Design of distributed wastewater treatment systems with multiple contaminants

Chemical Engineering Journal 228 (2013) 381–391

Contents lists available at SciVerse ScienceDirect

Chemical Engineering Journal journal homepage: www.elsevier.com/locate/cej

Design of distributed wastewater treatment systems with multiple contaminants Zhi-Hua Liu, Jing Shi  , Zhi-Yong Liu ⇑ Key Laboratory of Green Chemical Technology and Efficient Energy Saving of Hebei Province, Hebei University of Technology, Tianjin 300130, China

h i g h l i g h t s  Design of distributed wastewater treatment is investigated in this paper.  New method to reduce unreasoning stream-mixing in design is proposed.  Processes divided into three-unit groups based on minimum-mixing rule proposed.  Precedence of the units in the group selected is determined by heuristic rules.  It is shown that the design method proposed is simple and effective.

a r t i c l e

i n f o

Article history: Received 16 January 2013 Received in revised form 26 April 2013 Accepted 30 April 2013 Available online 16 May 2013 Keywords: Design Environment Water Systems engineering Pinch method Process synthesis

a b s t r a c t In design of a distributed wastewater treatment system, wastewater degradation caused by unreasoning stream-mixing will increase the total treatment flowrate, and this will often increase treatment cost. Therefore, it is necessary to reduce unreasoning stream-mixing as much as possible in the design procedure. This paper proposes a new method to reduce unreasoning stream-mixing in the design of distributed wastewater treatment networks. The design procedure includes following steps: (1) the main function of each treatment unit is identified; (2) the minimum treatment amount of each unit for its main contaminant, without considering other contaminants, is obtained with pinch method; (3) for the systems with many treatment units, a three-unit-group is selected and the precedence order of the units in the group is determined with the heuristic rules proposed in this paper. The above procedure will continue till the number of the units left is equal to or less than three. Some literature examples are investigated, and the results obtained in this work are compared to that obtained in the literature. It is shown that the design method proposed is simple and effective. Ó 2013 Elsevier B.V. All rights reserved.

1. Introduction Water is an important resource for process industry and agriculture. Water network synthesis is an important research aspect in process synthesis. Researchers have presented considerable work for design of water-using networks based on the water pinch analysis and/or the mathematical optimization. The development of water-using network design can be found in recent reviews [1,2]. The design for effluent treatment system is another important aspect of water network integration. Eckenfelder et al. [3] and Lankford et al. [4] addressed that distributed treatment systems had obvious advantages over centralized systems, because streams are treated separately or mixed when appropriate in the former

⇑ Corresponding author. Tel./fax: +86 22 6020 2047.  

E-mail address: [email protected] (Z.-Y. Liu). Current address: Datang Inner Mongolia Duolun Coal Chemical Co. Ltd., China.

1385-8947/$ - see front matter Ó 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.cej.2013.04.112

system. This can reduce the treatment amount and cost of the system. For distributed treatment system design, the methods proposed can be classified into pinch analysis methods, mathematical optimization methods and other methods. The first work based on pinch analysis was proposed by Wang and Smith [5]. In their work, before design, the minimum flowrate target was determined by pinch analysis method. Kuo and Smith [6] proposed a concept of mixing exergy loss to measure the wastewater degradation when the streams are mixed. After 1990, a few insight-based approaches were proposed for the wastewater treatment network integration [7,8]. The mathematical optimization methods have become increasingly popular to design complex systems with many processes. Galan and Grossmann [9] applied a non-convex nonlinear procedure to find the global optimum. Hernandez-Suarez et al. [10] presented an optimization approach to break a complex superstructure down into a number of basic network superstructures. Karuppiah and

382

Z.-H. Liu et al. / Chemical Engineering Journal 228 (2013) 381–391

Nomenclature Si fi RRA ci,A mi,A min,A mout,A C dis A TPj M rem TPj;A

stream Si flowrate of stream Si removal ratio of contaminant A concentration of contaminant A in Si mass load of contaminant A in Si mass load of contaminant A at TPj inlet mass load of contaminant A at TPj outlet concentration of contaminant A in discharging stream treatment unit j removed mass of contaminant A by TPj

