- Email: [email protected]

Contents lists available at ScienceDirect

Energy journal homepage: www.elsevier.com/locate/energy

Economic system optimization of air-cooled organic Rankine cycles powered by low-temperature geothermal heat sources €l Walraven a, c, Ben Laenen b, c, William D'haeseleer a, c, * Danie a

University of Leuven (KU Leuven) Energy Institute, TME Branch (Applied Mechanics and Energy Conversion), Celestijnenlaan 300A Box 2421, B-3001 Leuven, Belgium b Flemish Institute for Technological Research (VITO), Boeretang 200, B-2400 Mol, Belgium c EnergyVille (Joint Venture of VITO and KU Leuven), Dennenstraat 7, B-3600 Genk, Belgium

a r t i c l e i n f o

a b s t r a c t

Article history: Received 24 May 2014 Received in revised form 30 September 2014 Accepted 13 November 2014 Available online 12 December 2014

In this paper, an economic system optimization of an air-cooled organic Rankine cycle, powered by geothermal heat, is performed. The goal is to ﬁnd the conﬁguration of the ORC (organic Rankine cycle) which gives the highest possible net present value for the project. The cycle parameters, heat exchanger geometry and air-cooled condenser geometry are optimized together. The developed method is applied to a case study, based on a Belgian geothermal project, after which the most important parameters are varied. It is shown that the air-cooled condenser is a crucial component from both a thermodynamic and economic point of view. The discount rate, electricity price, brine inlet temperature and annual electricity price evolution have a strong inﬂuence on the conﬁguration and efﬁciency of the ORC and on the economics of the project. © 2014 Elsevier Ltd. All rights reserved.

Keywords: ORC (organic Rankine cycle) Geothermal Air-cooled condenser Economics Optimization

1. Introduction Low-temperature geothermal heat sources are widely available [1], but the heat-to-electricity conversion efﬁciency is low due to the low temperature. Much research has been performed to maximize this conversion efﬁciency by the use of ORCs (organic Rankine cycles) [2e5], but most of these studies focus on maximizing the thermodynamic conversion efﬁciency while ignoring the economics of the power plant. Only a limited number of studies are available in the open literature which perform a detailed economic optimization of organic Rankine cycles. Quoilin et al. [6] have minimized the speciﬁc cost of small ORCs. Recently Astolﬁ et al. [7] have used the same objective function for large ORCs powered by geothermal heat. However, many components of the ORC are often assumed to be ideal or modeled very simplistically; pinch-point-temperature differences are ﬁxed and the condenser temperature is ﬁxed. All these parameters depend in fact on the conﬁguration of the cycle

* Corresponding author. University of Leuven (KU Leuven) Energy Institute, TME branch (Applied Mechanics and Energy Conversion), Celestijnenlaan 300A Box 2421, B-3001 Leuven, Belgium. Tel.: þ32 16 32 25 11; fax: þ32 16 32 29 85. E-mail addresses: [email protected] (D. Walraven), [email protected] (B. Laenen), [email protected] (W. D'haeseleer). http://dx.doi.org/10.1016/j.energy.2014.11.048 0360-5442/© 2014 Elsevier Ltd. All rights reserved.

and the choice of the working ﬂuid and have a strong inﬂuence on the performance and the total cost of the ORC. The necessity of taking into account the inﬂuence of the sizing of the components has already been investigated in the literature. Madhawa Hettiarachchi et al. [8] have minimized the ratio of the total heat exchanger surface and the net power produced by the cycle of ﬁxed heat exchanger conﬁgurations. Franco and Villani [9] have ﬁrst optimized the cycle parameters, followed by optimizing the heat exchangers. They have performed an iteration between both optimizations. Walraven et al. [10] have optimized the conﬁguration of shell-and-tube heat exchangers together with the conﬁguration of the cycle. In this paper we go one step further, whereby economics and system optimization are combined to perform an economic system optimization. In such an optimization, the cycle and all the components are optimized together, so that the economic performance of the total installation is maximized. To obtain this goal, all components should be modeled in detail and cost-correlations for all components are necessary. In our previous work [10] we have already modeled most of the components, except for the cooling system and the turbine. The cooling system is a very important part of the installation, because the conversion efﬁciency of the cycle is low and a large part of the heat added to the ORC has to be dumped in the condenser. The cooling can be of the dry or the wet type. In this paper we will focus

