What is meant by Process Integration? In its broadest sense it could be taken to mean the achievement of more than one process objective in a single item of equipment. An example would be reactive distillation where both chemical reaction and product purification are achieved in a column. The simplest example is probably the heat recovery exchanger where the required cooling of one stream and the required heating of another are achieved in a single exchanger. This latter example leads to the most restricted sense of the term (and that is the sense relevant here) where it refers to the use of heat recovery for cost reduction in process plants.
Because of this emphasis on energy, process integration is often viewed as solely relating to heat recovery network design. This is a misconception. The technology, when properly applied, can have implications for reactor design, separator design and overall process optimization in any plant in which energy is a significant factor.
A heat recovery problem involving just one hot and one cold stream can easily be solved using a Temperature-Heat Load ( ) Plot.
The heat supply or demand associated with a stream is given by the relationship
where is the heat load, CP is the heat capacity flowrate, TS is the stream's "supply" temperature and TT is the stream's "target" temperature.
If the heat capacity flow-rate is independent of temperature, this relationship would appear as a straight line on a plot of Temperature against Heat Load.
Since in heat recovery problems heat load differences are the important factor, the line may be placed in any position along the heat load axis. However, its position on the temperature axis is fixed.
When the line representing the hot stream is superimposed on that representing the cold stream such that they are separated by a specified minimum temperature difference (see Figure 1), the following observations can be made:
where the lines overlap heat can be transferred from the hot stream to the cold stream at or above the specified minimum temperature difference. This 'overlap' indicates the scope for heat recovery.
where the hot line "overshoots" the cold line there is insufficient temperature driving force to permit heat recovery and the heat must be rejected to a cold utility.
the full heat content of the hot stream has now been accounted for, and any remaining cold stream heating duty— represented by the overshoot of the cold line—must be supplied through the use of hot utility.
The scope for heat recovery is dependent on the allowable temperature approach. As the minimum temperature approach is increased the lines move apart, the overlap reduces and the utility demands increase (Figure 2).
The approach described above can be applied to systems involving a multitude of hot and cold streams. Here, rather than deal with individual streams, overall supply and demand relationships are used. These relationships are called "Composite Curves" and were introduced by Huang and Elshout (1976). A methodology for generating such curves was presented by Umeda, Itoh and Shiroko (1978).
The Hot Composite Curve represents the overall heat supply within the processes as a function of temperature. It consists of a series on connected straight lines (Figure 3). Each change in slope represents a change in overall hot stream heat capacity flow-rate and is generally associated with either the "arrival" or "departure" of a stream from the temperature field. Where a stream arrives the overall heat capacity flow-rate increases and the slope of the Composite decreases.
The Cold Composite Curve represents the heat demands made within the process as a function of temperature.
Hot and Cold Composites can be superimposed in the manner described above for the two stream problem. Then, the overlap shows the scope for heat recovery within the process. The cold end overshoot shows the subsequent heat rejection to cold utility. The hot end overshoot shows the subsequent minimum quantity of heat required to operate the process.
With a two stream problem the close approach between the supply and demand lines occurs at the end of one of the lines. With Composite Curves the close approach often occurs at an intermediate point. This point was named the Pinch by Umeda, Itoh and Shiroko (1978).
The Pinch has significance for both the design of the heat recovery network and for the engineering of energy saving process modifications.
Consider the decomposition of the overall problem into two parts; an Above Pinch problem and a Below Pinch problem. Three key observations can be made.
Below the Pinch the heat supplied by the process equates with that demanded by the process plus the minimum cold utility requirement.
Above the Pinch the heat demanded by the process equates with that supplied from within the process plus the minimum hot utility requirement.
Any transfer of heat across the Pinch results in both heat balances being affected with both hot and cold utility requirements increasing by the quantity
These observations have very significant implications. First, they lead to a significant simplification in heat recovery network design methodology. A heat recovery network design ensuring minimum utility usage results from following just one simple expedient: prevent heat flow across the Pinch. Second, as described by Umeda et al. (1979), the overall energy needs of a process can be reduced by introducing process changes that shift heat demands from above to below the Pinch or shift heat supplies from below to above the Pinch.
The observation that the elimination of heat flow across the Pinch results in a network requiring minimum utility consumption resulted in the development of a powerful network design methodology called the Pinch Design Method [Linnhoff and Hindmarsh (1983)].
Heat flow across the Pinch can be eliminated by simply dividing the design problem into two parts: an above Pinch design and a below Pinch design. The results can then be merged in order to obtain an overall design.
The constraint that the designer now has to contend with is the minimum allowable temperature approach. This is most critical around the Pinch itself, where the temperature driving forces are smallest. The closest approach occurs at the Pinch itself and heat recovery matches must be selected such that the exchanger temperature profiles diverge as the streams move away from the Pinch. This is achieved if the matches obey the following rule:
This rule generally applies solely to matches made at the Pinch. Temperature driving forces in regions away from the Pinch are often large enough to allow matches not conforming to this rule.
