering, he is the author of the upcoming book, “Chemical Process Design and Integration.” E-mail him at firstname.lastname@example.org or email@example.com.
1. Hohman, E.C., “Optimum Networks of Heat Exchange,” Ph.D. Thesis, University of Southern California, Los Angeles (1971).
2. Huang, F. and R.V. Elshout, “Optimizing the Heat Recovery of Crude Units,” Chem. Eng. Progr., Vol. 72, p. 68 (Month needed 1976).
3. Linnhoff, B., D.R. Mason and I. Wardle, “Understanding Heat Exchanger Networks,” Comp. Chem. Eng., Vol. 3, p. 295 (1979).
4. Umeda, T., J. Itoh and K. Shiroko, “Heat Exchange System Synthesis,” Chem. Eng. Progr., Vol. 74, p. 70 (Month Needed 1978).
5. Umeda, T., T. Harada and K. Shiroko, “A Thermodynamic Approach to the Synthesis of Heat Integration Systems in Chemical Processes,” Comp. Chem. Eng., Vol. 3, p. 273 (1979).
6. Umeda, T., K. Niida and K. Shiroko, “A Thermodynamic Approach to Heat Integration in Distillation Systems,” A.I.Ch.E.J., Vol. 25, p. 423 (1979).
7. Itoh, J., K. Shiroko and T. Umeda, “Extensive Application of the T-Q Diagram to Heat Integrated System Synthesis,” presented at Intl. Conf. on Proc. Systems Eng., Kyoto, Japan (1982).
8. Linnhoff, B. and E. Hindmarsh, “The Pinch Design Method of Heat Exchanger Networks,” Chem. Eng. Sci., Vol. 38, p. 745 (1983).
9. Townsend, D.W. and B. Linnhoff, “Heat and Power Networks in Process Design,” A.I.Ch.E.J., Vol. 29, p. 742 (1983).
10. Klemes, J., V.R. Dhole, K. Raissi, S.J. Perry and L. Puigjaner, “Targeting and Design Methodology for Reduction of Fuel, Power and CO2 on Total Sites,” J. Appl. Therm. Eng., Vol. 17, p. 993 (1997).
11. Biegler, L.T., I.E. Grossmann and A.W. Westerberg, “Systematic Methods of Chemical Process Design,” Prentice Hall, Englewood Cliffs, N.J. (1997).
12. Varbanov, P.S., S. Doyle and R. Smith, “Modelling and Optimization of Utility Systems,” Trans. Inst. Chem. E., Vol. 82A, p. 561 (2004).
13. Varbanov, P.S., S. Perry, Y. Makwana, X.X. Zhu and R. Smith, “Top-level Analysis of Site Utility Systems,” Trans. Inst. Chem. E., Vol. 82A, p. 784 (2004).
14. Asante, N.D.K. and X.X. Zhu, “An Automated and Interactive Approach for Heat Exchanger Network Retrofit,” Trans. Inst. Chem. E., Vol. 75, p. 349 (1997).
15. Zhu, X.X. and N.D.K. Asante, “Diagnosis and Optimization Approach for Heat Exchanger Network Retrofit,” A.I.Ch.E.J., Vol. 45, p. 1,488 (1999).
16. El-Halwagi, M.M. and V. Manousiouthakis, “Synthesis of Mass Exchange Networks,” A.I.Ch.E.J., Vol. 8, p. 1,233 (1989).
17. Takama, N., T. Kuriyama, K. Shiroko and T. Umeda, “Optimal Water Allocation in a Petroleum Refinery,” Comp. Chem. Eng., Vol. 4, p. 251 (1980).
18. Wang, Y.P. and R. Smith, “Wastewater Minimization,” Chem. Eng. Sci., Vol. 49, p. 981 (1994).
19. Doyle, S.J. and R. Smith, “Targeting Water Reuse with Multiple Contaminants,” Trans. Inst. Chem. E., Vol. B75, p. 181 (1997).
20. Alves, J.J. and G.P. Towler, “Analysis of Refinery Hydrogen Distribution Systems,” Ind. Eng. Chem. Res., Vol. 41, p. 5,759 (2002).
21. Gadalla, M., M. Jobson and R. Smith, “Increase Capacity and Decrease Energy in Existing Refinery Distillation Columns,” Chem. Eng. Progr., Vol. 99, p. 44 (Apr. 2003).
22. Kaibel, G., “Distillation Columns With Vertical Partitions,” Chem. Eng. Technol., Vol. 10, p. 92 (1987).
23. Triantafyllou, C. and R. Smith, “The Design and Optimization of Fully Thermally-coupled Distillation Columns,” Trans. Inst. Chem. E., Vol. 70A, p. 118 (1992).
24. Becker, H., S. Godorr, H. Kreis and J. Vaughan, “Partitioned Distillation Columns â€“ Why, When and How,” Chem. Eng., Vol. 108, p. 68 (Jan. 2001).
25. Shultz, M.A., D.G. Stewart, J.M. Harris, S.P. Rosenblum, M.S. Shakur and D.E. O’;Brien, “Reduce Costs with Dividing-Wall Columns,” Chem. Eng. Progr., Vol. 98, p. 64 (May 2002).
26. Doherty, M.F. and M.F. Malone, “Conceptual Design of Distillation Systems,” McGraw-Hill, New York (2001).