D. Walraven et al. / Energy 80 (2015) 104e113

105

Fig. 1. Scheme of a single-pressure, recuperated (a) and double-pressure, simple (b) ORC.

Fig. 2. Geometry of an A-frame air-cooled condenser (a) and the bundle geometry of ﬂat tubes with corrugated ﬁns (b).

on the dry type which, in contrast to wet cooling, can be used at all locations. Different tube and ﬁn shapes are possible. Correlations for round tubes with plain, louvered, slit and wavy ﬁns exist [11e14]. Often ﬂat tubes are used instead of round tubes. Correlations for plain, louvered and slit ﬁns exist in the literature [15e18]. Flat tubes with plain ﬁns are often used in power plants and this type is implemented in this work, based on the work from Yang et al. [15]. In this paper, a model for an axial turbine [19] and a model for an air-cooled condenser [15] are added to our previous work [10], together with cost-correlations for all the components. The net present value of a case study based on a geothermal project in Belgium is maximized and the inﬂuence of different economic parameters are investigated. 2. Physical model 2.1. Organic Rankine cycle The cycles can be simple or recuperated, subcritical or transcritical and can have one or two pressure levels. Two examples are given in Fig. 1, in which the scheme of a single-pressure, recuperated ORC and a double-pressure, simple ORC are shown. All the possible heat exchangers (economizer, evaporator, superheater, desuperheater, condenser and recuperator) are shown in the ﬁgure, but are not always necessary. In all conﬁgurations it is assumed that state 1 is saturated liquid and that the isentropic efﬁciencies of the pump is 80%. More information about the modeling of the cycle can be found in our previous work [4,10].

2.2. Air-cooled condenser 2.2.1. Geometry Different types of ACCs (air-cooled condensers) exist, but in this paper only the A-frame air-cooled condenser with ﬂat tubes and corrugated ﬁns is used. This type is often used in power plants because the pressure drop is lower than the one in ACCs with round tubes [15,20]. Fig. 2 shows the geometry of such an A-frame aircooled condenser and the bundle geometry of ﬂat tubes with corrugated ﬁns. The tube bundle geometry is determined by the tubes' small width Ws, the ﬁn height H, the ﬁn pitch S, the tubes' large width Wl and the length of the tubes Lt. In an A-frame ACC the tube bundles are placed at an angle q with the horizontal. The vapor/two-phase ﬂuid enters the condenser at the top in the vapor duct, ﬂows down the tubes, in which it condenses, and the condensate is collected at the bottom in the condensate head. A fan at the bottom blows air over the tube bundles.

2.2.2. Model description The correlation of Yang et al. [15] is used to model the heattransfer coefﬁcient and the pressure drop at the air side of the heat exchanger. The Nusselt correlation is given as1:

1 Formula (13) in the original paper of Yang et al. [15] gives a Nusselt number which is 10 times to high, which is conﬁrmed in a personal communication by professor Yang.