Another important factor affecting the cost of a heat recovery network is the number of heat exchangers used. It was recognized that this is minimized if the load on a match is maximized.
So, the Pinch Design Method consists of the following steps:
Divide the overall problem into two parts at the Pinch.
Develop solutions for the individual parts. In each case start the design at the Pinch, making matches conforming to the CP rule. Maximize the heat load on these matches, thereby removing a stream from the problem.
Complete the subproblem design by making remaining heat recovery matches and by using utility units.
Merge the two designs to provide an overall solution.
The design achieved in this way will generally use more exchangers than necessary. This provides scope for further refinement. Exchangers can often be removed by shifting heat loads around the network using a technique called "loops and paths exploitation". The result is generally an infringement of the minimum temperature approach. The designer then has two options: the infringement can be accepted (in which case more area but fewer exchangers are used), or the temperature approach is restored through energy relaxation (in which case more energy but fewer exchangers are used).
As shown by Hohmann (1971) the minimum number of heat exchanger units required for a heat recovery network is given by
Nh = number of hot streams,
Nc = number of cold streams,
Nu = number of utility streams,
Ns = number of separate heat balanced systems.
In addition to developing an understanding of the thermodynamics of network design Umeda et al. (1978) showed how the area needs of a network could be estimated prior to design. They recognized that, in the case where hot and cold streams have uniform film heat transfer coefficients, α, the network area was minimized if the network exhibits pure counterflow. They then developed a network structure that exhibited this property. This structure was subsequently the basis on which Linnhoff and Ahmad (1990) derived an equation for network area estimation:
In practice, differing process streams can be expected to exhibit differing film heat transfer coefficients. One factor that greatly influences the coefficient actually achieved is the stream's allowable pressure drop. A means of predicting area requirements on the basis of stream pressure drop has been developed by Polley and Panjeh Shahi (1991).
These algorithms only provide approximate estimations for network area. In practice, as pointed out by Umeda et al., pure counterflow cannot be achieved in network design without using an inordinately large number of heat exchangers. The use of fewer exchangers leads to poorer use of temperature driving force and consequently to larger network areas. This "penalty" can often be compensate by judicious selection of stream matches. Where there are marked differences in individual stream coefficients, matches between streams having similar values should be favored. Then savings can be made through using higher temperature driving forces on matches having the lower overall heat transfer coefficients.
A reasonable prediction of the heat exchanger capital cost of a network can be obtained from:
where, Ca, Cb, and Cc are the cost factors normally applied in the estimation of the cost of a single heat exchanger.
As seen above, the minimum temperature approach has an effect on both network energy consumption and network area. The closer the temperature approach the lower the energy consumption but the higher the area required for the heat recovery (note: not necessarily total network area). This presents an optimization problem.
Umeda et al. (1978) proposed the following approach:
Develop composite curves.
Set energy recovery level.
Determine minimum area.
Reduce total capital cost by simplifying network structure.
Calculate total annual cost.
Repeat steps 2 to 5 to determine optimum network.
The approach suggested by Linnhoff and Ahmad (1990) can be summarized as follows:
Develop composite curves.
Set energy recovery level through specification of minimum allowable temperature approach.
Determine minimum network area.
Estimate minimum number of units.
Estimate subsequent capital cost.
Calculate total annual cost.
Repeat steps 2 to 6 to determine optimum energy recovery level.
Use the Pinch Design Method to subsequently design net work at the identified minimum temperature approach.
The procedure suggested by Umeda et al. could be undertaken by hand but is best suited to computer application. That suggested by Linnhoff and Ahmad is a hand method. Since it allows the designer to control the development of the network, it permits the engineer to both exercise art and to introduce practical constraints into the design. Consequently, it is possibly to be preferred. However, it should be recognized that Umeda's procedure could produce better results.
Figure 6. Minimum flue gas consumption. (a) Exhaust at acid dew point temperature, (b) Exhaust temperature constrained by process pinch. (c) Exhaust temperature constrained by process.
In 1982, Itoh, Shiroko, and Umeda introduced the Heat Demand and Supply Diagram. The HDS diagram is prepared by plotting the horizontal distance between the Composite Curves as a function of temperature. As such it shows the variation of heat supply and demand within the process (Figure 4). This allows the designer to determine the ways in which the residual heating and cooling duties can be met by available utilities. For instance, Figure 5 shows how utility steam levels can be set; Figure 6 shows how the quantity of flue gas necessary to operate a process can be determined; and, Figure 7 shows how a process' heating needs can be satisfied through a mixture of steam and flue gas.
