Abstract
An extended Pinch-based targeting and synthesis procedure in the domain
of multiple constraints (e.g. multi-contaminants) for material
recycle/reuse networks has been developed in this contribution. The main
steps are the identification of the ranking order of material
sinks/demands or sources/supplies. This is dependent on the limiting
contaminants of sinks. Each contaminant/constraint is assigned a Pinch
Diagram (Load vs Flowrate), and the Source and Sink Composite Curves
(CC) are plotted for each diagram. The Source CC should be shifted until
its lines could form a ’polygon’ with the Sink CC, for which the points
represent the vertices of the polygons. The sequential approach used is
to first identify a preliminary resource target and allocated sources
for each sink, then to follow certain heuristics for further reduction
of the freshwater. The proposed approach provides a minimum resource
target and a network design that achieves the targeted fresh resource
along with graphical representation.
Keywords : Process Integration; Multi-contaminant Water Pinch
Analysis; Material Recycle/Reuse; Water Network Design
Introduction
To progress toward Circular Economy and to reduce the reliance on
natural resources, the process industries have pursued material
conservation as a key approach, which is mainly based on the concept of
Process Integration [1].
El-Halwagi and Manousioutakhis [2]
used the analogy with Heat Integration and introduced the problem of
mass exchange network (MEN) for mass transfer operations in the process.
Mass transfer processes include absorption, stripping, extraction,
leaching, adsorption or ion exchange. They used the analogy to develop
the related Composite Curves (CC) and Composition Interval Table (CIT)
to determine the minimum water target, and the final network again is
using the analogy from the Heat Grid Diagram [3] for design
heuristics aided with Water Grid Diagram. Wang and Smith [4] again
extended the Heat Pinch Approach to target the minimum freshwater
consumption and wastewater discharged by the transfer of contaminants
from process streams to water streams. A simplified design procedure was
later introduced by Olesen and Polley [5], which results in better
constructions for regeneration reuse and recycles designs. Hallale
[6] introduced an extended targeting method for water minimisation.
Gomes et al. [7] presented a heuristic algorithmic procedure, water
source diagram (WSD), to synthesise water mass exchange networks.
El-Halwagi and Manousiouthakis [8] showed that it is possible to
target the minimum usage of external lean streams using again systematic
representations passed on the physical understanding such as Composite
Curves. El-Halwagi et al. [9] provided a single-stage targeting
method to identify minimum resources for a single contaminant water
network, with the solution strategy which identified through rigorous
analysis. Both methods apply to mass exchange processes, and they rely
on the basic principle of concentration driving force. For a review of
the historical development of Water Pinch Analysis see, e.g. Foo
[10]. Klemeš and Kravanja [11] provided an overview of the
development of Process Integration prior to 2013, and Klemeš et al.
[12] had more recently conducted a comprehensive overview of various
extensions of PA in Mass Integration, including water and hydrogen
integration.
However, a complicated issue is a
solution in the case of multi-contaminant water flows. Various other
strategies have been proposed to tackle this problem. Alva-Argaez et al.
[13] developed Mixed-Integer Linear Programming (MILP) based
multi-contaminant transhipment model used for targeting, particular for
mass exchange networks and wastewater minimisation problems. Their model
could tackle the general problem of mass exchange networks, but it could
not evaluate the mixing effect of multi-contaminant problems.
Gomes
et al. [7] developed an
algorithmic
and graphical procedure, which is called Water Source Diagram (WSD) for
a fixed load single contaminant system. This method is flexible and
simple to use. It aims at minimising water consumption and provides the
corresponding system network simultaneously directly. A series of works
related to the use of WSD with multiple contaminants have been
developed. An extension of the WSD for
multiple contaminant systems was presented by
Ulson
de Souza et al. [14] to reduce freshwater consumption in an
oil
refinery. As the reference contaminant, they chose the one which was
critical in the process, that is, the
one which needs more water in operation.
Karthick
et al. [15] used the WSD with
a mathematical
method and generated a hybrid procedure to deal with the minimisation
of freshwater
for wastewater
treatment. The WSD was used to provide starting points for
the mathematical
model.
Gomes
et al. [16] extended this method for maximum reuse and FL
operations. This approach requires the choice of a reference contaminant
and a reference operation to carry out the adjustments of concentrations
in the required operations. Calixto et
al. [17] developed a decomposition approach to be used with WSD,
where the objective is to prevent the need of calculations to avoid
violations in flowrate and/or concentration from the proposed network.