106

D. Walraven et al. / Energy 80 (2015) 104e113

Nu ¼ 0:05922Re0:9172

S Ws

0:9993

H Ws

0:3706

;

(1)

2.4. Heat exchangers

with Re the Reynolds number deﬁned as:

Re ¼

rair VAmin Ws ; mair

(2)

with rair the density of the air, VAmin the velocity at the minimum ﬂow area and mair the dynamic viscosity of the air. The ﬂuid properties are all calculated at the average temperature and pressure of the air. The heat-transfer coefﬁcient obtained from equation (1) is based on the total heat exchanger surface area, including the area of the ﬁns. The friction factor while neglecting natural convection is:

f ¼

238:8552 0:6684 S Re 2 Ws

1:4129

H Ws

0:1496

2 rair VAmin : 2

:

Dpair m_ air ; rair hfan hel;fan

3. Economics 3.1. Net present value

(3) NPV ¼ CEPC þ

(4)

The electrical power consumption of the fan is calculated as:

_ W fan ¼

The heat exchangers used in this paper are all shell-and-tube heat exchangers because of the availability of correlations for the performance and cost. More information about the modeling and optimization of this type of heat exchangers for the use in ORCs can be found in Walraven et al. [10].

The NPV (net present value)2 is calculated as

The factor 1/2 in equation (3) arrives from the fact that Yang et al. [15] developed their correlation for another type of ﬁns. This other type of ﬁns results in a pressure drop which is about twice as high and a heat-transfer coefﬁcient which is about the same as the conﬁguration investigated in this paper [15]. The pressure drop of the air ﬂowing through the tube bundle is then:

Dpair ¼ f

outlet state. The original prediction was given as a ﬁgure, which is curve-ﬁtted in Appendix A.

(5)

with m_ air the mass ﬂow rate of air, hfan the fan efﬁciency and hel,fan the efﬁciency of the fan engine. The product of the last two parameters is assumed to be 60%. Equations (1) and (3) are valid for q ¼ 60 , Wl ¼ 219 mm, Ws ¼ 19 mm and for

S 0:06 0:16 Ws

700 Re 14500: The correlations of Petukhov and Popov [21] and Gnielinski [22] are used to calculate the pressure drop and heat-transfer coefﬁcient for the single-phase ﬂow in the pipes, respectively. For the twophase ﬂow of the working ﬂuid in the pipes, the CISE correlation [23], the correlation of Chisholm [24] and the correlation of Shah [25] are used to calculate the void fraction, pressure drop and heattransfer coefﬁcient, respectively.

2.3. Turbine efﬁciency Three different types of turbines are used in ORCs: the axial ﬂow turbine, the centripetal turbine and the radial-inﬂow, radialoutﬂow turbine. In this paper, we focus on the ﬁrst type because of the availability of correlations. Macchi and Perdichizzi [19] used the Craig and Cox-model [26] to optimize the performance of axial ﬂow-turbines. This resulted in an efﬁciency prediction of a turbine stage based on the inlet state of the turbine and the isentropic

It

t t¼1 ð1 þ iÞ

;

(6)

with CEPC the EPC (engineering, procurement & construction overnight cost) of the installation,3 tLT the lifetime of the installation, It the income in year t and i the discount rate. The EPC cost consists of two parts: the cost of the drilling Cdrilling and the cost of the ORC CORC (see Section 3.2). The income in year t can be calculated as:

_ net p N C ; It ¼ W OM elec

(7)

_ net the net electric power output, which takes a generwith W ator efﬁciency of 98% into account, expressed in MWe, pelec the price obtained per MWh of produced electricity, N the number of full load hours per year (an availability of 95% is assumed) and COM the cost of operation and maintenance of the power plant, which is assumed to be 2.5% of the investment cost of the ORC per year [28]. 3.2. Cost of ORC The overnight EPC investment cost of the ORC, CORC, can be calculated as:

CORC ¼ H 1:25 0:75 Ws

tLT X

X

fM;i fP;i fT;i þ fI CE;i ;

