Applied Thermal Engineering 26 (2006) 374381 www.elsevier.

com/locate/apthermeng

Global energy targets and optimal operating conditions for waste energy recovery in Bisphenol-A plant
M.B. Noureldin *, A.K. Hasan
Department of Materials and Process Engineering, The University of Waikato, Private Bag 3105, Hamilton, New Zealand Received 24 December 2004; accepted 23 June 2005 Available online 19 August 2005

Abstract The worldwide increase in energy demands and the uncertainty of future energy costs make energy conservation a major element in controlling operating costs of any chemical process. This paper presents new opportunities for energy saving inside Bisphenol-A (BPA) plant through the optimal selection of process operating temperatures that enhance plants energy recovery system. These optimal operational modications are systematically selected from a set of allowable process operating conditions without enumeration using novel interval constraint satisfaction model-based software called TEM_iconsTM. The paper shows that simple in-process modications for the sake of energy utilities saving in an actual BPA industrial facility can result in more than 17% saving in hot utility consumption. 2005 Elsevier Ltd. All rights reserved.
Keywords: Operational optimization; Heat integration; Bisphenol-A; TEM_iconsTM; Interval constraint satisfaction; Constraint logic propagation

1. Introduction BPA is a very important raw material for plastic industry. Its production is an energy intensive process. Keeping the BPA energy consumption bill under control needs more aggressive energy recovery schemes than just straightforward heat integration via in-process modications that enhance the waste energy recovery. Many methods and tools for analyzing the energy dimension in industrial chemical plants have been developed during the last two decades. These methods used mainly mathematical programming and pinch technology to determine the optimal process conditions that result in ecient use of energy in chemical process industries.

Corresponding author. Tel.: +64 7 8384466x6795; fax: +64 7 8384835. E-mail address: bahy@waikato.ac.nz (M.B. Noureldin). 1359-4311/$ - see front matter 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.applthermaleng.2005.06.017

Two main challenges face the industrial community in adopting these methods. The rst challenge is related to the plant-wide model development time and cost as well as the combinatorial nature of the customized full-scale model of a real industrial facility. The second one is the need to search the space of all possibilities in process operating conditions modications without exhaustive enumeration and without using heuristics that render sub-optimal solution. Furthermore the lack of a systematic technique to rigorously quantify the potential benets, in energy saving due to in-process modications, does not help justify the decision for very involved energy study. This paper addresses these two challenges systematically using new interval constraint satisfactionbased technology known as TEM_iconsTM. It also demonstrates that signicant energy saving can be realized in BPA plants through optimal in-process modications.

M.B. Noureldin, A.K. Hasan / Applied Thermal Engineering 26 (2006) 374381

375

2. Background of energy eciency optimization methods and tools 2.1. State-of-the-art in energy integration State-of-the-art energy integration research through the late 1980s has been reviewed [1]. Much work has been done on energy integration since then, but it has lacked the break through inuence of research done in the late seventies and the eighties [24]. This is especially true for commercial software development, where pinch technology is still the technique most widely used for energy integration applications. Mixed-integer linear and non-linear programming (MILP and MINLP) have been extensively used in academic software packages [59]. These MILP and MINLP are now being incorporated in some commercial packages. Such widely used energy integration software packages include HX-NET by Hyprotech (recently acquired by AspenTech), SUPERTARGET by Linnho March (recently acquired by KBC) and HEXTRAN by Simulation Science Inc. (acquired by Foxboro, Inc.). In general, these packages lack an embedded user-friendly systematic tool to automatically include all possible process parametric and structural modications. They also lack the ability to automatically include the energy systems operability requirements, which are necessary to dene optimal design and operation of any energy system during the targeting phase [10,11]. 2.2. Pinch technology and the optimization of process operating conditions Linnho and Vredeveld introduced the term Pinch Technology in the early eighties of the last century [12]. A number of basic principles and tools were then added and have become the foundation of Pinch technology. A general strategy for process modications was established by Linnho and Vredeveld [12]. It was referred to as the Plus/Minus principle. That means to increase plus heat available above Pinch and/or heat demand below Pinch or to reduce minus heat demand above Pinch and/or heat available below Pinch. It has also been proposed to change temperatures across the pinch that will shift heat duties from one part of the process to the other that is often possible. This principle is simply treated, as the ultimate tool that provides a definite reference for any adjustment in process heat duties and indicates which process modication would be benecial or detrimental. The problem with this principle is that it fails in some specic cases as shown by the author in another paper [13] due to the change in pinch location upon process modications. Besides, it needs enumeration to select best set of operating conditions that can be benecial to the plants energy problem. This task

