- Email: [email protected]
Optimal model development of energy hub to supply water, heating and electrical demands of a cement factory
Accepted Manuscript Optimal model development of energy hub to supply water, heating and electrical demands of a cement factory Mostafa Mostafavi Sani, Alireza Noorpoor, Majid Shafie-Pour Motlagh PII:
S0360-5442(19)30447-5
DOI:
https://doi.org/10.1016/j.energy.2019.03.043
Reference:
EGY 14879
To appear in:
Energy
Received Date: 11 July 2018 Revised Date:
5 March 2019
Accepted Date: 7 March 2019
Please cite this article as: Sani MM, Noorpoor A, Shafie-Pour Motlagh M, Optimal model development of energy hub to supply water, heating and electrical demands of a cement factory, Energy (2019), doi: https://doi.org/10.1016/j.energy.2019.03.043. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Optimal model development of energy hub to supply water, heating and electrical demands of a cement factory
1
RI PT
Mostafa Mostafavi Sani1, Alireza Noorpoor1*, Majid Shafie-Pour Motlagh1 School of Environment, College of Engineering, University of Tehran, Tehran, Iran
Abstract
This paper uses the energy hub concept to meet the electricity, heat and water demand of a
SC
cement plant. Given the cement plant’s high potential for waste heat recovery so the use of recycled waste heat was compared to a state where such waste heat was not used by taking
M AN U
exergy, economic and environmental analyzes into account. The proposed energy hub consists of electricity, heat and desalinated water production equipment. Energy hub is optimized based on the Genetic Algorithm (GA) with the goal of supplying demand, as well as reducing costs and pollutants and increasing the exergy efficiency which ultimately will be selected using the concept of an energy hub at its optimal capacity. By comparing the two energy supplying
TE D
systems of the current case study and optimal energy hub, results showed that the Total Annual Cost (TAC) level decreased by about 37,000 $/year and exergy efficiency increased by 36%. CO2 emission will also decrease by about 21,620 tons/year. Moreover, in the energy hub concept, the waste heat recovery and ignoring waste heat scenarios were compared. Results
EP
indicate that upon optimization, the average of TAC when waste heat is recovered, is 27% of the condition when waste heat is ignored.
AC C
Keywords: Energy hub; Multi-Effect Desalination (MED); Reverse Osmosis (RO); Total Annual Cost (TAC); Genetic Algorithm.
*
Corresponding author: Alireza Noorpoor, Associate Professor, School of Environment, College of Engineering, University of Tehran, Tel: +98-021-66476132 E-mail address: [email protected]
ACCEPTED MANUSCRIPT
Nomenclature List of acronyms
v
Dispatch factors (-)
Combined-Cycle Power Plant
Vw
Permeate velocity [m/s]
CHP
Combined Heat and Power
wp
Pump Power (kW)
CRF
Capital Recovery Factor
Z&
Capital cost rate ($/year)
CTGs
Combustion Turbine Generators
Zk
Component purchase cost ($)
EW
Energy consumption per Water production in RO
ZO&M
Operation and Maintenance Cost
GOR
Gained Output Ratio
Subscripts
GT
Gas Turbine
amb
LHV
Lower Heat Value (kJ/kg)
b
MED
Multi-Effect Desalination
ORC
Organic Rankine Cycle
PAR
Peak to Average Ration
SC Ambient brine
M AN U
NSGA
RI PT
CCPP
boi
Boiler
bp
Brine Pump
CC
Combustion Chamber
ch
Chemical
CV
Control Volume
Non dominated elitist and rapid Sorting Genetic
Algorithm Natural Gas
MSF
Multi-stage flash distillation
D
Distraction
SDI
Silt Density Index
dmn
Demand
TE D
NG
Total Annual Cost ($/year)
e
Exergy
WHO
World Health Organization
em
emission
extra
Waste heat from cement plant
Water permeability coefficient [kg/m2·s·Pa]
f
Feed
Solute transport coefficient [kg/m2·s]
h
Heat
Coupling matrix
hot
Hot water
Salt concentration [ppm]
hpp
High Pressure Pump
Cw
Brine concentration [kg/m3]
i
inlet
CC
RO capital cost subsystems
m
Material
d
Feed channel equivalent diameter [m]
max Boi Maximum Boiler
A B Cij C
AC C
List of symbols
EP
TAC
ACCEPTED MANUSCRIPT
ds
Distance from the ideal point
NG
Natural Gas
Ds
Solute diffusivity [m2/s]
p
Permeate stream
e, ex
Exergy (kJ/kg)
pu
Ex
Exergy flow (kW)
Sc
EWRO
Electrical consumption per water production
pipeline Pipe line
Fij
Non-dimensionalize objective indexes
Ph
Physical
Q
Heat Sell
Material correction factor
S
Jw
Local permeate flux [kg/m2·s]
ST
Js
Local solute flux [kg/m2·s]
SW
K
Mass transfer coefficient [m/s]
L
Output loads of the hub
LDe
Output electrical loads of the hub (kW)
LDh
Output heat loads of the hub (kW)
LDw
Steam Turbine Sea Water
M AN U
f
RI PT
Interest rate
Schmidt number
SC
int
Pump
Sea Water Intake and Pre-treatment
T
Turbine
trf
Transfer
u
Number of exergy carrier
Output water loads of the hub (kW)
x
objective’s index
N
Operation hours in a year
w
Membrane wall
mf
Feed water flow rate (kg/s)
z
index for each Pareto front point
mhot
Hot water flow rate (L/s)
0
P
Input carriers of the hub
Greek symbols
PE
Exchanging electricity of the hub (kW)
α,β,ω,ε
Type of energy carrier
Exchanging heat of the hub (kW)
η
Efficiency
Exchanging water of the hub (kW)
ρ
Density (kg/ m3), Integer Numbers
Input vector of the hub
ξ
Chemical exergy/energy ratio
Output vector of the hub
ε
Total cycle exergy efficiency (-)
Q
Heat flow (kW)
τ
Hours of operation per year (h)
QB
Brine flow rate (m3/h)
φf
Fuel cost ($/kg)
QP
permeate flow rate (m3/h)
φe,s
Selling electricity cost ($/kWh)
PW Pi Pj
EP
AC C
PH
TE D
SWIP
Standard Situation
ACCEPTED MANUSCRIPT
Recovery (%)
ρp
Density of the permeate [kg/m3]
Re
Reynolds number
π
Osmotic pressure [MPa]
Membrane area (m2)
φe,b
Buying electricity cost ($/kWh)
Thot
Hot water temperatures (˚C)
ψem
Pollutant emission cost ($/kg)
RI PT
RRO
1. Introduction
The point of confrontation energy carriers, conversion equipment, energy storage and energy
SC
demand is known as the energy hub. The energy hub is a new and promising concept to
M AN U
optimally manage multiple energy carrier systems. The energy hub has significant potential for realizing energy system models and moves towards sustainable multi-energy systems [1]. Energy hub utilization will lead to 1) reduction in energy consumption 2) simplicity and high accuracy in multiple system carrier analysis and assessment 3) increased reliability 4) multi-dimensional analysis of energy systems (technical, economic, environmental). The main element in defining
TE D
energy networks is the energy hub. Numerous researches have been conducted on the hub. A residential energy hub model was proposed by Brahman et al. [2] that obtains solar radiation, natural gas (NG) and electricity as inputs to provide cooling, heating and electricity at its output
EP
port. An optimization issue was dealt with regarding electrical and thermal energy management
AC C
with the aim of reducing total energy costs of customer’s input in terms of NOx, CO2 and SOx emissions. Mohammadi et al. [3] addressed the advantages of integrating options such as smart technologies, storage systems, multi-generation systems, renewable energy resources, distributed energy resources and demand side management by proposing the smart energy hubs concept. In order to utilize zero emissions renewable sources into dispatchable combined cycle power plants, Bahrami et al. [4] fabricated a real time energy hub scheduling issue within a dynamic pricing market. The interaction of the energy hub is modelled in the form of a precise potential game for
ACCEPTED MANUSCRIPT
energy hub payment optimization purposes concerning gas and electricity utilities. The results of the simulations indicate that the suggested algorithm may enhance the average payoff of the energy hub by 18.8%. Moreover, companies that provide energy services may enhance energy
RI PT
network technical performances through a reduction of the peak to average ration i.e. PAR for gas and electricity by 7% and 27%, correspondingly. AIRafea et al. [5] proposed a modern energy hub. For this energy hub, hydrogen is introduced as an energy vector to replace coal fired
SC
Nanticoke power plant. With the addition of hydrogen into natural gases as a fuel for combined cycle power plants, additional costs and annual revenue reduction will be the consequences. CO2
M AN U
and NOx emission credits will also increase. In order to start increasing annual revenue, electricity should be sold at 0.12 $/kWh with hydrogen at 3.7% in fuel of combined-cycle power plant (CCPP). To maximize the employment of waste heat from coal fired thermal plants for the benefit of factories in the vicinity, Togawa et al. [6] proposed a stimulation process model that
TE D
integrates land use control, technology systems and spatial analysis. The acquired results showed that waste heat systems exhibited several environmental advantages compared to the individual boiler system. By employing waste heat, the consumption of fuel oil is reduced by 16.05 TJ/year
EP
and there is a CO2 emissions reduction of 1204 tons per year. Another solution to reduce energy consumption and environmental consequences is to use waste
AC C
heat from factories and to implement a comprehensive integrated system in remote areas suffering from water shortages. This paper addresses waste heat from a cement plant. The process of producing cement is extremely energy intensive with approximately 3-4 GJ energy consumption for each ton of produced cement. Furthermore, Iran is the fourth largest cement manufacturer in the world after China, India and the USA with an annual production of 75 million tons [7]. 40% to 60% of cement production costs are accredited to energy consumption
ACCEPTED MANUSCRIPT
[8]. 3% of energy production is accredited to the cement industry in Iran [9]. Within a cement factory, electrical energy is 25% of the total energy consumption while heat energy accounts for approximately 75% of total energy consumption [10]. In the case of integrated systems and the
RI PT
use of waste heat, the work of Salimi and Amidpour [11] can be referred to, which, based on the R-curve, examined the integration of different water desalination systems. Mokhtari et al. [12] proposed the gas turbine (GT) +MED+RO water desalination method based on exergy and
SC
thermo-economics analysis for the Persian Gulf region. The system was designed on the basis of the simultaneous generation of electricity and water and could reduce total water production
M AN U
costs. Muhammad Wakil Shahzad et al. [13] proposed a combination of absorption and RO systems based on low temperature industrial heat sources or heat from the sun to address lack of water in the Persian Gulf region, which resulted in 81% efficiency and energy consumption of 1.76 kWh/m3. Arash Nemati et al. [14] utilized waste heat from a large diesel ship engine to
TE D
produce desalinated seawater. On the basis of the radiation curve amidst optimization, the obtained efficiency was 37% and the cost of exergy destruction was 60 $/GJ. Reynolds et al. [15] suggested a district energy management optimization strategy for the optimization of district heat
EP
generated from a multi vector energy center. The outcomes show the possible gain when energy is managed in a holistic state, taking both supply and demand into account as well as multiple
AC C
energy vectors. Moeini-Aghtaie et al. [16] proposed an online economic dispatch for multiple energy carriers to achieve a probabilistic economic dispatch optimization model. This model is addressed using a stable optimization mechanism i.e. multi-agent genetic algorithm with an advantageous property of finding the global optima of the economic dispatch issue. Ko et al. [17] recommended a method for the optimization of an energy system whilst concurrently reducing life cycle cost and increasing renewable energy penetration as well as reducing annual
ACCEPTED MANUSCRIPT
greenhouse gas discharge. A non-dominated elitist and rapid Sorting Genetic (NSGA) is used for the purpose of multi objective optimization. The Pareto-optimal solution evolution based on the variations in the technical, economic and environmental objective functions via generations is
RI PT
examined. Kampouropoulose [18] suggested a method and associated software tools to derive the optimal energy flow pertaining to a multi-carrier system to meet energy needs and reduce a number of defined optimization criterion on the basis of the Genetic Algorithm.
SC
Given that fifty-nine percent of drinking water is supplied from the sea through desalination processes [19], various methods have been developed for the production of freshwater from the
M AN U
sea across the world. Energy expenditures (thermal, electricity) in water desalination systems are as follows: RO
TE D
this capacity [21, 22]. The energy used for RO system is dependent on the salinity of the water as well as the system used for energy recovery. For seawater, the energy consumption with and without the recycling device is 12 and 4 kWh/m3, respectively. However, with the development
EP
and improved performance of RO systems, the price of desalinated water has decreased from 2 $/m3 to roughly 0.8 $/m3[19, 23]. However, the MED thermal water desalination system is still
AC C
developing and has retained its position across the world. MED system has a heating energy consumption of 71.7 kWh and electricity consumption of 2.3 kWh [24]. Considering the high energy consumption of water desalination systems as well as rising fresh water demand (from 77.4 in 2014 to 92 million in 2018), this technology has become a major source of CO2 emissions [24]. In this regard, CO2 production is estimated at 100 million tons per year [19, 24]. In other words, for each cubic meter of fresh water, about 3 kg of greenhouse gas is produced
ACCEPTED MANUSCRIPT
[25, 26]. The increasing trend in 2040 will bring greenhouse gas emissions to 218 tons per year and an approximate temperature increase of 2 °C [27, 28]. Due to the high energy consumption of RO and MED, in order to 1) reduce costs 2) reduce greenhouse gas production 3) increase
RI PT
productivity and 4) reduce gas and oil consumption in energy rich areas (such as the Persian Gulf), waste heat from factories can be used to produce desalinated water and meet regional demands (power generation and water heating). The existence of rich gas and oil resources has
SC
led to industrial development in this region. Also, population dispersion in the Persian Gulf due to weather conditions as well as the existence of various islands in the area has become an issue
M AN U
for governments to meet the existing energy demand of residents. In regard to the water production along with other energy requirements is an energy-intensive process, it is logical to define an energy hub.
In the case study, the consumption requirements of various energy carriers in the cement factory
TE D
and in the nearby city should be satisfied. Thus, energy production and consumption management as well as the selection of equipment with optimal capacity are considered. In addition, the priority in energy systems is to maximize efficiency. In this research, the optimal
EP
utilization of waste heat in the cement factory which increases efficiency and reduces costs and emissions is taken into account. However, due to the lack of water desalination systems in
AC C
the current state, drinking water is transported to the place of consumption through tankers, which, in addition to production and transportation costs, entails accessibility and sanitation issues. In this study, by constructing desalination systems, electricity and heat generation at the place of consumption has not only significantly contributed to the supply of energy carriers, but also improved energy efficiency whilst reducing costs and emissions through optimal operation. In regard to the aforementioned points and the necessity to supply multiple energy carriers
ACCEPTED MANUSCRIPT
including electricity, heat and desalinated water, the case study is modelled using the concept of energy hub. In order to supply energy carriers with minimum costs and emissions along with maximum exergy efficiency, a methodology is presented in Fig. 1. According to this figure,
RI PT
the decision making variables and inputs amidst the modelling section are entered to the system
energy carriers in each of the energy hub equipment.
Weather condition
Inputs
Dispatch factors
Decision Variables
RRO
M AN U
Text
Design Hub
Demands
System of equations (Table 2)
Text
Mhot
EWRO QBoiler
Design factors
TE D
Thot
CHP Model
Outputs
PeCHP, PeST, PeRO QCHP, QST, Qb mRO, mMED Text
Genetic Algorithm
SC
of equation to determine production and consumption variables for electricity, heat, and water
PE, PSW, PH Total investment cost
O&M Cost
Objective function
Capacities
TAC
Boiler Model
Objective function
AC C
EP
Exergy efficiency
MED Model RO Model ST Model CO2 Emission Number of Equipment
Fig. 1. Description of modelling methodology
ACCEPTED MANUSCRIPT
Upon modelling each of the equipment and determining energy production and consumption,
RI PT
quantity and capacity of equipment, operation and maintenance costs, and quantity of produced emissions, the TAC can be created. Moreover, exergy efficiency may also be extracted. Non dominated elitist and rapid Sorting Genetic Algorithm (NSGA-II) was implemented to present a
SC
series of Pareto multiple optimal solutions. Changes in optimal value sensitivity analysis, exergy efficiency and TAC as well as changes in design factors were assessed. Finally, upon integrated
M AN U
optimization of the energy hub with the purpose of reducing costs and increasing exergy efficiency, the optimal capacity of the equipment and assigned energy carriers are determined. By assessing various papers on energy hubs and separating the subject areas of these papers (Table 1), some of the research novelties are mentioned below:
Coupling matrix parameters such as efficiencies, recoveries and gained output ratio
TE D
•
(GOR)are included variables by considering design parameters; •
Water flow is defined as an energy carrier in the energy hub, such that two water
EP
desalination systems consisting of MED and RO are simultaneously utilized to provide the means for optimal electricity and heat allocation to produce the required desalinated
•
AC C
water upon optimization;
Considering multi-objective optimization in terms of exergy efficiency and total annual
cost in an energy hub system with self-consumption.
ACCEPTED MANUSCRIPT
Table 1. review of various papers on energy hubs* Coupling matrix arrays
SC M AN U TE D EP
AC C
*
Industrial
2018
Residential
2018
Applications Optimization
Energy
2018
Economical
2018
Exergy
2018
Environmental
2017
Reliability
2016
Constant
2015
Variable
2015
Natural gas
2015
Hydrogen
2015
Water
2014
Cooling
2014
M. Geidl et al. [41] R. Evins et al. [29] A. Sharif et al. [30] M. Rastegar et al. [31] A. Maroufmashat et al. [32] A. Maroufmashat et al. [33] F. Brahman et al. [2] A. Najafi et al. [34] M. Majidi et al. [35] V. Davatgaran et al. [36] S. Amiri et al. [37] J. Teodoro et al. [38] R. Ghaffarpour et al. [39] M. Aghamohamadi et al. [40]
Heating
2007
Authors
Power
Year
Analyses
RI PT
Energy hub carriers
Objective functions
Cost & Emission Emission Cost Cost Cost Cost & Emission Cost & Emission Cost Cost & Emission Cost Cost & Reliability
In the table above, the filled in boxes represent the studies conducted in that field. As evident, variable arrays for the coupling matrix are not considered in any previous studies.
Cost Cost Cost
ACCEPTED MANUSCRIPT
2. Thermal modeling 2.1.Energy Hub In order to achieve an integrated system, a modeling and analysis framework for multi-carrier
RI PT
energy systems must be developed. The current model is between converter devices based on the efficiency of the output-to-input ratio. When there are numerous outputs, a conversion matrix or coupling matrix is defined which represents the transformation of energy carriers from the input
SC
to the hub outputs. If necessary, storage devices are used that take time and energy into account [41]. The energy conversion between different energy hub ports is typically shown using a
M AN U
coupling matrix (C). The relationship between different input and output ports for an example hub is shown in Fig. 2.
piα pψi pi
• • •
TE D
ω
pαj p βj • • •
pωj
EP
Fig. 2. Relationship between the input and output ports of an energy hub with a Cij coupling matrix [41]
AC C
The exchangeable power in inlet and outlet ports are expressed by Pi and Pj vectors respectively and these power vectors are related via the coupling matrix by the following equations. In this equation, the Cij is the coupling matrix array which according to the characteristics of the existing elements, associate the hub inputs and outputs (Eq. (1)) [41]. ⋯ " ⋱ ⋮ * ⋮ + = 0 ⋮ + ⋮ ⋯ $%%%%&%%%%'
()
)
(1)
ACCEPTED MANUSCRIPT
There are two important properties in the coupling matrix arrays: 1- According to the energy conservation law, energy is converted from one form to another, but in the process of conversion, there is also some energy wasted, so all coupling matrix arrays fall
0 ≤ ≤ 1
∀(3, 5)78
/
and
RI PT
into the maximum or minimum values according to the following equation (Eq. (2)) [41]. 9, : 7 ;
SC
In this equation:
∀(3, 5) 7 8 = <(1,2), (3,4), … A , 9, :, B 7 ; =
(2)
(3)
M AN U
2- According to the total energy conservation law, all outputs converted from a given input must be less than or equal to the input value. Therefore, the total arrays of each column in the coupling matrix must be less than or equal to the relation below (Eq. (4)) [41]. /
≤1
∀(3, 5)78
9, : 7 ;
(4)
TE D
0 ≤ ∑/PQ
If the output loads of the hub are represented by L and the input carriers of the hub are represented by P, the previous equation can be rewritten as follows:
S
(
EP
R ⋯ ⋱ ⋮ +* ⋮ + * ⋮ +=* ⋮ R ⋯ $%%%%&%%%%'
(5)
AC C
As long as the coupling matrix elements, known as an efficiency of energy hub equipment, have constant and independent values, Eq. (5) is linear. However, considering the dependence of the equipment efficiency on the input energy, this equation will be nonlinear [41]. Since one type of input energy into the hub may be split into several parts and enters various converters within the hub, coefficients named dispatch factors are defined which represent the amount of input energy assigned to each of the hub convertors. In multi-input and multi-output
ACCEPTED MANUSCRIPT
hubs, the number of coupling matrix elements depends on the dispatch factors as well as the converter equipment efficiency. Fig. 3 shows the industrial plant configuration evaluation for this section. Commercially available water oriented industry technology is the basis of this design.