(8)

i

with CE,i the delivered equipment cost of component i and fM,i, fP,i and fT,i correction factors (all 1) for non-standard material, pressure and temperatures, respectively. fI is an average installation-cost factor [29]. This installation-cost factor includes the costs for erection, instrumentation and control of the power plant and is about 0.6 [29,30]. Correlations for the equipment cost CE,i are given in Table 1. Most of the cost-correlations in Table 1 are valid for carbon steel, for design temperatures between 0 and 100 C and for design pressures between 0.5 and 7 bar. Such “normal” designs are good enough for most heat exchangers in a low-temperature ORC. Only the heat exchangers between brine and working ﬂuid work at higher pressures and temperatures and have a higher risk for corrosion. For these heat exchangers, the values of Table 1 are adjusted using the above mentioned correction factors; the tubes are made from stainless steel (fM ¼ 1.7), work at higher pressures

2 In this paper it is chosen to maximize the NPV of the power plant. It is also possible to minimize the LCOE (Levelized Cost of Electricity) [27]. 3 There is a caveat: CEPC is not the total investment cost. Some reﬂections on the investment cost are given in Appendix B.

D. Walraven et al. / Energy 80 (2015) 104e113

107

Table 1 Cost correlation for the different components. The data from Smith [29] and Towler and Sinnott [31] are adapted taking into account that 1 V ¼ 1.35 $ and with a CE (Chemical Engineering)-index of 564 in July 2013. Component

Capacity measure

Size range

Shell-and-tube heat exchanger

A [m2]

80e4000 m2

Cost correlation 3:50 104

_ pump [kW] W

4e700 kW

Air-cooled heat exchanger

Bare-tube A [m2]

200e2000 m2

Fan (incl. motor)

_ W fan [kW]

50e200 kW

_ W turbine [kW]

0.1e20.0 MW

Centrifugal pump (incl. motor)

Turbine

A

80

10:51 103 1:67 105

(fP ¼ 1.5) and at higher temperatures (fT ¼ 1.6) [29]. A correction factor fP ¼ 1.5 is used for the pump. The correlation for the turbine is one for a condensing steam turbine and no pressure-correction factor is used because steam turbines work at higher pressures. 4. Optimization The goal of the optimization is to maximize the net present value of the installation. This optimization is performed with the use of the CasADi [32] and WORHP [33] software. The models themselves are developed in Python and the ﬂuid properties are obtained from REFPROP [34], using the system optimization methodology of our earlier work [10]. 4.1. Optimization variables and constraints In case of a single-pressure simple cycle, the cycle is determined by the temperature before the turbine, the saturation temperature at the pressure before the turbine, the pressure at the inlet of the pump and the mass ﬂow of the working ﬂuid. The effectiveness of the recuperator is added when a recuperated cycle is used. For double-pressure cycles, the temperature before the second turbine, the saturation temperature at the pressure before the second turbine and the mass ﬂow rate through the second turbine are added. More information about these optimization variables can be found in Walraven et al. [10]. The optimization variables of each shell-and-tube heat exchangers are the shell diameter Ds, tube outside diameter do, tube pitch pt, bafﬂe cut lc and the distance between the bafﬂes Lb,c. The lower and upper bounds on these variables and the constraints used in this paper are given in Table 2. We refer again to our previous work [10] for more information. The geometry of the air-cooled condenser is determined by the tubes' small width Ws, the ﬁn height H, the ﬁn pitch S, the tubes' large width Wl, the length of the tubes Lt and the angle q of the tube bundle with the horizontal, as explained in Section 2.2. Wl, Ws and q are ﬁxed and Lt is the result of the model. So, H and S are necessary to determine the conﬁguration of the tube bundle. Extra variables

Table 2 Optimization variables and constraints used for shell-and-tube heat exchangers and their lower and upper bounds. Optimization variable

Lower bound

Upper bound

Shell diameter Ds Tube outside diameter do Relative tube pitch pt/do Relative bafﬂe cut lc/Ds Bafﬂe spacing Lb,c

0.3 m 5 mm 1.15 0.25 0.3 m

2m 50 mm 2.5 0.45 5m

Ratio of tube diameter to shell diameter do/Ds

/

0.1

Ref

0:55 W_ pump 4 0:89

A 200

[29]

[V2013]

[29]

[V2013]

[29]

[V2013]

