Sunteți pe pagina 1din 13

Energy Conversion and Management 185 (2019) 130–142

Contents lists available at ScienceDirect

Energy Conversion and Management


journal homepage: www.elsevier.com/locate/enconman

Comparative analysis of an organic Rankine cycle with different turbine T


efficiency models based on multi-objective optimization

Peng Li, Zhonghe Han , Xiaoqiang Jia, Zhongkai Mei, Xu Han, Zhi Wang
Key Lab of Condition Monitoring and Control for Power Plant Equipment, School of Energy, Power and Mechanical Engineering, North China Electric Power University,
Baoding 071000, China

A R T I C LE I N FO A B S T R A C T

Keywords: The radial-inflow turbine is a key component of the organic Rankine cycle (ORC) system, and its efficiency is
Organic Rankine cycle related to working fluid properties and working conditions. In this paper, multi-objective algorithm was em-
Constant turbine efficiency ployed to conduct a comparative analysis of an ORC system with different turbine efficiency models (constant
Variable turbine efficiency and variable). The system thermal efficiency and multi-objective optimization results between the constant
System thermal efficiency
turbine efficiency ORC system (CTORC) and the variable turbine efficiency ORC system (VTORC) were com-
Multi-objective optimization
pared, and the reasons for the difference between the CTORC and VTORC system were analyzed. Sensitivity
analysis was performed to compare the difference between the CTORC system and the VTORC system at different
waste flue gas inlet temperature. The results show that the system thermal efficiency of both the CTORC and the
VTORC increases with the increasing evaporation temperature and decreases with the increasing condensation
temperature, while the variation rate of system thermal efficiency is different between CTORC system and
VTORC system. The predicted turbine efficiency is significantly different for different working fluids and dif-
ferent working conditions. Considering the system comprehensive performance, R236ea is the optimal working
fluid for the CTORC system, while R365mfc is for the VTORC system. For different working fluids, the error
caused by using constant turbine efficiency model is different. With the increasing waste flue gas inlet tem-
perature, the error caused by using constant turbine efficiency increases for R236ea, while the error caused by
using constant turbine efficiency increases first and then decreases for R365mfc.

1. Introduction found that isentropic working fluids provide a higher thermal efficiency
and a higher specific heat capacity is advantages to improve the net
With the rapid development of the economy and society, the de- power output. Zhang et al. [13] performed a sustainability evaluation of
mand for fossil fuels has been increasing daily, and environmental an ORC system adopting an emergy analysis and life cycle method.
pollution has aroused a growing concern. The utilization of renewable They found that the sustainability of the ORC system is less than that of
energy and the recovery of waste heat are two effective solutions to renewable source (wind, hydro and geothermal) power plants, but
alleviate the energy risk and the environment deterioration [1–3]. Or- much greater than that of fossil fuel (oil and coal) power plants. In
ganic Rankine cycle (ORC) exhibits a high potential in conversion of order to maximize the utilization of the waste heat source and si-
low-grade heat source into electricity due to its simple structure, eco- multaneously keep the specific investment cost low, Garg et al. [14]
nomic feasibility and easy maintenance [4–6]. proposed a novel objective function that can reveal the tradeoff be-
In the last few decades, many studies have been performed in- tween the utilization extent of waste heat and the specific investment
vestigating the ORC system [7–10]. Bademlioglu et al. [11] compared cost. The ORC system was optimized by using particle swarm optimi-
different parameters’ impact weights on the ORC thermal efficiency by zation algorithm based on the novel objective function. Li et al. [15]
employing statistical methods, and found that the influence of the investigated the effects of the evaporation temperature, the pinch point
evaporation temperature, condensation temperature and turbine effi- temperature difference in the evaporator and condenser on the ORC
ciency on the ORC thermal efficiency is more significant. Javanshir system economic performance. The electricity production cost was se-
et al. [12] conducted a thermodynamic analysis of a simple subcritical lected as an evaluation criterion to optimize the parameters for dif-
and supercritical ORC system with different working fluids. It can be ferent working fluids at different heat source temperatures.


Corresponding author.
E-mail address: hanzhonghe@ncepu.edu.cn (Z. Han).

https://doi.org/10.1016/j.enconman.2019.01.117
Received 6 December 2018; Accepted 31 January 2019
0196-8904/ © 2019 Elsevier Ltd. All rights reserved.
P. Li, et al. Energy Conversion and Management 185 (2019) 130–142

Nomenclature ξ loss coefficient


φ Nozzle velocity coefficient
A area, m2 ν specific volume, m3 kg−1
c average cost rate per unit exergy, $ GJ−1; absolute velo- ψ rotor velocity coefficient
city, m s−1 α absolute velocity angle
Ċ cost rate, $ h−1 β relative velocity angle
D diameter, m ϕ maintenance factor
D̄2 hub-diameter ratio ρ degree of reaction
Ė exergy rate, kW
h specific enthalpy, kJ kg−1 Subscripts
L depth of rotor, m
m mass flow rate, kg s−1 0, 1, 2, 3, 4, 5, 6, 7, 9, 10 state points
M molecular weight, kg mol−1 con condenser
Q heat, kW eva evaporator
TCI total capital investment, $ k kth component
u peripheral velocity, m s−1 pump pump
w relative velocity, m s−1 s isentropic
W power, kW the thermal
Ż capital cost rate, $ h−1 tur turbine
ū1 peripheral velocity ratio wf working fluid

Greek letters

η efficiency

Multi-objective optimization can make a trade-off between the thermodynamic and economic performance indicators, respectively,
thermodynamic performance and economic performance to achieve a and the optimization of the ORC with four different working fluids was
better comprehensive performance [16–19]. Yang et al. [20] estab- performed using non-dominated sorting genetic algorithm.
lished a multi-objective model of an ORC system to maximize the net Almost all of the theoretical studies related to ORC system consider
power output and minimize the total capital cost. The model was solved a constant turbine efficiency model. It is evident that different working
using genetic algorithm to screen the optimal working fluid and de- fluid properties and different system operating conditions would lead to
termine the corresponding optimal operating parameters. Gimelli et al. different turbine efficiency. Thus, the results from the conventional
[21] addressed a multi-objective optimization problem of an ORC CTORC system are not necessarily accurate. Da Lio et al. [24] calculated
system using the electricity efficiency and the overall heat exchangers the turbine efficiency based on a mean-line model of the radial-inflow
area as performance indicators, with MDM as working fluid. In Pareto turbine. R245fa with real gas properties was selected as the working
optimal front solutions, the range of the electricity efficiency is from fluid, and it was found that different working conditions lead to dif-
14.1% to 18.9%, and the overall heat exchangers area is between ferent turbine efficiency. Rosset et al. [25] calculated the isentropic
446 m2 and 1079 m2. To find the optimum point of an ORC system with efficiency of a radial-inflow turbine based on a preliminary design map
R123 as the working fluid, multi-objective optimization with respect to accounting for the effect of the pressure ratio. Constrained multi-ob-
the exergy efficiency and the total product unit cost as performance jective optimization of the ORC system was conducted with the net
criteria was conducted based on genetic algorithm [22]. Özahi et al. power output and the heat exchange area as objective functions. Rahbar
[23] employed net power output and total cost rate as the et al. [26] proposed an optimized model of the VTORC system, and the

Fig. 1. The schematic diagram of a basic ORC.