Process integration is a wide ranging technology. Its application in retrofits has required the development of separate techniques [see Polley, Jegede, and Panjeh Shahi (1990)]. Developments that extend the technology to batch processes have been made by Kemp and Deakin (1989). Umeda, Niida and Shiroko (1979) showed how composite curves based on Carnot Factor rather than absolute temperature could be developed. Carnot factor curves have been used by Dhole and Linnhoff (1992) for the selection of refrigeration levels in cryogenic plants.
Huang, F. and Elshoot, R. (1976) Optimizing the heat recovery of crudeunits, Chem. Eng. Prog., 68-74, July 1976.
Umeda, T., Itoh, J., and Shiroko, K. (1978) Heal exchange system synthesis, Chem. Eng. Prog., 70-76, July.
Umeda, T., Niida, K., and Shiroko, K. (1979) A thermodynamic approach to heat integration in distillation systems, AICheJ, 25(3), 423-429.
Linnhoff, B. and Hindmarsh, E. (1983) The pinch design method for heat exchanger networks, Chem. Eng. Sci., 38, 745-763. DOI: 10.1016/0009-2509(83)80185-7
Linnhoff, B. and Ahmad, S. (1990) Cost optimum heat exchanger networks, Computers Chem. Engng., 14(7), 729-750. DOI: 10.1016/0098-1354(90)87083-2
Polley, G. T. and Panjeh Shahi, M. H. (1991) Interfacing heat exchanger network synthesis and detailed heat exchanger design, Chem. Eng. Res. Dev., 69A, 445-457.
Hohmann, E. C. (1971) Optimum Networks for Heat Exchange, Ph.D.Thesis, University of Southern California.
Itoh, J., Shiroko, K., and Umeda, T. (1982) Extensive applications of the T-Q diagram to heat integrated system synthesis, Int. Symp. on Process Systems Engineering, Kyoto, 1982, Computers. Chem. Eng., 10, 59-66, 1986. DOI: 10.1016/0098-1354(86)85046-3
Polley, G. T., Panjeh Shahi, M. S., and Jegede, F. O. (1990) Pressure drop considerations in the retrofit of heat exchanger networks, Chem. Eng. Res. Dev., 68A, 211-219.
Kemp, I. C. and Deakin, A. W. (1989) The cascade analysis for energy and process integration of batch processes, Chem. Eng. Res. Dev., 67A, 495-525.
Linnhoff, B. and Dhole, V. (1992) Shaftwork targets for low temperature process design, Chem. Eng. Sci., 47(8), 2081-2091. DOI: 10.1016/0009-2509(92)80324-6
- Huang, F. and Elshoot, R. (1976) Optimizing the heat recovery of crudeunits, Chem. Eng. Prog., 68-74, July 1976.
- Umeda, T., Itoh, J., and Shiroko, K. (1978) Heal exchange system synthesis, Chem. Eng. Prog., 70-76, July.
- Umeda, T., Niida, K., and Shiroko, K. (1979) A thermodynamic approach to heat integration in distillation systems, AICheJ, 25(3), 423-429. DOI: 10.1002/aic.690250306
- Linnhoff, B. and Hindmarsh, E. (1983) The pinch design method for heat exchanger networks, Chem. Eng. Sci., 38, 745-763. DOI: 10.1016/0009-2509(83)80185-7
- Linnhoff, B. and Ahmad, S. (1990) Cost optimum heat exchanger networks, Computers Chem. Engng., 14(7), 729-750. DOI: 10.1016/0098-1354(90)87083-2
- Polley, G. T. and Panjeh Shahi, M. H. (1991) Interfacing heat exchanger network synthesis and detailed heat exchanger design, Chem. Eng. Res. Dev., 69A, 445-457.
- Hohmann, E. C. (1971) Optimum Networks for Heat Exchange, Ph.D.Thesis, University of Southern California.
- Itoh, J., Shiroko, K., and Umeda, T. (1982) Extensive applications of the T-Q diagram to heat integrated system synthesis, Int. Symp. on Process Systems Engineering, Kyoto, 1982, Computers. Chem. Eng., 10, 59-66, 1986. DOI: 10.1016/0098-1354(86)85046-3
- Polley, G. T., Panjeh Shahi, M. S., and Jegede, F. O. (1990) Pressure drop considerations in the retrofit of heat exchanger networks, Chem. Eng. Res. Dev., 68A, 211-219.
- Kemp, I. C. and Deakin, A. W. (1989) The cascade analysis for energy and process integration of batch processes, Chem. Eng. Res. Dev., 67A, 495-525.
- Linnhoff, B. and Dhole, V. (1992) Shaftwork targets for low temperature process design, Chem. Eng. Sci., 47(8), 2081-2091. DOI: 10.1016/0009-2509(92)80324-6