Extension of the water sources diagram method to systems with simultaneous fixed flowrate and fixed load processes

https://doi.org/10.1016/j.cherd.2015.10.024Get rights and content

Highlights

  • Water source diagram (WSD) extension including fixed flowrate (FF) operations.

  • Case studies show WSD versatility to yield good water system designs.

  • Application of WSD algorithm in water system problems formulated by sources and sinks.

  • Water target and water system structure obtained simultaneously.

Abstract

In this article, the water source diagram (WSD) (Gomes et al., 2007) is extended to the design of water networks involving both fixed flowrate and fixed contaminant load, as well as water loss/gain operations. The algorithm targets minimum external water consumption while simultaneously synthesizing the corresponding water system structure. In addition, it is shown that the WSD can be applied to water allocation problems (WAP) based only on water sources and sinks, maintaining its good performance. To illustrate the methodology, case studies handling hybrid water system are presented, including a zero wastewater discharge discussion and data from a Brazilian pulp mill.

Introduction

Chemical and petrochemical plants use a large quantity of water. Water scarcity, restricting environmental laws, as well as rising costs of energy and effluent treatment suggests adopting strategies of water management. In this context, reuse, recycle, regeneration with recycle, and regeneration with reuse of water have been extensively studied aiming at reducing water consumption.

Several procedures have been proposed to design the water allocation problems (WAPs). In general, these procedures can be divided into three majors groups: conceptual engineering (i.e. pinch analysis, water pinch), algorithmic, and mathematical optimization-based procedures. Comprehensive descriptions of these procedures can be found in Bagajewicz (2000), El-Halwagi, 2012, El-Halwagi, 2006, Foo, 2012, Foo, 2009, Jeżowski (2010) and Klemeš (2012). These methodologies are part of process integration, an area of process system engineering. In particular, these methodologies are aimed at systematically reducing impacts on the environment through the reduction of the consumption of resources or harmful emissions (Klemeš et al., 2013).

Water using operations in chemical processes can be divided into two groups: (1) quality controlled and (2) quantity controlled (Dhole et al., 1995, Polley and Polley, 2000, Hallale, 2002, Manan et al., 2004). Quality-controlled operations are represented by fixed load (FL) operations and the main feature is that the water using units are modeled as mass transfer process with a fixed amount of contaminant that is transferred from a process stream to water (e.g. extraction, absorption and scrubbing). The inlet and outlet stream flowrates are typically equal and hence in this kind of operations there are no water losses or gains.

Quantity-controlled operations are represented by fixed flow (FF) operations where the focus is the flowrate through the operation (e.g. cooling towers, boilers, chemical reaction with water as reagent or product), and these water using units are usually not modeled as mass transfer process

The principal characteristic for the FF operations is that water loss or gain may takes place in the operation. This kind of problems also can be characterized as water source and sinks that consumes or generates a fix quantity of water. The inlet stream are bounded by permissible upper values of concentration while the outlet stream must leave the operation at the given maximum value of concentration and are thus independent on the inlet concentrations (Fan et al., 2012, Teles et al., 2009). Aiming fresh water consumption minimization, Prakash and Shenoy (2005) stated that the outlet stream should leave the operation at the given maximum value of concentration, while the inlet stream must have the maximum specified value, in both types of problems (FL or FF).

As reported by Foo (2009) a growing emphasis to synthesize water network with FF problem was lately observed. However, as described above, a limited number of works to design systems with FF using a conceptual approach have been reported in the literature. We now focus on reviewing the work on water systems with fixed flowrate (FF) operations: originally, Wang and Smith (1995) suggested the use of splitting and local recycling of water to meet the flowrate constraints in FF problems with multiple sources of water of varying quality. To account for water loss/gain the authors neglecting changes in water flowrate and then accounted the changes in the freshwater line. Next, Dhole et al. (1995) presented a targeting methodology for WAPs with FF operations based on a graphical approach. In their graphical representation of the problem, every inlet stream is treated as a demand and every outlet stream as a source. They also suggested that stream mixing and bypassing could be proposed to reduce the fresh water consumption. Polley and Polley (2000) noted a problem in Dhole et al. (1995) method: an incorrect stream mixing option could change the composite curve and lead to an apparent target higher than the true minimum fresh water consumption. In addition, Hallale (2002) also showed that the targeting procedure of Dhole et al. (1995) does not give correct targets because it relies on one chosen mixing option and therefore they could be wrong. In the same article, Hallale (2002) suggested a graphical procedure to find the absolute targets in water systems with FF operations based on a water surplus diagram (a diagram equivalent to the source and sink composite curve). However, the plotting procedure of the water surplus diagram is iterative and turns this task in a tedious and cumbersome work of trial-and-error steps. In addition, it has limitations when generating accurate targets because of its graphical nature. In addition, the methodology cannot handle multiple water supply sources. To overcome and eliminate the iterative steps of water surplus diagram, El-Halwagi et al. (2003) proposed a rigorous targeting approach applied to FF and FL problems based on source and sink composite curves. A numerical version of source and sink composite curves was developed by Almutlaq et al. (2005), called algebraic targeting approach. This approach uses the load interval diagram (LID) (Almutlaq and El-Halwagi, 2007). Another work based on LID was published by Aly et al. (2005) who presented a systematic procedure for water minimization based on two steps. In the first step, the water target is obtained using the load problem table (LPT), which is an adapted form from the LID. The second step, the design step, uses the pinch location and some guidelines to generate the water network through a special strategy of mixing the water sources in order to satisfy the respective water demands. This approach needs the construction of a table where the cascade analysis is performed, first finding the infeasible target and lately the true target. For the network design step, it is required to make the correct link between the source and demands in each concentration interval. This approach is time consuming because it involves a trial-and-error solution to link the sources and demands.