131
P. Li, et al. Energy Conversion and Management 185 (2019) 130–142

DIRECT algorithm was used to maximize the ratio of the cycle net into a cost stream Ċ.
electric power output to the turbine oversize. They found that the
Ċ = cE ̇ (5)
predicted turbine efficiency varies with working fluids, and the max-
imum difference in the turbine efficiency between R245fa and iso- For each component of the ORC system, the cost balance equation is
butane is 6.13%. However, the influence of turbine efficiency model established according to the exergy balance equation:
(constant and variable) selection on the ORC system comprehensive CṖ = CḞ − CḊ + Z ̇ (6)
performance (thermodynamic and economic) based on multi-objective
optimization are not extensively studied. The cost balance equations and the corresponding auxiliary equa-
In the present study, a variable turbine efficiency model is presented tions for each component of the ORC system are listed in Table 1. Ż
to calculate the turbine efficiency, which is unique for each set of op- represents the cost rate for each component of the ORC system, which
erating conditions and working fluid properties. The system thermal includes the total capital investment (TCI) and the operation and
efficiency of the CTORC and the VTORC were compared. Then, the maintenance (O&M) costs. The equations of TCI for different compo-
influence of the turbine efficiency model selection on the ORC system nents are given in Table 1. The capital cost rate can be calculated as:
comprehensive performance was investigated. Considering the ther- CRFϕ
modynamic performance and economic performance, system thermal Ż = TCI
τ (7)
efficiency and the cost per exergy unit of the product are selected as
objective functions. Non-dominated sorting genetic algorithm-II where CRF is the capital recovery factor; ϕ represents the maintenance
(NSGA-II) was employed to optimize the CTORC system and the VTORC factor; and τ represents the annual plant operation hours (7000 h). The
system. Taking R365mfc and R236ea as examples, the difference in the CRF is the function of the average annual interest rate, i (10%) and the
optimization results between the CTORC system and the VTORC system plant economic life, n (20 years).
were analyzed in detail. Moreover, sensitivity analysis of the ORC i (1 + i)n
CRF =
system was conducted with respect to the heat source temperature, and (1 + i)n − 1 (8)
the difference in the results between the CTORC and the VTORC was
also compared. In this study, the cost per exergy unit of the product is selected as an
optimization criterion for the ORC system.
2. Modelling and analysis methods ̇
Cp,total ̇
(c F,total EF,total + ∑k Z ̇ k)
c p,total = =
̇
Ep,total ̇
Ep,total (9)
2.1. Thermodynamic and exergoeconomic modelling

Conceptually, the operating principles of the organic Rankine cycle 2.2. Analysis and calculation of the turbine efficiency
are similar to the conventional steam Rankine cycle, except that it
employs low boiling point organic materials as working fluids instead of A constant turbine efficiency model is employed in the conventional
steam. A basic ORC system consists of four main components, including ORC system analysis process, but, as mentioned in the introduction, this
a pump, an evaporator, a turbine (expander) and a condenser, as pre- assumption is not necessarily accurate. In this study, to investigate the
sented in Fig. 1. The liquid working fluid from the condenser is pres- influence of the turbine efficiency model selection on the ORC system
sured by the pump and enters the evaporator, where it is heated by the performance, the constant turbine efficiency model is replaced by the
waste flue gas and becomes saturated vapor. The saturated working variable turbine efficiency model to predict the turbine efficiency that is
fluid vapor is directed into the turbine and expands to generate the unique for each set of operating conditions and working fluids. Both
mechanical power of the shaft, which is used to drive the generator. ORC analysis procedures are presented in Fig. 3. As shown in Fig. 3, the
Then, the vapor exhausted from the turbine is condensed to liquid by calculation model of the radial-inflow turbine efficiency and the ORC
the condenser, and the liquid working fluid re-enters the pump. The system are coupled with each other.
corresponding T-s diagram of a basic ORC system is presented in Fig. 2. A one-dimensional assumption is adopted in the calculation model
In this study, each component of the ORC system is considered to be of the radial-inflow turbine efficiency. Fig. 4 shows the schematic dia-
a control volume. The equations that are used to calculate the heat gram of the radial-inflow turbine (volute and diffuser are neglected) in
absorbed or rejected by the organic working fluid and the power gen- the meridional plane. The simplified flow process in the radial-inflow
eration or consumption are based on the mass and energy conservations turbine is described in Fig. 5. It can be found that the incoming working
and are given in Table 1. In addition to these equations, the net power fluid sequentially expands in the nozzle (0–1) and rotor (1–2), and the
output and the thermal efficiency of the ORC system are calculated as
follows:
Wnet = Wtur − Wpump (1)

ηthe = Wnet / Qeva (2)


Based on the second law of thermodynamics, the exergy balance
equations on the ORC system components can be expressed as:
EḊ = EḞ − EṖ − EL̇ (3)
where ĖF and ĖP represent the exergy rate of the fuel and the product,
respectively. ĖL describes the exergy rate that is lost to the ambient or
leaves the system. ĖD is the exergy destruction rate. The total exergy of
a working fluid stream is expressed as follows:
E ̇ = m [(h − henv ) − Tenv (s − senv )] (4)
The exergoeconomic analysis method integrates the thermodynamic
(exergy) aspect and economic principle. In this method, the exergy
stream Ė is multiplied by an average cost per exergy unit c to convert it Fig. 2. The T-s diagram of a basic ORC.

132
P. Li, et al. Energy Conversion and Management 185 (2019) 130–142

Table 1
Thermodynamic and cost balance equations for each component of the ORC system.
Components Thermodynamic equations Cost equations Total capital investment

Wpump = mwf (h5 − h4 ) C5̇ = C4̇ + CW


̇ pump + Zpump
̇ cWpump = cWtur 0.71 1 − 0.8
TCIpump = 422Wpump [1.41 + 1.41( )]
1 − ηpump
= mwf (h5s − h4 )/ ηpump

Pump
Qeva = mwf (h 0 − h5) C8̇ + C0̇ = C5̇ + C7̇ + Zeva
̇ c8 = c 7 TCIeva = 10, 000 + 324A0.91

Evaporator
Wtur = mwf (h 0 − h2) C2̇ + CẆ tur = C0̇ + Ztur
̇ 0.7
TCItur = 6000Wtur
= mwf (h 0 − h2s)·ηtur c2 = c0

Turbine
Qcon = mwf (h2 − h4 ) C4̇ + C10
̇ = C2̇ + C9̇ + Zcon
̇ c2 = c4 c9 = 0 TCIcon = 10, 000 + 324A0.91

Condenser

enthalpy of the working fluid is converted into the mechanical power of (c12 − c22) + (u12 − u22) + (w22 − w12)
ηu =
the radial-inflow turbine. The initial state of the organic vapor (from C02 (10)
the evaporator) at the radial-inflow turbine inlet is represented by point
0. In practical operations, due to the flow losses in the passages, the The relative velocity is defined as the ratio of each velocity to the
state of the organic vapor at the nozzle outlet changes from point 1s to ideal velocity, and the ideal velocity can be calculated as follows:
point 1, and the state of the organic vapor at the rotor outlet changes
from point 2s to point 2. Δhs is the total isentropic enthalpy drop of the c0 = 2Δhs (11)
organic working fluid across the entire radial-inflow turbine, and Δh is
the total actual enthalpy drop. Fig. 6 shows the velocity triangles, Therefore, the peripheral efficiency of the radial-inflow turbine can
which are used to express the distribution of absolute velocity, relative be expressed as:
velocity and velocity angles in the nozzle and rotor.
The peripheral efficiency of the radial-inflow turbine is determined ηu = c¯12 − c¯22 + u¯ 12 − u¯ 22 + w¯ 12 − w¯ 22
from: = 2(¯c1u u¯1 − c¯2u u¯2) (12)

Fig. 3. ORC system analysis procedure with constant and variable turbine efficiency.