RI PT
This plant consists of a MED for potable water production, RO plant, steam turbine, and combustion turbine generators (CTGs) and a boiler. The plant scope is identified and specified according to the aim of this research and multigenerational factors. An alternative energy source
SC
and high efficiency are proven properties of this plant to minimize CO2 emissions. Fig. 3 presents thermal plant overall system flow. Three varying potential energy sources were taken
M AN U
into account. Waste heat in cement plant is the first source. The second source is exhausted gas from the GT stack. The third source is heat energy from the boiler. Electrical load requirements for the energy hub can be met using energy sources such as electrical grid, steam turbine, and GT systems. The dispatch factors vMED, vBoi, vCHP, vST and vRO are in interval at all times [0, 1]. As
TE D
previously stated, the energy hub model is considered for multi-energy carrier systems and represents continuity equations by considering a control volume such that each input energy carrier to the system is multiplied by the coupling matrix coefficient, exhibiting its influence on
EP
the considered output load so that the sum of these inputs determines the output load by taking the energy hub productions and consumptions into account. The important point is that the
AC C
output loads are known as specified quantity in each energy hub which ultimately, by expanding the equations (Table 2) and implementing decision variables and assumptions for each equipment and energy converters, the required input energy to the system, the coupling matrix coefficients, and hence the capacities of each equipment can be determined. The decision variables for each equipment and dispatch factors are related to the considered objective function in optimization and will be determined accordingly. Thermal, electrical and water loads along
ACCEPTED MANUSCRIPT
with waste heat from the cement plant are shown as LDH, LDe, LDW and Qextra respectively which are determined according to Fig. 5 and section 5. Section 7.1 shows decision parameter values, vMED, vBo, vST, vRO, RRO, Thot, Mhot, EWRO, Qmax,Boi which are determined by the GA
RI PT
optimization method given their limits. The values of GT efficiency (ηCHP) and steam cycle efficiency (ηST) are obtained according to the reference curve of the power plant characteristics. The GOR value will also be determined from section 2.3 from the Thot and Mhot decision
SC
variables. According to Table 2, 12 variables can be determined from 12 equations and the energy hub can be obtained. In order to further highlight the aforementioned points, Fig. 1
M AN U
depicts a schematic of the paper’s modelling process. Based on this diagram, the weather condition along with the water, electricity, and heat demands as well as the decision variables are entered to the system equation shown in Table 2 as known values. All energy vectors pertaining to the energy equipment are specified in the hub. These values are entered in the equipment
TE D
modelling process. The output determines the equipment’s capacity, their quantity and other required parameters relevant to the TAC and exergy efficiency. Finally, by optimizing this model, the design factor, equipment and their quantities for the optimal objective functions will
EP
be determined.
In Table 2, row 1 of the energy continuity equation for electric current is presented in Fig. 3. In
AC C
GT power generators and steam turbines, these inputs should be equal to electrical consumption of RO and the power consumption of the hub. Rows 2 and 3 are auxiliary equations that are extracted according to their definition. In row 4 of this table, the hub water affinity equation is presented. The water produced by the RO and MED system should be equal to the demand for water.
ACCEPTED MANUSCRIPT
RI PT
Ex& Heat , χ
AC C
EP
TE D
M AN U
Ex& Power , χ
Ex&Water , χ
Ex&dmn ,1
SC
Ex&NG
LD H
Fig. 3. Schematic view of energy hub used in case study
LDe
Ex&dmn,2
LDW
Ex&dmn ,3
ACCEPTED MANUSCRIPT
Table 2. Design equation for energy hub model [6].
4 5 6 7
8 9 10
11
^YZ + ^_W` + Ua − R[ = 0
^YZ , ^_W` , Ua
^YZ = ]YZ (^_W` + Ua ) ^YZ = ]_W` (^YZ + Ua )
b(T − bUV + bc + T − b_W` + bWdef" − R[T = 0 bc = ]gh (b(T − bUV + T − b_W` + bWdef" ) DYZ = KNYZ ^YZ b(T =
(1 − i(T ) D(T i(T
jkl =
^_W` 2326 n5 b_W` 1 nH
bUV =
DUV iUV
](T
R[
b(T , bUV , bc , T
bWdef" , R[T
i(T
]UV
]YZ
]_W` ]gh
KNYZ
iUV
jkl
EP
12
DUV = ]UV (D(T + W − DYZ )
Decision variables
RI PT
3
D(T = ](T (DUV + W − DYZ )
D(T , DUV , W , DYZ
Specified quantity R[\
SC
2
D(T + DUV + W − DYZ − R[\ = 0
Unspecified quantity
M AN U
1
Equations
TE D
Row
AC C
In row 7, the heat produced by the GT, boiler and waste heat from the cement plant are heat generators in the hub, which should provide the heat required by the steam turbine, MED and hub thermal demand. Also, the current (PE), water (PSW) and heat (PH) input to the hub from the grid or the output from the hub to the grid can be placed with a positive and negative sign respectively. All equations of rows 2, 3, 4, 5, 6, and 8 are defined by dispatch factors vCHP, vST, vRO, vMED and vBoi which are decision variables and determine the amount of energy distribution in
each of the energy carriers within the hub equipment.
ACCEPTED MANUSCRIPT
2.2.RO modeling Two modeling methods for the RO system are proposed by FILMTEC [44]. The first method is
RI PT
element to element. In the second method, if the feed quality, temperature, and intensity of the purified water flow and the number of elements are known, the average value is used to calculate the feed pressure and quality of purified water. If instead of the number of elements, the feed pressure is specified, the number of elements is obtained through several integration equations.
SC
The design equations for the FILMTEC SW30 8-inch elements are given in the references [44].
M AN U
In this paper, considering the demand for water and the maximum water quality required based on the World Health Organization (WHO) standard for drinking water [45], the required water recovery, 6 elements per pressure vessel, membrane type selection (SW30XLE-400i) and water flow intensity and a silt density index (SDI) of under 3 [44], and the amount of power assigned to the RO can be determined based on the amount of feed pressure. The mean value method is
TE D
considered. In regard to the hub, by determining the amount of power, the pressure can be determined based on the equation below [44]. ^qr (r − "sc ) 8iop
(6)
EP
q = Nop
In this model, the Pelton turbine has been used as an energy recovery system. According to Eq.
[44].
AC C
(7), the amount of energy required by the RO system can be determined in accordance to Eq. (8)
Nq V =
^q c (c − op ) 8hpe iV
Nqt\e = Nqop − NqV
The basic equations in designing RO system are available in Appendix A.
(7) (8)
ACCEPTED MANUSCRIPT
2.3.MED modeling According to reference [43] where a multi-effect distillation system is assessed as experimental characterization, a number of empirical equations are extracted in order to assess the
RI PT
performance of the system under off-design conditions. In addition, the impact of the key parameters variation where the MED plant is controlled (inlet hot water flow rate (mhot) and temperature (Thot)) on GOR and freshwater production were evaluated with an empirical
SC
campaign of 82 tests. The MED plant steady state results working at off-design circumstances are addressed in a state which presents: The Thot and the mhot effect on the production of distillate (mMED) and on the plant GOR. The MED modeling equations are presented in this
M AN U
section. Eq. (9) and Eq. (10) are valid for 60 ≤ uvhe ≤ 74 ℃ and 7 ≤ ^vhe ≤ 14 U [43]. } ) ^_W` = −0.273 + (0.008409 uvhe ) − (0.04452 ^vhe ) + (0.0003093 uvhe − } ) (0.002485 ^vhe + (0.001969 uvhe ^vhe )
S
(9)
TE D
} ) jkl_W` = 648.2 − (26.74 uvhe ) − (16.45 ^vhe ) + (0.3842 uvhe + ~ ) } ) (0.3137 uvhe ^vhe ) + (0.5995 ^vhe − (0.001835 uvhe −
(10)
EP
~ ) } } (0.002371 uvhe ^vhe ) − (0.01844 ^vhe − (0.0001411 uvhe ^vhe )
AC C
The latter parameter determines the thermal efficiency of a MED plant (GOR) and it is expressed in Table 2. (row 12). Where mMED is the distillate production (L/s); Gain origin ration (GOR) (-); hot water temperatures (Thot), and the hot water flow rate (mhot); Qhot Heat consumed by MED. Moreover, in Table 3, the operation conditions for each system in the optimal scenario involving waste heat recovery are presented.
ACCEPTED MANUSCRIPT
Table 3. The operating conditions of the energy hub [57, 63 and 64]
GT
Boiler
1 64.35 11.54 9.56 115 1 86 10 0.098 90 36 4028.7 48550 0.65 2.75 0.0982
Pipeline length
m
1900
Ambient temperature Ambient pressure Salinity of Seawater
°C kPa ppm [mg/kg]
298 100 42000
TE D
ST
Fuel
-
1200 6
AC C
EP
Water transfer Ambient
L/s (-) (-) kJ/kg (-) (-) kg/s % % kW kJ/kg kg/m3 kgCO2/kgfuel kg/s
$
Value 0.9965 37.2
M AN U
MED
Unit (-) m2
RI PT
RO
Parameters Salt rejection Active RO membrane area Membrane element cost, estimation Element (type SW30XLE-400) Number of membranes in each pressure pipe Number of steps in RO Hot water inlet temperature Hot water flow rate GOR Air-fuel ratio Hot gas Specific Heat Isotropic turbine efficiency Compressor pressure ratio Fuel consumption Efficiency in full load mode Efficiency in full load mode Heat Consumption Lower Heating Value (LHV) NG density m CO2 Fuel consumption
SC
Component
2.4.Pumping system
A pumping system is necessary to transfer water. Eq. (11) shows the system’s energy consumption in Ppu (Mj/h) [6]. op =
H × ^qefr × C × 1000} × 60} ηop
(11)
ACCEPTED MANUSCRIPT Such that g is the gravitational acceleration (m/s2), ηop denotes pumping efficiency ratio i.e.
0.75, ^qefr represents media flux in terms of kg/s and H is the lifting height in terms of (m). According to the Darcye-Weisbach equation, Eq. (12) expresses the lifting height [6]: R ]} × +n [ 2H
RI PT
C=×
(12)
Such that L is the length of the pipeline (m), D is the internal diameter of the pipeline (m), v
SC
denotes media velocity (m/s), g denotes gravitational acceleration (m/s2), f represents coefficient of friction loss, and k denotes friction loss via local resistance.
M AN U
3. Exergy analysis
Exergy analysis can aid in developing strategies and guidelines for the effective use of energy in various systems. Exergy can be divided into four categories. Two prominent types of exergy are physical exergy and chemical exergy. In this study, potential exergy and kinetic exergy are
TE D
hypothesized as negligible. The physical exergy equals the maximum workable flow of matter and is determined by a process from the initial state or an environment that only involves heat exchange with its surroundings. Chemical exergy occurs by the exit of chemical compounds
EP
from a system of chemical equilibrium [46]. Chemical exergy is part of the exergy in the combustion process. By using the first and second law of thermodynamics, the following exergy
AC C
equilibrium is obtained [47-49].
K( u ( = (1 − ) bq − Nq − + ^q D − ^q \ D\ − Kq ` F u F \
(13)
Or Exq + mq ex = mq ex + Exq + Exq
(14)
ACCEPTED MANUSCRIPT
The χ parameter indicates that when exergy flow is outside the control volume shown in Fig. 3 and enters the system (χ=Buy), it is considered positive in equations and if exergy flow exits volume control (χ=Cell) it is shown as negative in equations.
RI PT
In Fig. 3, exergy flow is calculated as: Kq = ^qr,Vhe" RC
Where parameter ξ is defined as follows for NG [47]:
ov v Kqa"e\f = Kqa"e\f + Kqa"e\f
M AN U
Water flow exergy is equal to [50]:
SC
0.0698 = 1.033 + 0.0169 −
(15)
(16)
(17)
v Where Kqa"e\f and Kqa"e\f are water physical exergy and chemical exergy respectively and are ov
Kq = (1 −
uh )b q uv v
Kq h\f = Nq
TE D
calculated in [50]. Heat exergy and electricity are calculated based on Eq. (18) and Eq. (19) [51]:
(18) (19)
~ q ∑ ∑ e Ko,e e(∑p K st,p ) ¡e = q ∑ ∑ e(^qr,(T RC + Kc ) ¡e e Kr,e
AC C
;=
EP
Finally, hub exergy efficiency is calculated based on Eq. (20) [51]:
(20)
If u=1, heat exergy of the demand heat flow is calculated as follows [52]: Kq st,p = 1 −
u Cq uv st
(21)
If u=2, electricity demand exergy is calculated as follows [51]: Kq st,} = Kq¢
(22)
ACCEPTED MANUSCRIPT
If u=3, water demand exergy is calculated as follows [50]: st Kq st,~ = Kqa"e\f = ^q"e\f Kq"e\f
(23)
RI PT
4. Economic analysis Economic analysis is required to select equipment and coefficients stated in the hub section. In economic analysis, equipment costs and maintenance costs are calculated according to the utilized equipment. Eq. (24) is used to estimate MED costs [52]:
SC
_W` = £_W` lÅ + £Z&_
(24)
M AN U
In this equation, C = 81430.27 $ according the reference [53] and £Z&_ is calculated by Eq. (25)
[25].
£_W` = ^qÆf\¢v "e\f
RO system costs are set according to Table 4.
TE D
Table 4. RO system costs equations [25].
AC C
EP
Component Capital cost of the seawater intake and pre-treatment Capital cost of high pressure pump and pre-treatment
Pump
CCUa¦ = 996 (mr 24).§
Equation
CC¨©© = 52 (mvoo ΔPr )
CC¬© = a (mVΔP)¯° f . ϕ² 1 − η³ ¯´ ϕ² = 1 + 1 − η cast iron = 1 f = µ Steel = 1.41
f :Material correction factor, here: f = 1.41
ϕ² : First law efficiency correction factor $
a = 549.13 ÀÁÂ.ÃÄ ،a} = 0.71 ،a~ = 3 ،η³ = 0.8
(25)
ACCEPTED MANUSCRIPT
Other hub costs are set according to Table 5: Table 5. The cost functions for each equipment of energy hub [54, 55].
GT
Cost function ZÈÉ = (−0.014 V + 600)V ZÊÉ = 1200 UV
Steam Turbine
.§Í ZËÌ = 205bqgh \f
Boiler
RI PT
Component
SC
Economic parameters are determined based on reference [54, 55] shown in Table 5. The cost of the transmission line is also determined by Eq. (26) with the unit of $/m [6].
where D is the media pipe diameter (m).
M AN U
efr,o o\ t\ = 0.9734 (32000 + 213000 [)
(26)
By using assumptions for Eq. (11) and Eq. (12), the necessary costs, efr , CO2 emissions flow
correspondingly:
EP
efr,op = L\ × op
TE D
pertaining to the functionality of a pumping system are represented by Eq. (27) and Eq. (28),
^h},efr = D\s × op
(27)
(28)
AC C
Such that ce denotes generation cost within a power plant in terms of ($/MJ), eem represents the CO2 emission factor in terms of (kgCO2/MJ). Finally, TAC is calculated based on Eq. (29) [54]: uÎ Ï$ÐDEMÑ = _W` + YZ + V + UV + gh + efr + ∑ eÒKc Ó\,c − K¢ Ó\,¢ + ^qr,ehe" RCrp\ Ór + 3600 ^h} Ô\s Õ ¡
(29)
ACCEPTED MANUSCRIPT
Table 6. Values of economic parameters [60, 61, 62] Unit %
Value 10
Operation years (n)
Year
20
Hour
8760
3
$/m $/kWh $/kWh $/kWh $/m3 $/kg CO2
0.5 0.015 0.045 0.001 0.002 0.02086
SC
Operation hours in a year (N) Cost of fuel Electricity sales price Electricity purchase price Heating cost Desalinated water cost CO2 emission factor (ψem)
RI PT
Economic parameters Interest rate (int)
M AN U
Table 6 presents cost parameters. τ is an hour time period and N is equal to 8760 hours. Capital recovery factor (CRF) is dependent on equipment estimated life and interest rates and is written
lÅ =
3GF (1 + 3GF)Ö\"f (1 + 3GF)Ö\"f − 1
(30)
TE D
as [52]:
Where ‘year’ is the system service life and int is the interest rate.
EP
5. Case study
A diagram of the cement plant in the city of Qeshm is presented in Fig. 4. Fig. 4-a and Fig. 4-
AC C
b show the clinker production process and the cement plant’s curing stage. In Fig. 4-a, upon combining and grinding the primary raw material, it is entered from point 1 to the preheater and after passing through the cyclones, its moisture is removed and the limestone is calcified. Then, it enters the kiln at point 2. The raw materials are cooked in the kiln and converted to a clinker prior to be transferred to a clinker cooler to cool down (clinker point 3). There are fans installed at the bottom of the clinker cooler which absorb most of the hot clinker heat and exhaust it through its upper duct. Also, in Fig. 4-b, the potential points for heat recovery from the cement
ACCEPTED MANUSCRIPT
factory are shown. One of the points where it is possible to utilize waste heat is the suspension preheater boiler (SP boiler), where the hot gas from the kiln runs upwards in the opposite direction of the raw materials, transferring some of its heat. However, a large part of the heat in
RI PT
the SP boiler is still recoverable. The resulting gases thermal energy (317 °C) that may be recovered using SP boiler is 1532 kW. Clinker has to be cooled upon being baked at 1200 °C temperature. Gases from the clinker cooler exhibit a second heat source which is recovered using
SC
an after quenching chamber boiler (AQC) (approximately 4088 kW) [57, 58]. Table 7 describes the characteristics of these points. Since the city of Khaledin, located on the island of Qeshm is
M AN U
close to the cement plant, its water demand to supply its 1396 population is shown in Fig. 5. The demand for consumed water (kg/s) for each hour during the year for Khaledin is determined which should be supplied by the energy hub. Also, Fig. 5 presents the cement factory’s consumption of electricity (kW) for each hour during the year. Since the variances in
TE D
consumption in various hours of the year are insignificant in relation to the consumption level, the average annual consumption for all hours can be considered. Also, the distance between the cement factory and the city of Thurian with 2339 populations is approximately 1.9 km, which
EP
will allow the transfer and sales of excess water to the city. In regard to the case study, the lack of implementing technology and the inconsistency between investment costs and local prices are
AC C
the main reasons for not using renewable energies. Also, due to inaccessibility to sufficient historical data in this paper in addition to considering the constant loading profile and national grid as a backup in storing energy, the uncertainty was not considered. Another reason that diminishes the need for uncertainty analysis is that some equipment such as desalination systems utilize variety of technologies (MED and RO) and the use of an electricity energy carrier for the RO system and thermal energy carrier for the MED system will aid in decreasing uncertainty.
ACCEPTED MANUSCRIPT
Table 7. Qeshm cement plant gas condition [57] Mass flow rate (kg/s)
290
21.4
311 215 93 99
15.9
15.9
15.9
21.4
AC C
EP
TE D
M AN U
Clinker cooler exhaust Preheater exhaust SP boiler exhaust Stack exhaust AQC boiler exhaust
Temperature (˚C)
RI PT
Parameter
SC
Point
Fig. 4. Schematic of heat utilization potential in Qeshm cement plant
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 5. Values of electrical loads and water consumption
M AN U
6. Optimization
GA (Evolution) is an algorithm based on repetition. The strategy of this algorithm is to apply a random search to determine an optimal solution. In principle, the simple method of biological evolution is to emulate the evolutionary characteristics of a population of individuals in which
TE D
the individual consists of decision variables values for a potential optimized solution. The decision variables and optimization variables are given in Table 8. The objective functions of this optimization are Eq. (20) and Eq. (29) which determine the hub’s exergy efficiency and
EP
economic parameter, respectively.