Grossmann [11,12] proposed a superstructure model that included all feasible alternatives for wastewater treatment, reuse and recycle. Castro et al. [13] adopted an algorithm, which is composed of two parts, to find global optimal solutions. Statyukha et al. [14] designed wastewater treatment network with a hybrid approach, which is a sequential method integrating insight-based techniques and mathematical programming. Bandyopadhyay [15] proposed an algebraic methodology based on process integration to determine the minimum treatment flowrate. Shi and Liu [16] proposed a new concept, the total treatment flowrate potential (TTFP), to reduce the unreasoning mixing of streams in the design procedure. Liu et al. [17] designed the wastewater treatment networks with single contaminant by optimizing the treatment amount function. Soo et al. [18] extended the graphical targeting technique to determine the minimum treatment flowrate for the single and two contaminant system with multiple processes. In design of a distributed wastewater treatment system, wastewater degradation caused by unreasoning stream-mixing will increase the total treatment flowrate, this will often increase treatment cost. Therefore, it is necessary to reduce unreasoning stream-mixing as much as possible in the design procedure [5,6,16,17]. In this paper, the reducing of stream-mixings is considered as a major factor to decrease the total treatment amount of the system. A new heuristic (minimum-mixing) rule is proposed to determine the precedence order of the treatment units. Pinch method is used to calculate the minimum treatment amount of each unit for its main contaminant, without considering other contaminants. The main contaminant of unit is identified by its removal ability to each contaminant. Some literature examples are investigated, and the results obtained in this work are compared to that obtained in the literature. It is shown that the design method proposed is simple and effective. 2. Problem statement Given is a group of wastewater streams, which contain certain contaminants with known concentrations. Given is also a group of treatment units and each unit can remove one or a few contaminants. It is assumed that there is no flowrate loss in the treatment process. The task is to design a distributed treatment system for removing certain contaminant(s) to meet the environmental limiting with minimum total treatment amount, because the treatment amount is often proportional to the treatment cost [5,6].

MTPj,A STPj SUj SUj-P FTPj,in clim Env ;A M lim Env ;A

mass load of contaminant A at TPj inlet collection of streams treated by TPj outlet stream of TPj bypassed flow rate of pinch stream for TPj treatment amount of TPj for its main contaminant environment limit concentration of contaminant A environment mass load limit of contaminant A

3.1. The main function of treatment unit In a multiple contaminant system, each unit can often remove one or a few contaminants. For the system in which each unit can remove only one contaminant, the design procedure is simple. For the system in which some units can remove more than one contaminant, we need to identify the main contaminant of the units. For a treatment unit TPj with contaminant removal ratios of RRA, RRB. . ., the contaminant which corresponds to the largest removal ratio (say RRA) will be the main contaminant of TPj. In this situation, the main function of TPj is to remove contaminant A. If we only consider its main contaminant without considering the influence of other contaminants, the treatment amount of TPj can be obtained by pinch method. 3.2. Obtaining the streams treated by each unit for its main contaminant and determining the elements of STPj Pinch point is one of the key issues for design of the effluent treatment networks. According to Wang and Smith [5], the streams above the pinch of system are entirely treated, the pinch stream is partially treated, and those below the pinch are fully bypassed. Let us consider the situation when the main function of TPj is to remove contaminant A and this contaminant in all the streams is removed by TPj, without considering the influence of the other treatment units and contaminants. As shown in Table 1, all the streams are rearranged as decreasing concentration order of contaminant A, from S1 to Snk, where fi is the flowrate of Si, and ci,A and mi,A are the concentration and mass load of contaminant A in Si, respectively The removal ratio RRA is defined as RRA = (min,A - mout,A)/min,A, where min,A and mout,A are the mass loads of contaminant A at inlet and outlet of TPj, respectively. clim Env ;A is the environment limit concentration of contaminant A. For the system shown in Fig. 2, if stream Sp is the pinch stream, the minimum mass load of contaminant A removed by TPj should be:

Mrem TPj;A ¼

nk X mi;A  M lim Env ;A

ð1Þ

i¼1

Pnk where mass load of contaminant A in all i¼1 mi;A is the total P nk lim streams, and M lim ¼ c  i¼1 fi is the limiting mass load of conEnv ;A Env ;A taminant A in the discharging stream. j

The mass load of contaminant A at the inlet of TP will be: 3. Design procedures for the systems with multiple treatment units and contaminants The design flow chart is shown in Fig. 1. The detailed discussion of each step shown in Fig. 1 will be given in the following sections.

MTPj;A ¼

nk X mi;A  M lim Env ;A

!,

RRA

i¼1

Then, the treated amount of pinch stream Sp will be:

ð2Þ

Z.-H. Liu et al. / Chemical Engineering Journal 228 (2013) 381–391

383

Fig. 1. Design flow chart for distributed wastewater treatment systems with multiple contaminants.

Table 1 Determining the pinch by the mass loads of the streams.