133
P. Li, et al. Energy Conversion and Management 185 (2019) 130–142

w¯ 12 = c¯12 + u¯ 12 − 2¯c1 u¯1 cos α1 (17)

c¯1 = φ 1 − ρ (18)
Substitute Eqs. (17) and (18) into Eq. (16):

w¯ 2t = ρ + φ2 (1 − ρ) + D¯ 22 u¯ 12 − 2u¯1 φ cos α1 1 − ρ (19)


Therefore, the peripheral efficiency of the radial-inflow turbine can
be calculated by a dimensionless equation:

ηu = 2u¯1 [φ cos α1 1 − ρ
− D¯ 22 u¯1 + D¯ 2 ψ cos β2 ρ + φ2 (1 − ρ) + D¯ 22 u¯ 12 − 2u¯1 φ cos α1 1 − ρ ]
(20)
It can be determined that the peripheral efficiency of the radial-
inflow turbine is influenced by seven different dimensionless para-
meters. The peripheral velocity ratio ū1 is the ratio of the rotor per-
ipheral velocity to the ideal velocity, and the degree of reaction ρ is the
ratio of the enthalpy drop in the rotor to the enthalpy drop in the entire
radial-inflow turbine. The two parameters have a greater effect on the
peripheral efficiency of the radial-inflow turbine [27]. The peripheral
Fig. 4. Schematic diagram of the radial-inflow turbine in the meridional plane. velocity ratio and degree of reaction are optimized using a Lagrange
multiplier method with the peripheral efficiency of the radial-inflow
turbine as the objective. The optimal peripheral velocity ratio and de-
gree of reaction can be calculated as follows [28]:
ψ
u¯1opt =
cos β2 (1 − m2ψ4 ) φ2 cos2 α1
D¯ 22 ( 2 2 − ψ2 ) +
m m2ψ2 (1 − φ2) (21)

(1 − mψ2) φ cos α1 u¯1 2


ρopt = 1 − [ ]
mψ2 (1 − φ2) (22)
where m is a combined parameter that can be expressed as:

1 D¯ 22 (1 − φ2)(1 − cos2 β2 ψ2)


m= [1 ± ]
ψ2 cos2 α1 φ2 + D¯ 22 ((1 − φ2) (23)
From Eq. (16), w̄12 can be expressed as:
w¯ 12 = w¯ 2t2 − ρ + (u¯ 12 − u¯ 22) (24)
Substitute Eqs. (18) and (24) into Eq. (12), another expression of the
peripheral efficiency of the radial-inflow turbine can be presented as
follows:
Fig. 5. Flow process of the working fluid in the radial-inflow turbine. ηu = 1 − (1 − ρ)(1 − φ2) − w¯ 2t2 (1 − ψ2) − c¯22
= 1 − ξ n − ξr − ξ e (25)
where ξn = (1 − ϕ2)(1
− ρ ) , ξr = − w¯ 2t2 (1
and ξe = ψ2 )
represent the c̄22
nozzle loss coefficient, rotor loss coefficient and leaving velocity loss
coefficient, respectively.
In addition to the three types of aforementioned losses, two other
types of losses (friction loss and leakage loss) need to be considered in
the calculation of the turbine efficiency.
The friction loss coefficient can be calculated as:
D12 u1 3 1 1
ξf = f · ·( )· ·
Fig. 6. Velocity triangles of the radial-flow turbine. ν1 100 1.36 m f ·Δhs (26)
The leakage loss coefficient can be calculated as:
D
u¯2 = 2 u¯1 = D¯ 2 u¯1 δ
⎧1 − 1.3· 1m , 0.015 <
δ
⩽ 0.05
D1 (13) 1m
ξ1 =
⎨ 0.95 − 0.31· δ , δ
> 0.05
c¯1u = c¯1 cos α1 = φ 1 − ρ cos α1 ⎩ 1m 1m (27)
(14)
Differing from the other four loss coefficients, the leakage loss
c¯2u = u¯2 − w¯ 2 cos β2 coefficient does not represent the ratio of the leakage loss to the isen-
= D¯ 2 u¯1 − ψw¯ 2t cos β2 (15) tropic enthalpy drop. Thus, the radial-inflow turbine efficiency can be
expressed as:
w¯ 2t = ρ + w¯ 12 − u¯ 12 + D¯ 2 u¯ 12 (16) ηtur = (ηu − ξf )·ξl = (1 − ξn − ξr − ξe − ξf )·ξl (28)

134
P. Li, et al. Energy Conversion and Management 185 (2019) 130–142

An overview of the calculation model of the radial-inflow turbine selection, crossover and mutation operators. The parent population and
efficiency is presented in Fig. 7. A real gas formulation is used in this the offspring population are combined to form a new population
model, and Matlab software is used to solve the calculation model. Rt = Pt ∪ Qt, whose size is 2 N. (3) Since Rt includes all the members in
Since there are not sufficient experimental data for a radial-inflow the parent population and the offspring population, elitism is preserved.
turbine using organic working fluid, an air radial-inflow turbine is used Non-dominated sorting is conducted for population Rt, then population
to validate the calculation model of the radial-inflow turbine efficiency. Rt is classified into different non-dominated fronts F1, F2…, a higher
Jones [29] performed rig test of a radial-inflow turbine with air as ranking means a larger distance from the optimal solutions. (4) A new
working fluid and drew the turbine efficiency characteristic curve population Pt+1 with a size of N is created, and it is filled with members
(experimental curve of the turbine efficiency vs the peripheral velocity of different non-dominated fronts. The new population only has N slots,
ratio), as shown in Fig. 8. The turbine efficiency characteristic curve which is less than the population Qt size 2 N. The filling begins from the
obtained by the calculation model of the radial-inflow turbine effi- members of F1; if the number of members of F1 is less than N, the filling
ciency was also presented in Fig. 8 to compare the calculation results
with the experimental results. The turbine efficiency increases first and
then decreases with the increasing peripheral velocity ratio, which is
accord with Ref. [27]. In addition, it can be seen that the predicted
turbine efficiency coincide with the experimental value well with a
maximum error of 0.573%. Therefore, the radial-inflow turbine effi-
ciency calculation model is considered accurate and can be used in this
study.

2.3. Optimization model and simulation method

The system thermal efficiency and the cost per exergy unit of the
product are selected as objective functions to investigate the influence
of the turbine efficiency model selection on the multi-objective opti-
mization results of the ORC system. NSGA-II is used to maximize the
thermal efficiency and minimize the cost per exergy unit of the product
simultaneously. Multi-objective optimization model of the ORC system
can be described as:

⎧ max(ηthe) = f1 (T0, T4 )
⎩ min(c p,total ) = f2 (T0, T4 )
⎨ (29)

The physical constraints for the multi-objective optimization pro-


blem are listed as follows:

(1) The evaporation temperature should less than the heat source
temperature and critical temperature of the working fluid, and the
pinch point temperature difference in the evaporator should higher
than 5 K.