AC C
Table 8. Design parameters and their range of variation for the optimization procedure.
CHP Power dispatch factor MED Water dispatch factor Boiler Heat dispatch factor Steam Turbine Power dispatch factor RO Water dispatch factor Water recovery rate (RO) Hot water inlet temperature (MED) Hot water flow rate (MED) Power consumption per Water Desalinated (RO) Maximum Heat Produced by Boiler
Parameter vCHP (-) vMED (-) vBoi (-) vST (-) vRO (-) RRO (%) Thot (°C) Mhot (L.s-1) EWRO(kJ/kg)
Lower bound 0 0 0 0 0 0.3 60 7 0.1
Upper bound 1 1 1 1 1 0.85 74 14 8
Qmax,Boi (kW)
100
3000
ACCEPTED MANUSCRIPT
A decision making process should be employed to select a single optimal point from the present points on the Pareto front. Such a process is typically based on engineering experiences where the importance of every decision making objective is highlighted. An assumed point taking the
RI PT
role of an ideal point is usually used to determine the final decision making process. When two objective functions are individually optimized whilst disregarding other objective functions, the compositions of these values symbolizes the ideal point or ideal objective point. Fig. 6 presents
SC
the usual Pareto front in addition to the ideal point to maximize the first objective whilst minimizing the second objective. It is evident that it is not possible to achieve both objectives at
M AN U
optimal points concurrently. The closest Pareto frontier point to the ideal point may be chosen as the final optimal solution because the ideal point is not a solution that is placed on the Pareto frontier. Prior to this, the objectives are non-dimensionalized. LINMAP, linear programming techniques for multidimensional preferences analysis method is implemented to non-
t Å×d =
Å×d ss Ø∑×(Å×d )}
°
TE D
dimensionalize the objectives in this paper by employing the following equation [59]: (31)
Such that z is the index for each Pareto front point, x is each objective’s index and mm displays
EP
the Pareto front’s number of points. Hence, the distance for each Pareto front point from the ideal
AC C
point is attained [59]: }
t × = Ù(Å×d − Å t \",d )} d
(32)
Such that ideal is ideal objective function index. Thus, the value of distance for each Pareto frontier point is derived via the equation stated above and optimal point leading to minimum d is chosen as final optimal solution.
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 6. Pareto optimal front for energy hub According to [56], some of the essential characteristics of genetic optimization algorithms are:
TE D
(1) The generality of the formulation.
(2) Robustness, in the sense of avoiding local optima, looking for global optimal points. (3) The capability of handling multiple objectives.
EP
(4) Computational efficiency decision variables.
AC C
6.1. Parameter tuning
In this paper, the decision parameters are assumed values which can be changed within an acceptable and specific range as well as exerting significant effect on objective functions in order to minimize the TAC and maximize the exergy efficiency upon optimization. These values are initially determined from the concept model and related formulas. Ultimately, their effects on the objectives are assessed such that through final selection, by making changes in a specified range, the optimized acceptable values can converge to the results of the equations. Dispatch factors are
ACCEPTED MANUSCRIPT
among the variables used in all energy hub studies to achieve optimal energy flows in control volume associated with energy hubs. Rows 1, 2, 3, 5 and 6 of Table 2 in the paper, are related to the design equation for the energy hub model, which is dedicated to decision variables vCHP, vST,
RI PT
vRO, vMED and vBoiler, respectively. Each dispatch factor represents the production or consumption capacity share of the system in relation to the overall energy carrier production or consumption share and thus is a determining parameter in the energy hub’s internal equipment capacities and
SC
input energy carriers. For this purpose, the sensitivity analysis for each of these variables in
and TAC are provided in Appendix B.
M AN U
relation to the objective function and key values of energy hub output such as exergy efficiency
In this study, the results of the objective function from two-objective optimization of the energy hub for different algorithms parameters in this case study are presented in Table B-1. The resulting values from the genetic algorithm parameters in Table B-1 are extracted according to
TE D
reference [61] to analyze the sensitivity of the objective functions. Each result is presented upon best in 3 times run. As evident, the best results are obtained at levels 4 and 5 which have the least TAC with highest exergy efficiency. From these two levels, level 5 exhibited the most optimal
EP
results with population size 150, mutation 0.03 and crossover 0.8 and its parameters were
AC C
selected for system outputs.
7. Validation
Validation indicates that the model is a logical representation of an actual system in order to assess the actual system activity with enough fidelity to analyze the main goals. Each model typically analyzes a specific problem and presents various system components in different levels of abstraction. As a result, there is a model of different levels of validity for various system components, which together, examines the system behavior. There are three separate aspects for
ACCEPTED MANUSCRIPT
most models which should be assessed during model validation: Assumptions, Input parameter and Output values. For this purpose, in order to achieve full validation of the energy hub, validation of each equipment model should be considered to attain validation of the integrated
RI PT
energy hub system which does not yet exist. Thus, each equipment should be validated separately. 7.1. RO
SC
For the purpose of validating the written code, ROSA software was implemented for plug system. Validation was achieved via a system where water volumetric flow rate and salt
Table 9 presents the validation results.
M AN U
concentration are 40 m3/h and 4049 ppm, respectively. The sample was modeled in two stages.
Table 9. The validation results of ROSA software Modeling results
TE D
Parameters
AC C
EP
Produced freshwater in stage 1 (ppm) Produced freshwater in stage 2 (ppm) Production of fresh water in each stage 1 (m3/h) Production of fresh water in each stage 2 (m3/h) Feed pressure (MPa) Recovery
ROSA software results
Error (%)
21.94
23.50
7.1
59.52
65.58
10.1
40.12
40.83
1.7
19.87
19.17
3.5
3.53
3.77
6.7
46.0
46.0
0
7.2. MED
Parametric equations were successfully achieved in reference [43] prior to be statistically validated to predict the GOR and distillation production as a function of various input parameters via an extensive validation range. Moreover, the presented empirical model in the section 2.3
ACCEPTED MANUSCRIPT
stemming from the extensive empirical results has been compared with reference [52] (Fig. 7). This comparison was conducted under conditions where the values of mf and Tf were within the acceptable range of the empirical model. This comparison shows that the empirical model has
RI PT
minor differences with the numerical results, and the relevant errors are acceptable in terms of empirical model application. 7.2. Gas Turbine and Steam Turbine
EP
TE D
M AN U
SC
The model applied in the gas turbine and steam turbine was based on reference [48].
AC C
Fig. 7. MED model validation with a valid model in various hot water inlet temperatures
8. Results
The Pareto curve is derived from two-objective optimization, and in order to achieve optimal decision variables and optimal objective function values, the two-objective functions will be TAC and exergy efficiency. The genetic algorithm categorizes the solution space into a feasible domain and infeasible domain. The frontier between these two regions is the Pareto curve. Each
ACCEPTED MANUSCRIPT
of the points created result from the decision variables that leads to the determination of the objective functions’ values depicted on the Pareto curve. Each resulting point is an optimal point. The ideal point on this curve is the point with the lowest TAC and highest exergy efficiency.
RI PT
This point is in the infeasible domain. The point of the curve that is closest to this point is called the best point. Fig. 6-a shows the optimum Pareto front results for two objective functions.
The acquired results show that two chosen objective functions i.e. energy efficiency and TAC
SC
differ in the optimal state. Attempts to enhance energy efficiency will increase TAC in both cases and vice versa. The confliction trend in the optimal state should be highlighted via multi-
M AN U
objective optimization. Increases in energy efficiency increases optimal TAC with the increment slope significantly increasing for maximum energy efficiency. Thus, the energy hub is not economically viable for maximum energy efficiency. The developer should choose the final optimal point according to the significance of each objective. Additionally, the optimum Pareto
TE D
point is influenced by optimal results acquired in the energy hub. Table 8 lists ten design factors i.e. decision variables along with relevant variation ranges in order to maximize energy efficiency value while minimizing TAC. Fig. 6-b illustrates optimal Pareto fronts with and
EP
without waste heat recovery stemming from the cement plant. The acquired results show that when waste heat is included, the optimal Pareto front is completely dominated by the optimal
AC C
results that exclude heat recovery. According to Fig. 6-b, exergy efficiency in all locations with waste heat recovery is greater compared to the without waste heat recovery whilst exhibiting lower TAC. According to Table 10, the exergy efficiency increased by an average of 20% in the case of without waste heat recovery compared to waste heat recovery case while its TAC function decreased by 72.6%. Moreover, the effect of optimization with the concept of energy hub was shown in a case study in Table 10. In this regard, the effect of the optimal energy hub
ACCEPTED MANUSCRIPT
system was compared with the current system (with considering energy recovery from the cement factory). In this table, the exergy efficiency and TAC for the current system are shown along with the energy recovery from the cement plant. According to this comparison, in the
RI PT
optimal mode, the exergy efficiency increased by 29.6% and TAC decreased by 19.3%. This illustrates the impact of using the concept of energy hub with the optimal allocation of energy carriers and the addition of new energy equipment.
M AN U
Pareto optimal fronts for input values given in Table 9.
SC
Table 10. The optimum values of exergy efficiency and the TAC for the design points A–C in
With using energy hub concept Parameter ηex (-)
A 0.78 2553476
B (Best Point) 0.7445 1104138
C 0.727 1059092
Without heat recovery Best Point 0.616 4036266.47
TE D
TAC ($/ year)
With Waste heat recovery
Current system with heat recovery 0.5741 1368827.47
The use of heat loss improves both objective functions. In regard to the TAC function, it should be stated that the high potential of heat loss in the cement plant has led the energy hub to utilize
EP
this cheap energy to not only self-sufficiently reduce part of the electricity, heat, and fresh water production costs, but also reduce O&M costs. Concerning the exergy efficiency function, it
AC C
should be stated that directly reducing fuel consumption according to Eq. (20) will effect exergy efficiency and increase this objective function. A decision making process is implemented to select the final solution from optimal Pareto front points. This process is conducted on the basis of engineering experiences and the significance of every objective for the decision makers. Fig. 6-a shows that at design point A (78%) there is maximum exergy efficiency exists and the TAC is highest at this point. Furthermore, at design
ACCEPTED MANUSCRIPT
point C, the minimum TAC is achieved (1059092 $/year) along with the lowest exergy efficiency value (72.7%). The optimal exergy efficiency state as a single objective function is at design point A. The optimal condition where TAC is a single objective function is at design
RI PT
point C. Table 10 lists the optimal values for two objectives regarding five usual points from AC (Pareto optimal fronts) for the input values stated in Table 11. Also, Table 11 shows that, moving from the point C to point A, the vMED value is approximately 6% of its initial value,
SC
which indicates a decrease in annual costs and an increase in exergy efficiency for the use of MED desalinating water entailing less use of the RO system.
M AN U
Table 11. The optimum values of design parameters for optimum selected points A–C in Pareto optimum front. With Waste heat recovery Parameter
TE D
B (Best Point) 1 0.015 0.444 0.995 0.494 0.459 64.35 11.54 3.76 1777.6
C 1 0.316 0.48 0.999 0.234 0.468 64.7 11.98 3.8 1768.3
AC C
EP
vCHP (-) vMED (-) vBoi (-) vST (-) vRO (-) RRO (%) Thot (°C) Mhot (L.s-1) EWRO(kJ/( kg)) Qmax,Boi (kW)
A 0.997 0.02 0.547 0.76 0.51 0.45 64.4 11.59 4.28 1810.3
Without heat recovery Best Point 0.997 0.039 0.61 0.577 0.72 0.34 66.65 10.56 3.2 1511.66
In the RO system, high pressure water should be supplied between 60 bar to 80 bar to desalinate seawater. A greater power allocation to the RO system in the steady state water supply system increases the water supply pressure and increases the difference between water supply and osmotic pressure flux of the flowing water in each membrane, thus entailing increased system recovery and increased fresh water flow. The genetic algorithm has increased the production share of the system by increasing the power allocated to the RO.
ACCEPTED MANUSCRIPT
Also, a 25% reduction in the use of a steam turbine to generate electricity will increase TAC and exergy efficiency. The 42% reduction in the use of the steam turbine to generate electricity and an increase of 38% in the use of the auxiliary boiler indicates that in order to meet electricity
RI PT
demand, electricity should be bought from the grid and in order to meet heat demand, fuel consumption should be increased and bought from the natural gas grid.
Since the city of Khaledin that is located in the vicinity of the cement plant, has a specific
SC
population, as a result, the highest demand for energy carriers and heat is limited, resulting in the maximum sales of these two energy carriers through a hub of 6980 kW for electricity and 5723
M AN U
kW for heat. In the energy hub, urban water demand is considered as the demand for energy hubs. Results indicate that due to the high costs of water transfer and the distance from the nearest city (apart from the adjoining city) to the cement plant, there are no water sales outside the hub. Upon optimization and using the energy hub concept, the energy purchased from the
TE D
factory has dropped from 3.2317 kg/s to 2.114 kg/s, indicating that part of the water is produced in the hub and utilized in the adjacent city (1.11 kg/s). Also, the energy hub is self-sustained in electricity maintenance, with the small amount exception, the remaining 2988 kW supplied by
EP
the GT and steam turbine. Due to the 5630 kW heat recovery from the cement plant, the majority (4856 kW) is sold outside the hub. Also, in order to supply the NG required by the GT and
AC C
boiler, 0.0982 kg/s is consumed in the hub. In Table 11, it is noteworthy that the use of gas turbines is maximally utilized and due to the increased efficiency of the GT system, lower costs and reduced environmental pollutants, it is considered feasible.
Another factor in reducing TAC can be accredited to the utilization of the gas turbine. One of the features of this apparatus is its electricity and heat supply based on an initial production cost.
ACCEPTED MANUSCRIPT
Therefore, the genetic algorithm increases the role of this system in the energy hub leading to a reduction in the installation costs pertaining to other equipment such as a boiler. On the other hand, the confrontation between the heat loss of the cement plant and elimination of a boiler
TE D
M AN U
SC
the gas turbine has led to a general reduction of TAC.
RI PT
from the system compared to the non-utilization of waste heat and increased fuel consumption in
temperature
AC C
EP
Fig. 8. The rate of exergy efficiency in energy hub at hot water flow rate and hot water
As shown in Fig. 8, it is evident that energy efficiency increases with Thot as expected. Considering the impact of QMED on energy efficiency, it is clear that energy efficiency slightly decreases for Low QMED (from 100 kW to 200 kW) while decreasing considerably (approximately 2.2%) for larger QMED. Such progressive decreases may be because of the evaporative process via the tube bundle that is closer to its hot water temperature design value
ACCEPTED MANUSCRIPT
(11.54 L/s). Due to the increases in thermal energy consumed, which is greater than the distillate production, an energy efficiency reduction is experienced by QMED that is over 200 kW. Also, by increasing the rate of Thot (from 60 °C to 74 °C), the exergy efficiency increases by
RI PT
1.5%. Fig. 9 shows vRO, vMED variations in terms of TAC for both waste heat recovery and without waste heat recovery. In this figure, with vMED variations from 0 to 1, TAC increases by an average of 3.1%. Also, vRO variations from 0 to 1 will increase TAC by an average of 2.5%
SC
and up from 1049362 $/year. The percentage of TAC variations for without waste heat recovery
AC C
EP
TE D
compared to waste heat recovery.
M AN U
remains the same. Moreover, the average TAC increase in without waste heat mode is 3.8%
Fig. 9. The rate of TAC in energy hub at hot water flow rate and hot water temperature
By increasing the allocated share of fresh water production to the RO system, it is evident that TAC costs have increased. In order to produce more water, the RO system requires higher water pressure from the intake water supply as well as higher water pressure. By allocating more
ACCEPTED MANUSCRIPT
energy to this component, the water pressure will increase. The main factor in increasing the RO costs according to Table 4 is the pre-treatment costs. In this state, by increasing the pre-treatment
TE D
M AN U
SC
RI PT
costs and its effect on the RO system, the TAC will increase.
Fig. 10. The rate of TAC in energy hub at hot water flow rate and hot water temperature
EP
Fig. 10 shows the variation in the price of water ($/m3) for three modes of production by RO,
AC C
MED and the entire hub system, taking into account the share of each of the two water producers. From 0 to 0.4, the UPC change is 0.65%, while changes up to 1 will result in a 40% drop. By changing the value from 0 to 0.3, the UPC will drop by 23% and will change from 0.54 to 0.416. On the other hand, its reduction from 0.3 to 1 would be 11%. It is clear that changes in the share of water produced by each desalination system will not affect the production cost of other systems, and the cost of producing water will be higher if the unit is smaller. The process of changing the final price of water production in the energy hub system indicates that increasing
ACCEPTED MANUSCRIPT
the share of water production in both MED and RO systems reduces water production costs. Also, Fig. 10 shows that the price changes of the MED system are significantly high compared to the RO system. The MED system is considered for a specific capacity with equivalent
RI PT
investment costs whereas if it utilizes less heat than its nominal capacity, the production of fresh water is reduced in a fixed investment. Due to the reduction in fresh water production, the final water costs will significantly increase, in which case the investment cost is considered for the
SC
MED, but the system will not be utilizing its nominal capacity. This greatly effects the final price. On the other hand, the RO system can only control fresh water flow by reducing feed
M AN U
water pressure and the valves in the brine water pathway. As a result, in the RO system, price changes are only dependent on energy consumption prices and O&M, thus pertaining to price
AC C
EP
TE D
changes in this limited range.
Fig. 11. Variation of TAC versus ratio of price and energy carrier in energy hub
ACCEPTED MANUSCRIPT
Fig. 11 shows the sensitivity analysis conducted on key factors such as water price, heating and electricity. The results from this analysis highlighted the energy carrier with negative slope that may be supplied with other sources and acquired with a positive slope. The achieved results
RI PT
show that when water selling price increases, the TAC is significantly raised with constantly varying costs. The optimal system tends to function on the basis of water sales. The TAC will increase as much as 6.6% because of the raise in water price ration. In addition, Fig. 11 depicts
SC
that with higher heating sales price ration from -60% to 60%, overall system TAC will be reduced by 4%. Based on the data shown in Fig. 11, it is evident that an increase in electricity
M AN U
price has less influence on TAC in comparison to other energy carriers.
In the current system, the following resources are used to meet the demand of the consumption section. Electricity supply from the national grid (conventional centralized power plants on the island), fresh water produced from the RO system and heating based on current heat generators is
TE D
dispersed for each consumer. Also, since the systems are designed to just meet the demands of consumers, purchases and sales with the grid are not considered. Table 12 presents the results of the basic parameters of the hub both in an optimal state and current state. In the optimal scenario,
EP
by recovery waste heat from the cement plant, heating, power and water generation in addition to optimizing energy flows and capacity of utilized equipment, fuel consumption will be reduced by
AC C
over half whilst exergy efficiency will increase by 36%. Also, CO2 emission will drop by 21624 ton/year and TAC will decrease by about 37,000 $/year. In a more detailed assessment on each equipment, the results on Table 13 indicate that in the current state, the total amount of fresh water is supplied through the RO system and the electricity is supplied by the grid whilst by adopting the energy hub concept, 18% by MED, 33% by the RO system and 49% is supplied by the water grid. Moreover, part of the heat produced by GT and waste heat is consumed by the
ACCEPTED MANUSCRIPT
steam turbine and MED and the remainder is sold to the grid. On the other hand, about 50% of the plant’s electricity is supplied by the GT and 50% by the steam turbine. Fig. 12 presents changes in CO2 emissions for both cases of including and without energy hub for water and
RI PT
electricity energy carriers. In Iran, CO2 emissions resulting from grid electricity is 0.67 kg/kWh and CO2 emission resulting from natural gases is approximately 0.23 kg/kWh [65]. Also, the amount of CO2 emissions needed to produce heating was 0.38 kg/kWh [73] (an
M AN U
to produce each cubic meter of water [26].