MTPj;A 

Streams

fi

ci,A

mi,A

S1 S2 ... Sp-1 Sp Sp+1 ... Snk

f1 f2 ... fp-1 fp fp+1 ... fnk

c1,A c2,A ... cp-1,A cp,A cp+1,A ... cnk,A

m1,A m2,A ... mp-1,A mp,A mp+1,A ... mnk,A

F treated TPj;pinch ¼

p1 X mi;A i¼1

cp;A

ð3Þ

The bypass amount of the pinch stream will be:

F bypass TPj;pinch

¼ fp  F treated TPj;pinch

The total treatment amount of TPj will be:

Fig. 2. The treatment amount of a unit for its main contaminant A.

ð4Þ

384

F TPj;in ¼ F treated TPj;pinch þ

Z.-H. Liu et al. / Chemical Engineering Journal 228 (2013) 381–391 p1 X fi

ð5Þ

and B, respectively. The relevance of STP1 and STP2 can be divided into the following three conditions.

i¼1

For the system shown in Fig. 2, streams S1, S2, . . . , SP-1 and part of SP should be treated by TPj. These streams (S1, S2, . . . , SP) are called as set STPj.

3.3. The relevance between treatment units For convenience, let us consider two treatment units (say TP1 and TP2) in an effluent treatment system, in which the flowrates and contaminant concentrations of streams S1, S2,. . .,Sn are known. The main functions of TP1 and TP2 are to remove contaminant A

3.3.1. TP1 and TP2 being irrelevant units The intersection of two sets, STP1 and STP2, shown in Fig. 3a, is an empty set (STP1 \ STP2 ¼ Ø). For this situation, we call that TP1 and TP2 are irrelevant units. As shown in Fig. 3a, the main function of TP1 is to remove contaminant A. The streams above pinch are S1, S2, S3 and pinch stream is S4. If TP1 is executed first, the stream to be treated is the mixing of S1, S2, S3 and part of S4. The above mixing will not affect TP2 at all because the stream to be treated by TP2 is the mixing of S5, S6, S7 and part of S8. Vice versa, if TP2 is executed first, the treatment amount of TP1 will not be affected either. In Fig. 3, SU1 is the stream of outlet at TP1 and SU1-P is the by-

Fig. 3. The relevance between treatment units.

Z.-H. Liu et al. / Chemical Engineering Journal 228 (2013) 381–391 Table 2 Stream and treatment unit data for Example 1. Stream

(a) Stream data 1 2

Concentration (ppm) A

B

100 15

20 200

Treatment unit

(b) Treatment unit data 1 2

Flowrate fi (t/h)

40 40

Removal ratio RRi (%) A

B

95 0

0 97.6

passed flowrate of pinch stream for TP1. For the two sets whose intersection is an empty set (STP1 \ STP2 ¼ Ø), whatever TP1 or TP2 is executed first, the treatment amount of the follow-up unit will not increase. 3.3.2. TP1 and TP2 being relevant units The intersection of sets, STP1 and STP2, shown in Fig. 3b, is a nonempty set (STP1 \ STP2 –Ø and STP1 å STP2 ). For this situation, we call that TP1 and TP2 are relevant units. The intersection of STP1 and STP2 is {S1, S2}. As shown in Fig. 3b, the main function of TP1 is to remove contaminant A. If TP1 is executed first, the stream to be treated is the mixing of S1, S2, S3 and part of S4. After TP1, the treatment amount of TP2 will be larger than that shown in Fig. 3b because streams S1 and S2 have been mixed in TP1. Vice versa, if TP2 is executed first, the treatment amount of TP1 will be larger than that shown in Fig. 3b as well. For the two sets whose intersection is a non-empty set (and one set is not a proper subset of the other set), whatever TP1 or TP2 is executed first, the treatment amount of the follow-up unit will increase. 3.3.3. TP1 is included in TP2 One set (say, STP1) is a proper subset of another set (say, STP2), as shown in Fig. 3c. For this situation, we call that TP1 is included in TP2. In Fig. 3c, the set STP1 is the collection of streams S1, S2, S3 (STP1 = {S1, S2, S3}). The set STP2 is the collection of streams S1, S2 S3 . . . , Sj (STP2 = {S1, S2 S3, . . . , Sj}, j > 3). It can be seen from Fig. 3c that STP1 is a proper subset of STP2. When TP1 is executed first, the treatment flowrate of TP2 will not increase. But when TP2 is executed first, the treatment flowrate of TP1 will increase. From the above discussion, the relevance between treatment units can be summarized as: irrelevant, relevant and included. 3.4. Grouping rules From the situations discussed above, it can be seen that if some streams are mixed in a treatment unit unreasonably, the treatment flowrate of the subsequent units will be increased. This will increase the total treatment amount of the system. In this paper,