This method also was applied to FL systems. Francisco et al. [18]
extended this tool to fixed flowrate problem with multiple contaminants.
The main steps involve the determination of the correct reference
contaminants and operations, and later to adjust the inlet or outlet
concentrations accordingly. Calixto et al. [19] provided an overview
of the WSD and its application. Other attempts include the work of
Castaño and Higuita [20], who used the property of turbidity (which
sums a number of contaminants) in the design of water networks. The
authors regarded turbidity as the key measured parameter and linear
correlations of it were made with the concentration of the suspended
solids. Mabitla and Majozi [21] presented a hybrid of graphical and
mathematical approaches in solving multi-contaminant water and
regeneration networks. The graphical approach involves the
pre-processing steps to identify minimum water target and optimal
regenerator removal ratios.
Another concept proposed by Liu et al. [22] is the concentration
potential for a multi-contaminant problem. The concepts are presented
based on the overall allocating possibility of source streams to demand
streams. The concept is analogous to the single contaminant water
network as it identifies the concentration order of the streams. Fan et
al. [23] extended the concepts of concentration potential to the
fixed flowrate operations. Li et al. [24] provide a review on this
approach with their extension and applicability, and Zhao et al.
[25] utilised the concept in designing heat-integrated water
networks.
The Water Source Diagram (WSD) is a powerful tool and is widely used for
industrial implementations due to its graphical visualisation platform
[19]. Their algorithm to determine the resource target and network
design require the adjustment on the concentration of the contaminant.
In most of their works, they assumed the linear mass load ratio from
Wang and Smith [26]. This might not be practical in realistic mass
transfer operations. Calixto et al. [17] also explained that
additional algorithm is needed to predict contaminant violation for the
design framework. Francisco et al. [18] implemented this to the WSD
algorithm, but the overall algorithm can be complicated. The potential
concentration concept from Liu et al. [22] is another successful
implementation, but it relies on the prioritisation of the sources. Chin
et al. [27] mentioned that the prioritisation of the sources might
not be the optimal allocation strategy, but to satisfy all the
contaminant limits of the sinks as much as possible. This is applicable
for fixed flowrate operations.
Using Mathematical Programming in MEN
synthesis,
Short et al.[28] formulated a comprehensive mixed-integer nonlinear program for
MEN design. Oladosu et al. [29] presented an algebraic approach for
simultaneous targeting and design of MENS with streams splitting. The
proposed segregated Composition Interval Table (SECIT) was used to
determine MEN Pinch Point and simultaneously target and design MENS.
Yanwarizal et al. [30] adopted the Pinch based graphical tools and
plotted the Composite Curves separating the streams instead of
compositing the streams with similar concentration level. This strategy
allows the simultaneous targeting and design of the MEN with a minimum
fresh resource. However, their study was limited for single contaminant
only.
In spite of the benefits of using MP in synthesising MEN network dealing
with multi-contaminant problems, the approach is less favoured by design
engineers due to the full automation of the solution. The superstructure
formulation is demanding when the reasonable confidence in the proposed
solutions and to produce solutions away from the optima. The solution
strategy also provides restricted physical insights and analysis of the
problem, making the internal bottlenecks unknown. This is true;
especially when an authentic problem involves a lot of contaminants or
streams properties constraints. The development of PA is an essential
step for industrial application because it reveals the inherent system
limitations – targets of the resource supply and the internal
bottlenecks. These targets are then used in the detailed design model
setting the optimisation strategy and providing bounds on the key
variables. This is the logic which has led to the Process Integration
strategy to identify the target for minimum usage of fresh resources
ahead of detailed design, e.g. Heat Integration [3].
Based on the analysis, the objective of this paper is to develop a
Pinch-Based simultaneous targeting and design of minimum fresh resource
using interactive graphical representations. The specific problem of
water using systems with multiple contaminants is studied in this paper.
The proposed method involves the ranking of different resources to be
reuse/ recycling according to its composition of various contaminants as
well as assigning sinks to the proper contaminant cascades. The
graphical plots using the load vs flowrate diagram can show the Pinch
Point for different contaminants, maximum mass recovery, source
allocation and minimum external resource targets for individual sinks
simultaneously. The plot result can then be translated into source
allocation network with the optimal design without the need to perform
calculations to check for mass transfer feasibility.