SC
efficiency of 65% was assumed for the boilers). On the other hand, 3 kg of CO2 will be released
AC C
EP
TE D
Table 12. Comparison of results between basic parameters of the optimal scenario and current system on Qeshm Island Parameter Current system Optimized System UPC ($/m3) 0.6738 0.429 Fuel Consumption 4871668.6 2121120 (kg/year) CO2 Emission 224295000 202671000 (kgCO2/year) 38.1 74.45 ηex (%) TAC ($/year) 1141060 1104138
ACCEPTED MANUSCRIPT
224,295 208,000 193,236
RI PT
16,016
202,671
9,169
183
Water
Plant
Total
M AN U
Electricity
279
SC
CO2 Emission (ton/year)
240,000 220,000 200,000 180,000 160,000 140,000 120,000 100,000 80,000 60,000 40,000 20,000 0
Without Hub
With Hub
Fig. 12. Variation of CO2 emissions (kgCO2/year) and energy carrier with energy hub and current system
*
AC C
EP
TE D
Table 13. Assessment of the allocated demand in the hub and comparison with the current system Energy With Waste heat Component Current system carrier recovery MED 0.5828 Water (kg/s) RO 3.2317 1.0686 Grid 1.5804 Boiler 0 GT 3272.3 Heat (kW) Waste heat 5625 Grid -4700* GT 1493.7 Power (kW) Steam Turbine 1489.6 Grid 2988 7.76 In this table, the negative numbers indicate the amount of sales of energy carriers to the grid.
According to the reference [66], 900 tons of carbon dioxide are produced, for every 1,000 tons of cement, about half of which is related to the burning of fuel in the kiln and the other half to the
ACCEPTED MANUSCRIPT
calcining reaction in the preheater. As the amount of cement production in Qeshm cement factory is 231 thousand tons, about 208 thousand tons of carbon dioxide will be produced annually. Regarding heat recovery (5620 kW) at the cement plant which is equivalent to the
RI PT
reduced carbon dioxide for the recovered waste heat, carbon dioxide is declined by 7%. In the case of without energy hub, the CO2 gas produced is considered based on electricity, and water demand from the grid. The demand for electricity based on Fig. 5 is 2988 kW which produces
SC
16015 tons of CO2 per year. With the energy hub, upon optimization and heat recovery from the hub, the consumption of NG is 0.0982 kg/s equivalent to 9210 tons of CO2 per year for power
M AN U
generation in GT which is the main CO2 producer when using the energy hub. On the other hand, the primary energy of desalination systems is captured from the power and heat generated within the energy hub, it is therefore considered to be equivalent to the change in carbon dioxide emissions in the primary energy changes. Then, the amount of reduced water received from the
TE D
grid and utilization of local desalinating systems was assessed. Also, the amount of CO2 generated through water transfer is addressed (Eq. (28)). Due to reduced water purchases from 3.2317 kg/s to 2.114 kg/s, and by taking into account the CO2 generated by the water
EP
transportation, the amount of generated CO2 from the water was reduced by 96 ton CO2 per year. Considering all energy carriers before and after optimization, the environmental impact of
year.
AC C
producing 224 thousand tons of CO2 per year has been reduced to 202 thousand tons of CO2 per
9. Conclusions
In this paper, an amalgamation of water, power and heating generation systems were utilized to optimally develop a modern operational strategy called energy hub. The TAC for the energy hub and energy efficiency was optimized by taking into account the waste heat recovery from the
ACCEPTED MANUSCRIPT
cement plant. The proposed energy hub has the potential to supply of electrical, thermal and water loads. Results from this design were compared to the cases that without waste heat recovery. Moreover, all equipment will be selected using the concept of an energy hub at its best
RI PT
capacity. High energy consumption of major desalinated production processes, reduction of economic costs, reducing greenhouse gases in an integrated water, power and heat system and
factory as an energy hub.
•
After optimization, the exergy efficiency and TAC increased by 20% and decreased by 72.6% respectively.
•
CO2 emissions reduced from 16295 tons per year to 5329 tons per year upon utilizing the
TE D
energy hub concept. •
M AN U
The acquired research results are presented below:
SC
increase the exergy efficiency, are the main reasons for utilizing and optimizing the cement
TAC increases and decreases moderately with increases in water and heating price while it is negligible due to the small amount of grid electricity purchased. The lower the share of water production, the higher the price of water per cubic meter
EP
•
•
AC C
The following may be regarded as suggestions for the continuation of this research: Optimizing the use of the organic Rankine cycle (ORC) system and a comparison of
various working fluids to determine the optimal fluid for the purpose of electricity production from the cement plant’s heat recovery with the concept of an energy hub;
•
The use of hydrogen fuel or biogas to generate power and heat in a CHP and boiler;
•
Combining renewable systems such as wind turbines along with an energy hub for the cement plant to evaluate reduced costs.
ACCEPTED MANUSCRIPT
Appendix A- RO system equations In modeling the reverse osmosis system according to the mass transfer equations, it is evident
RI PT
that the water flow pressure and salt from the membrane are as presented in Eq. (A-1) and Eq. (A-2) and also the mean velocity in each membrane element is determined by Eq. (A-3). The salt concentration of the produced water is determined by Eq. (A-4). Also, the salt concentration in
SC
proximity to the wall is obtained according to the film theory by Eq. (A-5). In regard to the conjunction equation, Eq. (A-6) and Eq. (A-7) can be used for water and Eq. (A-8) can be used
Ú = Î × u Å[r − o − Ú¢ = àá − o â
∆r − (Ý − Ýo )] × 10ß 2
Ú + Ú¢ 8 Ú¢ o = × 1000 ã r + c = o + ( − o )D ä 2 bo = × bg = bÆ − bo bÆ r − bo o g = bg
(A-1) (A-2) (A-3) (A-4) (A-5) (A-6) (A-7) (A-8)
AC C
EP
TE D
=
M AN U
for salt [12]:
In the equations above, A, B and Sw are the permeability coefficients for water, salt, membrane and its area, respectively. The equations that can be used to reduce the number of unknowns are as follows:
[¢ 0.0033 × b" × R × æ ∆r = N × ~ n = 0.04 × lD .Íå × L .~~ ×
(A-9) (A-10)
ACCEPTED MANUSCRIPT bc + br 2 0.2641 × (u + 273) Ý= 10ß − b" =
ã ç è
and Ds is the permeability of salt (m /s), d is the distance between feeds, Qa is 2
RI PT
Where lD =
(A-11)
the average flow determined by Eq. (A-10). Lm is the membrane length and N is number of elements at each PV. π is osmosis pressure and C is salt concentration. In order to estimate the
SC
average pressure drop, the Hagan-Poiseuille equation is used. The Sherwood number is calculated by using Eq. (A-9) and Eq. (A-10) which is used to calculate the polarization of the
Appendix B-sensitivity analysis
M AN U
concentration.
The results from the sensitivity analysis in determining decision parameters in the genetic
AC C
EP
TE D
algorithm are presented below.
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
Fig. B-1. Variation of objective functions versus dispatch factors in energy hub
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. B-2. Variation of objective functions versus decision variables in energy hub
AC C
EP
Table B-1. The results of optimum objective functions for different algorithms parameters in operating condition Level 1 2 3 4 5 6 7 8 Population 50 50 100 100 150 150 200 200 Size Mutation 0.01 0.01 0.02 0.02 0.03 0.03 0.04 0.04 Probability Crossover 0.8 0.9 1 1.1 0.8 0.9 1 1.1 Probability TAC ($/year) 1212562 1197736 1277446 1168324 1104138 1122893 1121454 1158685 Exergy 73.16 77.51 73.61 75.308 74.45 72.302 73.582 76.25 Efficiency (%)
ACCEPTED MANUSCRIPT
References [1] Mohammadi M, Noorollahi Y, Mohammadi-ivatloo B, Hosseinzadeh M, Yousefi H, Torabzadeh Khorasani S. Optimal management of energy hubs and smart energy hubs – A
RI PT
review. Renewable and Sustainable Energy Reviews 2018; 89:33–50. [2] Brahman F, Honarmand M, Jadid S. Optimal electrical and thermal energy management of a residentialenergy hub, integrating demand response and energy storage system. Energy and Buildings 2015; 90:65–75.
[3] Mohammadi S, Moradi-Dalvand M, Ghasemi H, Ghazizadeh M.S. Optimal Design of
SC
Multicarrier Energy Systems Considering Reliability Constraints. IEEE Transaction on Power Delivery 2015; 30:878-886.
M AN U
[4] Bahrami S, Toulabi M, Ranjbar S, Moeini-Aghtaie M, Ranjbar A. A Decentralized Energy Management Framework for Energy Hubs in Dynamic Pricing Markets. IEEE 2017;9:19493053.
[5] AlRafea K, Fowler M, Elkamel A, Hajimiragha A. Integration of renewable energy sources into combined cycle power plants through electrolysis generated hydrogen in a new designed energy hub. International journal of hydrogen energy 2016; 41:16718-16728.
TE D
[6] Togawa T, Fujita T, Dong L, Fujii M, Ooba M. Feasibility assessment of the use of power plant-sourced waste heat for plant factory heating considering spatial configuration. Journal of Cleaner Production 2014; 81:60-69.
[7] Amiri A, Vaseghi M.R. Waste Heat Recovery Power Generation, Systems for Cement
EP
Production Process. IEEE Transactions on Industry Applications 2015; 51:13 – 19. [8] Kabir G, Abubakar A.I, El-Nafaty U.A. Energy audit and conservation opportunities for
AC C
pyroprocessing unit of a typical dry process cement plant. Energy 2010; 35:1237–1243. [9] Renewable Energy Organization of Iran. www.satba.gov.ir. (Access time: 2017) [10] Madlool N.A, Saidur R, Hossain M.S, Rahim N.A. A critical review on energy use and savings in the cement industries, Renew. Sustain. Energy Rev. 2011; 15:2042–2060. [11] Salimi M, Amidpour M. Investigating the integration of desalination units into cogeneration systems utilizing R-curve tool. Desalination 2017; 419:49-59. [12] Mokhtari H, Sepahvand M, fasihfar A. Thermoeconomic and exergy analysis in using hybrid
systems
(GT+MED+RO)
Gulf. Desalination 2016; 399:1-15.
for
desalination
of
brackish
water
in
Persian
ACCEPTED MANUSCRIPT
[13] Shahzad M. W, Burhan M, Ng K.C. Pushing desalination recovery to the maximum limit: Membrane and thermal processes integration. Desalination 2017; 416:54-64. [14] Nemati A, Sadeghi M, Yari M. Exergoeconomic analysis and multi-objective optimization of a marine engine waste heat driven RO desalination system integrated with an organic Rankine
RI PT
cycle using zeotropic working fluid, Desalination. 2017; 422:113-123.
[15] Reynolds J, Ahmad M. W, Rezgui Y, Hippolyte J. Operational supply and demand optimization of a multi-vector district energy system using artificial neural networks and a genetic algorithm. Applied Energy 2019; 235:699–713.
SC
[16] Moeini-Aghtaie M, Fotuhi-Firuzabad M. Multiagent Genetic Algorithm: An Online Probabilistic View on Economic Dispatch of Energy Hubs Constrained by Wind Availability.
M AN U
IEEE Transactions on sustainable energy 2014; 5:699–708.
[17] Ko M. J, Kim Y. Sh, Chung M. H, Jeon H. Ch. Multi-Objective Optimization Design for a Hybrid Energy System Using the Genetic Algorithm. Energies 2015; 8:2924-2949. [18] Kampouropoulose K. Multi-objective optimization of an energy hub using artificial intelligence, A thesis for the degree of Doctor of Philosophy in Electrical Engineering, Universitat Politècnica de Catalunya, Barcelona, Catalonia, Spain, 2018.
TE D
[19] Stewart B, Hoang M, Zarzo D, Olewniak F, Campos E, Bolto B, Barron O. Desalination techniques: A review of the opportunities for desalination in agriculture. Desalination 2015; 364:2–16.
[20] Lattemann S, Kennedy M.D, Shippers J.C, Amy G. Chapter 2 Global Desalination
EP
Situation. Sustainability Science and Engineering 2010; 2:7–39. [21] IDA Desalination Yearbook 2016–2017, Water Desalination Report. Global water
AC C
intelligence 2016.
[22] Voutchkov N. Energy use for membrane seawater desalination – current status and trends. Desalination 2018; 431:2-14. [23] Shalaby S.M. Reverse osmosis desalination powered by photovoltaic and solar Rankine cycle power systems: A review. Renewable and Sustainable Energy Reviews 2017; 73:789–797. [24] Ng K.C, Shahzad M.W. Sustainable desalination using ocean thermocline energy. Renewable and Sustainable Energy Reviews 2018; 82:240–246. [25] Mokhtari H, Ahmadisedigh H, Ebrahimi I. Comparative 4E analysis for solar desalinated water production by utilizing organic fluid and water. Desalination 2016; 377:108–122.
ACCEPTED MANUSCRIPT
[26] Salcedo R, Antipova E, Boer D, Jiménez L, Guillén-Gosálbez G, Multi-objective optimization of solar Rankine cycles coupled with RO desalination considering economic and life cycle environmental concerns. Desalination 2012; 286:358–371. [27] Francey R.J, Trudinger C.M, Schoot M, Law R.M, Krummel P.B. Atmospheric verification
RI PT
of anthropogenic CO2 emission trends. Nature Climate Change 2013; 3:520–524.
[28] Peters G.P, Andrew R.M, Boden T, Canadell J.G, Ciais P, Quéré C. Le, Marland G, Raupach M.R, Wilson C, the challenge to keep global warming below 2 °C. Nature Climate Change 2013; 3:4–6.
address operational constraints. Energy 2014; 73:387–98.
SC
[29] Evins R, Orehounig K, Dorer V, Carmeliet J. New formulations of the energy hub model to
M AN U
[30] Sharif A, Almansoori A, Fowler M, Elkamel A, Alrafea K. Design of an energy hub based on natural gas and renewable energy sources. International Journal Energy Research 2014; 38:363–73.
[31] Rastegar M, Fotuhi-Firuzabad M. Load Management in a Residential Energy Hub with Renewable Distributed Energy Resources, Energy and Buildings 2015;107:234-242. [32] Maroufmashat A, Fowler M, Khavas S.S, Elkamel A, Roshandel R, Hajimiragha A. Mixed
TE D
integer linear programing based approach for optimal planning and operation of a smart urban energy network to support the hydrogen economy. International Journal of Hydrogen Energy 2015;41: 7700-7716.
[33] Maroufmashat A, Elkamel A, Fowler M, Sattari S, Roshandel R, Hajimiragha A. Modeling
EP
and optimization of a network of energy hubs to improve economic and emission considerations. Energy 2015; 93:2546–2558.
AC C
[34] Najafi A, Falaghi H, Contreras J, Ramezani M. Medium-term energy hub management subject to electricity price and wind uncertainty. Applied Energy 2016; 168:418–433. [35] Majidi M, Nojavan S, Zare K, A cost-emission framework for hub energy system under demand response program, Energy 2017;134;157-166. [36] Davatgaran V, Saniei M, Mortazavi S. Optimal bidding strategy for an energy hub in energy market, Energy 2018;148:482-493. [37] Amiri S, Honarvar M, sadegheih A. Providing an Integrated Model for Planning and Scheduling Energy Hubs and Preventive Maintenance, Energy 2018;163:1093-1114.
ACCEPTED MANUSCRIPT
[38] Ramos-Teodoro J, Rodríguez F, Berenguel M, Torres J, Heterogeneous resource management in energy hubs with self-consumption: Contributions and application example, Applied Energy 2018;229:537–550. [39] Ghaffarpour R, Mozafari B, Ranjbar A, Torabi T. Resilience oriented water and energy hub
RI PT
scheduling considering maintenance constraint, Energy 2018;158:1092-1104.
[40] Aghamohamadi M, Samadi M, Rahmati I. Energy Generation Cost in Multi-Energy Systems; an Application to a Nonmerchant Energy Hub in Supplying Price Responsive Loads, Energy 2018;161:878-891.
SC
[41] Geidl M, Koeppel G, Favre-Perrod P, Klöckl B .The Energy Hub – A Powerful Concept for Future Energy Systems. Third Annual Carnegie Mellon Conference on the Electricity Industry
M AN U
2007:13 – 14.
[42] Goldstein L, Hedman B, Knowles D, Freedman S.I, Woods R, Schweizer T. Gas-Fired Distributed Energy Resource Technology Characterizations, National Renewable Energy Laboratory (NREL), 2003. [43]
Chorak
A,
Palenzuela
P,
Alarcón-Padilla
D.C, Abdellah
A. B.
Experimental
characterization of a multi-effect distillation system coupled to a flat plate solar collector
TE D
field: empirical correlations. Applied Thermal Engineering 2017; 120:298-313. [44] Water & Process Solutions, FILMTEC™ Reverse Osmosis Membranes Technical, Dow Water & Process Solutions - The Dow Chemical Company. (Access time: 2011) [45] Guideline for drinking Water Quality, Fourth edition incorporation the first addendum,
EP
World Health Organization. 2017.
[46] Asgari S, Noorpoor A.R, Boyaghchi F.A. Parametric assessment and multi-objective
AC C
optimization of an internal auto-cascade refrigeration cycle based on advanced exergy and exergoeconomic concepts. Energy 2017; 125:576-590. [47] Rosen M, Dincer I, Kanoglu M. Role of exergy in increasing efficiency and sustainability and reducing environmental impact. Energy Policy 2008; 36:128-137. [48] Ameri M, Mokhtari H, Mostafavi Sani M. 4E analyses and multi-objective optimization of different fuels application for a large combined cycle power plant. Energy 2018; 156:371-386. [49] Boyaghchi F. A, Chavoshi M, Sabeti V. Multi-generation system incorporated with PEM electrolyzer and dual ORC based on biomass gasification waste heat recovery: Exergetic, economic and environmental impact optimizations. Energy 2018; 145:38-51.
ACCEPTED MANUSCRIPT
[50] Sharqawy M.H, Lienhard J.H, Zubair S.M. On exergy calculations of seawater with applications in desalination systems, International Journal of Thermal Sciences 2011;50:187196. [51] Kotas T. J. The Exergy Method of Thermal Plant Analysis. 1985.
RI PT
[52] Ghasemi A, Heidarnejad P, Noorpoor A.R. A novel Solar-Biomass Based Multi-Generation Energy System Including Water Desalination and Liquefaction of Natural Gas System: Thermodynamic and Thermoeconomic optimization. Journal of Cleaner Production 2018; 196:424-437.
SC
[53] Piacentino A. Application of advanced thermodynamics, thermoeconomics and exergy costing to a Multiple Effect Distillation plant: In-depth analysis of cost formation process.
M AN U
desalination 2015; 371:88–103.
[54] Hajabdollahi H. Investigating the effects of load demands on selection of optimum CCHPORC plant. Applied Thermal Engineering 2015; 87:547-558
[55] Breeze P. The Cost of Power Generation The current and future competitiveness of renewable and traditional technologies. 2010.
[56] Sanaye S, Khakpaay N. Simultaneous use of MRM (maximum rectangle method) and
TE D
optimization methods in determining nominal capacity of gas engines in CCHP (combined cooling, heating and power) systems. Energy 2014;72:145-158. [57] Web Site of Cement Qeshm Factory, www.qeshmcement.com. (Access time: 2016) [58] Web Site of Iranian Cement Factories, www.irancement.com. (Access time: 2017)
EP
[59] Boyaghchi F. A, Chavoshi M. Monthly assessments of exergetic, economic and environmental criteria and optimization of a solar micro-CCHP based on DORC. Solar Energy
AC C
2018; 166: 351-370.
[60] Web Site of Iranian Ministry of Energy, www.moe.gov.ir. (Access time: 2018) [61] Hajabdollahi H, Ganjehkaviri A, Jaafar M.N. Assessment of new operational strategy in optimization of CCHP plant or different climates using evolutionary algorithms. Applied Thermal Engineering 2015; 75:468-480. [62] Mabrouk A, Nafey A.S, Fath H.E.S. Thermoeconomic analysis of some existing desalination processes. Desalination 2007; 205:354–373. [63] Web Site of Iranian Fuel Conservation Organization, www.ifco.org. (Access time:2018)
ACCEPTED MANUSCRIPT
[64] Du Y, Xie L, Liu J, Wang Y, Xu Y, Wang S. Multi-objective optimization of reverse osmosis networks by lexicographic optimization and augmented epsilon constraint method. Desalination 2014; 333:66–81 [65] Energy balance sheet of year 94, Power and Energy Affairs of Iran, Office of Planning for
RI PT
Electricity and Energy. 2015.
AC C
EP
TE D
M AN U
SC
[66] United States Environmental Protection Agency, www.epa.gov. (Access time:2018)
ACCEPTED MANUSCRIPT
Highlights •
Proposing of the modified energy hub model to analyze economic and environmental
RI PT
performance Finding the optimum range of variables based on TAC and exergy efficiency
•
Water is incorporated as an energy stream by simultaneously utilizing MED and RO.