⎧ T0 < T7
T0 < Tcr

⎩ ΔTeva > 5 (30)

(2) The condensation temperature should less than the evaporation


temperature and higher than the ambient temperature. In this
study, the condensation temperature varies in the range of
313.15–333.15 K.

⎧ Tamb < T4 < T0


⎩313.15 < T4 < 333.15
⎨ (31)

NSGA-II is a computationally efficient multi-objective optimization


algorithm, that was proposed by Deb et al. [30] on the basis of natural
selection in the biological genetic process. Two special mechanisms,
including non-dominated sorting and crowding distance, are used in
NSGA-II. Such mechanisms are advantageous to ensure both the con-
vergence and diversification of the population. NSGA-II has been
widely used by many researchers in different research fields, including
power system optimization [31,32], heat exchanger design [33], and
green building optimization design [34,35].
The main procedures of the NSGA-II are summarized as follows: (1)
Initial parent population P0 is created randomly with a size of N. To
conveniently describe the procedures of the algorithm, the tth genera-
tion was used in the follow description. (2) The offspring population Qt
is generated from the parent population Pt through binary tournament Fig. 7. Flowchart of the turbine efficiency calculation.

135
P. Li, et al. Energy Conversion and Management 185 (2019) 130–142

0.90 increment for the CTORC system is larger than that for the VTORC
0.89 Experimental value(Ref. [29]) system, resulting in the increasing rate of the system thermal efficiency
Predicted turbine efficiency for the CTORC system is higher than that for the VTORC system. For
0.88
both the CTORC system and the VTORC system, the difference in the
0.87 system thermal efficiency for different working fluids is small in the low
Turbine efficiency

evaporation temperature range, and the difference value gradually in-


0.86
creases with the increasing evaporation temperature. In addition,
0.85 compared with the CTORC system, the difference in the system thermal
0.84 efficiency for different working fluids in the VTORC system is much
higher at a high evaporation temperature. Among the six working
0.83 fluids, R236ea always has the lowest thermal efficiency for the CTORC
0.82 system, while for the VTORC system, the lowest thermal efficiency
working fluids change from R236ea to ipentane when the evaporation
0.81
temperature exceeds 376.15 K. The sequence order of some working
0.80 fluids is also different between the CTORC system and the VTORC
0.58 0.60 0.62 0.64 0.66 0.68 0.70 0.72 0.74 0.76 0.78 0.80 0.82
system. For the CTORC system, the sequence order is pentane, ipentane
Velocity ratio and R365mfc in terms of thermal efficiency, while the sequence order
for the same three working fluids changes to R365mfc, pentane and
Fig. 8. Validation of the radial-inflow turbine efficiency calculation model.
ipentane for the VTORC system.
The turbine efficiency with different working fluids at various
continued with the members of F2, F3…. Say that Fl is the last non- evaporation temperatures is shown in Fig. 12. The turbine efficiency
dominated front that can be accommodated by population Pt+1. In decreases with the increasing evaporation temperature. For different
general, the number of solutions in Fl is more than the remaining slots working fluids the predicted turbine efficiency is different, and the
in the population Pt+1. Thus, a novel comparison operator called difference increases with the increment of the evaporation temperature.
crowding distance was proposed to sort the solutions in Fl. In the same The radial-inflow turbine with R365mfc as working fluid has the largest
rank, the solutions with a larger crowding distance are better and are turbine efficiency among the six working fluids, and the maximum
chosen prior to fill population Pt+1. (5) If the stopping conditions are turbine efficiency is 0.827 when the evaporation temperature is
not met, a new offspring population Qt+1 is created, and the procedure 353.15 K. The radial-inflow turbine with ipentane as working fluid has
returns to step (2). The iteration ends once the stopping conditions are the smallest turbine efficiency, and the minimum turbine efficiency is
met. The basic procedures of NSGA-II are shown in Fig. 9. 0.678 when the evaporation temperature is 412.15 K. The difference
between the maximum turbine efficiency and the minimum turbine
3. Results and discussion efficiency is 0.149, which indicates that using constant turbine effi-
ciency for different working fluids and different operating conditions
3.1. Simulation conditions and working fluids will lead to a serious error.
To analyze and compare the differences between the CTORC system
In this study, an ORC system utilizing the waste flue gas is in- and the VTORC system, the system thermal efficiency difference is
vestigated. The waste flue gas inlet temperature is assumed to be defined as follows:
433.15 K (a common temperature of the waste flue gas) [5,36]. To
Δηthe = ηthe, CTORC − ηthe, VTORC (32)
prevent the low-temperature corrosion, the waste flue gas outlet tem-
perature should exceed the acid dew point, thus the waste flue gas The system thermal efficiency difference between the CTORC and
outlet temperature is assumed to be 363.15 K [37,38]. Additionally, to the VTORC is presented in Fig. 13. For pentane and ipentane, the
compare the comprehensive performance of different working fluids system thermal efficiency difference becomes higher with the in-
more accurate, the utilized total waste heat is assumed to be constant at creasing evaporation temperature, and the variation rate gradually in-
1 MW [39]. The constant turbine efficiency is assumed to be 0.8 in this creases. The reason for this outcome is that only in a low evaporation
study [40,41]. Six organic working fluids, R114, pentane, R236ea, temperature range the turbine efficiency for pentane and ipentane are
ipentane, R365mfc and hexane, were analyzed and compared between slightly higher than the constant turbine efficiency 0.8, and the re-
the CTORC and VTORC. Basic thermodynamic properties of the selected maining turbine efficiency for pentane and ipentane are lower than the
working fluids are given in Table 2. constant turbine efficiency 0.8. Additionally, the higher the evaporation
temperature is, the lower the turbine efficiency is, and the larger the
3.2. Comparison of the CTORC system and the VTORC system results deviation between the predicted turbine efficiency and the constant
turbine efficiency is. Therefore, a higher evaporation temperature
To investigate the influence of the turbine efficiency model selection
on the ORC system, a thermodynamic analysis of both the CTORC
system and the VTORC system with different working fluids was con-
ducted at various operation conditions. The variation of the system
thermal efficiency with the increasing evaporation temperature for the
CTORC system and the VTORC system are presented in Figs. 10 and 11,
respectively. The system thermal efficiency increases with the incre-
ment of the evaporation temperature for both the CTORC and the
VTORC, and the increasing rate for both systems is significant different.
For the CTORC system, the turbine efficiency is constant, and the en-
thalpy drop in the radial-inflow turbine increases with the increasing
evaporation temperature. For the VTORC system, the turbine efficiency
decreases with the increment of the evaporation temperature (just as
presented in Fig. 12), reducing the increment of the enthalpy drop in
the radial-inflow turbine. Therefore, the amount of net power output Fig. 9. Basic procedures of NSGA-II.