385

we decide the precedence order of the treatment units in following way: the less the mixing a treatment unit causes, the higher the priority of the unit. The precedence order of the treatment units is in the following order: units causing no or less mixing, irrelevant units, relevant units and the units causing serious mixing. This is called as the minimum-mixing rule. In the design procedure, if all the treatment units are considered together, the problem will be very complex. In this paper, we propose a new design strategy: select three treatment units as a group prior to other units based on the minimum-mixing rule proposed above. Here, the number of treatment units in the selected group can be increased if the number of units in system is too large. In summary, the rules to select three units as a prior group are as follows: (1) the unit(s) which will not cause any stream-mixing should be selected first. (2) Then the units, which are irrelevant or will cause less stream-mixing, are considered. (3) The units, which are relevant and will cause large amount of stream-mixing, will be selected last. The precedence order of treatment units in the selected group will be determined by the rules in Section 3.5. 3.5. The rules to determine precedence order of treatment units in the selected group The precedence of treatment units in the selected group also needs to be determined according to the minimum-mixing rule. In addition, the effect of non-main contaminant should also be considered (as will be discussed in Section 3.5.2). 3.5.1. Rules to determine precedence of units removing one contaminant For units, which can only remove one contaminant, the precedence order of units can be determined based on the minimummixing rules discussed as above, in the following way: (1) The units not causing any stream-mixing. (2) The units causing less stream-mixing. (3) The units causing large stream-mixing. 3.5.2. Rules to determine precedence of units removing multiple contaminants For units, which can remove multiple contaminants, besides the effect of the main contaminant, the effect of non-main contaminant should also be considered when determining the precedence order of the treatment units. Let us take an example of two units to show how to consider the effect of non-main contaminants of units. As shown in Fig. 3, if the removal ratios of TP1 for contaminants A and B are 90% and 60%, respectively, and the removal ratios of TP2 for contaminants A and B are 0% and 90%, respectively. The main contaminants of TP1 and TP2 are contaminants A and B, respectively. The non-main contaminant of TP1 is contaminant B. Now we will consider the effect of non-main contaminant in two situations.

Fig. 4. The pinch streams and treatment amounts of TP1 and TP2 in Example 1.

386

Z.-H. Liu et al. / Chemical Engineering Journal 228 (2013) 381–391

Fig. 5. The design for Example 1.

4. Case studies

Table 3 Stream and treatment unit data for Example 2. Stream

(a) Stream data 1 2 3

Concentration (ppm) A

B

C

600 400 200

500 200 1000

500 100 200

Treatment unit

(b) Treatment unit data 1 2 3

Flowrate fi (t/h)

4.1. Example 1

20 15 5

The data for Example 1 are shown in Table 2, taken from Wang and Smith [5]. The environment limit for each contaminant is 10 ppm. The design procedure is as follows.

Removal ratio RRi (%) A

B

C

90 0 0

0 99 0

0 0 80

1. Identify the main function of each process. From Table 2, it can be seen that each unit can only remove one contaminant. 2. Obtain the streams treated by each unit for its main contaminant

For the situation when the two units are irrelevant, as shown in Fig. 3a, if TP1 is prior to TP2, the treatment amount of TP2 will be less than that shown in Fig. 3a, because TP1 has removed some amount of contaminant B, and this will reduce the treatment flowrate TP2. However, if TP2 is prior to TP1, the treatment amount of TP1 will not be less than that shown in Fig. 3a. For the situation when the two units are relevant or included, the precedence order of treatment units can be determined by the following method. If TP1 is prior to TP2, the treatment amounts of two units (for their main contaminants) can be calculated according to Section 3.2. For the sequence of TP1 and TP2, the total amount is denoted as F1,2. Similarly, for the sequence of TP2 and TP1, the total amount is denoted as F2,1. If F1,2 < F2,1, TP1 will be prior to TP2, vice versa. The design procedure proposed is illustrated with a few literature examples as follows.

According to the procedure in Section 3.2, the pinch stream and the treatment amount of each unit can be obtained. Take TP1 as an example, which is specified to remove its main contaminant A. Arrange two streams in descending order of concentration of A: S1, S2, as shown in Fig. 4a. It can be obtained by the procedure in Section 3.2 that the pinch stream of TP1 is S1, and the treatment amount of TP1 is 40t/h. In Fig. 4a, SU1 is the outlet stream of TP1. Similarly, as shown in Fig. 4b, it can be obtained by the procedure in Section 3.2 that the pinch stream of TP2 is S1, and the streams to be treated by TP2 are S2 (40 t/h) and part of S1 (9.85 t/h). In Fig. 4b, SU2 is the outlet stream of TP2 and SU2-P is the bypass flowrate of pinch stream. 3. Execute the units in turn and obtain final design.

Fig. 6. The design for Example 2.

Z.-H. Liu et al. / Chemical Engineering Journal 228 (2013) 381–391 Table 4 The stream data after TP3 in Example 2. Stream

From Table 3a, it can be seen that the main functions of TP1, TP2 and TP3 are to remove contaminants A, B and C, respectively.

