Available online at www.sciencedirect.
com
ScienceDirect
Energy Procedia 36 (2013) 243 – 252
TerraGreen 13 International Conference 2013 - Advancements in Renewable Energy
and Clean Environment
Modeling and optimization of material/energy flow exchanges
in an eco-industrial park
Chao GUa,b,c, Sébastien LEVENEURb, Lionel ESTELb, Adnan YASSINEa,c,*
a
Université du Havre, LMAH(Laboratoire de Mathématiques Appliquées du Havre), 25 rue Philippe Lebon, 76063, le Havre,
France
b
Institut National des Sciences Appliquées de Rouen, LSPC(Laboratoire de Sécurité des Procédés Chimiques), Avenue l’Université,
76801, Saint Etienne du Rouvray, France
c
Institut Supérieur d’Etudes Logistiques(ISEL), quai Frissard, 76600, le Havre, France
Abstract
Nowadays, industrial symbiosis is a key concept of industrial ecology, which studies material and energy
exchange flows in the local industrial systems to reduce the costs, e.g., the wastes treatment cost, and to reduce the
pollution, e.g., greenhouse gas emissions. An industrial park is a set of manufacturing businesses producing different
products and by-products located at the same place (city, region, etc.). As the concept of this model encourages the
development of synergy and leverage of resource networks, to the advantage of all of the enterprises present in an
industrial park, a general mathematical model has been proposed. The aims of this general model are: to maximize
total quantity of exchanges flows, to maximize total economic benefice of an industrial park, and to reduce relative
environmental pollution, industrial waste treatment cost and delivery cost. This model can assure a win-win situation
for industries and environment. There are rigorous mathematical models for specific ecological industrial parks [1].
To the best of our knowledge, there is no currently other general mathematical model for designing and optimizing an
ecological industrial park. In addition, there is no currently complete ecological industrial park in France.
© 2013The
© 2013 Chao GU, Sébastien
Authors. Published byLEVENEUR,
Elsevier Ltd. Lionel ESTEL, Adnan YASSINE. Published by Elsevier Ltd.
Selection and/or
Selection and/or peer-review
peer-review underunder responsibility
responsibility of TerraGreen
of the TerraGreen Academy.
Academy
Keywords: Sustainable development; Modeling; Optimization; Industrial ecology; Industrial symbiosis; Environmental protection;
Pollution control; Waste treatment; Industrial development; Environmental impact.
________
* Corresponding author. Tel.:+3-023-274-4374; fax: +3-023-274-4310.
E-mail address:
[email protected]1876-6102 © 2013 The Authors. Published by Elsevier Ltd.
Selection and/or peer-review under responsibility of the TerraGreen Academy
doi:10.1016/j.egypro.2013.07.028
244 Chao Gu et al. / Energy Procedia 36 (2013) 243 – 252
1. Introduction
1.1 Sustainable development and industrial ecology
According to the Brundtland Commission [1], the definition of the term ‘sustainable development’ was
stated as “development that meets the needs of the present without compromising the ability of future
generations to meet their own needs”.
The concept of industrial ecology provides a useful systems perspective to support sustainable
development while assuring the shareholders’ value creation. We can say that, industrial ecology is a
response of industries to sustainable development.
For the first time, the notion of Industrial Ecology was developed in 1989 by Robert Frosh and
Nicholas Gallipolis [2] from General Motors Research Laboratories. Robert Frosh made a second
contribution in another paper, where he claimed that the product designer plays a pivotal role in materials
cycling [3].
Broadly speaking, the Industrial Ecology is committed to solve questions related to resource usage in
technological societies, with the purpose of adding to the part of knowledge necessary to begin evaluating
related environmental quality issues and resource availability questions [4].