can become impossible in huge chemical complex where the number of alternatives is enormous. 2.3. TEM_ICONSTM software at a glance TEM_iconsTM is an interval constraint satisfactionbased software. It contains unique capability for optimal energy systems targeting and design. It is the rst user-friendly software and up to our knowledge is the only one that can globally calculate, without exhaustive enumeration, energy utility targets under all possible combinations of process parametric and structural modications. It can also: (1) rigorously calculate these targets under all known possible combinations of process disturbances and uncertainties, and (2) dene the optimal process conditions so that optimal utility consumption can be specied. It globally solves the minmin and maxmax energy problems for industrial applications [14]. Solving these two problems gives the global minimum and maximum heating and cooling utilities for any plant. These global energy targets often are not attained together due to the fact that the process conditions that lead to the global minimum in heating utility are not the same one that lead to the global minimum cooling utilities. TEM_iconsTM like any other modern window-based software uses the main bar of the opening window to select the task to be implemented. Double clicking the ow task in the main bar leads to the software main window. This window consists of four sub-templates named ow, network variables, ROMC (region of minimum choice) and utilities with and without integration. The rst two of these four sub-templates are used for process conditions input data display and the second two are used for the calculated data display. The ow sub-template is used to display the process data entered through an interactive window named ow. After the ll-in-the-space task of such input template, the ow sub-template displays all the process streams associated with its names, types, lower values of the supply and target temperatures and the upper values of the supply and target temperatures, status of the streams (active or in-active) and the streams ow-specic heat values. The network variables sub-template displays the desired minimum temperature dierence between the hot streams to be cooled and the cold streams to be heated and the desired level of accuracy in temperature. The calculated values of global minimum heating and cooling utilities and the lower and upper values of pinch temperature region are displayed respectively in the two sub-templates named utilities with and without integration and the ROMC. Upon entering all the process conditions, in form of intervals, and specifying the desired temperature accuracy as well as the hot and cold streams minimum temperature dierence, the global energy consumption values will be calculated

376

M.B. Noureldin, A.K. Hasan / Applied Thermal Engineering 26 (2006) 374381

and displayed in the utilities sub-template in form of intervals. If the problem under such given process conditions exhibits pinch region, the boundaries of this region (lower and upper bound) on a temperature scale will be displayed in the ROMC sub-template. To specify the exact process conditions that result in desired global minimum heating or cooling utility consumption, the desired utility objective is double-clicked and the process conditions leading to the selected objective target is automatically calculated and displayed, in form of collapsed intervals, in the ow sub-template [15].

