A Process Graph Approach to Industrial Symbiosis

K. B. Avisoa,b, A.S.F. Chiuc, K. D. S. Yud, ,

M. A. B. Promentillaa,b L.F. Razona,b, A.T. Ubandob,e,

C. L. Syc and R. R. Tana,b

a Chemical Engineering DepartmentbCenter for Engineering and Sustainable Development Research

cIndustrial Engineering Departmentd School of Economics

e Mechanical Engineering DepartmentDe La Salle University Manila, Philippines

PRES 15

Kuching, Malaysia

August 23 – 27, 2015

Introduction

� Population growth coupled with climate

change are expected to aggravate issues on

resource scarcity

� Freshwater is a key resource for human

sustainability

� Industrial Ecology provides a systematic

framework to achieve sustainability

2

Industrial Ecology

• Popularised in 1989 by Frosch and Gallopoulos

• It utilizes an analogy between the industrial system and natural ecosystems (metabolism and symbiosis) to achieve sustainability

• Waste materials from one industry become inputs of another industry (Industrial symbiosis)

• IE is a systems approach towards sustainability

Reference: Frosch and Gallopoulos, 1989, Scientific American, 261, 94 - 102

Industrial

System

ComponentResources

Products

By-Products

Waste

Industrial

System

ComponentResources

Products

By-Products

Waste

Industrial

System

ComponentResources

Products

By-Products

Waste

Industrial Ecology

Industrial

System

Component

Industrial

System

Component

Material and

Energy

Exchange

Industrial

System

Component

Industrial System Industrial Eco-system

PRES 15

Kuching, Malaysia

August 23 – 27, 2015

Industrial Symbiosis

Kalundborg Eco-industrial Park, Denmark4

Reference: Ecodecision, Spring 1996 (20)

PRES 15

Kuching, Malaysia

August 23 – 27, 2015

Industrial Symbiosis (IS)

• The symbiotic relationships in industrial systems are encouraged by geographical proximity as in eco-industrial parks (EIP)(Ehrenfeld and Chertow, 2002)

• The exchange of common utilities such as energy and water are precursors to full-blown IS (Chertow, 2007)

• Optimization models prescribe designs to maximize benefits in IS (e.g. Lovelady and El-Halwagi, Chew and Foo, 2009)

5

PRES 15

Kuching, Malaysia

August 23 – 27, 2015

11 134 417

879

938

530

1817

934

7

1287

Process Systems Engineering (PSE)

in the Design of Water Exchange

Networks

6

1

3

2

4

FW

WW

Optimized Network

Plant A

Plant B

Plant E

SR1

SK1

Plant C

SK3

SR2

SK2

Plant D

SR3

SR4

SK4

SR5

200 t/h 1,221.38 t/h

422.53 t/h

78.62 t/h

1,000 t/h

3,500 t/h

2,501.15 t/h

512.07 t/h

1,987.93 t/h

Centralized

Regeneration

UnitCR = 500 ppm

FW

1,000 t/h

498.85 t/h

WW12.07 t/h

78.62 t/h

PRES 15

Kuching, Malaysia

August 23 – 27, 2015

Issues on Industrial Symbiosis

• IS lends itself to uncertainties in the reliability of the

exchange networks (Liao et al., 2007)

• Formerly independent units are now highly

interconnected

• Variability in process streams exist due to seasonal

variations

• Risk assessment and management strategies should be

developed to handle system variability and reliability

7

PRES 15

Kuching, Malaysia

August 23 – 27, 2015

Input-Output Modelling

S-1

S-2

S-3

Wastes and Pollutants

Fin

al O

utp

uts

Reso

urc

e I

np

uts

System Boundary

8

PRES 15

Kuching, Malaysia

August 23 – 27, 2015

Input-Output Modelling

S-

1

S-

2

S-3

Wastes and Pollutants

Fin

al

Ou

tpu

ts

Reso

urc

e

Inp

uts

System Boundary

9

Interdependencies in IS

networks can be modelled using

Input-Output Analysis

PRES 15

Kuching, Malaysia

August 23 – 27, 2015

Problem Statement

� Given n resource sources, m resource sinks

� What is the optimal resource exchange network to reduce

fresh resource consumption? Minimize annual costs?