•
The results of optimization are compared to the conventional mode.
AC C
EP
TE D
M AN U
SC
•
S0360-5442(19)30447-5
DOI:
https://doi.org/10.1016/j.energy.2019.03.043
Reference:
EGY 14879
To appear in:
Energy
Received Date: 11 July 2018 Revised Date:
5 March 2019
Accepted Date: 7 March 2019
Please cite this article as: Sani MM, Noorpoor A, Shafie-Pour Motlagh M, Optimal model development of energy hub to supply water, heating and electrical demands of a cement factory, Energy (2019), doi: https://doi.org/10.1016/j.energy.2019.03.043. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Optimal model development of energy hub to supply water, heating and electrical demands of a cement factory
1
RI PT
Mostafa Mostafavi Sani1, Alireza Noorpoor1*, Majid Shafie-Pour Motlagh1 School of Environment, College of Engineering, University of Tehran, Tehran, Iran
Abstract
This paper uses the energy hub concept to meet the electricity, heat and water demand of a
SC
cement plant. Given the cement plant’s high potential for waste heat recovery so the use of recycled waste heat was compared to a state where such waste heat was not used by taking
M AN U
exergy, economic and environmental analyzes into account. The proposed energy hub consists of electricity, heat and desalinated water production equipment. Energy hub is optimized based on the Genetic Algorithm (GA) with the goal of supplying demand, as well as reducing costs and pollutants and increasing the exergy efficiency which ultimately will be selected using the concept of an energy hub at its optimal capacity. By comparing the two energy supplying
TE D
systems of the current case study and optimal energy hub, results showed that the Total Annual Cost (TAC) level decreased by about 37,000 $/year and exergy efficiency increased by 36%. CO2 emission will also decrease by about 21,620 tons/year. Moreover, in the energy hub concept, the waste heat recovery and ignoring waste heat scenarios were compared. Results
EP
indicate that upon optimization, the average of TAC when waste heat is recovered, is 27% of the condition when waste heat is ignored.
AC C
Keywords: Energy hub; Multi-Effect Desalination (MED); Reverse Osmosis (RO); Total Annual Cost (TAC); Genetic Algorithm.
*
Corresponding author: Alireza Noorpoor, Associate Professor, School of Environment, College of Engineering, University of Tehran, Tel: +98-021-66476132 E-mail address: [email protected]
ACCEPTED MANUSCRIPT
Nomenclature List of acronyms
v
Dispatch factors (-)
Combined-Cycle Power Plant
Vw
Permeate velocity [m/s]
CHP
Combined Heat and Power
wp
Pump Power (kW)
CRF
Capital Recovery Factor
Z&
Capital cost rate ($/year)
CTGs
Combustion Turbine Generators
Zk
Component purchase cost ($)
EW
Energy consumption per Water production in RO
ZO&M
Operation and Maintenance Cost
GOR
Gained Output Ratio
Subscripts
GT
Gas Turbine
amb
LHV
Lower Heat Value (kJ/kg)
b
MED
Multi-Effect Desalination
ORC
Organic Rankine Cycle
PAR
Peak to Average Ration
SC Ambient brine
M AN U
NSGA
RI PT
CCPP
boi
Boiler
bp
Brine Pump
CC
Combustion Chamber
ch
Chemical
CV
Control Volume
Non dominated elitist and rapid Sorting Genetic
Algorithm Natural Gas
MSF
Multi-stage flash distillation
D
Distraction
SDI
Silt Density Index
dmn
Demand
TE D
NG
Total Annual Cost ($/year)
e
Exergy
WHO
World Health Organization
em
emission
extra
Waste heat from cement plant
Water permeability coefficient [kg/m2·s·Pa]
f
Feed
Solute transport coefficient [kg/m2·s]
h
Heat
Coupling matrix
hot
Hot water
Salt concentration [ppm]
hpp
High Pressure Pump
Cw
Brine concentration [kg/m3]
i
inlet
CC
RO capital cost subsystems
m
Material
d
Feed channel equivalent diameter [m]
max Boi Maximum Boiler
A B Cij C
AC C
List of symbols
EP
TAC
ACCEPTED MANUSCRIPT
ds
Distance from the ideal point
NG
Natural Gas
Ds
Solute diffusivity [m2/s]
p
Permeate stream
e, ex
Exergy (kJ/kg)
pu
Ex
Exergy flow (kW)
Sc
EWRO
Electrical consumption per water production
pipeline Pipe line
Fij
Non-dimensionalize objective indexes
Ph
Physical
Q
Heat Sell
Material correction factor
S
Jw
Local permeate flux [kg/m2·s]
ST
Js
Local solute flux [kg/m2·s]
SW
K
Mass transfer coefficient [m/s]
L
Output loads of the hub
LDe
Output electrical loads of the hub (kW)
LDh
Output heat loads of the hub (kW)
LDw
Steam Turbine Sea Water
M AN U
f
RI PT
Interest rate
Schmidt number
SC
int
Pump
Sea Water Intake and Pre-treatment
T
Turbine
trf
Transfer
u
Number of exergy carrier
Output water loads of the hub (kW)
x
objective’s index
N
Operation hours in a year
w
Membrane wall
mf
Feed water flow rate (kg/s)
z
index for each Pareto front point
mhot
Hot water flow rate (L/s)
0
P
Input carriers of the hub
Greek symbols
PE
Exchanging electricity of the hub (kW)
α,β,ω,ε
Type of energy carrier
Exchanging heat of the hub (kW)
η
Efficiency
Exchanging water of the hub (kW)
ρ
Density (kg/ m3), Integer Numbers
Input vector of the hub
ξ
Chemical exergy/energy ratio
Output vector of the hub
ε
Total cycle exergy efficiency (-)
Q
Heat flow (kW)
τ
Hours of operation per year (h)
QB
Brine flow rate (m3/h)
φf
Fuel cost ($/kg)
QP
permeate flow rate (m3/h)
φe,s
Selling electricity cost ($/kWh)
PW Pi Pj
EP
AC C
PH
TE D
SWIP
Standard Situation
ACCEPTED MANUSCRIPT
Recovery (%)
ρp
Density of the permeate [kg/m3]
Re
Reynolds number
π
Osmotic pressure [MPa]
Membrane area (m2)
φe,b
Buying electricity cost ($/kWh)
Thot
Hot water temperatures (˚C)
ψem
Pollutant emission cost ($/kg)
RI PT
RRO
1. Introduction
The point of confrontation energy carriers, conversion equipment, energy storage and energy
SC
demand is known as the energy hub. The energy hub is a new and promising concept to
M AN U
optimally manage multiple energy carrier systems. The energy hub has significant potential for realizing energy system models and moves towards sustainable multi-energy systems [1]. Energy hub utilization will lead to 1) reduction in energy consumption 2) simplicity and high accuracy in multiple system carrier analysis and assessment 3) increased reliability 4) multi-dimensional analysis of energy systems (technical, economic, environmental). The main element in defining
TE D
energy networks is the energy hub. Numerous researches have been conducted on the hub. A residential energy hub model was proposed by Brahman et al. [2] that obtains solar radiation, natural gas (NG) and electricity as inputs to provide cooling, heating and electricity at its output
EP
port. An optimization issue was dealt with regarding electrical and thermal energy management
AC C
with the aim of reducing total energy costs of customer’s input in terms of NOx, CO2 and SOx emissions. Mohammadi et al. [3] addressed the advantages of integrating options such as smart technologies, storage systems, multi-generation systems, renewable energy resources, distributed energy resources and demand side management by proposing the smart energy hubs concept. In order to utilize zero emissions renewable sources into dispatchable combined cycle power plants, Bahrami et al. [4] fabricated a real time energy hub scheduling issue within a dynamic pricing market. The interaction of the energy hub is modelled in the form of a precise potential game for
ACCEPTED MANUSCRIPT
energy hub payment optimization purposes concerning gas and electricity utilities. The results of the simulations indicate that the suggested algorithm may enhance the average payoff of the energy hub by 18.8%. Moreover, companies that provide energy services may enhance energy
RI PT
network technical performances through a reduction of the peak to average ration i.e. PAR for gas and electricity by 7% and 27%, correspondingly. AIRafea et al. [5] proposed a modern energy hub. For this energy hub, hydrogen is introduced as an energy vector to replace coal fired
SC
Nanticoke power plant. With the addition of hydrogen into natural gases as a fuel for combined cycle power plants, additional costs and annual revenue reduction will be the consequences. CO2
M AN U
and NOx emission credits will also increase. In order to start increasing annual revenue, electricity should be sold at 0.12 $/kWh with hydrogen at 3.7% in fuel of combined-cycle power plant (CCPP). To maximize the employment of waste heat from coal fired thermal plants for the benefit of factories in the vicinity, Togawa et al. [6] proposed a stimulation process model that
TE D
integrates land use control, technology systems and spatial analysis. The acquired results showed that waste heat systems exhibited several environmental advantages compared to the individual boiler system. By employing waste heat, the consumption of fuel oil is reduced by 16.05 TJ/year
EP
and there is a CO2 emissions reduction of 1204 tons per year. Another solution to reduce energy consumption and environmental consequences is to use waste
AC C
heat from factories and to implement a comprehensive integrated system in remote areas suffering from water shortages. This paper addresses waste heat from a cement plant. The process of producing cement is extremely energy intensive with approximately 3-4 GJ energy consumption for each ton of produced cement. Furthermore, Iran is the fourth largest cement manufacturer in the world after China, India and the USA with an annual production of 75 million tons [7]. 40% to 60% of cement production costs are accredited to energy consumption
ACCEPTED MANUSCRIPT
[8]. 3% of energy production is accredited to the cement industry in Iran [9]. Within a cement factory, electrical energy is 25% of the total energy consumption while heat energy accounts for approximately 75% of total energy consumption [10]. In the case of integrated systems and the
RI PT
use of waste heat, the work of Salimi and Amidpour [11] can be referred to, which, based on the R-curve, examined the integration of different water desalination systems. Mokhtari et al. [12] proposed the gas turbine (GT) +MED+RO water desalination method based on exergy and
SC
thermo-economics analysis for the Persian Gulf region. The system was designed on the basis of the simultaneous generation of electricity and water and could reduce total water production
M AN U
costs. Muhammad Wakil Shahzad et al. [13] proposed a combination of absorption and RO systems based on low temperature industrial heat sources or heat from the sun to address lack of water in the Persian Gulf region, which resulted in 81% efficiency and energy consumption of 1.76 kWh/m3. Arash Nemati et al. [14] utilized waste heat from a large diesel ship engine to
TE D
produce desalinated seawater. On the basis of the radiation curve amidst optimization, the obtained efficiency was 37% and the cost of exergy destruction was 60 $/GJ. Reynolds et al. [15] suggested a district energy management optimization strategy for the optimization of district heat
EP
generated from a multi vector energy center. The outcomes show the possible gain when energy is managed in a holistic state, taking both supply and demand into account as well as multiple
AC C
energy vectors. Moeini-Aghtaie et al. [16] proposed an online economic dispatch for multiple energy carriers to achieve a probabilistic economic dispatch optimization model. This model is addressed using a stable optimization mechanism i.e. multi-agent genetic algorithm with an advantageous property of finding the global optima of the economic dispatch issue. Ko et al. [17] recommended a method for the optimization of an energy system whilst concurrently reducing life cycle cost and increasing renewable energy penetration as well as reducing annual
ACCEPTED MANUSCRIPT
greenhouse gas discharge. A non-dominated elitist and rapid Sorting Genetic (NSGA) is used for the purpose of multi objective optimization. The Pareto-optimal solution evolution based on the variations in the technical, economic and environmental objective functions via generations is
RI PT
examined. Kampouropoulose [18] suggested a method and associated software tools to derive the optimal energy flow pertaining to a multi-carrier system to meet energy needs and reduce a number of defined optimization criterion on the basis of the Genetic Algorithm.
SC
Given that fifty-nine percent of drinking water is supplied from the sea through desalination processes [19], various methods have been developed for the production of freshwater from the
M AN U
sea across the world. Energy expenditures (thermal, electricity) in water desalination systems are as follows: RO
TE D
this capacity [21, 22]. The energy used for RO system is dependent on the salinity of the water as well as the system used for energy recovery. For seawater, the energy consumption with and without the recycling device is 12 and 4 kWh/m3, respectively. However, with the development
EP
and improved performance of RO systems, the price of desalinated water has decreased from 2 $/m3 to roughly 0.8 $/m3[19, 23]. However, the MED thermal water desalination system is still
AC C
developing and has retained its position across the world. MED system has a heating energy consumption of 71.7 kWh and electricity consumption of 2.3 kWh [24]. Considering the high energy consumption of water desalination systems as well as rising fresh water demand (from 77.4 in 2014 to 92 million in 2018), this technology has become a major source of CO2 emissions [24]. In this regard, CO2 production is estimated at 100 million tons per year [19, 24]. In other words, for each cubic meter of fresh water, about 3 kg of greenhouse gas is produced
ACCEPTED MANUSCRIPT
[25, 26]. The increasing trend in 2040 will bring greenhouse gas emissions to 218 tons per year and an approximate temperature increase of 2 °C [27, 28]. Due to the high energy consumption of RO and MED, in order to 1) reduce costs 2) reduce greenhouse gas production 3) increase
RI PT
productivity and 4) reduce gas and oil consumption in energy rich areas (such as the Persian Gulf), waste heat from factories can be used to produce desalinated water and meet regional demands (power generation and water heating). The existence of rich gas and oil resources has
SC
led to industrial development in this region. Also, population dispersion in the Persian Gulf due to weather conditions as well as the existence of various islands in the area has become an issue
M AN U
for governments to meet the existing energy demand of residents. In regard to the water production along with other energy requirements is an energy-intensive process, it is logical to define an energy hub.
In the case study, the consumption requirements of various energy carriers in the cement factory
TE D
and in the nearby city should be satisfied. Thus, energy production and consumption management as well as the selection of equipment with optimal capacity are considered. In addition, the priority in energy systems is to maximize efficiency. In this research, the optimal
EP
utilization of waste heat in the cement factory which increases efficiency and reduces costs and emissions is taken into account. However, due to the lack of water desalination systems in
AC C
the current state, drinking water is transported to the place of consumption through tankers, which, in addition to production and transportation costs, entails accessibility and sanitation issues. In this study, by constructing desalination systems, electricity and heat generation at the place of consumption has not only significantly contributed to the supply of energy carriers, but also improved energy efficiency whilst reducing costs and emissions through optimal operation. In regard to the aforementioned points and the necessity to supply multiple energy carriers
ACCEPTED MANUSCRIPT
including electricity, heat and desalinated water, the case study is modelled using the concept of energy hub. In order to supply energy carriers with minimum costs and emissions along with maximum exergy efficiency, a methodology is presented in Fig. 1. According to this figure,
RI PT
the decision making variables and inputs amidst the modelling section are entered to the system
energy carriers in each of the energy hub equipment.
Weather condition
Inputs
Dispatch factors
Decision Variables
RRO
M AN U
Text
Design Hub
Demands
System of equations (Table 2)
Text
Mhot
EWRO QBoiler
Design factors
TE D
Thot
CHP Model
Outputs
PeCHP, PeST, PeRO QCHP, QST, Qb mRO, mMED Text
Genetic Algorithm
SC
of equation to determine production and consumption variables for electricity, heat, and water
PE, PSW, PH Total investment cost
O&M Cost
Objective function
Capacities
TAC
Boiler Model
Objective function
AC C
EP
Exergy efficiency
MED Model RO Model ST Model CO2 Emission Number of Equipment
Fig. 1. Description of modelling methodology
ACCEPTED MANUSCRIPT
Upon modelling each of the equipment and determining energy production and consumption,
RI PT
quantity and capacity of equipment, operation and maintenance costs, and quantity of produced emissions, the TAC can be created. Moreover, exergy efficiency may also be extracted. Non dominated elitist and rapid Sorting Genetic Algorithm (NSGA-II) was implemented to present a
SC
series of Pareto multiple optimal solutions. Changes in optimal value sensitivity analysis, exergy efficiency and TAC as well as changes in design factors were assessed. Finally, upon integrated
M AN U
optimization of the energy hub with the purpose of reducing costs and increasing exergy efficiency, the optimal capacity of the equipment and assigned energy carriers are determined. By assessing various papers on energy hubs and separating the subject areas of these papers (Table 1), some of the research novelties are mentioned below:
Coupling matrix parameters such as efficiencies, recoveries and gained output ratio
TE D
•
(GOR)are included variables by considering design parameters; •
Water flow is defined as an energy carrier in the energy hub, such that two water
EP
desalination systems consisting of MED and RO are simultaneously utilized to provide the means for optimal electricity and heat allocation to produce the required desalinated
•
AC C
water upon optimization;
Considering multi-objective optimization in terms of exergy efficiency and total annual
cost in an energy hub system with self-consumption.
ACCEPTED MANUSCRIPT
Table 1. review of various papers on energy hubs* Coupling matrix arrays
SC M AN U TE D EP
AC C
*
Industrial
2018
Residential
2018
Applications Optimization
Energy
2018
Economical
2018
Exergy
2018
Environmental
2017
Reliability
2016
Constant
2015
Variable
2015
Natural gas
2015
Hydrogen
2015
Water
2014
Cooling
2014
M. Geidl et al. [41] R. Evins et al. [29] A. Sharif et al. [30] M. Rastegar et al. [31] A. Maroufmashat et al. [32] A. Maroufmashat et al. [33] F. Brahman et al. [2] A. Najafi et al. [34] M. Majidi et al. [35] V. Davatgaran et al. [36] S. Amiri et al. [37] J. Teodoro et al. [38] R. Ghaffarpour et al. [39] M. Aghamohamadi et al. [40]
Heating
2007
Authors
Power
Year
Analyses
RI PT
Energy hub carriers
Objective functions
Cost & Emission Emission Cost Cost Cost Cost & Emission Cost & Emission Cost Cost & Emission Cost Cost & Reliability
In the table above, the filled in boxes represent the studies conducted in that field. As evident, variable arrays for the coupling matrix are not considered in any previous studies.
Cost Cost Cost
ACCEPTED MANUSCRIPT
2. Thermal modeling 2.1.Energy Hub In order to achieve an integrated system, a modeling and analysis framework for multi-carrier
RI PT
energy systems must be developed. The current model is between converter devices based on the efficiency of the output-to-input ratio. When there are numerous outputs, a conversion matrix or coupling matrix is defined which represents the transformation of energy carriers from the input
SC
to the hub outputs. If necessary, storage devices are used that take time and energy into account [41]. The energy conversion between different energy hub ports is typically shown using a
M AN U
coupling matrix (C). The relationship between different input and output ports for an example hub is shown in Fig. 2.
piα pψi pi
• • •
TE D
ω
pαj p βj • • •
pωj
EP
Fig. 2. Relationship between the input and output ports of an energy hub with a Cij coupling matrix [41]
AC C
The exchangeable power in inlet and outlet ports are expressed by Pi and Pj vectors respectively and these power vectors are related via the coupling matrix by the following equations. In this equation, the Cij is the coupling matrix array which according to the characteristics of the existing elements, associate the hub inputs and outputs (Eq. (1)) [41]. ⋯ " ⋱ ⋮ * ⋮ + = 0 ⋮ + ⋮ ⋯ $%%%%&%%%%'
()
)
(1)
ACCEPTED MANUSCRIPT
There are two important properties in the coupling matrix arrays: 1- According to the energy conservation law, energy is converted from one form to another, but in the process of conversion, there is also some energy wasted, so all coupling matrix arrays fall
0 ≤ ≤ 1
∀(3, 5)78
/
and
RI PT
into the maximum or minimum values according to the following equation (Eq. (2)) [41]. 9, : 7 ;
SC
In this equation:
∀(3, 5) 7 8 = <(1,2), (3,4), … A , 9, :, B 7 ; =
(2)
(3)
M AN U
2- According to the total energy conservation law, all outputs converted from a given input must be less than or equal to the input value. Therefore, the total arrays of each column in the coupling matrix must be less than or equal to the relation below (Eq. (4)) [41]. /
≤1
∀(3, 5)78
9, : 7 ;
(4)
TE D
0 ≤ ∑/PQ
If the output loads of the hub are represented by L and the input carriers of the hub are represented by P, the previous equation can be rewritten as follows:
S
(
EP
R ⋯ ⋱ ⋮ +* ⋮ + * ⋮ +=* ⋮ R ⋯ $%%%%&%%%%'
(5)
AC C
As long as the coupling matrix elements, known as an efficiency of energy hub equipment, have constant and independent values, Eq. (5) is linear. However, considering the dependence of the equipment efficiency on the input energy, this equation will be nonlinear [41]. Since one type of input energy into the hub may be split into several parts and enters various converters within the hub, coefficients named dispatch factors are defined which represent the amount of input energy assigned to each of the hub convertors. In multi-input and multi-output
ACCEPTED MANUSCRIPT
hubs, the number of coupling matrix elements depends on the dispatch factors as well as the converter equipment efficiency. Fig. 3 shows the industrial plant configuration evaluation for this section. Commercially available water oriented industry technology is the basis of this design.