Applying the interval constraint satisfaction-based user-friendly TEM_iconsTM software for the BPA plant the above questions are answered systematically and without enumeration. 4.1. BPA process description overview Bisphenol-A (BPA) is an important intermediate used mainly in the manufacturing of polycarbonates and epoxy resins. Polycarbonate is the largest and the fastest growing market for BPA, accounting for about 60% of the world demand. The second largest end user (about 30%) for bisphenol-A is epoxy resins. Approximately 10% of BPA consumption goes into various specialty applications such as ame-retardants, unsaturated polyester resins, polysulfone resins, polyetherimides, polyacrylates, etc. Acetone reaction with fresh and recycle phenol in a catalytic stripping reactor utilizing ion exchange resin as catalyst renders BPA. The outlet reaction mixture consisting of un-reacted phenol and saturated with BPA is essentially free of acetone and contains very low levels of water and impurities. This mixture ows directly to the crystallization section where BPA adduct crystals are obtained. Adduct crystals are melted and distilled under vacuum to produce raw BPA. Un-reacted acetone, water and some phenol are separated from this reaction mixture. The acetone is recycled to BPA reactor, the water is eciently discharged, and the phenol is puried by distillation and mixed with fresh feed phenol. Alternate devices to form shapes such as akes and pellets solidify the molten BPA. The solidied BPA is conveyed to bagging and storage facilities. The mother liquor containing impurities, phenol and dissolved BPA is mostly recycled to the BPA reactor. Part of the mother liquor is sent to the purge recovery system where these impurities are partially decomposed and recombined to form BPA. The euents are mixed with the mother liquor and recycled to the BPA reactor. The undesirable impurities are condensed at the purge recovery system and discharged as tarry materials that can be used as fuel. The optimal purge ratio from the mother liquor is determined to control the product quality while keeping the consumption of raw materials low. 4.2. BPA process data extraction for energy consumption analysis The BPA plant studied in this paper consists of 22 hot and 15 cold streams as shown in Fig. 1. These streams are described in details in Table 1. Such streams are the ones allowed to be included in our heat integration study by the BPA process owners using the new technique presented earlier. The heating and cooling requirements are tabulated including identications and compositions.

3. Problem statement Given an industrial BPA plant that consists of several processing units, determine systematically (i) the global minimum of energy consumption for both heating and cooling utilities under all possible combination of allowed process operating conditions modications and (ii) the set of process conditions that render desired energy consumption scenario(s) without exhaustive enumeration.

4. Solution methodology Develop an energy path model for the BPA plant with all possible combinations of process modications in a constraint logic propagation environment with interval solver. This step helps capturing the eect of process streams interactions on the interval values of streams ow specic heats and supply and target temperatures due to process disturbances. This step can be avoided via obtaining the list of the allowed in-process modications for the sake of energy saving from the BPA process owners. Dene the global target/theoretical benchmark for minimum and maximum energy consumptions under interval-based process conditions. Find the process solution that achieves a desired energy consumption target via setting up the desired energy consumption target(s) and screening the interval-based process conditions that result in the desired energy consumption target. These procedures are formulated, automatically generated and implemented in the novel TEM_iconsTM software environment [15]. The questions addressed in this paper are (1) What are the global minima for both heating and cooling utilities in a real BPA plant via allowing set(s) of inprocess conditions modication? and (2) What are the values of these process conditions that will enable us realize these minima or at least one global minimum of an objective subject to a local minimum of the other objective?

M.B. Noureldin, A.K. Hasan / Applied Thermal Engineering 26 (2006) 374381

377

Phenol Acetone Catalyst

Formulation Tank 1 2, 3, 4 BPA Reactors

BPA Crystallizers BPA Melters Rotary Vacuum Filters

19, 20, 21, 22 Recycled

Phenol Storage

Phenol Desorbers

5 1

Anion Bed 3 Water Stripper

10 Phenol Nitrogen Phenolic Waster Water Feed Tank

6, 7, 8, 9 13, 14, 15

Recycled

BPA Product Flakers Lights to Incineration 8 BPA Rec. Column 10

To Glitsch Unit 4 7 Recycle Phenol to BPA Surge Tank Recycle Phenol to Phenol Wash Isomer Column

Recovered BPA to BPA Surge Tank

2 16 Purge to Tar Cracker Isomerization Reactor

Phenol Distillation Column 11

12 11

Mother Liquor Storage

12 Anion Bed

13 Tars to Cracker

Fig. 1. BPA plant data extraction ow sheet.