136
P. Li, et al. Energy Conversion and Management 185 (2019) 130–142

Table 2 0.84
Basic thermal properties of the selected working fluids.
0.82
Working fluids M/g·mol−1 Tbolling/K Pcritical/MPa Tcritical/K
0.80
R114 170.92 276.741 3.25 145.68
Pentane 72.15 309.21 3.37 469.70 0.78

Turbine efficiency
R236ea 152.04 279.34 3.50 412.44
Ipentane 72.15 300.98 3.34 460.35 0.76
R365mfc 148.07 313.3 3.27 460.00
Hexane 86.18 341.86 3.03 507.82 0.74 R114
pentane
0.72
R236ea
13
0.70 ipentane
R365mfc
0.68 hexane
12
System thermal efficiency/%

0.66
350 360 370 380 390 400 410 420
11 Evaporation temperature/K
Fig. 12. Turbine efficiency at different evaporation temperatures.
10 R114
pentane 2.5
9
R236ea
R114
ipentane

System thermal efficiency difference/%


2.0 pentane
R365mfc
R236ea
8 hexane
ipentane
1.5 R365mfc
7 hexane
350 360 370 380 390 400 410 420
Evaporation temperature/K 1.0

Fig. 10. Variation of system thermal efficiency with the evaporation tempera-
ture for the CTORC. 0.5

13 0.0

12 -0.5
350 360 370 380 390 400 410 420
System thermal efficiency/%

Evaporation temperature/K
11
Fig. 13. Variation of the system thermal efficiency difference with the eva-
poration temperature.
10
R114 11.0
pentane
9
R236ea 10.5
ipentane
10.0
R365mfc
System thermal efficiency/%

8
hexane 9.5

7 9.0
350 360 370 380 390 400 410 420
Evaporation temperature/K 8.5
R114
pentane
Fig. 11. Variation of system thermal efficiency with the evaporation tempera- 8.0 R236ea
ture for the VTORC. ipentane
7.5 R365mfc
would lead to a larger system thermal efficiency difference between the 7.0
hexane
CTORC and the VTORC. For R365mfc, R114, hexane and R236ea, when
the evaporation temperature is low, the turbine efficiency is higher than 6.5
312 314 316 318 320 322 324 326 328 330 332 334
the constant turbine efficiency 0.8, thus the CTORC system thermal Condensation temperature/K
efficiency is lower than the VTORC system thermal efficiency. As the
evaporation temperature increases gradually, the system thermal effi- Fig. 14. Variation of system thermal efficiency with the condensation tem-
ciency difference decreases due to the decreasing turbine efficiency, as perature for the CTORC.
shown in Fig. 13. When the evaporation temperature reaches a value
that makes the predicted turbine efficiency equal 0.8, the system to increase. Additionally, the higher the evaporation temperature is, the
thermal efficiency difference equals 0. With a further increasing eva- larger the system thermal efficiency difference is.
poration temperature, the turbine efficiency continues to decrease, thus Figs. 14 and 15 present the variation of the system thermal effi-
the CTORC system thermal efficiency is higher than the VTORC system ciency with the increasing condensation temperature for the CTORC
thermal efficiency, and the system thermal efficiency difference begins and the VTORC, respectively. The system thermal efficiency decreases

137
P. Li, et al. Energy Conversion and Management 185 (2019) 130–142

11.0 12
Ideal
10.5 point

10.0 11
System thermal efficiency/%

System thermal efficiency/%


9.5
10
9.0
R114
8.5 R114 pentane
pentane 9 R236ea
8.0
R236ea ipentane
7.5 ipentane R365mfc
R365mfc 8 hexane
7.0 hexane
6.5
312 314 316 318 320 322 324 326 328 330 332 334 7
15 16 17 18 19 20 21 22 23 24 25 26
Condensation temperature/K
Cost per exergy of the product/$·GJ-1
Fig. 15. Variation of system thermal efficiency with the condensation tem-
perature for the VTORC. Fig. 18. Pareto optimal front for the CTORC system.

0.83 12

0.82
11
0.81
System thermal efficiency/%
0.80
Turbine efficiency

10
0.79

0.78 R114
R114 9 pentane
0.77 pentane R236ea
R236ea ipentane
0.76
ipentane 8 R365mfc
0.75 R365mfc hexane
hexane
0.74 7
312 314 316 318 320 322 324 326 328 330 332 334 15 16 17 18 19 20 21 22 23 24 25 26
Condensation temperature/K Cost per exergy of the product/$·GJ-1
Fig. 16. Turbine efficiency at different condensation temperatures. Fig. 19. Pareto optimal front for the VTORC system.

0.8 11.5 0.820


Constant turbine efficiency Turbine efficiency
0.7 R114
Dynamic turbine efficeincy
System thermal efficiency difference/%

pentane 11.0
0.6 0.815
R236ea
System thermal efficiency/%

10.5
0.5 ipentane
R365mfc
Turbine efficiency
0.4 10.0 0.810
hexane
0.3
9.5 0.014
0.2 0.805
9.0
0.1

0.0 8.5 0.800


0.003
-0.1 8.0
-0.2 0.795
15 16 17 18 19 20 21 22 23
-1
-0.3 Cost per exergy of the product/$·GJ
312 314 316 318 320 322 324 326 328 330 332 334
Condensation temperature/K Fig. 20. Comparison of the Pareto optimal front between the CTORC system
and the VTORC system for R365mfc.
Fig. 17. Variation of system thermal efficiency difference with the condensa-
tion temperature.
VTORC decreased 2.58%. With the increasing condensation tempera-
ture, the enthalpy drop in the radial-inflow turbine decreases, while the
linearly with the increasing condensation temperature for the CTORC
turbine efficiency increases, as shown in Fig. 16, reducing the decre-
and the VTORC, where the decreasing rate for the VTORC is lower than
ment of the enthalpy drop in the radial-inflow turbine. Therefore, the
that for the CTORC. Taking ipentane as an example, in the investigated
decrement of the net power output for the VTORC system is less than
condensation temperature range the system thermal efficiency for the
that for the CTORC, leading to the decreasing rate of the system thermal
CTORC decreased 3.20%, while the system thermal efficiency for the

138
P. Li, et al. Energy Conversion and Management 185 (2019) 130–142

12.0 0.80 390


Constant turbine efficiency
Dynamic turbine efficeincy CTORC system
11.5
Turbine efficiency VTORC system

Optimal evaporation temperature/K


0.78
System thermal efficiency/%

0.035
11.0 385
0.077

Turbine efficiency
0.76
10.5

380
10.0
0.74

9.5
0.72 375
9.0

8.5 0.70
16 17 18 19 20 21 22
370
Cost per exergy of the product/$·GJ-1

Fig. 21. Comparison of the Pareto optimal front between the CTORC system
and the VTORC system for R236ea.
322.0
CTORC system

Optimal condensation temperature/K


410 VTORC system
CTORC system
Optimal evaporation temperature/K

405 VTORC system 321.5

400

321.0
395

390
320.5

385

380 320.0

375 412 416 420 424 428 432 436 440 444 448
Waste flue gas inlet temperature/K
370
323.5 Fig. 23. Optimal evaporation temperature and condensation temperature at
CTORC system different waste flue gas inlet temperatures for R365mfc.
Optimal condensation temperature/K

323.0 VTORC system


0.82
322.5

0.014
322.0
0.80
0.010
0.014
321.5
Turbine efficiency

0.78
321.0 R365mfc
R236ea
320.5 0.76
0.076
320.0
0.74
319.5
412 416 420 424 428 432 436 440 444 448
Waste flue gas inlet temperature/K 0.72