0:76 W_ fan [V2013] 50 0:8 4 _ 10 þ 716W turbine [V2013]

1:31 104 1:66

0:68

[29] [31]

Table 3 Optimization variables and constraints used for the air-cooled condenser and their lower and upper bounds. Optimization variable

Lower bound

Upper bound

Fin height H Fin pitch S Maximum air velocity VAmin Number of tubes ntubes

14.25 mm 1.14 mm 1.5 m/s 1000

23.75 mm 3.04 mm 10 m/s 10,000

Tube length Lt

/

20 m

are the velocity of the cooling air at the minimum cross section VAmin and the number of tubes ntubes. A non-linear constraint is used to limit the length of the tubes. The values for the lower and upper bounds are given in Table 3.

5. Results 5.1. Reference case In this subsection the parameters of our “reference” case are deﬁned, after which a variation of the parameters is performed. The investigated “reference” case is based on a proposed geothermal demonstration project in Belgium and the “reference” parameters are given in Table 4. In the next subsections, the inﬂuence of many of these parameters on the performance of the ORC is investigated. In this paper, we only focus on electricity production from the geothermal heat source and do not take into account heating purposes. We should stress that many economic parameters used in this paper are based on the literature and that the economic analysis is therefore not detailed enough to be used for a business plan.

Table 4 Parameters of the reference case. Well parameters Brine wellhead temperature Brine production Consumption well pumps Cost wells

125 C 194 kg/s 600 kWe 27.5 MV

Economic parameters Electricity price Lifetime plant Discount rate Electricity price increase

50 V/MWhe 30 years 4%/year 5%/year

Environmental conditions Average dry bulb temperature Air pressure

10.3 C 1016 hPa

108

D. Walraven et al. / Energy 80 (2015) 104e113

Fig. 3. Net present value of a single-pressure simple (a) and a single-pressure recuperated (b) ORC powered by geothermal heat, for various thermodynamic-cycle ﬂuids and for varying inlet temperature of the brine.

Fig. 4. Speciﬁc cost of a single-pressure simple (a) and a single-pressure recuperated (b) ORC powered by geothermal heat, for various thermodynamic-cycle ﬂuids and for varying inlet temperature of the brine.

5.2. Inﬂuence of the brine inlet temperature Fig. 3 shows the NPV of a simple and a recuperated ORC powered by geothermal heat, depending on the inlet temperature of the brine. The NPV increases with increasing brine inlet temperature, as a consequence of the decreasing speciﬁc overnight investment cost of the ORC with increasing brine inlet temperature as shown in Fig. 4. Two conﬂicting effects determine the evolution of the speciﬁc overnight ORC investment cost. The absolute cost of the ORC increases with increasing brine inlet temperature (larger turbine and larger heat exchangers), but the net electric power production (see Fig. 5) also increases. This last effect is larger so that the speciﬁc

overnight investment cost of the ORC decreases with increasing brine inlet temperature. The strong increase of the net electric power production with increasing brine inlet temperature can be explained by two effects. On the one hand, the gross electric power output of the cycle increases with increasing brine inlet temperature (a simple consequence of the thermodynamic efﬁciency e Carnot). On the other hand, less electrical power (relatively speaking) is needed to drive the fans of the cooling system because the heat added to the cycle is converted more efﬁciently into mechanical power. So, the gross electric power output increases and the electricity consumption of the auxiliaries decreases (relatively).

Fig. 5. Net electric power production of a single-pressure simple (a) and a single-pressure recuperated (b) ORC powered by geothermal heat, for various thermodynamic-cycle ﬂuids and for varying inlet temperature of the brine.

D. Walraven et al. / Energy 80 (2015) 104e113

109

Fig. 6. Distribution of the speciﬁc cost (a) and the absolute cost (b) for single-pressure, simple and recuperated ORCs with isobutane as the working ﬂuid for a brine inlet temperature of 125 C. The cost of the superheater is very small in all cases and is not shown in the ﬁgure.