5. Results and discussion Without heat integration the BPA plant needs about 24.8 MBtu/h and 34.7 MBtu/h to satisfy its heating and cooling requirements respectively. Conducting heat integration between the hot and cold streams with a DTmin = 20 F and a pinch point at 233/213 F the heating and cooling utilities requirements are dramatically reduced to about 8.4 MBtu/h and 18.3 MBtu/h respectively. These results show the clear importance of heat integration in such chemical facility. Considering a set of possible process modication in the range of 5 F for each sensible heat streams supply and target temperatures we obtain new impressive achievable targets for energy consumption for the BPA plant. These new targets of heating and cooling utilities are about 6.9 MBtu/h and 14.1 MBtu/h respectively as shown in Fig. 2. It exhibits more than 17% savings in the heating utilities and about 23% savings in the cooling utilities over the targets realized using pinch technology for a given BPA plant data in Table 1. It is instructive to mention that only simple modications in the supply and/or targets temperature in BPA plant streams in the range of 5 F can result in such signicant saving of about 17% in the BPA plant heating utilities. It is important to note here that these benchmark calculations of heating and cooling utilities are global and performed under all possible combinations of modica-

tions in sensible heat streams supply and target temperatures all-at-once-change without enumeration. Upon selecting desired utility objective TEM_ iconsTM is used again to calculate the optimal process conditions that lead to it. The optimal supply and target temperatures of the BPA plant that results in global minimum heating utility of 6.9 MBtu/h are listed in Table 2. The modications in the BPA plant hot and cold streams supply and target temperatures do not render global minima for both heating and cooling utilities simultaneously. Since the heating utility is always more important in BPA plants the results discussed here start with the calculation of the global minimum heating utility subject to a local minimum cooling utility. The target of a global minimum heating utility, that equals to 6,919,401.64 Btu/h and exhibits 17.29% saving in heating utilities, can be realized with only a local minimum for the cooling utility equal to 14,250,457.52 Btu/h as shown in Fig. 3. This local minimum cooling utility is a little higher than the global one of 14,086,441 Btu/h calculated earlier and shown in Fig. 2. Even though heating utility minimization is always the desired energy target in BPA plants, it is important for the process owners to know the cooling utility benchmark under possible process modications to avoid any potential bottlenecks. The calculation of the global minimum cooling utility in the BPA plant under a set of possible process modication of 5 F for each sensible heat streams supply and target temperatures in the BPA

378

M.B. Noureldin, A.K. Hasan / Applied Thermal Engineering 26 (2006) 374381

Table 1 BPA plant streams data for heat integration Stream no. Stream identication Composition of stream Flow Supply Target FCp rate, Ib/h temp. F temp. F (Btu/h F) 220,518 77,569 77,569 77,569 8960 175 185 185 185 200 143 163 163 163 120 114,669.36 41,111.57 41,111.57 41,111.57 6,003.20 (sen) 3,960,320.00 (latent) 7,173.00 (sen) 6,774,500.00 (latent) 6,243.12 (sen) 2,509,254.00 (latent) 577,800.00 112,255.80 Heat duty, (Btu/h) 3,669,419.52 904,454.54 904,454.54 904,454.54 4,440,576.00

Hot streams h1 Reactor feed h2 Crystallizer feed h3 Crystallizer feed h4 Crystallizer feed h5 Water stripper top product

82% 66% 66% 66% 68%

Phenol, Phenol, Phenol, Phenol, Phenol,

5% acetone, 10% BPA 25% BPA, 1.5% water 25% BPA, 1.5% water 25% BPA, 1.5% water 31% water, 2% acetone

h6

Azeotropic column top product Phenol column top product

88% Water, 2% acetone, 10% phenol 7970

212

205

6,824,711.00

h7

100% Phenol

12,006

193

116

2,989,974.24

h8 h9 h10

Isomer column top product Isom. column tops dowtherm loop BPA rec. col. top product