1.2 Industrial symbiosis and eco-industrial park (EIP)
Industrial symbiosis is a subset of industrial ecology, which has a particular focus on material and
energy exchange [5]. Agarwal A. and Strachan P. [6] said “Industrial symbiosis can be defined as sharing
of services, utility, and by-product resources among diverse industrial actors in order to add value, reduce
costs and improve the environment.”
An eco-industrial park (EIP) is a special kind of industrial park. In an EIP, the attempt of businesses
cooperate with the others is to reduce waste and pollution, efficiently share resources (e.g., materials,
energy, water and so on). It also helps achieve sustainable development, with the aim to increase
economic gains and to improve environmental quality.
The Eco-industrial Park Handbook [7] states that "An Eco-Industrial Park is a community of
manufacturing and service businesses located together on a common property. Members seek enhanced
environmental, economic, and social performance through collaboration in managing environmental and
resource issues."
Currently, there are rigorous mathematical models for specific ecological industrial parks [1]. Best to
our knowledge, there is no currently any other general mathematical model for constructing and
optimizing an ecological industrial park.
1.3 Industrial synergy and partner selection
Since the 1980s, thanks to the studies of Auster [8], Harrigan and Newman [9], Garg and al. [10], and
Olhager and al. [11], the problem of partner selection has been widely addressed in the contexts of
strategic alliances and supply chain management.
It has been proved that it’s always better for industries to work together than to work alone, on the
condition of the good choice of partners. The use of industrial synergies reduces the negative
environmental impacts, despite a growth of the economic activities as showed in Fig.1.
Chao Gu et al. / Energy Procedia 36 (2013) 243 – 252 245
Fig.1. Illustration of the advantage of the use of industrial synergies.
In fact, the primary motivation for the choice of partners can be described by “synergy” that is
represented as in equation (1). The equation (1) can be interpreted in the sense that the alliance between
the industries leads to the more gain compared to the case that every industry works independently [12].
k
The value v(...) denotes the function of the satisfaction alliance. S is represented an alliance with a
partner k.
v(S 1 S 2 ... S n ) t ¦k 1 v(S k )
n
(1)
1.5 The concept and the objective of this mathematical optimization model
Generally speaking, a mathematical optimization model was proposed to encourage industrial
symbiosis and resource network while assuring the economic gains of all the industries present in an
industrial park. Precisely, the objective of this model was as followed: to maximize total quantity of
exchanges flows, to maximize total economic gains of an industrial park, and to reduce relative
environmental pollution, industrial waste treatment cost and delivery cost.
The concept is shown as in Fig. 2. This mathematical model takes into account the actual situation of
an industrial park and optimizes the flow exchanges. It also can study the flow exchanges related with the
industries outside the industrial park. This concept encourages recycling wastes or by-products rather than
just throw them away. This model can assure a win to win situation for industries and environment. It
means the industries can increase economic gains and improve environmental quality in the same time.
Fig.2. Industrial Park-Le Havre Region: an illustration of our concept, where the arrows represents the exchange flows. The
different colors are for the different types of the materials/energies.
246 Chao Gu et al. / Energy Procedia 36 (2013) 243 – 252
2. Methodology and assumptions
2.1 Methodology
Firstly, an analysis of the different types of the materials’/energy’s flows of every industry in a targeted
industrial park must be done. Through the analysis, we can have basic information of this industrial park.
This information allows us to figure out clearly the flow exchanges’ possibilities.
Secondly, for every industry, input and output flows should be analyzed. The relative economic gains
were formulated by several expressions. The parameters used in the expressions are also explained
clearly.
Thirdly, production and delivery costs have been considered and included into this model.
Finally, an optimization model is constructed.
2.2 Assumptions
For the hypothesis, we assume that the model is based on steady state, which means, for every industry,
the total quantity of the input flows is equal to the total quantity of the output flows. It can be considered
that there is neither the storage nor the management concerning the storage in the industrial park.
All costs are expressed in euros.
All quantities are expressed in million tons.
All distances are expressed in kilometers.