RI PT
This plant consists of a MED for potable water production, RO plant, steam turbine, and combustion turbine generators (CTGs) and a boiler. The plant scope is identified and specified according to the aim of this research and multigenerational factors. An alternative energy source
SC
and high efficiency are proven properties of this plant to minimize CO2 emissions. Fig. 3 presents thermal plant overall system flow. Three varying potential energy sources were taken
M AN U
into account. Waste heat in cement plant is the first source. The second source is exhausted gas from the GT stack. The third source is heat energy from the boiler. Electrical load requirements for the energy hub can be met using energy sources such as electrical grid, steam turbine, and GT systems. The dispatch factors vMED, vBoi, vCHP, vST and vRO are in interval at all times [0, 1]. As
TE D
previously stated, the energy hub model is considered for multi-energy carrier systems and represents continuity equations by considering a control volume such that each input energy carrier to the system is multiplied by the coupling matrix coefficient, exhibiting its influence on
EP
the considered output load so that the sum of these inputs determines the output load by taking the energy hub productions and consumptions into account. The important point is that the
AC C
output loads are known as specified quantity in each energy hub which ultimately, by expanding the equations (Table 2) and implementing decision variables and assumptions for each equipment and energy converters, the required input energy to the system, the coupling matrix coefficients, and hence the capacities of each equipment can be determined. The decision variables for each equipment and dispatch factors are related to the considered objective function in optimization and will be determined accordingly. Thermal, electrical and water loads along
ACCEPTED MANUSCRIPT
with waste heat from the cement plant are shown as LDH, LDe, LDW and Qextra respectively which are determined according to Fig. 5 and section 5. Section 7.1 shows decision parameter values, vMED, vBo, vST, vRO, RRO, Thot, Mhot, EWRO, Qmax,Boi which are determined by the GA
RI PT
optimization method given their limits. The values of GT efficiency (ηCHP) and steam cycle efficiency (ηST) are obtained according to the reference curve of the power plant characteristics. The GOR value will also be determined from section 2.3 from the Thot and Mhot decision
SC
variables. According to Table 2, 12 variables can be determined from 12 equations and the energy hub can be obtained. In order to further highlight the aforementioned points, Fig. 1
M AN U
depicts a schematic of the paper’s modelling process. Based on this diagram, the weather condition along with the water, electricity, and heat demands as well as the decision variables are entered to the system equation shown in Table 2 as known values. All energy vectors pertaining to the energy equipment are specified in the hub. These values are entered in the equipment
TE D
modelling process. The output determines the equipment’s capacity, their quantity and other required parameters relevant to the TAC and exergy efficiency. Finally, by optimizing this model, the design factor, equipment and their quantities for the optimal objective functions will
EP
be determined.
In Table 2, row 1 of the energy continuity equation for electric current is presented in Fig. 3. In
AC C
GT power generators and steam turbines, these inputs should be equal to electrical consumption of RO and the power consumption of the hub. Rows 2 and 3 are auxiliary equations that are extracted according to their definition. In row 4 of this table, the hub water affinity equation is presented. The water produced by the RO and MED system should be equal to the demand for water.
ACCEPTED MANUSCRIPT
RI PT
Ex& Heat , χ
AC C
EP
TE D
M AN U
Ex& Power , χ
Ex&Water , χ
Ex&dmn ,1
SC
Ex&NG
LD H
Fig. 3. Schematic view of energy hub used in case study
LDe
Ex&dmn,2
LDW
Ex&dmn ,3
ACCEPTED MANUSCRIPT
Table 2. Design equation for energy hub model [6].
4 5 6 7
8 9 10
11
^YZ + ^_W` + Ua − R[ = 0
^YZ , ^_W` , Ua
^YZ = ]YZ (^_W` + Ua ) ^YZ = ]_W` (^YZ + Ua )
b(T − bUV + bc + T − b_W` + bWdef" − R[T = 0 bc = ]gh (b(T − bUV + T − b_W` + bWdef" ) DYZ = KNYZ ^YZ b(T =
(1 − i(T ) D(T i(T
jkl =
^_W` 2326 n5 b_W` 1 nH
bUV =
DUV iUV
](T
R[
b(T , bUV , bc , T
bWdef" , R[T
i(T
]UV
]YZ
]_W` ]gh
KNYZ
iUV
jkl
EP
12
DUV = ]UV (D(T + W − DYZ )
Decision variables
RI PT
3
D(T = ](T (DUV + W − DYZ )
D(T , DUV , W , DYZ
Specified quantity R[\
SC
2
D(T + DUV + W − DYZ − R[\ = 0
Unspecified quantity
M AN U
1
Equations
TE D
Row
AC C
In row 7, the heat produced by the GT, boiler and waste heat from the cement plant are heat generators in the hub, which should provide the heat required by the steam turbine, MED and hub thermal demand. Also, the current (PE), water (PSW) and heat (PH) input to the hub from the grid or the output from the hub to the grid can be placed with a positive and negative sign respectively. All equations of rows 2, 3, 4, 5, 6, and 8 are defined by dispatch factors vCHP, vST, vRO, vMED and vBoi which are decision variables and determine the amount of energy distribution in
each of the energy carriers within the hub equipment.
ACCEPTED MANUSCRIPT
2.2.RO modeling Two modeling methods for the RO system are proposed by FILMTEC [44]. The first method is
RI PT
element to element. In the second method, if the feed quality, temperature, and intensity of the purified water flow and the number of elements are known, the average value is used to calculate the feed pressure and quality of purified water. If instead of the number of elements, the feed pressure is specified, the number of elements is obtained through several integration equations.
SC
The design equations for the FILMTEC SW30 8-inch elements are given in the references [44].
M AN U
In this paper, considering the demand for water and the maximum water quality required based on the World Health Organization (WHO) standard for drinking water [45], the required water recovery, 6 elements per pressure vessel, membrane type selection (SW30XLE-400i) and water flow intensity and a silt density index (SDI) of under 3 [44], and the amount of power assigned to the RO can be determined based on the amount of feed pressure. The mean value method is
TE D
considered. In regard to the hub, by determining the amount of power, the pressure can be determined based on the equation below [44]. ^qr (r − "sc ) 8iop
(6)
EP
q = Nop
In this model, the Pelton turbine has been used as an energy recovery system. According to Eq.
[44].
AC C
(7), the amount of energy required by the RO system can be determined in accordance to Eq. (8)
Nq V =
^q c (c − op ) 8hpe iV
Nqt\e = Nqop − NqV
The basic equations in designing RO system are available in Appendix A.
(7) (8)
ACCEPTED MANUSCRIPT
2.3.MED modeling According to reference [43] where a multi-effect distillation system is assessed as experimental characterization, a number of empirical equations are extracted in order to assess the
RI PT
performance of the system under off-design conditions. In addition, the impact of the key parameters variation where the MED plant is controlled (inlet hot water flow rate (mhot) and temperature (Thot)) on GOR and freshwater production were evaluated with an empirical
SC
campaign of 82 tests. The MED plant steady state results working at off-design circumstances are addressed in a state which presents: The Thot and the mhot effect on the production of distillate (mMED) and on the plant GOR. The MED modeling equations are presented in this
M AN U
section. Eq. (9) and Eq. (10) are valid for 60 ≤ uvhe ≤ 74 ℃ and 7 ≤ ^vhe ≤ 14 U [43]. } ) ^_W` = −0.273 + (0.008409 uvhe ) − (0.04452 ^vhe ) + (0.0003093 uvhe − } ) (0.002485 ^vhe + (0.001969 uvhe ^vhe )
S
(9)
TE D
} ) jkl_W` = 648.2 − (26.74 uvhe ) − (16.45 ^vhe ) + (0.3842 uvhe + ~ ) } ) (0.3137 uvhe ^vhe ) + (0.5995 ^vhe − (0.001835 uvhe −
(10)
EP
~ ) } } (0.002371 uvhe ^vhe ) − (0.01844 ^vhe − (0.0001411 uvhe ^vhe )
AC C
The latter parameter determines the thermal efficiency of a MED plant (GOR) and it is expressed in Table 2. (row 12). Where mMED is the distillate production (L/s); Gain origin ration (GOR) (-); hot water temperatures (Thot), and the hot water flow rate (mhot); Qhot Heat consumed by MED. Moreover, in Table 3, the operation conditions for each system in the optimal scenario involving waste heat recovery are presented.
ACCEPTED MANUSCRIPT
Table 3. The operating conditions of the energy hub [57, 63 and 64]
GT
Boiler
1 64.35 11.54 9.56 115 1 86 10 0.098 90 36 4028.7 48550 0.65 2.75 0.0982
Pipeline length
m
1900
Ambient temperature Ambient pressure Salinity of Seawater
°C kPa ppm [mg/kg]
298 100 42000
TE D
ST
Fuel
-
1200 6
AC C
EP
Water transfer Ambient
L/s (-) (-) kJ/kg (-) (-) kg/s % % kW kJ/kg kg/m3 kgCO2/kgfuel kg/s
$
Value 0.9965 37.2
M AN U
MED
Unit (-) m2
RI PT
RO
Parameters Salt rejection Active RO membrane area Membrane element cost, estimation Element (type SW30XLE-400) Number of membranes in each pressure pipe Number of steps in RO Hot water inlet temperature Hot water flow rate GOR Air-fuel ratio Hot gas Specific Heat Isotropic turbine efficiency Compressor pressure ratio Fuel consumption Efficiency in full load mode Efficiency in full load mode Heat Consumption Lower Heating Value (LHV) NG density m CO2 Fuel consumption
SC
Component
2.4.Pumping system
A pumping system is necessary to transfer water. Eq. (11) shows the system’s energy consumption in Ppu (Mj/h) [6]. op =
H × ^qefr × C × 1000} × 60} ηop
(11)
ACCEPTED MANUSCRIPT Such that g is the gravitational acceleration (m/s2), ηop denotes pumping efficiency ratio i.e.
0.75, ^qefr represents media flux in terms of kg/s and H is the lifting height in terms of (m). According to the Darcye-Weisbach equation, Eq. (12) expresses the lifting height [6]: R ]} × +n [ 2H
RI PT
C=×
(12)
Such that L is the length of the pipeline (m), D is the internal diameter of the pipeline (m), v
SC
denotes media velocity (m/s), g denotes gravitational acceleration (m/s2), f represents coefficient of friction loss, and k denotes friction loss via local resistance.
M AN U
3. Exergy analysis
Exergy analysis can aid in developing strategies and guidelines for the effective use of energy in various systems. Exergy can be divided into four categories. Two prominent types of exergy are physical exergy and chemical exergy. In this study, potential exergy and kinetic exergy are
TE D
hypothesized as negligible. The physical exergy equals the maximum workable flow of matter and is determined by a process from the initial state or an environment that only involves heat exchange with its surroundings. Chemical exergy occurs by the exit of chemical compounds
EP
from a system of chemical equilibrium [46]. Chemical exergy is part of the exergy in the combustion process. By using the first and second law of thermodynamics, the following exergy
AC C
equilibrium is obtained [47-49].
K( u ( = (1 − ) bq − Nq − + ^q D − ^q \ D\ − Kq ` F u F \
(13)
Or Exq + mq ex = mq ex + Exq + Exq
(14)
ACCEPTED MANUSCRIPT
The χ parameter indicates that when exergy flow is outside the control volume shown in Fig. 3 and enters the system (χ=Buy), it is considered positive in equations and if exergy flow exits volume control (χ=Cell) it is shown as negative in equations.
RI PT
In Fig. 3, exergy flow is calculated as: Kq = ^qr,Vhe" RC
Where parameter ξ is defined as follows for NG [47]:
ov v Kqa"e\f = Kqa"e\f + Kqa"e\f
M AN U
Water flow exergy is equal to [50]:
SC
0.0698 = 1.033 + 0.0169 −
(15)
(16)
(17)
v Where Kqa"e\f and Kqa"e\f are water physical exergy and chemical exergy respectively and are ov
Kq = (1 −
uh )b q uv v
Kq h\f = Nq
TE D
calculated in [50]. Heat exergy and electricity are calculated based on Eq. (18) and Eq. (19) [51]:
(18) (19)
~ q ∑ ∑ e Ko,e e(∑p K st,p ) ¡e = q ∑ ∑ e(^qr,(T RC + Kc ) ¡e e Kr,e
AC C
;=
EP
Finally, hub exergy efficiency is calculated based on Eq. (20) [51]:
(20)
If u=1, heat exergy of the demand heat flow is calculated as follows [52]: Kq st,p = 1 −
u Cq uv st
(21)
If u=2, electricity demand exergy is calculated as follows [51]: Kq st,} = Kq¢
(22)
ACCEPTED MANUSCRIPT
If u=3, water demand exergy is calculated as follows [50]: st Kq st,~ = Kqa"e\f = ^q"e\f Kq"e\f
(23)
RI PT
4. Economic analysis Economic analysis is required to select equipment and coefficients stated in the hub section. In economic analysis, equipment costs and maintenance costs are calculated according to the utilized equipment. Eq. (24) is used to estimate MED costs [52]:
SC
_W` = £_W` lÅ + £Z&_
(24)
M AN U
In this equation, C = 81430.27 $ according the reference [53] and £Z&_ is calculated by Eq. (25)
[25].
£_W` = ^qÆf\¢v "e\f
RO system costs are set according to Table 4.
TE D
Table 4. RO system costs equations [25].
AC C
EP
Component Capital cost of the seawater intake and pre-treatment Capital cost of high pressure pump and pre-treatment
Pump
CCUa¦ = 996 (mr 24).§
Equation
CC¨©© = 52 (mvoo ΔPr )
CC¬© = a (mVΔP)¯° f . ϕ² 1 − η³ ¯´ ϕ² = 1 + 1 − η cast iron = 1 f = µ Steel = 1.41
f :Material correction factor, here: f = 1.41
ϕ² : First law efficiency correction factor $
a = 549.13 ÀÁÂ.ÃÄ ،a} = 0.71 ،a~ = 3 ،η³ = 0.8
(25)
ACCEPTED MANUSCRIPT
Other hub costs are set according to Table 5: Table 5. The cost functions for each equipment of energy hub [54, 55].
GT
Cost function ZÈÉ = (−0.014 V + 600)V ZÊÉ = 1200 UV
Steam Turbine
.§Í ZËÌ = 205bqgh \f
Boiler
RI PT
Component
SC
Economic parameters are determined based on reference [54, 55] shown in Table 5. The cost of the transmission line is also determined by Eq. (26) with the unit of $/m [6].
where D is the media pipe diameter (m).
M AN U
efr,o o\ t\ = 0.9734 (32000 + 213000 [)
(26)
By using assumptions for Eq. (11) and Eq. (12), the necessary costs, efr , CO2 emissions flow
correspondingly:
EP
efr,op = L\ × op
TE D
pertaining to the functionality of a pumping system are represented by Eq. (27) and Eq. (28),
^h},efr = D\s × op
(27)
(28)
AC C
Such that ce denotes generation cost within a power plant in terms of ($/MJ), eem represents the CO2 emission factor in terms of (kgCO2/MJ). Finally, TAC is calculated based on Eq. (29) [54]: uÎ Ï$ÐDEMÑ = _W` + YZ + V + UV + gh + efr + ∑ eÒKc Ó\,c − K¢ Ó\,¢ + ^qr,ehe" RCrp\ Ór + 3600 ^h} Ô\s Õ ¡
(29)
ACCEPTED MANUSCRIPT
Table 6. Values of economic parameters [60, 61, 62] Unit %
Value 10
Operation years (n)
Year
20
Hour
8760
3
$/m $/kWh $/kWh $/kWh $/m3 $/kg CO2
0.5 0.015 0.045 0.001 0.002 0.02086
SC
Operation hours in a year (N) Cost of fuel Electricity sales price Electricity purchase price Heating cost Desalinated water cost CO2 emission factor (ψem)
RI PT
Economic parameters Interest rate (int)
M AN U
Table 6 presents cost parameters. τ is an hour time period and N is equal to 8760 hours. Capital recovery factor (CRF) is dependent on equipment estimated life and interest rates and is written
lÅ =
3GF (1 + 3GF)Ö\"f (1 + 3GF)Ö\"f − 1
(30)
TE D
as [52]:
Where ‘year’ is the system service life and int is the interest rate.
EP
5. Case study
A diagram of the cement plant in the city of Qeshm is presented in Fig. 4. Fig. 4-a and Fig. 4-
AC C
b show the clinker production process and the cement plant’s curing stage. In Fig. 4-a, upon combining and grinding the primary raw material, it is entered from point 1 to the preheater and after passing through the cyclones, its moisture is removed and the limestone is calcified. Then, it enters the kiln at point 2. The raw materials are cooked in the kiln and converted to a clinker prior to be transferred to a clinker cooler to cool down (clinker point 3). There are fans installed at the bottom of the clinker cooler which absorb most of the hot clinker heat and exhaust it through its upper duct. Also, in Fig. 4-b, the potential points for heat recovery from the cement
ACCEPTED MANUSCRIPT
factory are shown. One of the points where it is possible to utilize waste heat is the suspension preheater boiler (SP boiler), where the hot gas from the kiln runs upwards in the opposite direction of the raw materials, transferring some of its heat. However, a large part of the heat in
RI PT
the SP boiler is still recoverable. The resulting gases thermal energy (317 °C) that may be recovered using SP boiler is 1532 kW. Clinker has to be cooled upon being baked at 1200 °C temperature. Gases from the clinker cooler exhibit a second heat source which is recovered using
SC
an after quenching chamber boiler (AQC) (approximately 4088 kW) [57, 58]. Table 7 describes the characteristics of these points. Since the city of Khaledin, located on the island of Qeshm is
M AN U
close to the cement plant, its water demand to supply its 1396 population is shown in Fig. 5. The demand for consumed water (kg/s) for each hour during the year for Khaledin is determined which should be supplied by the energy hub. Also, Fig. 5 presents the cement factory’s consumption of electricity (kW) for each hour during the year. Since the variances in
TE D
consumption in various hours of the year are insignificant in relation to the consumption level, the average annual consumption for all hours can be considered. Also, the distance between the cement factory and the city of Thurian with 2339 populations is approximately 1.9 km, which
EP
will allow the transfer and sales of excess water to the city. In regard to the case study, the lack of implementing technology and the inconsistency between investment costs and local prices are
AC C
the main reasons for not using renewable energies. Also, due to inaccessibility to sufficient historical data in this paper in addition to considering the constant loading profile and national grid as a backup in storing energy, the uncertainty was not considered. Another reason that diminishes the need for uncertainty analysis is that some equipment such as desalination systems utilize variety of technologies (MED and RO) and the use of an electricity energy carrier for the RO system and thermal energy carrier for the MED system will aid in decreasing uncertainty.
ACCEPTED MANUSCRIPT
Table 7. Qeshm cement plant gas condition [57] Mass flow rate (kg/s)
290
21.4
311 215 93 99
15.9
15.9
15.9
21.4
AC C
EP
TE D
M AN U
Clinker cooler exhaust Preheater exhaust SP boiler exhaust Stack exhaust AQC boiler exhaust
Temperature (˚C)
RI PT
Parameter
SC
Point
Fig. 4. Schematic of heat utilization potential in Qeshm cement plant
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 5. Values of electrical loads and water consumption
M AN U
6. Optimization
GA (Evolution) is an algorithm based on repetition. The strategy of this algorithm is to apply a random search to determine an optimal solution. In principle, the simple method of biological evolution is to emulate the evolutionary characteristics of a population of individuals in which
TE D
the individual consists of decision variables values for a potential optimized solution. The decision variables and optimization variables are given in Table 8. The objective functions of this optimization are Eq. (20) and Eq. (29) which determine the hub’s exergy efficiency and
EP
economic parameter, respectively.