Fig. 3a and b shows the NPV of simple and recuperated cycles, respectively. The addition of a recuperator results in a slightly higher NPV when dry ﬂuids (isobutane, R218, R227ea, R245fa and RC318) are used. This effect is only large enough to be seen in Fig. 3a and b when R218 is used. The NPV is determined by the net power output of the ORC and the cost of it. As seen from Fig. 5b, the net electric power output of the recuperated cycles is higher than the one of the simple cycles (for the dry ﬂuids). Less cooling is needed in recuperated cycles, so that, when all other conditions are the same, less electrical power is consumed by the fans and the cooling installation can be smaller. These two effects are more important than the extra cost for the recuperator as seen in Fig. 6a. From this ﬁgure it is seen that the ACC is by far the most expensive part of the ORC; about 80% of the total overnight investment cost of the ORC is credited to the ACC. Other expensive components are the turbine, the economizer and the recuperator. The evaporator is only a small part of the total cost and the cost of the pump is negligible. Fig. 6a and b give the distribution of the speciﬁc and absolute cost of simple and recuperated ORCs with isobutane as the working ﬂuid. It is chosen to plot the results for this ﬂuid because it is the dry ﬂuid which results in the highest NPV. Fig. 6a and b shows that the speciﬁc cost of the optimal simple and optimal recuperated are about the same, while the absolute overnight investment cost of the recuperated ORC is higher than the one of the simple ORC. This means that the net electric power production will be higher for the recuperated cycle than for the simple one. This is achieved by decreasing the condensing temperature of the recuperated ORC in comparison to the simple one. 5.3. Inﬂuence of the brine outlet temperature In some cases it is necessary to have a minimum brine outlet temperature to avoid scaling or to use the brine after the ORC for heating. In this paper, we focus only on electricity production and

do not take the possible direct use for heating into account, but it can be interesting to see the inﬂuence of the minimum allowed brine outlet temperature on the performance of the ORC. Fig. 7 shows the NPV as a function of the minimum brine outlet temperature for single-pressure, simple and recuperated cycles. No effect of the brine outlet temperature constraint is seen for temperatures up to about 50 and 60 C, which is the brine outlet temperature in case of no constraint, for the simple and recuperated cycles, respectively. For higher minimum brine outlet temperatures, the NPV of both the simple as the recuperated cycles starts to decrease. This decrease is stronger for the simple cycles than for the recuperated ones. It is therefore advantageous to use a recuperator for all ﬂuids, even wet ones, when the constraint on the brine outlet temperature is high. In the remainder of this paper, only the results of recuperated cycles will be shown. Simple cycles are in fact recuperated cycles with an effectiveness of the recuperator, which is an optimization variable, of zero. So, recuperated cycles are the most general ones. 5.4. Impact of 2 pressure levels Fig. 8a shows the net present value of double-pressure recuperated ORCs, depending on the inlet temperature of the brine. Comparison of Figs. 3b and 8a shows that the NPV increases when another pressure level is added for all investigated ﬂuids, except for R218. The maximum pressure in the optimal cycle for that ﬂuid, R218, is much higher than the critical point. The ﬁt between the brine cooling curve and the working ﬂuid heating curve in a temperature-heat diagram is very good, so adding another pressure level does not improve the cycle performance. The addition of an extra pressure level is especially proﬁtable for cycles in which the evaporation temperature in the single-pressure cycle is much lower than the critical temperature (isobutane and R245fa). The effect of the second pressure level is relatively small for the other ﬂuids.

Fig. 7. Net present value of an ORC powered by geothermal heat, depending on the outlet temperature of the brine. The brine inlet temperature is 125 C.

110

D. Walraven et al. / Energy 80 (2015) 104e113

Fig. 8. Net present value (a) and speciﬁc cost (b) of a double-pressure recuperated ORC powered by geothermal heat, depending on the inlet temperature of the brine.