BPA 100% Dowtherm G (BPA)

3210 261,060 4937

301 220 392

300 215 348

577,800.00 561,279.00

h11 h12 h13 h14 h15 h16 h17 h18 h19 h20

BPA rec. col. tops dowtherm loop BPA rec. col. top product 2nd cond. Phenol desorber bottoms Phenol desorber bottoms Phenol desorber bottoms Water stripper bottoms Vaccum seal water loop 113 F Tempered water loop Phenol desorber top product Phenol desorber top product

100% Dowtherm G 65% Phenol, 27% BPA, 9% others 100% BPA 100% BPA 100% BPA 80% Phenol, 11% BPA, 9% others 100% Water 113 F Tempered water 65% Phenol, 35% nitrogen 65% Phenol, 35% nitrogen

356,640 17,863 16,667 16,667 16,667 209,100 14,940 50,000 13,000 13,000

296 300 365 365 365 170 175 120 367 368

290 205 330 330 330 165 136 100 182 182

2,665.98 (sen) 1,005,963.12 888,660.00 (latent) 153,355.20 920,131.20 9,378.08 8,983.51 8,983.51 8,983.51 109,777.50 14,940.00 50,000.00 5,330.00 5,330.00 (sen) 1,755,000 (latent) 5,330.00 (sen) 1,755,000 (latent) 3,280.00 890,917.13 314,422.96 314,422.96 314,422.96 548,887.50 582,660.00 1,000,000.00 986,050.00 2,746,380.00

h21

Phenol desorber top product

65% Phenol, 35% nitrogen

13,000

368

182

2,746,380.00

h22

Phenol desorber top product

65% Phenol, 35% nitrogen

8000

367

182

606,800.00 34,728,988.95 4,592,575.05 6,195,000.00 439,016.16 1,672,000.00 48,132.00 95,607.29 95,607.29 95,607.29 95,607.29 195,773.76 4,430,000.00 419,000.00 902,000.00 2,520,000.00 2,988,000.00 24,783,270.11

Cold streams c1 Water stripper feed c2 Water stripper reboiler feed c3 Azetropic column feed c4 Azetropic column reboiler feed c5 Isomerization reactor feed c6 Phenol desorber feed c7 Phenol desorber feed c8 Phenol desorber feed c9 Phenol desorber feed c10 Phenol column reboiler feed c11 Phenol column reboiler feed c12 Isomer column reboiler feed c13 BPA column reboiler feed c14 176 F Tempered water loop c15 203 F Tempered water loop

78% Phenol, 11% BPA, 9% others 78% Phenol, 10% BPA, 11% others 65% Phenol, 33% water 98% Phenol, 2% water 81% Phenol, 13% BPA, 6% others 100% Nitrogen 100% Nitrogen 100% Nitrogen 100% Nitrogen 100% Phenol 71% BPA, 28% isomers 72% BPA, 27% isomers 80% BPA, 20% isomers 50% Glycol, 50% water 50% Glycol, 50% water

236,426 30,975 8976 8000 18,336 1873 1873 1873 1873 15,687 22,150 2095 4510 60,000 48,000

213 240 127 285 170 180 180 180 180 150 425 473 473 120 120

250 250 200 290 175 385 385 385 385 174 440 478 491 176 203

124,123.65 619,500.00 6,013.92 334,400.00 9,626.40 466.38 466.38 466.38 466.38 8,157.24 295,259.50 83,800.00 50,106.10 45,000.00 36,000.00

plant, shown in Fig. 4 is equal to 14,086,441.04 Btu/h. It exhibits 23% savings in minimum cooling utility we obtained through heat integration for the BPA plant data

shown in Table 1. Such global minimum cooling utility can be realized but with less attractive heating utility of 7,996,621.66 Btu/h. It is useful to mention here that this

M.B. Noureldin, A.K. Hasan / Applied Thermal Engineering 26 (2006) 374381

379

Fig. 2. BPA plant global utilities targets under all possible combinations of process modications.