Moreover, we assume that, if we would like to transport a flow from an industry j to an industry i, this
cost related can be a FOB (Free On Board) situation or a CIF (Cost Insurance Freight) situation. It
depends on the agreement between these two industries.
Generally speaking, this delivery cost includes one part of transport cost and another part of relative
insurance cost. If it is on the FOB situation, it means that buyers are going to pay this delivery cost. If it is
on the CIF situation, it is sellers that are going to pay this delivery cost.
Delivery cost is supposed to be linearly proportional to distance and to quantity, excepting some special
materials.
3. Variables and parameters
3.1 Variables
In this model, the variables are (Iij )i
k
1...N ; j 1...N ;k 1...K . , which can take real positive values. These
variables can be represented by ψ. In addition, the subscripts i, j, k, N, and K are the integers.
The quantity of an inside exchange flow, which is of a material/energy type named k, from an industry j
to an industry i can be represented as Iij .
k
In the contrary, the quantity of an inside exchange flow, which is of a material/energy type named k,
from an industry i to an industry j is showed as I ji .
k
3.2 Tensor Matrix S
The information, based on the first analysis of the types of the materials’/energy’s flows of every
industry in a targeted industrial park, allows us to figure out clearly the flow exchanges’ possibilities.
Moreover, a tensor matrix, named S, could be used to present these possibilities mathematically.
Chao Gu et al. / Energy Procedia 36 (2013) 243 – 252 247
We assume that [13], in the industrial park, the amount of the existing industries is N and the amount of
the existing materials’/energy’s type is K. Thus, the definition of matrix S is the following:
[S ] (sijk )i 1...N ; j 1...N ;k 1...K . , where sijk is the exchange’s possibility of a flow inside the industrial park.
The flow sijk is of a material/energy type named k, from the industry j to the industry i. In the contrary,
sijk has the same material/energy as sijk , but from the industry i to the industry j. As a result, the dimension
of the matrix S is K u N .
2
sijk is a binary parameter. It takes the value of either 1 or 0.
IiEki is considered as a flow that goes into an industry i of a material/energy type named ki .
I jSkj is considered as a flow that goes out off an industry j of a material/energy type named kj .
We will compare these two material/energy types, ki and kj . If they are the same, then sijk is equal to
1, where ki k . It means that there is an exchange possibility from an industry j to an industry i.
k
Otherwise, s is equal to 0, where ki k .
ij
ª § s11
1
· § s11 j · § s11N · º
« ¨ ¸ ¨ ¸ ¨ ¸»
« ¨ ¸ ¨ ¸ ¨ ¸»
«¨ sK ¸ ¨ sK ¸ ¨ sK ¸»
« © 11 ¹ © 1j ¹ © 1N ¹ »
« 1 »
« § si1 · § sij1 · § siN ·»
1
>S @ «¨ ¸ ¨ ¸ ¨ ¸»
«¨ ¸ ¨ ¸ ¨ ¸»
« ¨© siK1 ¸¹ ¨ sK ¸
© ij ¹
¨ sK ¸ »
© iN ¹
« »
« 1 »
«§¨ s N 1 ·¸ § s1Nj ·
¨ ¸
§ s1NN ·»
¨ ¸
«¨ ¸ ¨ ¸ ¨ ¸»
«¨ K ¸ ¨ sK ¸ ¨ s K ¸»
«¬© s N 1 ¹ © Nj ¹ © NN ¹»¼
Fig.3. The abstract form of a tensor matrix S
3.3 Input and output flows’ parameters
The values of input and output flows are known for any industry. If an exchange flow is just regarding
the industries inside the industrial park, the flow is characterized like an inside flow. If an exchange flow
involves any industry outside the industrial park, the flow is characterized like an outside flow.
3.3.1 Input flow parameters
The economical index of an inside flow, which is of a material/energy type named k, from an industry j
to an industry I is I ijk . It can take the integers among {-1, 0, 1}.