The speciﬁc cost of double-pressure recuperated ORCs is shown in Fig. 8b. Comparison with Fig. 4b shows that the speciﬁc overnight investment cost increases by a small amount (up to 5%) if an extra pressure level is added. The net electric power output increases too as shown in Fig. 8c. The effect of the increased net electric power output is more important than the increased investment cost of the ORC, due to the high investment cost of the wells; the increased investment cost of the ORC is small relative to the total investment cost. In the remainder of this paper, only the results of singlepressure cycles will be shown, but the trends are similar for double-pressure cycles. 5.5. Inﬂuence of the electricity price Fig. 9a shows the NPV of ORCs powered by geothermal heat as a function of the electricity price. The NPV increases strongly with increasing electricity price, because not only the price of the electricity increases, but the amount of electricity produced increases

Fig. 10. Speciﬁc cost of a single-pressure, recuperated ORC powered by geothermal heat, depending on the electricity price.

Fig. 9. Net present value (a) and net electric power production (b) of a single-pressure, recuperated ORC powered by geothermal heat, depending on the electricity price.

D. Walraven et al. / Energy 80 (2015) 104e113

Fig. 11. Net present value of a single-pressure, recuperated ORC powered by geothermal heat, depending on the discount rate (a) or the electricity price evolution (b).

too. The latter effect is shown in Fig. 9b, which shows the net electric power produced. The higher electric power production is explained by the fact that a higher electricity price, and thus a higher income during the lifetime of the power plant, allows to invest more in a more efﬁcient power plant. This increase in electric power production follows from a decreasing condenser temperature, decreasing pinch-point-temperature differences and a decreasing heat-source-outlet temperature, so more heat is added to the cycle. This results also in a higher speciﬁc cost of the ORC, which is shown in Fig. 10. 5.6. Inﬂuence of the discount rate and electricity price evolution Fig. 11a and b shows the net present value as a function of the discount rate and the annual electricity price evolution, respectively. The trend of a decreasing discount rate and an increasing electricity price evolution are analogous to the effect of an increasing electricity price (Fig. 9a). The effect of the discount rate is very strong, due to the high investment and low operational costs. 6. Conclusions A system optimization of an air-cooled ORC powered by geothermal heat is performed in this paper. The cycle parameters of the ORC, the geometry of the heat exchangers and the geometry of the air-cooled heat exchanger are optimized together in order to obtain the maximum net present value of the installation. It is shown that the brine inlet temperature, the brine outlet temperature, the electricity price, discount rate and electricity price evolution have a strong inﬂuence on the net present value of the geothermal power plant. For cycles with dry ﬂuids it is always useful to include a recuperator because this heat exchanger decreases the cooling load. A recuperator is advantageous for cycles with wet ﬂuids only when the brine outlet temperature is constrained. The performance (thermodynamic and economic) of singlepressure subcritical cycles in which the evaporation temperature is “much” lower than the critical temperature can improve strongly by addition of an extra pressure level, although the speciﬁc cost of the ORC increases. The air-cooled condenser is a very important component. It has a strong inﬂuence on the efﬁciency of the power plant through the condensing temperature and the power consumption of the fans and it accounts for a large part of the investment cost of the ORC (about 80% in the case study). Acknowledgments €l Walraven is supported by a VITO doctoral grant. Danie

Nomenclature

Greek

Dp h m r q

pressure drop [Pa] efﬁciency [e] dynamic viscosity [Pa s] density [kg/m3 tube bundle angle [ ]

Roman A C do Ds f f h H i I lc Lb Lt m_ N NPV Nu ntubes ORC p pt Q_

surface area [m2] cost [V] tube outside diameter [m] shell diameter [m] friction factor [e] correction factor [e] speciﬁc enthalpy [J/kg] ﬁn height [m] discount rate [%] income [V] bafﬂe cut length [m] bafﬂe spacing [m] length of the tubes (ACC) [m] mass ﬂow [kg/s] number of full load hours [e] net present value [V] Nusselt number [e] number of tubes [e] organic Rankine cycle price [V] tube pitch [m]