Table 2 Optimal operating conditions for global minimum heating utilities Stream no. Stream identication Composition of stream Initial Supply temp. F h1 h2 h3 h4 h5 h7 h10 h12 h13 h14 h15 h17 h18 h19 h20 h21 h22 C1 C3 C6 C7 C8 C9 c10 c14 c15 Reactor feed Crystallizer feed Crystallizer feed Crystallizer feed Water stripper top product Phenol column top product BPA rec. col. top product BPA rec. col. top product 2nd cond. Phenol desorber bottoms Phenol desorber bottoms Phenol desorber bottoms Vaccum seal water loop 113 F Tempered water loop Phenol desorber top product Phenol desorber top product Phenol desorber top product Phenol desorber top product Water stripper feed Azetropic column feed Phenol desorber feed Phenol desorber feed Phenol desorber feed Phenol desorber feed Phenol column reboiler feed 176 F Tempered water loop 203 F Tempered water loop 82% Phenol, 5% acetone, 10% BPA 66% Phenol, 25% BPA, 1.5% water 66% Phenol, 25% BPA, 1.5% water 66% Phenol, 25% BPA, 1.5% water 68% Phenol, 31% water, 2% acetone 100% Phenol (BPA) 65% Phenol, 27% BPA, 9% others 100% BPA 100% BPA 100% BPA 100% Water 113 F Tempered water 65% Phenol, 35% nitrogen 65% Phenol, 35% nitrogen 65% Phenol, 35% nitrogen 65% Phenol, 35% nitrogen 78% Phenol, 11% BPA, 9% others 65% Phenol, 33% water 100% Nitrogen 100% Nitrogen 100% Nitrogen 100% Nitrogen 100% Phenol 50% Glycol, 50% water 50% Glycol, 50% water 175 185 185 185 200 193 392 300 365 365 365 175 120 367 368 368 367 213 127 180 180 180 180 150 120 120 Target temp. F 143 163 163 163 120 116 348 205 330 330 330 136 100 182 182 182 182 250 200 385 385 385 385 174 176 203 New Supply temp. F 170 180 180 180 200 193 392 305 370 370 370 170 115 367 368 368 367 218 122 175 175 175 175 145 115 115 Target temp. F 148 168 168 168 125 121 343 210 325 325 325 141 105 187 187 187 187 245 205 380 380 380 380 179 181 208

local minimum heating utility target is worse than the global one obtained earlier by more than 15%. Comparing the global minimum heating utility objective solu-

tion, which exhibits BPA plant needs of about 6.9 MM Btu/h and 14.25 MM Btu/h heating and cooling utilities respectively, with the global minimum cooling utility

380

M.B. Noureldin, A.K. Hasan / Applied Thermal Engineering 26 (2006) 374381

Fig. 3. BPA global minimum heating utility target with a local minimum cooling utility.

Fig. 4. BPA global minimum cooling utility target with a local minimum heating utility.

objective solution, which exhibits utilities need of about 8 MM Btu/h and 14.09 MM Btu/h heating and cooling utilities respectively, It is clear that choosing the global heating solution as the desired objective renders the minimum utilities (heating and cooling) solution for this BPA plant. Whereas, BPA plants hot utilities such as high and medium pressures steam, are always sources of bottlenecks and more expensive than the cold utility, which is normally water, the global minimum heating objective solution will not only be the minimum utility cost solution but also the desired one for process debottlenecking.