248 Chao Gu et al. / Energy Procedia 36 (2013) 243 – 252
k
The internal unit cost of a material/energy type named k for an industry i is C EIi .
The quantity of a material/energy type named k, which is needed by an industry i for the manufacturing
is IEik .
k
The external unit cost of a material/energy type named k for an industry i is C EXi .
The economical index of an outside flow, which is of a material/energy type named k for an industry i
k
is I Ei . It can take the integers among {-1, 0, 1}.
3.3.2 Output flow parameters
Similarly, we have used the parameters below for the analysis concerning the output flows.
The economical index of an inside flow, which is of a material/energy type named k, from an industry i
to an industry j is I kji . It can take the integers among {-1, 0, 1}.
k
The internal unit cost of a material/energy type named k for an industry i is C SIi .
The quantity of a material/energy type named k, which is needed by an industry i for the manufacturing
is ISik .
k
The external unit cost of a material/energy type named k for an industry i is CSXi .
The economical index of an outside flow, which is of a material/energy type named k for an industry I
k
is I Si . It can take the integers among {-1, 0, 1}.
3.4 Parameters regarding production and delivery
Parameter C pi has been used for representing the production cost of an industry i. For transporting a
flow from i to j, which is of a material type named k, a delivery cost must be formulated.
In order to express this formula related, here are the related parameters.
The transport distances between the industry i and the industry j is rij .
k
The unit transport cost of a certain material/energy type named k is Cunit . The unit is euro per
kilometer (€/km).
The index of the total delivery cost for a flow to transport between two industries i and j, which is of a
k
material’s/energy’s type named k is I delivery ji . It can take the integer values 0 or 1.
4. Expressions
To calculate the total income of an industry i inside the industrial park studied, several parts have to be
determined as followed. If sum is positive, it means earning for an industry i. If it is negative, it means
spending for an industry i.
x Sum E: total sum of the incomes and the expenditures related with the input flows.
x Sum S: total sum of the incomes and the expenditures related with the output flows.
x The production cost of the industry i.
x The delivery cost that the industry i has to be take in charge.
The part of the sum related with the input and the output flows can be illustrated in the figure 4(a).
Chao Gu et al. / Energy Procedia 36 (2013) 243 – 252 249
And, the figure 4 illustrated the production and the delivery cost of an industry i.
Fig.4. (a) illustration of the sums related with the input flows and the output flows, as well as the production cost of the industry I;
(b) illustration of the delivery cost for a flow to transport from an industry i to an industry j.
4.1 Expressions regarding the sums E/S
The sum regarding the input flows of an industry i, RiE ǡ are a function of the variables
(Iijk )i 1...N ; j 1...N ;k 1...K . , and it is showed as in (2). It’s a sum of incomes and expenditures regarding
each type of materials or energies at the entrance. For each type of materials or energies, the sum includes
two parts: sum related with the inside flows and sum related with outsides flows.
¦ 1 ¦ j 1 ij ij ij EIi
s kI k I k C k ) (IEik ¦ j 1 sijkIijk ) u (CEXi
K N N
RiE (Iijk ) k
[( k
I Eik )] (2)
Similarly, the sum regarding the output flows of an industry i, RiS , are also a function of the
variables (I ) k
ij i 1...N ; j 1...N ;k 1...K . , and it can be represented as in (3). It’s a sum of incomes and
expenditures regarding each type of materials or energies at the exits.
¦ [(¦ j 1 s kjiI jik I kji CSIi ) (ISik ¦ j 1 s kjiI jik ) u (CSXi
K N N
RiS (I jik ) k 1
k k
I Sik )] (3)
4.2 Production costs
It should be noted that the production cost includes all the necessary costs for the manufacturing in
except of the primary resources’ costs. For example, the production cost includes the maintenance cost,
the labor cost, and etc. The production cost for an industry i can be represented as C pi .