Concentration (ppm)

SU3 SU3-P S2

Flowrate fi (t/h)

A

B

545.87 200 400

567.66 1000 200

2. Obtain the streams treated by each unit for its main contaminant.

23.13 1.87 15

SU3 is the outlet stream of TP3. SU3-P is the bypass flowrate of pinch stream.

Table 5 The stream data after TP2 in Example 2. Stream

Concentration (ppm) A

Flowrate fi (t/h)

SU2-P SU2 S2

545.87 518.33 400

1.51 23.49 15

SU2 is the outlet stream of TP2. SU2-P is the bypass flowrate of pinch stream.

From above analysis, TP1 will not cause any stream-mixing. Therefore, the precedence order is TP1 and TP2, according to Section 3.5.1. After TP1, as shown in Fig. 5a, SU1 and S2 are the current streams of TP2. It can be obtained by the procedure in Section 3.2 that the treatment amount of TP2 will be 49.85t/h. The final design is shown in Fig. 5b, which is the same as that obtained by Wang and Smith [5].

4.2. Example 2 The data for Example 2 are shown in Table 3, taken from Kuo and Smith [6]. The environment limit for each contaminant is 100 ppm. The design procedure is as follows. 1. Identify the main function of each process.

Concentration (g/m3) A

(a) Stream data 1 100 2 600 3 900 4 10 5 40 6 0 7 120 8 370 9 900 10 250 11 0 12 0 13 2000 14 0 15 1000

Flowrate fi (m3/s)

B

C

D

E

50 800 0 10 170 1100 10 20 350 270 1190 0 600 5 1510

350 1500 600 100 0 0 500 100 200 90 60 20 340 100 270

0 0 150 3000 500 200 2000 30 80 0 230 800 0 600 150

70 910 230 850 690 340 70 690 230 580 370 100 30 40 220

Treatment unit

(b) Treatment unit data 1 2 3 4 5

According to the procedure in Section 3.2, the pinch stream and the treatment amount of each unit can be obtained. Let us show the detailed calculation for TP3, whose main contaminant is C. Arrange all streams in descending order of concentration of C: S1, S3, S2, as shown in Fig. 6a. It can be obtained by the procedure in Section 3.2 that the pinch stream of TP3 is S3, and the treatment amount of TP3 is 23.13 t/h (FTP3,in = 23.13 t/h). In Fig. 6a, SU3 is the outlet stream of TP3 and SU3-P is the bypass flowrate of pinch stream S2. Similarly, the streams to be treated by TP1 are S1 (20 t/ h) and part of S2 (11.67 t/h), and the treatment amount of TP1 is 31.67 t/h (FTP1,in = 31.67 t/h). The streams to be treated by TP2 are S3 (5 t/h) and part of S1 (18.28 t/h), and the treatment amount of TP2 is 23.28 t/h (FTP2,in = 23.28 t/h). 3. Execute the units in turn and obtain final design. Because FTP1,in > FTP2,in > FTP3,in, TP3 should be executed first, according to Section 3.5.1. After TP3, as shown in Fig. 6a, SU3, SU3-P and S2 are the current streams of TP1 and TP2. The stream data after TP3 are listed in Table 4. From Table 4, it can be obtained by the procedure in Section 3.2 that the treatment amounts of TP1 and TP2 are 33.23t/h (FTP1,in) and 23.49 t/h (FTP2,in), respectively. Because FTP1,in > FTP2,in, TP2 should be executed before TP1 according to Section 3.5.1. After TP2, as shown in Fig. 6b, SU2, SU2-P and S2 are the current streams of TP1. The stream data after TP2 are listed in Table 5. From Table 5, it can be obtained by the procedure in Section 3.2 that the streams to be treated by TP1 are SU2-P (1.51 t/h), SU2 (23.49 t/h) and part of S2 (9.17 t/h). The treatment amount of TP1 is 34.17 t/h (FTP1,in). The final design is shown in Fig. 6c, which is the same as the best result obtained by Kuo and Smith [6].

4.3. Example 3

Table 6 Stream and treatment unit data for Example 3. Stream

387

36 24 15 25 18 35 9 2 3 23 89 1 5 41 8

The data for Example 3 are shown in Table 6, taken from Galan and Grossmann [19]. The environment limit for each contaminant is 100 g/m3. The design procedure is as follows. 1. Identify the main function of each process.