Simultaneously, Manan et al. (2004) proposed the water cascade analysis (WCA) technique, a numerical targeting tool that can be applied to obtain the minimum freshwater and wastewater targets for both FL and FF problems with single contaminant. This procedure is a numerical version of the water surplus diagram (Hallale, 2002), but without the iterative step; it also requires the construction of two diagrams, the water cascade and the pure water surplus cascade diagrams. These two diagrams are integrated by the interval water balance table. Foo (2007) extended the WCA to handle FF problems with multiple water supply sources. The proposed extension is based on the addition of three new steps to locate the minimum consumption of pure and impure fresh water sources. Finally, Foo et al. (2006) illustrated a process involving a zero liquid discharge network in a paper mill using the WCA. Parand et al. (2013a) proposed some adjustments in WCA to allow the correct identification of infeasible targets, which are the major iterative issues of the method.

Prakash and Shenoy (2005) developed the near neighbor algorithm (NNA). This algorithm is based in the use of the nearest source streams available in the neighborhood to satisfy a specific water demand in terms of concentration. In others words, the method creates a mix source that is just above and a source that is just below the specific demand to meet the demand value for FF problems. To be applied in FL problems it is necessary to first convert it into a FF problem in terms of sources and demands. This method cannot be used in problems with multiple water supply sources and with regeneration processes. In addition, it uses a graphical approach, the material recovery pinch diagram (MRPD), to determine the minimum freshwater consumption. An extension of NNA, the enhanced NNA (Shenoy, 2012), increased the applicability of the algorithm to FL problem giving priority to local-recycle matches. Later, Agrawal and Shenoy (2006) analyzed the capability of the NNA to target the minimum freshwater consumption in FF problems for a single contaminant. They extended the composite curve concept to create the composite table algorithm (CTA) to determine the minimum fresh water consumption, which is a hybrid graphical and numerical targeting technique. Parand et al. (2013a) demonstrated the applicability of the CTA for various water network synthesis problems (e.g. FL, mixed FL and FF, multiple pinch, and threshold problems) considering reuse/recycle schemes. Nevertheless, in integrated water networks, the graphical analysis of the limiting composite curves (LCC) can be very complicated. In turn, Deng and Feng (2011) extended the method proposing the improved problem table (IPT), to target conventional and property-based water networks with multiple resources. This extension needs additional calculation to consider more than one external water source, which turns the proposed new approach somewhat complicated and/or cumbersome.

Bandyopadhyay (2006) presented a hybrid approach based on numerical and graphical techniques, which is a generalized form of the early concept of source composite curve (Bandyopadhyay et al., 2006) to reduce the waste production for a sort of applications (water management, hydrogen management and material reuse/recycle). The numerical technique is based on similar assumptions than those of the WCA, but involves only a single cascading instead of a double one. Nevertheless, in this method the cascading is made from the highest to the lowest concentrations whereas in the WCA it is in the reverse direction. The graph obtained was named source composite curve and it is plotted using the numerical result obtained first. With this plot, it is possible to predict the outlet wastewater flowrate and concentration. The source composite curve has the drawbacks of the graphical methods, with their curves being tedious and time consuming to be drawn.

Alwi and Manan (2007) presented a new procedure and a set of new heuristics, which improve the source and sink composite curves (El-Halwagi et al., 2003), in order to establish the minimum flowrate targets involving multiple water sources of utilities in FF and FL problems. Parand et al. (2013b) showed that these heuristics were not appropriate to be used above the infeasible pinch point.