4.3 Delivery cost expressions
4.3.1 General case
k
The notion CTji is considered as the actual total delivery cost of an exchange flow from an industry i to
an industry j, which is of the material/energy type named k, as in (4). This delivery cost includes transport
cost and insurance cost. It can be formulated as a following function of the variables
(Iijk )i 1...N ; j 1...N ;k 1...K . .
If a flow is on FOB situation, then it is industry j that should pay this delivery cost. If a flow is on CIF
situation, then it is industry i that should pay this delivery cost.
250 Chao Gu et al. / Energy Procedia 36 (2013) 243 – 252
k
CTji (I jik ) rji s kjiI jik (Cunit
k
transport Cunitinsurance)
k
(4)
k
The notion CTi has been considered as all the transport costs that an industry i need to take in charge,
k
as in (5). I delivery ji is equal to 1 while it is on CIF situation. If else, it is equal to 0.
¦ ¦
N K
CTik (I jik ) j 1
k
I
k 1 delivery ji
k
CTji (I jik ) (5)
4.3.2 Special case
The transport cost is no more linear for some special materials and for some special transport mode. We
can take a flow CO2 as an example. There are many ways to transport CO2: by pipeline, by ship, by train
and by truck. In the year 2010, Duke University has published an article [14] concerning how to
generalize the transport cost of transporting CO2 by using pipelines. In this article, a formula was given
under some assumptions. We can integrate and reformulate this formula into this model as (6).
bco 2 pipeline
Cavg co 2 pipeline exp( aco 2 pipeline cco 2 pipeline u ln(I jik )) (6)
I jik
Where ݇ ൌ ܱܥଶ and aco 2 pipeline , bco 2 pipeline , cco 2 pipeline are positive real numbers.
We can have an expression for calculating the total transport cost of a flow CO2 from an industry i to an
industry j by pipeline as (7), where݇ ൌ ܱܥଶ .
k
CTji (I jik ) rji s kjiCavgco 2 pipeline (7)
k
The rest part, C is the same as that in the general case mentioned before.
Ti
4.4 Final expression for the total income of an industry
So, finally, we can say that, Ri , the total income of an industry i, has the formula as in (8). It is a
function of the variable ψ, with \ (Iijk )i 1...N ; j 1...N ;k 1...K .
Ri (\ ) RiE (Iijk ) RiS (I jik ) C pi CTi (Iijk ) (8)
5. Optimization model
It was mentioned before that this model aims to maximize the total economic benefit of the industrial
park studied and to maximize the total quantity of the exchange flows. Thus, the objective function of our
optimization program is as in (9) and as in (10).
Furthermore, the economic benefit of every industry inside this industrial park can be simultaneous
assured. As a result, the main constraints related are as in (11), which means the economic benefit of
every industry should be positive. The upper and lower bounds’ constraints are also included for each
variable, which are not shown in this work because of space.
Chao Gu et al. / Energy Procedia 36 (2013) 243 – 252 251
¦ R (\ )
N
f1 i 1 i (9)
¦ ¦ ¦
N N K
f2 i 1 j 1
I k
k 1 ij
(10)
Ri (\ ) ! 0, fori 1...N . (11)
So, finally, we can have an optimization program, as in (12).
max F (\ ) ( f1 (\ ), f 2 (\ ))
°s.c.
°
( P) ®
R (\ ) ! 0, fori 1...N . (12)
° i
°\ (Iijk )i 1...N , j 1...N ,k 1...K .
¯
6. Model applications
This mathematical optimization model can be used for simulating the exchange materials’/energies’
flows of an industrial park to have a maximal circulation of the resources.
Furthermore, the carbon tax is an environmental tax on emissions of carbon dioxide. It intended to
limit the greenhouse gas in order to control global warming. In France, the question whether the carbon
tax should be applied is under discussion. If the carbon tax is going to be applied in the future, our model
can be usable by changing the economical index of the flows CO2. Moreover, if the technology CCR
(Carbon Capture and Recycling) or CCS (Carbon Capture and Storage) will be widely used in the
industries, this model can also be applicable.