Removal ratio RRi (%) A

B

C

D

E

40 90 0 0 0

0 50 0 0 90

98 0 0 0 0

0 0 0 99 0

0 0 90 90 99

From the specifications of the treatment units, it can be seen that the main functions of TP1, TP2 . . . TP5 are to remove contaminants C, A, E, D and B, respectively. The analysis is as follows: TP1 can remove two contaminants, A and C, and the removal ratios are 40% and 98%, respectively. Therefore, the main function of TP1 is to remove contaminant C. Similarly, the main functions of TP2, TP3 and TP4 are to remove contaminants A, E and D, respectively. For TP5, although its removal ratio for contaminant E is higher than contaminant B, there is no treatment unit whose main function is to remove contaminant B. Therefore the main function of TP5 is determined as to remove contaminant B. 2. Obtain the streams treated by each unit for its main contaminant and determine the elements of STPj. Based on the procedure in Section 3.2, the elements of STPj and the treatment amount of each unit can be obtained. For example,

388

Z.-H. Liu et al. / Chemical Engineering Journal 228 (2013) 381–391

Table 7 The streams to be treated by each unit for its main contaminant in Example 3. Treatment unit (main function)

STPj

Treatment amount FTPj,in (m3/s)

1 2 3 4 5

{S2, S3, S7} {S13, S15, S3} {S2, S4, S8, S5, S10, S11} {S4, S7, S12, S14, S5} {S15, S11, S6, S2}

45.71 26.75 206.26 92.17 155.83

(C) (A) (E) (D) (B)

Table 8 The treatment unit data in the first three-unit group selected in Example 3. Treatment unit

1 2 4

Removal ratio RRi (%) A

C

D

40 90 0

98 0 0

0 0 99

the elements of set STP2 are S13, S15 and S3. Stream S3 is the pinch stream of TP2. The treatment amount of TP2 is 26.75 m3/s. Similarly, the treatment amounts of other units can be obtained, as shown in Table 7.

Table 9 The stream data after the first three-unit group completed in Example 3. Stream

Concentration (g/m3) B

E

S9 S6 S10 S11 SU2-P SU2 SU4-P SU4 S8 S1

350 1100 270 1190 1510 574.73 421.47 152.83 20 50

230 340 580 370 220 144.72 563.5 45.35 690 70

Flowrate fi (m3/s)

3 35 23 89 0.38 12.62 9.23 123.77 2 36

SU2 is the outlet stream of TP2; SU2-P is the bypass flowrate of pinch stream. SU4 is the outlet stream of TP4. SU4-P is the bypass flowrate of pinch stream.

Table 10 The stream data after TP5 in Example 3. Stream

Concentration (g/m3) E

Flowrate fi (m3/s)

SU5 SU5-P S8 S1

3.37 45.35 690 70

199.09 96.91 2 36

3. Analyze the relevance between units and find a three-unit group to be executed first.

SU5 is the outlet stream of TP5. SU5-P is the bypass flowrate of pinch stream.

From Table 7, it can be seen that the treatment amounts of TP1, TP2 and TP4 are obviously smaller than those of the others. According to the rules in Section 3.4, TP1, TP2 and TP4 are selected as the three-unit group to be executed first.

Now, we will determine the precedence order of TP1, TP2 and TP4. Table 8 lists the removal ratios of contaminants of the units in the selected group. From table 8, the non-main function of TP1 is to remove contaminant A with removal ration of 40%. TP2 and TP4 can remove their main contaminant only and these two units are irrelevant. Whatever TP2 or TP4 is executed first, the treatment amount of the follow-up unit will not increase. However

4. Execute the units in the selected group in turn according to rules in Section 3.5.

Fig. 7. The streams after the first group executed in Example 3.

Z.-H. Liu et al. / Chemical Engineering Journal 228 (2013) 381–391

389

Fig. 8. Final design for Example 3.

Table 11 Stream and treatment unit data for Example 4. Stream

Concentration (ppm) A

(a) Stream data 1 1100 2 40 3 200 4 60 5 400 Treatment unit

Flowrate fi (t/h)

B

C

D

E

F

500 0 220 510 170

500 100 200 500 100

200 300 500 200 300

800 910 150 780 900

100 200 0 100 0

19 7 8 6 17

Removal ratio RRi (%) A

(b) Treatment unit data 1 99 2 3 4 5

B

C

D

E

F

99

90 99

99

99 99

Fig. 9. The streams after the first group executed in Example 4.