AC C
Table 8. Design parameters and their range of variation for the optimization procedure.
CHP Power dispatch factor MED Water dispatch factor Boiler Heat dispatch factor Steam Turbine Power dispatch factor RO Water dispatch factor Water recovery rate (RO) Hot water inlet temperature (MED) Hot water flow rate (MED) Power consumption per Water Desalinated (RO) Maximum Heat Produced by Boiler
Parameter vCHP (-) vMED (-) vBoi (-) vST (-) vRO (-) RRO (%) Thot (°C) Mhot (L.s-1) EWRO(kJ/kg)
Lower bound 0 0 0 0 0 0.3 60 7 0.1
Upper bound 1 1 1 1 1 0.85 74 14 8
Qmax,Boi (kW)
100
3000
ACCEPTED MANUSCRIPT
A decision making process should be employed to select a single optimal point from the present points on the Pareto front. Such a process is typically based on engineering experiences where the importance of every decision making objective is highlighted. An assumed point taking the
RI PT
role of an ideal point is usually used to determine the final decision making process. When two objective functions are individually optimized whilst disregarding other objective functions, the compositions of these values symbolizes the ideal point or ideal objective point. Fig. 6 presents
SC
the usual Pareto front in addition to the ideal point to maximize the first objective whilst minimizing the second objective. It is evident that it is not possible to achieve both objectives at
M AN U
optimal points concurrently. The closest Pareto frontier point to the ideal point may be chosen as the final optimal solution because the ideal point is not a solution that is placed on the Pareto frontier. Prior to this, the objectives are non-dimensionalized. LINMAP, linear programming techniques for multidimensional preferences analysis method is implemented to non-
t Å×d =
Å×d ss Ø∑×(Å×d )}
°
TE D
dimensionalize the objectives in this paper by employing the following equation [59]: (31)
Such that z is the index for each Pareto front point, x is each objective’s index and mm displays
EP
the Pareto front’s number of points. Hence, the distance for each Pareto front point from the ideal
AC C
point is attained [59]: }
t × = Ù(Å×d − Å t \",d )} d
(32)
Such that ideal is ideal objective function index. Thus, the value of distance for each Pareto frontier point is derived via the equation stated above and optimal point leading to minimum d is chosen as final optimal solution.
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 6. Pareto optimal front for energy hub According to [56], some of the essential characteristics of genetic optimization algorithms are:
TE D
(1) The generality of the formulation.
(2) Robustness, in the sense of avoiding local optima, looking for global optimal points. (3) The capability of handling multiple objectives.
EP
(4) Computational efficiency decision variables.
AC C
6.1. Parameter tuning
In this paper, the decision parameters are assumed values which can be changed within an acceptable and specific range as well as exerting significant effect on objective functions in order to minimize the TAC and maximize the exergy efficiency upon optimization. These values are initially determined from the concept model and related formulas. Ultimately, their effects on the objectives are assessed such that through final selection, by making changes in a specified range, the optimized acceptable values can converge to the results of the equations. Dispatch factors are
ACCEPTED MANUSCRIPT
among the variables used in all energy hub studies to achieve optimal energy flows in control volume associated with energy hubs. Rows 1, 2, 3, 5 and 6 of Table 2 in the paper, are related to the design equation for the energy hub model, which is dedicated to decision variables vCHP, vST,
RI PT
vRO, vMED and vBoiler, respectively. Each dispatch factor represents the production or consumption capacity share of the system in relation to the overall energy carrier production or consumption share and thus is a determining parameter in the energy hub’s internal equipment capacities and
SC
input energy carriers. For this purpose, the sensitivity analysis for each of these variables in
and TAC are provided in Appendix B.
M AN U
relation to the objective function and key values of energy hub output such as exergy efficiency
In this study, the results of the objective function from two-objective optimization of the energy hub for different algorithms parameters in this case study are presented in Table B-1. The resulting values from the genetic algorithm parameters in Table B-1 are extracted according to
TE D
reference [61] to analyze the sensitivity of the objective functions. Each result is presented upon best in 3 times run. As evident, the best results are obtained at levels 4 and 5 which have the least TAC with highest exergy efficiency. From these two levels, level 5 exhibited the most optimal
EP
results with population size 150, mutation 0.03 and crossover 0.8 and its parameters were
AC C
selected for system outputs.
7. Validation
Validation indicates that the model is a logical representation of an actual system in order to assess the actual system activity with enough fidelity to analyze the main goals. Each model typically analyzes a specific problem and presents various system components in different levels of abstraction. As a result, there is a model of different levels of validity for various system components, which together, examines the system behavior. There are three separate aspects for
ACCEPTED MANUSCRIPT
most models which should be assessed during model validation: Assumptions, Input parameter and Output values. For this purpose, in order to achieve full validation of the energy hub, validation of each equipment model should be considered to attain validation of the integrated
RI PT
energy hub system which does not yet exist. Thus, each equipment should be validated separately. 7.1. RO
SC
For the purpose of validating the written code, ROSA software was implemented for plug system. Validation was achieved via a system where water volumetric flow rate and salt
Table 9 presents the validation results.
M AN U
concentration are 40 m3/h and 4049 ppm, respectively. The sample was modeled in two stages.
Table 9. The validation results of ROSA software Modeling results
TE D
Parameters
AC C
EP
Produced freshwater in stage 1 (ppm) Produced freshwater in stage 2 (ppm) Production of fresh water in each stage 1 (m3/h) Production of fresh water in each stage 2 (m3/h) Feed pressure (MPa) Recovery
ROSA software results
Error (%)
21.94
23.50
7.1
59.52
65.58
10.1
40.12
40.83
1.7
19.87
19.17
3.5
3.53
3.77
6.7
46.0
46.0
0
7.2. MED
Parametric equations were successfully achieved in reference [43] prior to be statistically validated to predict the GOR and distillation production as a function of various input parameters via an extensive validation range. Moreover, the presented empirical model in the section 2.3
ACCEPTED MANUSCRIPT
stemming from the extensive empirical results has been compared with reference [52] (Fig. 7). This comparison was conducted under conditions where the values of mf and Tf were within the acceptable range of the empirical model. This comparison shows that the empirical model has
RI PT
minor differences with the numerical results, and the relevant errors are acceptable in terms of empirical model application. 7.2. Gas Turbine and Steam Turbine
EP
TE D
M AN U
SC
The model applied in the gas turbine and steam turbine was based on reference [48].
AC C
Fig. 7. MED model validation with a valid model in various hot water inlet temperatures
8. Results
The Pareto curve is derived from two-objective optimization, and in order to achieve optimal decision variables and optimal objective function values, the two-objective functions will be TAC and exergy efficiency. The genetic algorithm categorizes the solution space into a feasible domain and infeasible domain. The frontier between these two regions is the Pareto curve. Each
ACCEPTED MANUSCRIPT
of the points created result from the decision variables that leads to the determination of the objective functions’ values depicted on the Pareto curve. Each resulting point is an optimal point. The ideal point on this curve is the point with the lowest TAC and highest exergy efficiency.
RI PT
This point is in the infeasible domain. The point of the curve that is closest to this point is called the best point. Fig. 6-a shows the optimum Pareto front results for two objective functions.
The acquired results show that two chosen objective functions i.e. energy efficiency and TAC
SC
differ in the optimal state. Attempts to enhance energy efficiency will increase TAC in both cases and vice versa. The confliction trend in the optimal state should be highlighted via multi-
M AN U
objective optimization. Increases in energy efficiency increases optimal TAC with the increment slope significantly increasing for maximum energy efficiency. Thus, the energy hub is not economically viable for maximum energy efficiency. The developer should choose the final optimal point according to the significance of each objective. Additionally, the optimum Pareto
TE D
point is influenced by optimal results acquired in the energy hub. Table 8 lists ten design factors i.e. decision variables along with relevant variation ranges in order to maximize energy efficiency value while minimizing TAC. Fig. 6-b illustrates optimal Pareto fronts with and
EP
without waste heat recovery stemming from the cement plant. The acquired results show that when waste heat is included, the optimal Pareto front is completely dominated by the optimal
AC C
results that exclude heat recovery. According to Fig. 6-b, exergy efficiency in all locations with waste heat recovery is greater compared to the without waste heat recovery whilst exhibiting lower TAC. According to Table 10, the exergy efficiency increased by an average of 20% in the case of without waste heat recovery compared to waste heat recovery case while its TAC function decreased by 72.6%. Moreover, the effect of optimization with the concept of energy hub was shown in a case study in Table 10. In this regard, the effect of the optimal energy hub
ACCEPTED MANUSCRIPT
system was compared with the current system (with considering energy recovery from the cement factory). In this table, the exergy efficiency and TAC for the current system are shown along with the energy recovery from the cement plant. According to this comparison, in the
RI PT
optimal mode, the exergy efficiency increased by 29.6% and TAC decreased by 19.3%. This illustrates the impact of using the concept of energy hub with the optimal allocation of energy carriers and the addition of new energy equipment.
M AN U
Pareto optimal fronts for input values given in Table 9.
SC
Table 10. The optimum values of exergy efficiency and the TAC for the design points A–C in
With using energy hub concept Parameter ηex (-)
A 0.78 2553476
B (Best Point) 0.7445 1104138
C 0.727 1059092
Without heat recovery Best Point 0.616 4036266.47
TE D
TAC ($/ year)
With Waste heat recovery
Current system with heat recovery 0.5741 1368827.47
The use of heat loss improves both objective functions. In regard to the TAC function, it should be stated that the high potential of heat loss in the cement plant has led the energy hub to utilize
EP
this cheap energy to not only self-sufficiently reduce part of the electricity, heat, and fresh water production costs, but also reduce O&M costs. Concerning the exergy efficiency function, it
AC C
should be stated that directly reducing fuel consumption according to Eq. (20) will effect exergy efficiency and increase this objective function. A decision making process is implemented to select the final solution from optimal Pareto front points. This process is conducted on the basis of engineering experiences and the significance of every objective for the decision makers. Fig. 6-a shows that at design point A (78%) there is maximum exergy efficiency exists and the TAC is highest at this point. Furthermore, at design
ACCEPTED MANUSCRIPT
point C, the minimum TAC is achieved (1059092 $/year) along with the lowest exergy efficiency value (72.7%). The optimal exergy efficiency state as a single objective function is at design point A. The optimal condition where TAC is a single objective function is at design
RI PT
point C. Table 10 lists the optimal values for two objectives regarding five usual points from AC (Pareto optimal fronts) for the input values stated in Table 11. Also, Table 11 shows that, moving from the point C to point A, the vMED value is approximately 6% of its initial value,
SC
which indicates a decrease in annual costs and an increase in exergy efficiency for the use of MED desalinating water entailing less use of the RO system.
M AN U
Table 11. The optimum values of design parameters for optimum selected points A–C in Pareto optimum front. With Waste heat recovery Parameter
TE D
B (Best Point) 1 0.015 0.444 0.995 0.494 0.459 64.35 11.54 3.76 1777.6
C 1 0.316 0.48 0.999 0.234 0.468 64.7 11.98 3.8 1768.3
AC C
EP
vCHP (-) vMED (-) vBoi (-) vST (-) vRO (-) RRO (%) Thot (°C) Mhot (L.s-1) EWRO(kJ/( kg)) Qmax,Boi (kW)
A 0.997 0.02 0.547 0.76 0.51 0.45 64.4 11.59 4.28 1810.3
Without heat recovery Best Point 0.997 0.039 0.61 0.577 0.72 0.34 66.65 10.56 3.2 1511.66
In the RO system, high pressure water should be supplied between 60 bar to 80 bar to desalinate seawater. A greater power allocation to the RO system in the steady state water supply system increases the water supply pressure and increases the difference between water supply and osmotic pressure flux of the flowing water in each membrane, thus entailing increased system recovery and increased fresh water flow. The genetic algorithm has increased the production share of the system by increasing the power allocated to the RO.
ACCEPTED MANUSCRIPT
Also, a 25% reduction in the use of a steam turbine to generate electricity will increase TAC and exergy efficiency. The 42% reduction in the use of the steam turbine to generate electricity and an increase of 38% in the use of the auxiliary boiler indicates that in order to meet electricity
RI PT
demand, electricity should be bought from the grid and in order to meet heat demand, fuel consumption should be increased and bought from the natural gas grid.
Since the city of Khaledin that is located in the vicinity of the cement plant, has a specific
SC
population, as a result, the highest demand for energy carriers and heat is limited, resulting in the maximum sales of these two energy carriers through a hub of 6980 kW for electricity and 5723
M AN U
kW for heat. In the energy hub, urban water demand is considered as the demand for energy hubs. Results indicate that due to the high costs of water transfer and the distance from the nearest city (apart from the adjoining city) to the cement plant, there are no water sales outside the hub. Upon optimization and using the energy hub concept, the energy purchased from the
TE D
factory has dropped from 3.2317 kg/s to 2.114 kg/s, indicating that part of the water is produced in the hub and utilized in the adjacent city (1.11 kg/s). Also, the energy hub is self-sustained in electricity maintenance, with the small amount exception, the remaining 2988 kW supplied by
EP
the GT and steam turbine. Due to the 5630 kW heat recovery from the cement plant, the majority (4856 kW) is sold outside the hub. Also, in order to supply the NG required by the GT and
AC C
boiler, 0.0982 kg/s is consumed in the hub. In Table 11, it is noteworthy that the use of gas turbines is maximally utilized and due to the increased efficiency of the GT system, lower costs and reduced environmental pollutants, it is considered feasible.
Another factor in reducing TAC can be accredited to the utilization of the gas turbine. One of the features of this apparatus is its electricity and heat supply based on an initial production cost.
ACCEPTED MANUSCRIPT
Therefore, the genetic algorithm increases the role of this system in the energy hub leading to a reduction in the installation costs pertaining to other equipment such as a boiler. On the other hand, the confrontation between the heat loss of the cement plant and elimination of a boiler
TE D
M AN U
SC
the gas turbine has led to a general reduction of TAC.
RI PT
from the system compared to the non-utilization of waste heat and increased fuel consumption in
temperature
AC C
EP
Fig. 8. The rate of exergy efficiency in energy hub at hot water flow rate and hot water
As shown in Fig. 8, it is evident that energy efficiency increases with Thot as expected. Considering the impact of QMED on energy efficiency, it is clear that energy efficiency slightly decreases for Low QMED (from 100 kW to 200 kW) while decreasing considerably (approximately 2.2%) for larger QMED. Such progressive decreases may be because of the evaporative process via the tube bundle that is closer to its hot water temperature design value
ACCEPTED MANUSCRIPT
(11.54 L/s). Due to the increases in thermal energy consumed, which is greater than the distillate production, an energy efficiency reduction is experienced by QMED that is over 200 kW. Also, by increasing the rate of Thot (from 60 °C to 74 °C), the exergy efficiency increases by
RI PT
1.5%. Fig. 9 shows vRO, vMED variations in terms of TAC for both waste heat recovery and without waste heat recovery. In this figure, with vMED variations from 0 to 1, TAC increases by an average of 3.1%. Also, vRO variations from 0 to 1 will increase TAC by an average of 2.5%
SC
and up from 1049362 $/year. The percentage of TAC variations for without waste heat recovery
AC C
EP
TE D
compared to waste heat recovery.
M AN U
remains the same. Moreover, the average TAC increase in without waste heat mode is 3.8%
Fig. 9. The rate of TAC in energy hub at hot water flow rate and hot water temperature
By increasing the allocated share of fresh water production to the RO system, it is evident that TAC costs have increased. In order to produce more water, the RO system requires higher water pressure from the intake water supply as well as higher water pressure. By allocating more
ACCEPTED MANUSCRIPT
energy to this component, the water pressure will increase. The main factor in increasing the RO costs according to Table 4 is the pre-treatment costs. In this state, by increasing the pre-treatment
TE D
M AN U
SC
RI PT
costs and its effect on the RO system, the TAC will increase.
Fig. 10. The rate of TAC in energy hub at hot water flow rate and hot water temperature
EP
Fig. 10 shows the variation in the price of water ($/m3) for three modes of production by RO,
AC C
MED and the entire hub system, taking into account the share of each of the two water producers. From 0 to 0.4, the UPC change is 0.65%, while changes up to 1 will result in a 40% drop. By changing the value from 0 to 0.3, the UPC will drop by 23% and will change from 0.54 to 0.416. On the other hand, its reduction from 0.3 to 1 would be 11%. It is clear that changes in the share of water produced by each desalination system will not affect the production cost of other systems, and the cost of producing water will be higher if the unit is smaller. The process of changing the final price of water production in the energy hub system indicates that increasing
ACCEPTED MANUSCRIPT
the share of water production in both MED and RO systems reduces water production costs. Also, Fig. 10 shows that the price changes of the MED system are significantly high compared to the RO system. The MED system is considered for a specific capacity with equivalent
RI PT
investment costs whereas if it utilizes less heat than its nominal capacity, the production of fresh water is reduced in a fixed investment. Due to the reduction in fresh water production, the final water costs will significantly increase, in which case the investment cost is considered for the
SC
MED, but the system will not be utilizing its nominal capacity. This greatly effects the final price. On the other hand, the RO system can only control fresh water flow by reducing feed
M AN U
water pressure and the valves in the brine water pathway. As a result, in the RO system, price changes are only dependent on energy consumption prices and O&M, thus pertaining to price
AC C
EP
TE D
changes in this limited range.
Fig. 11. Variation of TAC versus ratio of price and energy carrier in energy hub
ACCEPTED MANUSCRIPT
Fig. 11 shows the sensitivity analysis conducted on key factors such as water price, heating and electricity. The results from this analysis highlighted the energy carrier with negative slope that may be supplied with other sources and acquired with a positive slope. The achieved results
RI PT
show that when water selling price increases, the TAC is significantly raised with constantly varying costs. The optimal system tends to function on the basis of water sales. The TAC will increase as much as 6.6% because of the raise in water price ration. In addition, Fig. 11 depicts
SC
that with higher heating sales price ration from -60% to 60%, overall system TAC will be reduced by 4%. Based on the data shown in Fig. 11, it is evident that an increase in electricity
M AN U
price has less influence on TAC in comparison to other energy carriers.
In the current system, the following resources are used to meet the demand of the consumption section. Electricity supply from the national grid (conventional centralized power plants on the island), fresh water produced from the RO system and heating based on current heat generators is
TE D
dispersed for each consumer. Also, since the systems are designed to just meet the demands of consumers, purchases and sales with the grid are not considered. Table 12 presents the results of the basic parameters of the hub both in an optimal state and current state. In the optimal scenario,
EP
by recovery waste heat from the cement plant, heating, power and water generation in addition to optimizing energy flows and capacity of utilized equipment, fuel consumption will be reduced by
AC C
over half whilst exergy efficiency will increase by 36%. Also, CO2 emission will drop by 21624 ton/year and TAC will decrease by about 37,000 $/year. In a more detailed assessment on each equipment, the results on Table 13 indicate that in the current state, the total amount of fresh water is supplied through the RO system and the electricity is supplied by the grid whilst by adopting the energy hub concept, 18% by MED, 33% by the RO system and 49% is supplied by the water grid. Moreover, part of the heat produced by GT and waste heat is consumed by the
ACCEPTED MANUSCRIPT
steam turbine and MED and the remainder is sold to the grid. On the other hand, about 50% of the plant’s electricity is supplied by the GT and 50% by the steam turbine. Fig. 12 presents changes in CO2 emissions for both cases of including and without energy hub for water and
RI PT
electricity energy carriers. In Iran, CO2 emissions resulting from grid electricity is 0.67 kg/kWh and CO2 emission resulting from natural gases is approximately 0.23 kg/kWh [65]. Also, the amount of CO2 emissions needed to produce heating was 0.38 kg/kWh [73] (an
M AN U
to produce each cubic meter of water [26].