7. Conclusion and perspectives
7.1 Conclusion
An essential mathematical optimization model to construct an ecological industrial park has been made.
Through this work, the actual situation of the industrial park can be known. And, the potential exchange
possibilities, which are among these industries existing inside the industrial park, can be figured out
rapidly. In addition, this model can also point out the potential exchange possibilities with the industries
outside the industrial park. To conclude, this model can maximize the exchanges among the industries
towards a maximal recycling of the circulation. Furthermore, this work is still going to be applicable
when the carbon tax appears. If the technology CCR or CCS is widely used in the near future, this work
will also be applicable.
Using this model, the manufacturing of every industry inside the industrial park can be controlled. The
amount of every type of material/energy to be exchanged, the total amount of all the materials/energies to
be exchanged and the amount of the wastes that cannot be recycled can be calculated easily. Also, finally,
the amount of pollution and the total economic benefit can be figured out.
7.2 Discussion and perspectives
This model is a steady model and it might not be able to reflex actual situation of an industry park very
well.
252 Chao Gu et al. / Energy Procedia 36 (2013) 243 – 252
So, several aspects will be developed this optimization model in the near future in several aspects as
following.
x The concept of the model will be modified to make it usable for dynamic case.
x More expressions of the transport cost will be enhanced to have more choices. The aim is to be usable
for special cases.
x More simulations will be done in order to test the limit of this model.
x A simulation of our model will be applied on the region Le Havre.
Acknowledgement
This study was financed by the “CHAIRE CTSC” (http://www.chaire-ctsc.fr/).
References
[1] Kathleen BA et al. Bi-level fuzzy optimization approach for water exchange in eco-industrial parks.
Process Safety and Environmental Protection 2010; 88: 31–40.
[2] United Nations. Report of the world commission on environment and development , 1987.
[3] Frosch RA, Gallopolous NE. Strategies for manufacturing. Sci Am 1989; 261(3):144–152.
[4] Frosch RA. Industrial ecology: minimizing the impact of industrial waste. Phys Today 1994;
47(11):63–68.
[5] Harper EM, Graedel TE. Industrial ecology: a teenager’s progress. Technology in Society 2004;
26:433-445.
[6] Agarwal A, Strachan P. Is Industrial Symbiosis only a Concept for Developed Countries? The
Journal for Waste & Resource Management Professionals, 2008; 42
[7] Lowe Ernest A. Eco-industrial park handbook for Asian developing countries, 2001
[8] Auster E. Macro and strategic perspectives on interorganizational linkages: A comparative
analysisand review with suggestions for reorientation. Advances in Strategic Management 1994; 10(1):3-
40.
[9] Harrigan KR, Newman WH. Bases of interorganizational cooperation: Propensity, power,
persistence. Journal of Management Studies 1990; 27(4):417-434.
[10] Garg D, Narahari Y, Viswanadham N. Achieving sharp deliveries in supply chains through
variance pool allocation. European Journal of Operational Research 2006; 171(1):227-254.
[11] Olhager J, Selldin E. Supply chain management survey of Swedish manufacturing firms.
International Journal of Production Economics 2004; 89(3): 353-361.
[12] Huang JJ et al. A multi-objective programming model for partner selection-perspectives of
objective synergies and resource allocations. Expert Systems with Applications 2010 ; 37: 3530–3536.
[13] Mamoune A, Yassine A. Creating an inductive model of industrial development with optimized
flows for reducing its environmental impacts. Energy Procedia 2011; 6: 396-403.
[14] Munish KC, Lincoln FP, Eric Williams. Potential economies of scale in CO2 transport through use
of a trunk pipeline. Energy Conversion and Management 2010; 51(12): 2825-2834.