Table 12 The streams to be treated by units for their main contaminants in Example 4. Treatment unit (main contaminant)

STPj

1 2 3 4 5

{S1, {S4, {S4, {S5, {S2,

(A) (B) (C) (D) (E)

Treatment amount FTPj,in (t/h) S5} S1} S1} S3, S2} S5, S1, S4}

27.96 23.13 21.83 32 43.71

FTP2,in  FTP4,in (26.75 m3/s  92.17 m3/s), according to Section 3.5.1, TP2 will be prior to TP4. For the two relevant units TP1 and TP2, according to the procedure in Section 3.2, the total amount of two units can be obtained. Because TP1 can remove two contaminants, A and C, the prece-

dence order of these two units should be determined by the method in Section 3.5.2. For the sequence of TP1 and TP2, the total amount of two units are 58.33 m3/s (F1,2), and for the sequence of TP2 and TP1, the total amount of two units are 72.46 m3/s (F2,1). Because F1,2 < F2,1, TP1 will be prior to TP2 according to Section 3.5.2. From the above discussion, the precedence order can be TP1, TP2, and TP4. When the selected group is completed, the structure of the system is shown in Fig. 7. The stream data are shown in Table 9. 5. Execute the rest two units in turn according to rules in Section 3.5 Now, the number of the un-executed processes is 2. Therefore, it is not necessary to group any more. From Table 6b, it can be seen that TP5 can remove contaminants B and E. TP3 can remove contaminant E only. The precedence order of these two units should be determined by method in Section 3.5.2. According to the procedure in Section 3.2 the total amount of these two units can be obtained. For the sequence of TP5 and TP3, the total amount of two

390

Z.-H. Liu et al. / Chemical Engineering Journal 228 (2013) 381–391

Fig. 10. Final design for Example 4.

Table 13 The stream data after the first three-unit group completed in Example 4. Stream

Concentration (ppm)

Flowrate fi (t/h)

A

E

SU2-P SU2 SU4

1100 830.25 271.25

800 794.81 66.88

1.87 23.13 32

SU2 is the outlet stream of TP2; SU2-P is the bypass flowrate of pinch stream. SU4 is the outlet stream of TP4.

units are 199.09 m3/s (F5,3), and for the sequence of TP3 and TP5, the total amount of two units are 295.46 m3/s (F3,5). Because F5,3 < F3,5, TP5 will be prior to TP3 according to Section 3.5.2. After TP5, the current streams for TP3 are shown in Table 10, where SU5-P is the bypassing stream of TP5. From Table 10, it can be obtained that the total mass load of contaminant E in all the current streams is 28,453 g/s, which is less than the environment mass load limit of contaminant E, M lim Env ;E . Therefore, it is not necessary to execute TP3. 6. Obtain the final design The final design is shown in Fig. 8. The total treatment amount is FU1 + FU2 + FU4 + FU5 = 381.19 m3/s, which is smaller than the literature result (440 m3/s) [19]. From above results, it can be seen that the literature result (440 m3/s) might be a local optimum solution.

3. Analyze the relevance between units and find a three-unit group to be executed first. In Table 12, because the treatment amounts of TP2 and TP3 are smaller than the other units, TP2 and TP3 should be considered first. TP1 is relevant to TP2 and TP3. Therefore, TP1 will not be classified in the three-unit group to be executed first. The three-unit group to be executed first will include TP2, TP3 and TP4 according to the rules in Section 3.4. 4. Execute the units in the selected group in turn according to the rules in Section 3.5. Because FTP2 > FTP3, TP3 should be executed before TP2 according to Section 3.5.1. Both TP2 and TP3 are irrelevant to TP4, however the treatment amounts of TP2 and TP3 are less than that of TP4, the precedence order will be TP3, TP2, TP4 according to Section 3.5.1. 5. Execute the rest two units in turn according to the rules in Section 3.5. When the three-unit group is completed, the structure of the system is shown in Fig. 9, and the stream data are shown in Table 13. From Table 13, it can be obtained by the method mentioned in Section 3.2 that the treatment amount of TP5 (FTP5 = 20.92 t/h) is less than that of TP1 (FTP1 = 36.88 t/h). Therefore, TP5 should be executed before TP1 according to the minimum-mixing rule. 6. Obtain the final design.

4.4. Example 4 The data for Example 4 are shown in Table 11, taken from Castro et al. [13]. The environment limit for each contaminant is 100 ppm. The design procedure is as follows. 1. Identify the main function of each process. From the specifications of the units, it can be seen that the main functions of TP1 to TP5 are to remove contaminants A, B, C, D and E, respectively. It is not necessary to consider contaminant F, because its mass load is less than the environment mass load limit of contaminant F, Mlim Env ;F . 2. Obtain the streams treated by each unit for its main contaminant and determine the elements of STPj. In Table 12, the streams to be treated by each unit are arranged in descending concentration order. The last stream for each unit is the pinch stream.