Fig. 22. Optimal evaporation temperature and condensation temperature at


412 416 420 424 428 432 436 440 444 448
different waste flue gas inlet temperatures for R236ea.
Waste flue gas inlet temperature/K

efficiency for the VTORC being lower than that for the CTORC. With the Fig. 24. Predicted turbine efficiency at different waste flue gas inlet tempera-
increasing condensation temperature, the difference in the system ture.
thermal efficiency for different working fluids decreases, and the var-
iation of the difference in system thermal efficiency for different 0.57 for the VTORC system. The sequence order of some working fluids
working fluids in the VTORC system is slightly more violent in the in- in terms of the system thermal efficiency for the CTORC system is
vestigated condensation temperature range. When the condensation pentane > ipentane > R365mfc, and the sequence order of the same
temperature increases from 313.15 K to 333.15 K, the difference in the working fluids changes to R365mfc > pentane > ipentane for the
system thermal efficiency between hexane and R236ea changes from VTORC system.
0.74 to 0.51 for the CTORC system, while the difference in the system Fig. 16 shows the turbine efficiency with different working fluids at
thermal efficiency between hexane and R236ea changes from 1.00 to various condensation temperatures. As the condensation temperature

139
P. Li, et al. Energy Conversion and Management 185 (2019) 130–142

increases, the turbine efficiency increases linearly, and the difference of are different, as shown in Fig. 21. The reason for this outcome is that
turbine efficiency among different working fluids decreases. Among the the predicted turbine efficiency in the VTORC system is lower than the
six working fluids, R365mfc has the largest turbine efficiency, and constant turbine efficiency 0.8, and the maximum deviation reaches
ipentane has the smallest turbine efficiency. A maximum turbine effi- 0.077, as shown in Fig. 21; in addition, the maximum relative deviation
ciency of 0.821 is obtained by R365mfc when the condensation tem- reaches 9.62%. The low predicted turbine efficiency leads to the dif-
perature is 333.15 K, and a minimum turbine efficiency of 0.745 is ference in the Pareto optimal front between the CTORC system and the
obtained by ipentane when the condensation temperature is 313.15 K. VTORC system, and the difference increases with the decreasing pre-
The difference between the maximum turbine efficiency and the dicted turbine efficiency, as shown in Fig. 21. For the CTORC system,
minimum turbine efficiency is 0.076. Fig. 17 presents the system the turbine efficiency is the same for R365mfc and R236ea, and R236ea
thermal efficiency difference between the CTORC and the VTORC. For is better than R365mfc in terms of comprehensive performance. How-
pentane and ipentane, with the increasing condensation temperature, ever, for the VTORC system, the predicted turbine efficiency of
the system thermal efficiency difference decreases, and the decreasing R365mfc is higher than that of R236ea, making R365mfc better than
rate also gradually decreases. It is due to the fact that the turbine ef- R236ea in terms of comprehensive performance. Additionally, it can
ficiency for pentane and ipentane is lower than the constant turbine also be found that the influence of the turbine efficiency model selec-
efficiency 0.8, and as the condensation temperature increases, the tur- tion on the multi-objective optimization results of the ORC system is
bine efficiency increases; in addition, the deviation between the pre- related to the working fluid properties; for different working fluids, the
dicted turbine efficiency and the constant turbine efficiency decreases. influences are different.
For hexane, R114 and R365mfc, with the increasing condensation
temperature, the system thermal efficiency difference between the 3.3. Sensitivity analysis of the CTORC system and the VTORC system
CTORC and the VTORC increases, while the increasing rate gradually
decreases. The reason is that the turbine efficiency for hexane, R114 The waste flue gas inlet temperature varies with the factory and
and R365mfc is higher than the constant turbine efficiency 0.8, and environmental conditions. In this section, sensitivity analysis of the
with the increasing condensation temperature, the turbine efficiency CTORC system and the VTORC system were conducted to investigate
increases, and the deviation between the predicted turbine efficiency the influence of the turbine efficiency model selection on the ORC
and the constant turbine efficiency increases. For R236ea, the turbine system at different waste flue gas inlet temperatures. R365mfc and
efficiency is lower than the constant turbine efficiency 0.8 in the low R236ea were selected as working fluids. Multi-objective optimization of
condensation temperature range, and with the increasing condensation the CTORC and the VTORC were conducted at each waste flue gas inlet
temperature, the turbine efficiency gradually increases to over 0.8. temperature, and the aid of an ideal point method [40,42] was used to
Therefore, the system thermal efficiency difference between the CTORC determine the final optimal solution. The process of decision-making
and the VTORC decreases first and then increases. for the aid of an ideal point method is shown in Fig. 18. A hypothetical
NSGA-II was employed to conduct multi-objective optimization of point (ideal point) that maximizes the system thermal efficiency and
the CTORC system and the VTORC system based on the system thermal simultaneously minimizes the cost per exergy of the product is as-
efficiency and the cost per exergy unit of the product. The Pareto op- sumed, and the point of the Pareto optimal front that is nearest to the
timal fronts for the CTORC system and the VTORC system are presented ideal point is considered as the final optimal solution.
in Figs. 18 and 19, respectively. It is evident that there is a significant Figs. 22 and 23 show the optimal evaporation temperature and
difference in the distribution of the Pareto optimal front between the condensation temperature at different waste flue gas inlet temperatures
CTORC system and the VTORC system. For all working fluids except for for R236ea and R365mfc, respectively. As shown in Fig. 22, for R236ea,
R365mfc, the distribution ranges of the system thermal efficiency of the the optimal evaporation temperature and the optimal condensation
CTORC are not only higher than that for the VTORC, but also wider temperature for the CTORC system and the VTORC system increase
than that for the VTORC. Taking R236ea as an example, the distribution with the increasing waste flue gas inlet temperature, and the increasing
range of the system thermal efficiency for the CTORC is from 9.11% to rate gradually decreases. It is apparent that the optimal evaporation
11.77%, while the distribution range of the system thermal efficiency of temperature of the CTORC is always higher than that of the VTORC.
the VTORC is from 8.66% to 10.53%. Based on the comprehensive Moreover, as the waste flue gas inlet temperature increases, the dif-
performance, the sequence order of some working fluids for the CTORC ference in the optimal evaporation temperature between the CTORC
system is different from that for the VTORC system. For the CTORC system and the VTORC system increases, indicating that the error
system, R236ea is better than R365mfc, and ipentane is better than caused by using constant turbine efficiency becomes larger with the
hexane, while for the VTORC system, the working fluids sequence order increasing waste flue gas inlet temperature. The optimal condensation
is the opposite; that is, R365mfc is better than R236ea, and hexane is temperature of the CTORC system varies in the range of 320.0–321.5 K,
better than ipentane. The optimal working fluids for the CTORC system while the optimal condensation temperature of the VTORC system
and the VTORC system are also different. For the CTORC system, varies in a slightly higher range of 321.0–323.5 K. The difference in the
R236ea is the optimal working fluid among the six working fluids. optimal condensation temperatures between the CTORC system and the
However, for the VTORC system, R365mfc is the optimal working fluid. VTORC system also increases with the increment of the waste flue gas
To analyze the reason that lead to the optimal working fluids being inlet temperature. As shown in Fig. 24, for R236ea, with the increasing
different between the CTORC system and the VTORC system in detail, waste flue gas inlet temperature, the predicted turbine efficiency de-
Figs. 20 and 21 show a comparison of the Pareto optimal fronts between creases and the deviation between the predicted turbine efficiency and
the CTORC system and the VTORC system for R365mfc and R236ea, the constant turbine efficiency 0.8 increases. Therefore, the error from
respectively. Additionally, the predicted turbine efficiency corre- using the constant turbine efficiency in an ORC system with R236ea as
sponding to the Pareto optimal fronts of the VTORC system are also working fluid increases with the increment of the waste flue gas inlet
presented in Figs. 20 and 21. As shown in Fig. 20, for R365mfc, the temperature.
Pareto optimal front of the CTORC system and the VTORC system are As shown in Fig. 23, for R365mfc, as the waste flue gas inlet tem-
essentially in coincidence. That is due to the fact that there is a small perature increases, the optimal evaporation temperature and the op-
deviation between the predicted turbine efficiency in the VTORC timal condensation temperature for the CTORC system and the VTORC
system and the constant turbine efficiency 0.8. The maximum deviation system increase, and the increasing rate gradually increases. The op-
between the predicted turbine efficiency and the constant turbine ef- timal evaporation temperature of the CTORC system is lower than that
ficiency 0.8 is only 0.014, as shown in Fig. 20. However, for R236ea, of the VTORC system, while the optimal condensation temperature of
the Pareto optimal fronts of the CTORC system and the VTORC system the CTORC system is higher than that of the VTORC system in the low