Liu et al. (2007) presented a technique called modified concentration interval analysis (MCIA), which uses an equivalent process (called fictitious operation) to represent water loss and/or generation.

Alwi and Manan (2008) created a new graphical approach, called network allocation diagram (NAD), for simultaneous targeting and designing the water system for both FF and FL problems. The methodology is divided into four steps: (1) targeting; (2) allocation of sources to demands using heuristic rules; (3) design of the network using NAD, and (4) network evolution. This approach has the same drawbacks of the graphical approaches.

Foo (2009) presented a deep and ample review for WAP involving FF process for single impurity network relying on conceptual based approach. The author provides some comparisons between techniques developed for FL problems. The review covers targeting techniques for water reuse/recycle, regeneration, and wastewater treatment, along with the network design methods.

Deng et al. (2011) presented a new graphical approach to design the system and calculate the minimum fresh water consumption simultaneously. This approach is an extension of the limiting composite curve of Wang and Smith (1994), and can be applied in both types of problems (FF and FL). In problems that require uniform flowrate, the authors use local recycle to fulfill the flowrate constraint, and in problems with water loss/gain, the modified limiting water profile is used to find the specific value of minimum freshwater flowrate.

Fan et al. (2012) made an extension of the Liu et al. (2009) proposal, redefining the concept of concentration potential to consider recycling in fixed flowrate operations and to design the water system for both FF and FL cases. They divide the WAP into two groups of operations, fixed load and fixed flowrate operations. Initially, a design satisfying the needs of all fixed contaminant load operations is obtained followed by the addition of fixed flowrate operations to the design. They show that the results are close to the results obtained by mathematical programming. However, the procedure for hand calculations is somewhat long and tedious, needing a considerable effort to define which operation will be performed first and what stream will be used. When the calculation focused in a certain operation is completed, the initial procedure must be repeated before choosing another operation and stream to be performed.

Foo (2013) developed a generalized guideline for process changes for the synthesis of water network with FF operations. The author extended the plus–minus principle from heat exchanger network synthesis for targeting.

There are several authors who have studied processes with FF and/or FL operations including regeneration and targeting the minimum freshwater consumption (Bandyopadhyay et al., 2006, Ng et al., 2007a, Ng et al., 2007b, Ng et al., 2007c, Ng et al., 2008, Zhao et al., 2013, Parand et al., 2014). We do not review these efforts extensively because in this article, we do not study regeneration.

Furthermore, operational conditioning may fluctuate over time, leading to variations in actual mass load and/or water flowrate in the network operations. These fluctuations can lead to process disruptions affecting the operation effluents, which may become unacceptable for reuse in other operations, where they were previously acceptable. This feature makes the problem more complex. The presence of uncertainties in the water system analysis often makes it difficult to ensure feasible operation. More information about uncertainty in water systems can be found in, for example, Linninger et al. (2000), Koppol and Bagajewicz (2003), Karuppiah and Grossmann (2008), Feng et al. (2011) and Khor et al. (2014), using the mathematical programming approach, and Tan et al. (2007) and Zhang et al. (2009), adopting pinch analysis. Despite of the importance of this analysis, it will not be addressed in the present paper, because our scope, before analyzing the uncertainty effects, is to show how the WSD method can be extended to all kinds of water operations.

The majority of the methodologies previously described are focused on water consumption targets and therefore an additional approach to obtain the water system structure is needed. In the literature, there are some well-established design techniques for FL problems, e.g. water sources diagram (Gomes et al., 2007), water main method (Smith, 2005) and water grid diagram (Wang and Smith, 1994), and for FF problems, the source-sink mapping diagrams (El-Halwagi, 2006), NNA (Prakash and Shenoy, 2005) or heuristics rules (Aly et al., 2005). Among these set of methods, the WSD (Gomes et al., 2007) obtains both targets and system structure simultaneously in a direct simple manner for FL problems. It has similarities to an earlier method proposed by Gómez et al. (2001) and it has been extended to the case of multiple contaminants by Gomes et al. (2013) and to the analysis of hydrogen systems in petroleum refineries (Borges et al., 2012). It was also used in water systems with differentiated regeneration (Guelli Ulson de Souza et al., 2011). However, all the aforementioned applications/extensions of the WSD involve only the formulation of WAPs with fixed contaminant load (FL) operations.