The final design is shown in Fig. 10. The total treatment amount is FU1 + FU2 + FU3 + FU4 + FU5 = 134.75 t/h, which is 8.7% larger than the literature result (124.44 t/h) [13]. 5. Conclusions and discussions This article presents a new method to design the distributed effluent treatment systems. For a complex system, a three-unit group is selected and executed first based on the minimum-mixing rule proposed in this paper. Then the next three-unit group will be selected and executed till the number of unexecuted units is less than three. In the selected group, the precedence order of the units can be determined based on the heuristic rules proposed. Pinch method is used to calculate the minimum treatment amount of each unit for its main contaminant without considering other contaminants. It is shown that the method proposed in this work is simple and effective. The results obtained in this paper are close to the optimal results in literature. Compared to pinch methods, the approach proposed can design system with multiple contaminants. On the other hand, compared to mathematical programming

Z.-H. Liu et al. / Chemical Engineering Journal 228 (2013) 381–391

methods, the procedure proposed in this work is based on clear insights and can reduce the calculation effort. The method proposed cannot guarantee to provide optimal design for multiple contaminant systems, but the designs obtained can be used as initial structures for mathematical programming method. Acknowledgments This work is supported by the National Natural Science Foundation of China (Grant No. 21176057), Natural Science Foundation of Hebei Province, Hebei, China (Grant No. 11966725D) and State Key Laboratory of Chemical Engineering (Grant No. Open Research Project Skloche-K-2011-04). References [1] D.C.Y. Foo, State-of-the-art review of pinch analysis techniques for water network synthesis, Ind. Eng. Chem. Res. 48 (2009) 5125–5159. [2] J. Jezowski, Review of water network design methods with literature annotations, Ind. Eng. Chem. Res. 49 (2010) 4475–4516. [3] J.W.W. Eckenfelder, J. Patoczka, A.T. Watkin, Wastewater treatment, Chem. Eng. 92 (September 2) (1985) 60–74. [4] P.W. Lankford, W.W. Eckenfelder, K.D. Torrens, Reducing wastewater toxicity, Chem. Eng. 95 (November 7) (1988) 72–81. [5] Y.P. Wang, R. Smith, Design of distributed effluent treatment systems, Chem. Eng. Sci. 49 (1994) 3127–3145. [6] W.C.J. Kuo, R. Smith, Effluent treatment system design, Chem. Eng. Sci. 52 (1997) 4273–4290. [7] T.K. Zhelev, N. Bhaw, Combined water–oxygen pinch analysis for better wastewater treatment management, Waste Manage. (Oxford) 20 (2000) 665– 670.

391

[8] C.W. Hui, R. Smith, Targeting and design for minimum treatment flowrate for vent streams, Chem. Eng. Res. Des. 79 (2001) 13–24. [9] B. Galan, I.E. Grossmann, Optimal design of distributed wastewater treatment networks, Ind. Eng. Chem. Res. 37 (1998) 4036–4048. [10] R. Hernández-Suárez, J. Castellanos-Fernández, J.M. Zamora, Superstructure decomposition and parametric optimization approach for the synthesis of distributed wastewater treatment networks, Ind. Eng. Chem. Res. 43 (2004) 2175–2191. [11] R. Karuppiah, I.E. Grossmann, Global optimization for the synthesis of integrated water systems in chemical processes, Comput. Chem. Eng. 30 (2006) 650–673. [12] R. Karuppiah, I.E. Grossmann, Global optimization of multiscenario mixed integer nonlinear programming models arising in the synthesis of integrated water networks under uncertainty, Comput. Chem. Eng. 32 (2008) 145–160. [13] P.M. Castro, H.A. Matos, A.Q. Novais, An efficient heuristic procedure for the optimal design of wastewater treatment systems, Resour. Conserv. Recycl. 50 (2007) 158–185. _ [14] G. Statyukha, O. Kvitka, I. Dzhygyrey, J. Jezowski, A simple sequential approach for designing industrial wastewater treatment networks, J. Clean. Prod. 16 (2008) 215–224. [15] S. Bandyopadhyay, Targeting minimum waste treatment flow rate, Chem. Eng. J. 152 (2009) 367–375. [16] J. Shi, Z.Y. Liu, A simple method for design of distributed wastewater treatment systems with multiple contaminants, AIChE J. 57 (2011) 3226–3232. [17] Z.H. Liu, J. Shi, Z.Y. Liu, Design of wastewater treatment networks with single contaminant, Chem. Eng. J. 192 (2012) 315–325. [18] S.S.T. Soo, E.L. Toh, K.K.K. Yap, D.K.S. Ng, D.C.Y. Foo, Synthesis of distributed wastewater treatment networks for one-and two-contaminant systems, Chem. Eng. Res. Des. 91 (2013) 106–119. [19] B. Galan, I.E. Grossmann, Optimization strategies for the design and synthesis of distributed wastewater treatment networks, Comput. Chem. Eng. 23 (1999) S161–S164.