140
P. Li, et al. Energy Conversion and Management 185 (2019) 130–142

waste flue gas inlet temperature range. With the increasing waste flue turbine efficiency model selection on the ORC system at different waste
gas inlet temperature, the difference in the optimal evaporation tem- flue gas inlet temperatures.
perature and the difference in the optimal condensation temperature
between the CTORC system and the VTORC system decreases. When the (1) When the constant turbine efficiency is replaced with the variable
waste flue gas inlet temperature is higher than 435.9 K, the optimal turbine efficiency, the variation rate of the system thermal effi-
evaporation temperature of the CTORC system is higher than that of the ciency with evaporation temperature or condensation temperature
VTORC system, while the optimal condensation temperature of the decreases, and the sequence order of some working fluids also
CTORC system is lower than that of the VTORC system. With the in- changes. For different working fluids and different working condi-
creasing waste flue gas inlet temperature, both the difference in the tions, the predicted turbine efficiency is quite different.
optimal evaporation temperature and the difference in the optimal (2) Different turbine efficiency models lead to different multi-objective
condensation temperature between the CTORC system and the VTORC optimal results for ORC system. The distribution of Pareto optimal
system increase. The reason for the aforementioned results is that the fronts between the CTORC system and the VTORC is significantly
predicted turbine efficiency of R365mfc is higher than the constant different. R236ea is the optimal working fluid for CTORC system,
turbine efficiency 0.8 when the waste flue gas inlet temperature is low, while R365mfc is the optimal working fluid for VTORC system.
as shown in Fig. 24. With the increasing waste flue gas inlet tempera- (3) For different working fluids, the influence of turbine efficiency
ture, the predicted turbine efficiency decreases, and the deviation be- model selection on the ORC system is different. As the waste flue
tween the constant turbine efficiency 0.8 and the predicted turbine gas inlet temperature increases, the error caused by using constant
efficiency also decreases. When the waste flue gas inlet temperature is turbine efficiency increases for R236ea, while the error caused by
higher than 435.9 K, the predicted turbine efficiency is lower than the using constant turbine efficiency increases first and then decreases
constant turbine efficiency 0.8. With the increasing waste flue gas inlet for R365mfc. The predicted turbine efficiency of both R236ea and
temperature, the deviation between the constant turbine efficiency 0.8 R365mfc decreases with the increasing waste flue gas inlet tem-
and the predicted turbine efficiency begins to increase. Therefore, the perature.
error from using constant turbine efficiency in an ORC system with
R365mfc as working fluid increases first and then decreases with the Declaration of interests
increasing waster flue gas inlet temperature.
Fig. 24 shows the predicted turbine efficiency of R236ea and The authors declared that there is no conflict of interest.
R365mfc at different waste flue gas inlet temperatures. It is evident that
the predicted turbine efficiency of R365mfc is higher that of R236ea, Acknowledgements
and the difference between the two predicted turbine efficiency in-
creases in the investigated waste flue gas inlet temperature range. The This work was supported by the National Natural Science
predicted turbine efficiency of R236ea and R365mfc decreases with the Foundation of China (NO. 51306059) and the Fundamental Research
increasing waste flue gas inlet temperature. The variation rate of the Funds for the Central Universities in China (NO. 2017XS120).
predicted turbine efficiency of R236ea gradually decreases, while the
variation rate of the predicted turbine efficiency of R365mfc gradually References
increases. As shown in Figs. 22 and 23, the variation range of the op-
timal condensation temperature is obviously smaller than the variation [1] Nawi ZM, Kamarudin SK, Abdullah SS, Lam SS. The potential of exhaust waste heat
range of the optimal evaporation temperature; thus, the optimal eva- recovery (WHR) from marine diesel engines via organic rankine cycle. Energy
2019;166:17–31.
poration temperature is the main factor that influences the predicted [2] Landelle A, Tauveron N, Haberschill P, Revellin R, Colasson S. Organic Rankine
turbine efficiency. For R236ea, the increasing rate of the optimal eva- cycle design and performance comparison based on experimental database. Appl
poration temperature decreases in the high waste flue gas inlet tem- Energy 2017;204:1172–87.
[3] Zhang Z, Li H, Chang H, Pan Z, Luo X. Machine learning predictive framework for
perature range, thus the decreasing rate of the predicted turbine effi- CO2 thermodynamic properties in solution. J CO2 Util 2018;26:152–9.
ciency also decreases accordingly. However, for R365mfc, the [4] Li P, Han Z, Jia X, Mei Z, Han X, Wang Z. Analysis and comparison on thermo-
increasing rate of the optimal evaporation temperature increases in the dynamic and economic performances of an organic Rankine cycle with constant and
one-dimensional dynamic turbine efficiency. Energy Convers Manage
high waste flue gas inlet temperature range, so the decreasing rate of 2019;180:665–79.
the predicted turbine efficiency also increases accordingly. It can also [5] Mahmoudi A, Fazli M, Morad MR. A recent review of waste heat recovery by
be found that the deviation between the constant turbine efficiency 0.8 Organic Rankine Cycle. Appl Therm Eng 2018;143:660–75.
[6] Mahmoudzadeh Andwari A, Pesiridis A, Karvountzis-Kontakiotis A, Esfahanian V.
and the predicted turbine efficiency of R365mfc decreases first and then
Hybrid electric vehicle performance with organic rankine cycle waste heat recovery
increases with the increasing waste flue gas inlet temperature, and the system. Appl Sci 2017;7:437.
maximum deviation value is 0.014. The deviation between the pre- [7] Xu G, Song G, Zhu X, Gao W, Li H, Quan Y. Performance evaluation of a direct vapor
dicted turbine efficiency of R236ea and the constant turbine efficiency generation supercritical ORC system driven by linear Fresnel reflector solar con-
centrator. Appl Therm Eng 2015;80:196–204.
0.8 increases with the increasing waste flue gas inlet temperature, and [8] Quoilin S, Van Den Broek M, Declaye S, Dewallef P, Lemort V. Techno-economic
the maximum deviation value is 0.076. Therefore, it can be concluded survey of Organic Rankine Cycle (ORC) systems. Renew Sustain Energy Rev
that the error caused by using constant turbine efficiency is related to 2013;22:168–86.
[9] Pethurajan V, Sivan S, Joy GC. Issues, comparisons, turbine selections and appli-
the working fluid properties, and the error for R365mfc is smaller than cations – an overview in organic Rankine cycle. Energy Convers Manage
that for R236ea. 2018;166:474–88.
[10] Kazemi N, Samadi F. Thermodynamic, economic and thermo-economic optimiza-
tion of a new proposed organic Rankine cycle for energy production from geo-
4. Conclusions thermal resources. Energy Convers Manage 2016;121:391–401.
[11] Bademlioglu AH, Canbolat AS, Yamankaradeniz N, Kaynakli O. Investigation of
To investigate the influence of the turbine efficiency model selection parameters affecting Organic Rankine Cycle efficiency by using Taguchi and
ANOVA methods. Appl Therm Eng 2018;145:221–8.
on the organic Rankine cycle system, a comparative analysis on the [12] Javanshir A, Sarunac N. Thermodynamic analysis of a simple Organic Rankine
system thermal efficiency and the multi-objective optimization results Cycle. Energy 2017;118:85–96.
between the CTORC system and the VTORC system was conducted. [13] Zhang H, Guan X, Ding Y, Liu C. Emergy analysis of Organic Rankine Cycle (ORC)
for waste heat power generation. J Clean Prod 2018;183:1207–15.
Both the reasons for the difference of the system thermal efficiency and
[14] Garg P, Orosz MS. Economic optimization of Organic Rankine cycle with pure fluids
for the difference of the multi-objective optimization results were and mixtures for waste heat and solar applications using particle swarm optimi-
analyzed in detail. Sensitivity analysis of the CTORC system and the zation method. Energy Convers Manage 2018;165:649–68.
VTORC system were conducted to investigate the influence of the [15] Li YR, Du MT, Wu CM, Wu SY, Liu C, Xu JL. Economical evaluation and