In the present article, a simple procedure (an extension of the WSD) to determine the minimum fresh water consumption and simultaneously the system structure in water systems with FF and/or FL operations is presented; single or multiple water sources are also considered. The proposed extension does not resort to iterative procedures and cumbersome calculations when dealing with FF operations. It does not use recycles and can consider alternative network structures. It can also be used in processes including water loss/gain operations, and regeneration processes can be considered.

The extension proposed here can also be used in WAPs adopting the source/sink formulation. This type of representation is used in many methods described in the literature especially for FF problem, with water losses and/or gain. In a unit, an inlet stream may be treated as a demand (or sink) and an outlet stream may be considered as a source to other water using units (Dhole et al., 1995, Polley and Polley, 2000, Hallale, 2002, El-Halwagi et al., 2003, Prakash and Shenoy, 2005, Foo, 2009, Shenoy, 2011). Moreover, as the WSD procedure can generate different water system structures, its results can be easily adapted to industry constrains. In addition to the aforementioned two types of operations, one can use an alternative representation of sources and sinks. Indeed, an operation may also be divided in two parts (Dhole et al., 1995, Polley and Polley, 2000, Hallale, 2002, Prakash and Shenoy, 2005, Almutlaq et al., 2005, Bandyopadhyay, 2006, Alwi and Manan, 2007, Shenoy, 2011): (1) its inlet can be treated as a “sink” or a “demand”, and (2) its outlet can be treated as a “source” to others operations. In this context, “sink” can be defined as a stream that goes into an operation with a specific water quality requirement and “source” is a stream going out an operation carrying the contaminant in a specific concentration. These definitions allow process having multiple inlet and/or outlet streams to be easily modeled in WAPs.

The article is organized as follows: first, the use of equivalent operations in considering FF operations is discussed. Following, the extension of the WSD procedure initially proposed by Gomes et al. (2007) to include FF operations problems is presented. Finally, five case studies from the literature are used to highlight the performance of the proposed extended WSD algorithm: the first Case Study involves a process for the production of a chemical specialty in which all operations are of the FF type, two of them presenting water loss and one involving water gain. The second Case Study is an alumina plant, with water loss in all operations. The third Case Study also involves an alumina plant, but now with a regeneration operation (centralized treatment) to achieve zero water discharge (ZWD). The fourth Case Study shows the use of the WSD procedure in WAP formulations using the definition of water sources and sinks, only. The fifth Case Study involves a retrofit of a real Brazilian pulp mill.

Section snippets

Equivalent operations in FF operations

Because the WSD procedure was developed for FL operations, we will use the concept of fictitious units (Liu et al., 2007, Zheng et al., 2006), to take into account FF operations.

When there is water generation in an operation, the original operation is represented by two equivalent operations each with constant flowrate equal to the inlet flowrate of the original operation, with the corresponding inlet and outlet concentrations, and a new water source (see Fig. 1a). The flowrate of is new water

Extended WSD algorithm for a single contaminant

We now present the extended WSD algorithm for single contaminant processes (Gomes et al., 2013, Gomes et al., 2007). It consists of four steps:

  • 1.

    Introduction of equivalent operations and/or new external water sources.

  • 2.

    Generation of the WSD including all equivalent operations.

  • 3.

    Application of the WSD algorithm as originally proposed for FL operations.

  • 4.

    Use of an evolutionary step after the use of the original WSD algorithm. In this evolutionary step it is evaluated whether all the FF operation

Case studies

Case Study 1: This case study is taken form Wang and Smith (1995) and is a process for the production of a chemical specialty; it involves both FL operations (cyclone and steam system) and FF operations (reactor, filtration and cooling system). There are water losses and water gains (cooling system and reactor). Other authors (Deng et al., 2011, Jeżowski et al., 2003, Mann and Liu, 1999, Sorin and Bédard, 1999) used this example to show the performance of their methodologies. Its original

Conclusions

The extension of the WSD procedure for WAPs in single contaminant processes is presented and its application in five different Case Studies shows that this algorithm can be used successfully in several processes, involving fixed contaminant load operations (FL operations), fixed flowrate operations (FF operations), water loss/gain operations, hybrid water systems (with both types of operations), and also in water systems described only by sources and sinks. This extension involves the

Acknowledgments

Authors thanks CNPq, CAPES and Programa de Recursos Humanos da ANP – (PRH/EQ/13) for financial support.

References (66)

  • J.J. Klemeš et al.

    Recent developments in process integration

    Chem. Eng. Res. Des.

    (2013)
  • A.A. Linninger et al.

    Planning of waste reduction strategies under uncertainty

    Comput. Chem. Eng.

    (2000)
  • Y. Liu et al.