� Given a crisis event that results in the reduction in

capacity of one plant in the network, how should the

exchanges be modified to reduce system disruption or

failure?

10

PRES 15

Kuching, Malaysia

August 23 – 27, 2015

Optimization Model

�R���

���+ F� = D�∀j

�R���

���≤ S�∀i

�R��C� + F�C��

���≤ D�Q�∀j

11

min = �F�P��

���+ AC AC – annual costs

Ci – quality of source i

Dj – resource reqt of

demand j

Fj – amount of resource

delivered to sink j

PF – freshwater cost

Qj – required quality of

demand j

Rij – flowrate or recycle

stream

Si – available source

PRES 15

Kuching, Malaysia

August 23 – 27, 2015

P-graph Model

� Process graph or p-graph is a graph theoretic method

developed for process network synthesis

� P-graph utilizes 3 algorithms to identify the optimal

network structure

� MSG – maximal structure generation

� SSG – solution structure generation

� ABB – advanced branch and bound

� P-graph is a graphical representation of

matrix calculations such as MILP

12

RM1

P1

RM2OPERATING UNIT

PRES 15

Kuching, Malaysia

August 23 – 27, 2015

P-graph model of the IS

network

� Each plant is considered as a process unit

� Material/Energy flows are modeled as raw materials,

product or by-products

� Streams are pre-qualified based on process unit

requirements

Assumptions:

� Complete substitutability of available resources

13

PRES 15

Kuching, Malaysia

August 23 – 27, 2015

Case Study

� The case study is taken from Keckler and Allen (1998)

� The reuse and treatment of water is considered between 3

industrial plants in an EIP considering the establishment

of a water treatment facility

� A scenario on capacity reduction is investigated

14

11 134 417

879

938

530

1817

934

7

1287

PRES 15

Kuching, Malaysia

August 23 – 27, 2015

Water Limiting Data

Plant Water need

(cu m/d)

Input Quality

(ppm)

Output Quality

(ppm)

(TOC, TSS, TDS) (TOC, TSS, TDS)

M 42 25, 500, 2500 1928, 2639, 7824

O 3,600 25, 25, 200 484, 105, 904

P 4,940 5, 100, 500 8, 22, 276

Fresh water n/a n/a 0,1,140

15

PRES 15

Kuching, Malaysia

August 23 – 27, 2015

Water Quality of Treatment

Processes

16

Treatment

Step

Symbol Output Quality

(TOC, TSS, TDS)

Treatment

cost ($/cu m.)

Primary and

Secondary

A 20,30, 1000 1.45

Filtration and

Precipitation

B 5, 10, 500 0.11

Reverse

Osmosis

C 5, 1, 10 1.58

Freshwater S 0, 1, 140 n/a

Hub H n/a 0.53

Maximal structure 17

Optimal network for baseline

operation18

What happens if one of the plants

experiences a disruption?

19

Optimal network for baseline

operation20

60%

capacity

Optimal adjustment if P is 40%

inoperable21

Feasible Solutions 22

Freshwater for all

Plants

Feasible Solutions 23

Fewer Treatment

Steps

PRES 15

Kuching, Malaysia

August 23 – 27, 2015

Conclusion

� 9 feasible solutions have been generated

� The solution vary in degree of recycling and water

treatment

� The solutions provide insight on potential risk

management strategies to deal with failures in IS

networks

24

PRES 15

Kuching, Malaysia

August 23 – 27, 2015

Future Work

� Integration of additional criteria for evaluating sub-

optimal solutions

� Implementation of P-graph framework in consideration of

multiple product/by-product exchanges in IS networks

� Implement without pre-qualifying streams

25

PRES 15

Kuching, Malaysia

August 23 – 27, 2015

Acknowledgment

� The authors would like to thank the Department of

Science Technology for funding this research

26

PRES 15

Kuching, Malaysia

August 23 – 27, 2015

THANK YOUFor comments and suggestions you may also contact me at:

Tel. No.: + 632 – 5244611 loc 127

Email: [email protected]

27