141
P. Li, et al. Energy Conversion and Management 185 (2019) 130–142

optimization of subcritical organic Rankine cycle based on temperature matching Turbomach 1996;118:362–70.
analysis. Energy 2014;68:238–47. [30] Deb K, Pratap A, Agarwal S, Meyarivan TAMT. A fast and elitist multiobjective
[16] Cui Y, Geng Z, Zhu Q, Han Y. Multi-objective optimization methods and application genetic algorithm: NSGA-II. IEEE Trans Evol Comput 2002;6:182–97.
in energy saving. Energy 2017;125:681–704. [31] Ganjehkaviri A, Jaafar MM, Hosseini SE, Barzegaravval H. On the optimization of
[17] Bahlouli K. Multi-objective optimization of a combined cycle using exergetic and energy systems: results utilization in the design process. Appl Energy
exergoeconomic approaches. Energy Convers Manage 2018;171:1761–72. 2016;178:587–99.
[18] Tocci L, Pal T, Pesmazoglou I, Franchetti B. Small scale organic rankine cycle (orc): [32] Wang J, Yan Z, Wang M, Li M, Dai Y. Multi-objective optimization of an organic
a techno-economic review. Energies 2017;10:413. Rankine cycle (ORC) for low grade waste heat recovery using evolutionary algo-
[19] Zhang X, Bai H, Zhao X, Diabat A, Zhang J, Yuan H, et al. Multi-objective optimi- rithm. Energy Convers Manage 2013;71:146–58.
sation and fast decision-making method for working fluid selection in organic [33] Sanaye S, Hajabdollahi H. Thermal-economic multi-objective optimization of plate
Rankine cycle with low-temperature waste heat source in industry. Energy Convers fin heat exchanger using genetic algorithm. Appl Energy 2010;87:1893–902.
Manage 2018;172:200–11. [34] Yang MD, Lin MD, Lin YH, Tsai KT. Multiobjective optimization design of green
[20] Yang F, Zhang H, Song S, Bei C, Wang H, Wang E. Thermoeconomic multi-objective building envelope material using a non-dominated sorting genetic algorithm. Appl
optimization of an organic Rankine cycle for exhaust waste heat recovery of a diesel Therm Eng 2017;111:1255–64.
engine. Energy 2015;93:2208–28. [35] Nguyen AT, Reiter S, Rigo P. A review on simulation-based optimization methods
[21] Gimelli A, Luongo A, Muccillo M. Efficiency and cost optimization of a regenerative applied to building performance analysis. Appl Energy 2014;113:1043–58.
Organic Rankine Cycle power plant through the multi-objective approach. Appl [36] Wu S, Li C, Xiao L, Liu C, Li Y. A comparative study on thermo-economic perfor-
Therm Eng 2017;114:601–10. mance between subcritical and transcritical Organic Rankine Cycles under different
[22] Behzadi A, Gholamian E, Houshfar E, Habibollahzade A. Multi-objective optimi- heat source temperatures. Chin Sci Bull 2014;59:4379–87.
zation and exergoeconomic analysis of waste heat recovery from Tehran's waste-to- [37] Zhang C, Liu C, Xu X, Li Q, Wang S, Chen X. Effects of superheat and internal heat
energy plant integrated with an ORC unit. Energy 2018;160:1055–68. exchanger on thermo-economic performance of organic Rankine cycle based on
[23] Özahi E, Tozlu A, Abuşoğlu A. Thermoeconomic multi-objective optimization of an fluid type and heat sources. Energy 2018;159:482–95.
organic Rankine cycle (ORC) adapted to an existing solid waste power plant. Energy [38] Roy JP, Mishra MK, Misra A. Parametric optimization and performance analysis of a
Convers Manage 2018;168:308–19. waste heat recovery system using Organic Rankine Cycle. Energy 2010;35:5049–62.
[24] Da Lio L, Manente G, Lazzaretto A. A mean-line model to predict the design effi- [39] Chen Q, Xu J, Chen H. A new design method for Organic Rankine Cycles with
ciency of radial inflow turbines in organic Rankine cycle (ORC) systems. Appl constraint of inlet and outlet heat carrier fluid temperatures coupling with the heat
Energy 2017;205:187–209. source. Appl Energy 2012;98:562–73.
[25] Rosset K, Mounier V, Guenat E, Schiffmann J. Multi-objective optimization of turbo- [40] Feng Y, Zhang Y, Li B, Yang J, Shi Y. Sensitivity analysis and thermoeconomic
ORC systems for waste heat recovery on passenger car engines. Energy comparison of ORCs (organic Rankine cycles) for low temperature waste heat re-
2018;159:751–65. covery. Energy 2015;82:664–77.
[26] Rahbar K, Mahmoud S, Al-Dadah RK, Moazami N. Modelling and optimization of [41] Chen LX, Hu P, Zhao PP, Xie MN, Wang FX. Thermodynamic analysis of a High
organic Rankine cycle based on a small-scale radial inflow turbine. Energy Convers Temperature Pumped Thermal Electricity Storage (HT-PTES) integrated with a
Manage 2015;91:186–98. parallel organic Rankine cycle (ORC). Energy Convers Manage 2018;177:150–60.
[27] Li YS, Lu GL. Radial-inflow turbine and centrifugal compressor. Beijing: Machinery [42] Pierobon L, Nguyen TV, Larsen U, Haglind F, Elmegaard B. Multi-objective opti-
Industry Press; 1984. mization of organic Rankine cycles for waste heat recovery: application in an off-
[28] Ji GH. Turbine expander. Beijing: Machinery Industry Press; 1982. shore platform. Energy 2013;58:538–49.
[29] Jones AC. Design and test of a small, high pressure ratio radial turbine. J

142

S-ar putea să vă placă și