    Synthesis of water utilization system using concentration interval analysis method (I) non-mass-transfer-based operation

    Chin. J. Chem. Eng.

    (2007)
  • D.K.S. Ng et al.

    Extension of targeting procedure for “Ultimate Flowrate Targeting with Regeneration Placement” by Ng et al., Chem. Eng. Res. Des., 85 (A9): 1253–1267

    Chem. Eng. Res. Des.

    (2008)
  • D.K.S. Ng et al.

    Ultimate flowrate targeting with regeneration placement

    Chem. Eng. Res. Des.

    (2007)
  • R. Parand et al.

    Targeting water utilities for the threshold problem without waste discharge

    Chem. Eng. Res. Des.

    (2013)
  • G. Poplewski et al.

    Optimization-based method for calculating water networks with user specified characteristics

    Chem. Eng. Res. Des.

    (2010)
  • R. Prakash et al.

    Targeting and design of water networks for fixed flowrate and fixed contaminant load operations

    Chem. Eng. Sci.

    (2005)
  • U.V. Shenoy

    Enhanced nearest neighbors algorithm for design of water networks

    Chem. Eng. Sci.

    (2012)
  • U.V. Shenoy

    Unified targeting algorithm for diverse process integration problems of resource conservation networks

    Chem. Eng. Res. Des.

    (2011)
  • M. Sorin et al.

    The global pinch point in water reuse networks

    Process Saf. Environ. Prot.

    (1999)
  • R.R. Tan et al.

    Assessing the sensitivity of water networks to noisy mass loads using Monte Carlo simulation

    Comput. Chem. Eng.

    (2007)
  • J.P. Teles et al.

    MILP-based initialization strategies for the optimal design of water-using networks

    Chem. Eng. Sci.

    (2009)
  • Y.P. Wang et al.

    Wastewater minimisation

    Chem. Eng. Sci.

    (1994)
  • H.P. Zhao et al.

    Design of regeneration recycling water networks with internal water mains by using concentration potential concepts

    Chem. Eng. Sci.

    (2013)
  • Z. Zhang et al.

    Studies on resilience of water networks

    Chem. Eng. J.

    (2009)
  • P. Zheng et al.

    Water system integration of a chemical plant

    Energy Convers. Manag.

    (2006)
  • V. Agrawal et al.

    Unified conceptual approach to targeting and design of water and hydrogen networks

    AIChE J.

    (2006)
  • A.M. Almutlaq et al.

    An algebraic targeting approach to resource conservation via material recycle/reuse

    Int. J. Environ. Pollut.

    (2007)
  • A.M. Almutlaq et al.

    An algebraic approach to targeting waste discharge and impure fresh usage via material recycle/reuse networks

    Clean Technol. Environ. Policy

    (2005)
  • S.R.W. Alwi et al.

    Generic graphical technique for simultaneous targeting and design of water networks

    Ind. Eng. Chem. Res.

    (2008)
  • S.R.W. Alwi et al.

    Targeting multiple water utilities using composite curves

    Ind. Eng. Chem. Res.

    (2007)
  • S. Aly et al.

    A new systematic approach for water network design

    Clean Technol. Environ. Policy

    (2005)
  • Cited by (8)

    • Source diagram: A new approach for hybrid power systems design

      2021, Sustainable Energy Technologies and Assessments
      Citation Excerpt :

      In this context, the present paper aims to show a new approach called Hybrid Systems Sources Diagram (HSSD) to design optimal HPS with battery storage. HSSD is an extension of the heuristic-algorithmic methodology known as Sources Diagram (SD), which has been used in other fields such as water, to synthesize water networks considering single [25 26] and multiple contaminants [27], and cooling water system, to determine the minimum water flow rate in the cooling system without affecting the heat load removal from the process heat exchangers [28]. Also, SD can be applied in hydrogen management for minimum hydrogen consumption target calculation and, simultaneously, hydrogen network flowsheet synthesis [29], and carbon emissions reduction in electricity systems [30].

    • Cooling water sources diagram

      2021, Applied Thermal Engineering
    • Water sources diagram method in systems with multiple contaminants in fixed flowrate and fixed load processes

      2018, Journal of Cleaner Production
      Citation Excerpt :

      This is an important WSD feature, which allows to analyze different networks at the same time. In real systems, the choice can be oriented by the process constraints, such as: the distances among operations and the impossibility to use an effluent from a certain operation (da Silva Francisco et al., 2015). For problems described as sources/sinks, when the last interval of each sink is analyzed, it is necessary to check if its flowrate demand is reached.

    View all citing articles on Scopus
    View full text