SC
efficiency of 65% was assumed for the boilers). On the other hand, 3 kg of CO2 will be released
AC C
EP
TE D
Table 12. Comparison of results between basic parameters of the optimal scenario and current system on Qeshm Island Parameter Current system Optimized System UPC ($/m3) 0.6738 0.429 Fuel Consumption 4871668.6 2121120 (kg/year) CO2 Emission 224295000 202671000 (kgCO2/year) 38.1 74.45 ηex (%) TAC ($/year) 1141060 1104138
ACCEPTED MANUSCRIPT
224,295 208,000 193,236
RI PT
16,016
202,671
9,169
183
Water
Plant
Total
M AN U
Electricity
279
SC
CO2 Emission (ton/year)
240,000 220,000 200,000 180,000 160,000 140,000 120,000 100,000 80,000 60,000 40,000 20,000 0
Without Hub
With Hub
Fig. 12. Variation of CO2 emissions (kgCO2/year) and energy carrier with energy hub and current system
*
AC C
EP
TE D
Table 13. Assessment of the allocated demand in the hub and comparison with the current system Energy With Waste heat Component Current system carrier recovery MED 0.5828 Water (kg/s) RO 3.2317 1.0686 Grid 1.5804 Boiler 0 GT 3272.3 Heat (kW) Waste heat 5625 Grid -4700* GT 1493.7 Power (kW) Steam Turbine 1489.6 Grid 2988 7.76 In this table, the negative numbers indicate the amount of sales of energy carriers to the grid.
According to the reference [66], 900 tons of carbon dioxide are produced, for every 1,000 tons of cement, about half of which is related to the burning of fuel in the kiln and the other half to the
ACCEPTED MANUSCRIPT
calcining reaction in the preheater. As the amount of cement production in Qeshm cement factory is 231 thousand tons, about 208 thousand tons of carbon dioxide will be produced annually. Regarding heat recovery (5620 kW) at the cement plant which is equivalent to the
RI PT
reduced carbon dioxide for the recovered waste heat, carbon dioxide is declined by 7%. In the case of without energy hub, the CO2 gas produced is considered based on electricity, and water demand from the grid. The demand for electricity based on Fig. 5 is 2988 kW which produces
SC
16015 tons of CO2 per year. With the energy hub, upon optimization and heat recovery from the hub, the consumption of NG is 0.0982 kg/s equivalent to 9210 tons of CO2 per year for power
M AN U
generation in GT which is the main CO2 producer when using the energy hub. On the other hand, the primary energy of desalination systems is captured from the power and heat generated within the energy hub, it is therefore considered to be equivalent to the change in carbon dioxide emissions in the primary energy changes. Then, the amount of reduced water received from the
TE D
grid and utilization of local desalinating systems was assessed. Also, the amount of CO2 generated through water transfer is addressed (Eq. (28)). Due to reduced water purchases from 3.2317 kg/s to 2.114 kg/s, and by taking into account the CO2 generated by the water
EP
transportation, the amount of generated CO2 from the water was reduced by 96 ton CO2 per year. Considering all energy carriers before and after optimization, the environmental impact of
year.
AC C
producing 224 thousand tons of CO2 per year has been reduced to 202 thousand tons of CO2 per
9. Conclusions
In this paper, an amalgamation of water, power and heating generation systems were utilized to optimally develop a modern operational strategy called energy hub. The TAC for the energy hub and energy efficiency was optimized by taking into account the waste heat recovery from the
ACCEPTED MANUSCRIPT
cement plant. The proposed energy hub has the potential to supply of electrical, thermal and water loads. Results from this design were compared to the cases that without waste heat recovery. Moreover, all equipment will be selected using the concept of an energy hub at its best
RI PT
capacity. High energy consumption of major desalinated production processes, reduction of economic costs, reducing greenhouse gases in an integrated water, power and heat system and
factory as an energy hub.
•
After optimization, the exergy efficiency and TAC increased by 20% and decreased by 72.6% respectively.
•
CO2 emissions reduced from 16295 tons per year to 5329 tons per year upon utilizing the
TE D
energy hub concept. •
M AN U
The acquired research results are presented below:
SC
increase the exergy efficiency, are the main reasons for utilizing and optimizing the cement
TAC increases and decreases moderately with increases in water and heating price while it is negligible due to the small amount of grid electricity purchased. The lower the share of water production, the higher the price of water per cubic meter
EP
•
•
AC C
The following may be regarded as suggestions for the continuation of this research: Optimizing the use of the organic Rankine cycle (ORC) system and a comparison of
various working fluids to determine the optimal fluid for the purpose of electricity production from the cement plant’s heat recovery with the concept of an energy hub;
•
The use of hydrogen fuel or biogas to generate power and heat in a CHP and boiler;
•
Combining renewable systems such as wind turbines along with an energy hub for the cement plant to evaluate reduced costs.
ACCEPTED MANUSCRIPT
Appendix A- RO system equations In modeling the reverse osmosis system according to the mass transfer equations, it is evident
RI PT
that the water flow pressure and salt from the membrane are as presented in Eq. (A-1) and Eq. (A-2) and also the mean velocity in each membrane element is determined by Eq. (A-3). The salt concentration of the produced water is determined by Eq. (A-4). Also, the salt concentration in
SC
proximity to the wall is obtained according to the film theory by Eq. (A-5). In regard to the conjunction equation, Eq. (A-6) and Eq. (A-7) can be used for water and Eq. (A-8) can be used
Ú = Î × u Å[r − o − Ú¢ = àá − o â
∆r − (Ý − Ýo )] × 10ß 2
Ú + Ú¢ 8 Ú¢ o = × 1000 ã r + c = o + ( − o )D ä 2 bo = × bg = bÆ − bo bÆ r − bo o g = bg
(A-1) (A-2) (A-3) (A-4) (A-5) (A-6) (A-7) (A-8)
AC C
EP
TE D
=
M AN U
for salt [12]:
In the equations above, A, B and Sw are the permeability coefficients for water, salt, membrane and its area, respectively. The equations that can be used to reduce the number of unknowns are as follows:
[¢ 0.0033 × b" × R × æ ∆r = N × ~ n = 0.04 × lD .Íå × L .~~ ×
(A-9) (A-10)
ACCEPTED MANUSCRIPT bc + br 2 0.2641 × (u + 273) Ý= 10ß − b" =
ã ç è
and Ds is the permeability of salt (m /s), d is the distance between feeds, Qa is 2
RI PT
Where lD =
(A-11)
the average flow determined by Eq. (A-10). Lm is the membrane length and N is number of elements at each PV. π is osmosis pressure and C is salt concentration. In order to estimate the
SC
average pressure drop, the Hagan-Poiseuille equation is used. The Sherwood number is calculated by using Eq. (A-9) and Eq. (A-10) which is used to calculate the polarization of the
Appendix B-sensitivity analysis
M AN U
concentration.
The results from the sensitivity analysis in determining decision parameters in the genetic
AC C
EP
TE D
algorithm are presented below.
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
Fig. B-1. Variation of objective functions versus dispatch factors in energy hub
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. B-2. Variation of objective functions versus decision variables in energy hub
AC C
EP
Table B-1. The results of optimum objective functions for different algorithms parameters in operating condition Level 1 2 3 4 5 6 7 8 Population 50 50 100 100 150 150 200 200 Size Mutation 0.01 0.01 0.02 0.02 0.03 0.03 0.04 0.04 Probability Crossover 0.8 0.9 1 1.1 0.8 0.9 1 1.1 Probability TAC ($/year) 1212562 1197736 1277446 1168324 1104138 1122893 1121454 1158685 Exergy 73.16 77.51 73.61 75.308 74.45 72.302 73.582 76.25 Efficiency (%)
ACCEPTED MANUSCRIPT
References [1] Mohammadi M, Noorollahi Y, Mohammadi-ivatloo B, Hosseinzadeh M, Yousefi H, Torabzadeh Khorasani S. Optimal management of energy hubs and smart energy hubs – A
RI PT
review. Renewable and Sustainable Energy Reviews 2018; 89:33–50. [2] Brahman F, Honarmand M, Jadid S. Optimal electrical and thermal energy management of a residentialenergy hub, integrating demand response and energy storage system. Energy and Buildings 2015; 90:65–75.
[3] Mohammadi S, Moradi-Dalvand M, Ghasemi H, Ghazizadeh M.S. Optimal Design of
SC
Multicarrier Energy Systems Considering Reliability Constraints. IEEE Transaction on Power Delivery 2015; 30:878-886.
M AN U
[4] Bahrami S, Toulabi M, Ranjbar S, Moeini-Aghtaie M, Ranjbar A. A Decentralized Energy Management Framework for Energy Hubs in Dynamic Pricing Markets. IEEE 2017;9:19493053.
[5] AlRafea K, Fowler M, Elkamel A, Hajimiragha A. Integration of renewable energy sources into combined cycle power plants through electrolysis generated hydrogen in a new designed energy hub. International journal of hydrogen energy 2016; 41:16718-16728.
TE D
[6] Togawa T, Fujita T, Dong L, Fujii M, Ooba M. Feasibility assessment of the use of power plant-sourced waste heat for plant factory heating considering spatial configuration. Journal of Cleaner Production 2014; 81:60-69.
[7] Amiri A, Vaseghi M.R. Waste Heat Recovery Power Generation, Systems for Cement
EP
Production Process. IEEE Transactions on Industry Applications 2015; 51:13 – 19. [8] Kabir G, Abubakar A.I, El-Nafaty U.A. Energy audit and conservation opportunities for
AC C
pyroprocessing unit of a typical dry process cement plant. Energy 2010; 35:1237–1243. [9] Renewable Energy Organization of Iran. www.satba.gov.ir. (Access time: 2017) [10] Madlool N.A, Saidur R, Hossain M.S, Rahim N.A. A critical review on energy use and savings in the cement industries, Renew. Sustain. Energy Rev. 2011; 15:2042–2060. [11] Salimi M, Amidpour M. Investigating the integration of desalination units into cogeneration systems utilizing R-curve tool. Desalination 2017; 419:49-59. [12] Mokhtari H, Sepahvand M, fasihfar A. Thermoeconomic and exergy analysis in using hybrid
systems
(GT+MED+RO)
Gulf. Desalination 2016; 399:1-15.
for
desalination
of
brackish
water
in
Persian
ACCEPTED MANUSCRIPT
[13] Shahzad M. W, Burhan M, Ng K.C. Pushing desalination recovery to the maximum limit: Membrane and thermal processes integration. Desalination 2017; 416:54-64. [14] Nemati A, Sadeghi M, Yari M. Exergoeconomic analysis and multi-objective optimization of a marine engine waste heat driven RO desalination system integrated with an organic Rankine
RI PT
cycle using zeotropic working fluid, Desalination. 2017; 422:113-123.
[15] Reynolds J, Ahmad M. W, Rezgui Y, Hippolyte J. Operational supply and demand optimization of a multi-vector district energy system using artificial neural networks and a genetic algorithm. Applied Energy 2019; 235:699–713.
SC
[16] Moeini-Aghtaie M, Fotuhi-Firuzabad M. Multiagent Genetic Algorithm: An Online Probabilistic View on Economic Dispatch of Energy Hubs Constrained by Wind Availability.
M AN U
IEEE Transactions on sustainable energy 2014; 5:699–708.
[17] Ko M. J, Kim Y. Sh, Chung M. H, Jeon H. Ch. Multi-Objective Optimization Design for a Hybrid Energy System Using the Genetic Algorithm. Energies 2015; 8:2924-2949. [18] Kampouropoulose K. Multi-objective optimization of an energy hub using artificial intelligence, A thesis for the degree of Doctor of Philosophy in Electrical Engineering, Universitat Politècnica de Catalunya, Barcelona, Catalonia, Spain, 2018.
TE D
[19] Stewart B, Hoang M, Zarzo D, Olewniak F, Campos E, Bolto B, Barron O. Desalination techniques: A review of the opportunities for desalination in agriculture. Desalination 2015; 364:2–16.
[20] Lattemann S, Kennedy M.D, Shippers J.C, Amy G. Chapter 2 Global Desalination
EP
Situation. Sustainability Science and Engineering 2010; 2:7–39. [21] IDA Desalination Yearbook 2016–2017, Water Desalination Report. Global water
AC C
intelligence 2016.
[22] Voutchkov N. Energy use for membrane seawater desalination – current status and trends. Desalination 2018; 431:2-14. [23] Shalaby S.M. Reverse osmosis desalination powered by photovoltaic and solar Rankine cycle power systems: A review. Renewable and Sustainable Energy Reviews 2017; 73:789–797. [24] Ng K.C, Shahzad M.W. Sustainable desalination using ocean thermocline energy. Renewable and Sustainable Energy Reviews 2018; 82:240–246. [25] Mokhtari H, Ahmadisedigh H, Ebrahimi I. Comparative 4E analysis for solar desalinated water production by utilizing organic fluid and water. Desalination 2016; 377:108–122.
ACCEPTED MANUSCRIPT
[26] Salcedo R, Antipova E, Boer D, Jiménez L, Guillén-Gosálbez G, Multi-objective optimization of solar Rankine cycles coupled with RO desalination considering economic and life cycle environmental concerns. Desalination 2012; 286:358–371. [27] Francey R.J, Trudinger C.M, Schoot M, Law R.M, Krummel P.B. Atmospheric verification
RI PT
of anthropogenic CO2 emission trends. Nature Climate Change 2013; 3:520–524.
[28] Peters G.P, Andrew R.M, Boden T, Canadell J.G, Ciais P, Quéré C. Le, Marland G, Raupach M.R, Wilson C, the challenge to keep global warming below 2 °C. Nature Climate Change 2013; 3:4–6.
address operational constraints. Energy 2014; 73:387–98.
SC
[29] Evins R, Orehounig K, Dorer V, Carmeliet J. New formulations of the energy hub model to
M AN U
[30] Sharif A, Almansoori A, Fowler M, Elkamel A, Alrafea K. Design of an energy hub based on natural gas and renewable energy sources. International Journal Energy Research 2014; 38:363–73.
[31] Rastegar M, Fotuhi-Firuzabad M. Load Management in a Residential Energy Hub with Renewable Distributed Energy Resources, Energy and Buildings 2015;107:234-242. [32] Maroufmashat A, Fowler M, Khavas S.S, Elkamel A, Roshandel R, Hajimiragha A. Mixed
TE D
integer linear programing based approach for optimal planning and operation of a smart urban energy network to support the hydrogen economy. International Journal of Hydrogen Energy 2015;41: 7700-7716.
[33] Maroufmashat A, Elkamel A, Fowler M, Sattari S, Roshandel R, Hajimiragha A. Modeling
EP
and optimization of a network of energy hubs to improve economic and emission considerations. Energy 2015; 93:2546–2558.
AC C
[34] Najafi A, Falaghi H, Contreras J, Ramezani M. Medium-term energy hub management subject to electricity price and wind uncertainty. Applied Energy 2016; 168:418–433. [35] Majidi M, Nojavan S, Zare K, A cost-emission framework for hub energy system under demand response program, Energy 2017;134;157-166. [36] Davatgaran V, Saniei M, Mortazavi S. Optimal bidding strategy for an energy hub in energy market, Energy 2018;148:482-493. [37] Amiri S, Honarvar M, sadegheih A. Providing an Integrated Model for Planning and Scheduling Energy Hubs and Preventive Maintenance, Energy 2018;163:1093-1114.
ACCEPTED MANUSCRIPT
[38] Ramos-Teodoro J, Rodríguez F, Berenguel M, Torres J, Heterogeneous resource management in energy hubs with self-consumption: Contributions and application example, Applied Energy 2018;229:537–550. [39] Ghaffarpour R, Mozafari B, Ranjbar A, Torabi T. Resilience oriented water and energy hub
RI PT
scheduling considering maintenance constraint, Energy 2018;158:1092-1104.
[40] Aghamohamadi M, Samadi M, Rahmati I. Energy Generation Cost in Multi-Energy Systems; an Application to a Nonmerchant Energy Hub in Supplying Price Responsive Loads, Energy 2018;161:878-891.
SC
[41] Geidl M, Koeppel G, Favre-Perrod P, Klöckl B .The Energy Hub – A Powerful Concept for Future Energy Systems. Third Annual Carnegie Mellon Conference on the Electricity Industry
M AN U
2007:13 – 14.
[42] Goldstein L, Hedman B, Knowles D, Freedman S.I, Woods R, Schweizer T. Gas-Fired Distributed Energy Resource Technology Characterizations, National Renewable Energy Laboratory (NREL), 2003. [43]
Chorak
A,
Palenzuela
P,
Alarcón-Padilla
D.C, Abdellah
A. B.
Experimental
characterization of a multi-effect distillation system coupled to a flat plate solar collector
TE D
field: empirical correlations. Applied Thermal Engineering 2017; 120:298-313. [44] Water & Process Solutions, FILMTEC™ Reverse Osmosis Membranes Technical, Dow Water & Process Solutions - The Dow Chemical Company. (Access time: 2011) [45] Guideline for drinking Water Quality, Fourth edition incorporation the first addendum,
EP
World Health Organization. 2017.
[46] Asgari S, Noorpoor A.R, Boyaghchi F.A. Parametric assessment and multi-objective
AC C
optimization of an internal auto-cascade refrigeration cycle based on advanced exergy and exergoeconomic concepts. Energy 2017; 125:576-590. [47] Rosen M, Dincer I, Kanoglu M. Role of exergy in increasing efficiency and sustainability and reducing environmental impact. Energy Policy 2008; 36:128-137. [48] Ameri M, Mokhtari H, Mostafavi Sani M. 4E analyses and multi-objective optimization of different fuels application for a large combined cycle power plant. Energy 2018; 156:371-386. [49] Boyaghchi F. A, Chavoshi M, Sabeti V. Multi-generation system incorporated with PEM electrolyzer and dual ORC based on biomass gasification waste heat recovery: Exergetic, economic and environmental impact optimizations. Energy 2018; 145:38-51.
ACCEPTED MANUSCRIPT
[50] Sharqawy M.H, Lienhard J.H, Zubair S.M. On exergy calculations of seawater with applications in desalination systems, International Journal of Thermal Sciences 2011;50:187196. [51] Kotas T. J. The Exergy Method of Thermal Plant Analysis. 1985.
RI PT
[52] Ghasemi A, Heidarnejad P, Noorpoor A.R. A novel Solar-Biomass Based Multi-Generation Energy System Including Water Desalination and Liquefaction of Natural Gas System: Thermodynamic and Thermoeconomic optimization. Journal of Cleaner Production 2018; 196:424-437.
SC
[53] Piacentino A. Application of advanced thermodynamics, thermoeconomics and exergy costing to a Multiple Effect Distillation plant: In-depth analysis of cost formation process.
M AN U
desalination 2015; 371:88–103.
[54] Hajabdollahi H. Investigating the effects of load demands on selection of optimum CCHPORC plant. Applied Thermal Engineering 2015; 87:547-558
[55] Breeze P. The Cost of Power Generation The current and future competitiveness of renewable and traditional technologies. 2010.
[56] Sanaye S, Khakpaay N. Simultaneous use of MRM (maximum rectangle method) and
TE D
optimization methods in determining nominal capacity of gas engines in CCHP (combined cooling, heating and power) systems. Energy 2014;72:145-158. [57] Web Site of Cement Qeshm Factory, www.qeshmcement.com. (Access time: 2016) [58] Web Site of Iranian Cement Factories, www.irancement.com. (Access time: 2017)
EP
[59] Boyaghchi F. A, Chavoshi M. Monthly assessments of exergetic, economic and environmental criteria and optimization of a solar micro-CCHP based on DORC. Solar Energy
AC C
2018; 166: 351-370.
[60] Web Site of Iranian Ministry of Energy, www.moe.gov.ir. (Access time: 2018) [61] Hajabdollahi H, Ganjehkaviri A, Jaafar M.N. Assessment of new operational strategy in optimization of CCHP plant or different climates using evolutionary algorithms. Applied Thermal Engineering 2015; 75:468-480. [62] Mabrouk A, Nafey A.S, Fath H.E.S. Thermoeconomic analysis of some existing desalination processes. Desalination 2007; 205:354–373. [63] Web Site of Iranian Fuel Conservation Organization, www.ifco.org. (Access time:2018)
ACCEPTED MANUSCRIPT
[64] Du Y, Xie L, Liu J, Wang Y, Xu Y, Wang S. Multi-objective optimization of reverse osmosis networks by lexicographic optimization and augmented epsilon constraint method. Desalination 2014; 333:66–81 [65] Energy balance sheet of year 94, Power and Energy Affairs of Iran, Office of Planning for
RI PT
Electricity and Energy. 2015.
AC C
EP
TE D
M AN U
SC
[66] United States Environmental Protection Agency, www.epa.gov. (Access time:2018)
ACCEPTED MANUSCRIPT
Highlights •
Proposing of the modified energy hub model to analyze economic and environmental
RI PT
performance Finding the optimum range of variables based on TAC and exergy efficiency
•
Water is incorporated as an energy stream by simultaneously utilizing MED and RO.
•
The results of optimization are compared to the conventional mode.
AC C
EP
TE D
M AN U
SC
•