It is interesting to note that the BPA hot and cold streams supply and target temperatures that render the global minimum cooling utility are exactly the same optimal conditions that render the global minimum heating utility for the BPA plant, listed in Table 2, except one single modication in the C1 stream supply temperature. The new C1 stream supply temperature has to be 208 F to achieve global minimum cooling utility in the BPA plant under the allowed simple modications of 5 F for each sensible heat streams supply and target temperatures. It is the intention of the authors to close this section with this important result that shows

M.B. Noureldin, A.K. Hasan / Applied Thermal Engineering 26 (2006) 374381

381

the eect of changing C1 stream supply temperature on the BPA plant hot and cold utility consumption. It shows that making the C1 cold stream colder and increasing the total cold streams duty below the pinch, as per the Pinch technology Plus/Minus principle, while can bring in savings in cooling utility it hurts the heating utility requirement. In depth discussion of process situations in which pinch technology plus/minus principle fails, to systematically lead to denite savings in both heating and cooling utilities, will be the topic of another paper [13].

References
[1] T. Gundersen, L. Naess, The synthesis of cost optimal heat exchanger networks: An industrial review of the state of the art, Computer and Chemical Engineering 12 (1988) 503530. [2] E.C. Hohmann, Optimum networks for heat exchanger, Ph.D. Thesis, University of Southern California, Los Angeles, 1971. [3] B. Linnho, E. Hindmarsh, The pinch design method for heat exchanger networks, Chemical Engineering Science 38 (1983) 745 763. [4] B. Linnho, Pinch analysis-state of the art overview, Transactions Institution Chemical Engineering Research and Design 71 (A5) (1993) 503522. [5] C.A. Floudas, A.R. Ciric, I.E. Grossmann, Automatic synthesis of optimum heat exchanger network congurations, AIChE J. 32 (1986) 276290. [6] C.A. Floudas, A.R. Ciric, Strategies for overcoming uncertainties in heat exchanger network synthesis, Computers and Chemical Engineering 13 (1989) 11171128. [7] A.K. Saboo, M. Morari, R.D. Colberg, RESHEXan interactive software package for the synthesis and analysis of resilient heat exchanger networksI and II, Computers and Chemical Engineering 10 (1986) 577591. [8] A.R. Ciric, C.A. Floudas, Heat exchanger network synthesis without decomposition, Computers and Chemical Engineering 15 (1991) 385396. [9] B. Bolio, A. Turkay, T.F. Yee, I.E. Grossmann, SYNHEAT Manual, Carnegie Mellon University, Pittsburgh, 1994. [10] S.A. Papoulias, I.E. Grossmann, A structural optimization approach in process synthesisII. Heat recovery networks, Computers and Chemical Engineering 7 (6) (1983) 707721. [11] L.T. Biegler, I.E. Grossmann, A.W. Westerberg, Systematic Methods of Chemical Process Design, Prentice Hall, New Jersey, USA, 1997. [12] B. Linnho, D.R. Vredeveld, Pinch technology has come of age, Chemical Engineering Progress 80 (7) (1984) 3340. [13] M.B. Noureldin, How and how not to check pinch technology principles, Appl. Therm. Eng., in preparation. [14] M.B. Noureldin, Improved System and Computer Software for Modeling Energy Consumption, New Zealand Patent No. 527244, 2003. [15] M.B. Noureldin, TEM_iconsTM 1.2 Users manual Report, Department of Materials and Process Engineering, University of Waikato, Hamilton, New Zealand, 2003.

6. Conclusions Heat integration of hot and cold streams is mandatory to cut energy cost in BPA plants. Signicant energy saving can be obtained through simple in-process modications. The 17% savings obtained in the BPA plants heating utility presented in this paper justies exploring more aggressive energy recovery opportunities in BPA plants through more involved in-process changes. Interval constraint satisfaction technology-based software introduced in this paper, known as TEM_ iconsTM, can systematically denes opportunities for energy saving through the all-at-once-changes in BPA process plants without enumeration. It can also be used to search systematically for optimal process conditions that render desired energy consumption targets.

Acknowledgement The nancial support of the New Zealand foundation of Research, Science and Technology, contract number UOWX0302 (solutions for energy eciency at work and at home) is acknowledged.