Re S t VAmin _ W Ws Wl Sub-and air drilling E el EPC fan

volume ﬂow [m3/s] Reynolds number [e] ﬁn pitch [m] time [year] velocity at minimum ﬂow area [m/s] mechanical power [kW] tube small width [m] tube large width [m] superscripts air drilling equipment electrical engineering, procurement and construction fan

111

112

I in is LT M net OM ORC out P pump T turbine

D. Walraven et al. / Energy 80 (2015) 104e113

installation inlet isentropic lifetime material nett operation and maintenance ORC outlet pressure Pump temperature turbine

Appendix B. Deﬁnition investment cost In this paper only the engineering, procurement & construction cost (CEPC) is taken into account, but the total investment cost consists also of other parts. Neither owner costs nor provisions for contingency nor ﬁnancing costs (or interest during construction) are considered here. The most widely used delineation can be found in D'haeseleer [27]: Total investment cost e Overnight construction cost * Owner's cost * Engineering, procurement & construction cost * Contingency provision e Interest during construction

Appendix C. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.energy.2014.11.048. Appendix A. Turbine-stage-efﬁciency prediction The turbine-stage-efﬁciency prediction is given as a function of qﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 1=4 the dimensional parameter X ¼ ln½ Q_ out is =Dh and the speciﬁc is

volume variation across the turbine stage in an isentropic process Y ¼ Q_ out is =Q_ in $Q_ in and Q_ out is are the volume ﬂow rate at the inlet of the turbine stage and outlet of the turbine stage in case of an isentropic turbine expansion, respectively. Dh is is the enthalpy drop across the turbine stage in case of an isentropic turbine expansion. Equation (A.1) gives the expression for the curve-ﬁt of the efﬁciency prediction in Macchi and Perdichizzi [19].

hturbine ¼ 0:892 9:08 102 X 1:03 102 Y 7:73 102 X 2 þ 9:79 105 Y 2 9:61 104 XY 2:34 102 X 3 þ 3:02 103 X 2 Y þ 9:68 105 Y 2 X 2:55 103 X 4 þ 1:49 103 X 3 Y þ 1:77 104 X 4 Y (A.1)

Figure A.12Curve ﬁt (d) of the efﬁciency prediction for a turbine stage according to equation (A.1). Data points () from Macchi and Perdichizzi [19].

References [1] Tester J, Anderson B, Batchelor A, Blackwell D, DiPippo R, Drake E, et al. The future of geothermal energy: impact of enhanced geothermal systems (EGS) on the United States in the 21st century. Tech. Rep. Massachusetts, USA: Massachusetts Institute of Technology; 2006. [2] Dai Y, Wang J, Gao L. Parametric optimization and comparative study of organic Rankine cycle (ORC) for low grade waste heat recovery. Energy Convers Manag 2009;50(3):576e82. [3] Saleh B, Koglbauer G, Wendland M, Fischer J. Working ﬂuids for lowtemperature organic Rankine cycles. Energy 2007;32(7):1210e21. [4] Walraven D, Laenen B, Dhaeseleer W. Comparison of thermodynamic cycles for power production from low-temperature geothermal heat sources. Energy Convers Manag 2013;66:220e33. [5] Astolﬁ M, Romano MC, Bombarda P, Macchi E. Binary ORC (organic Rankine cycles) power plants for the exploitation of mediumelow temperature geothermal sourcesepart A: thermodynamic optimization. Energy 2014a;66: 423e34. [6] Quoilin S, Declaye S, Tchanche BF, Lemort V. Thermo-economic optimization of waste heat recovery organic Rankine cycles. Appl Therm Eng 2011;31(14): 2885e93. [7] Astolﬁ M, Romano MC, Bombarda P, Macchi E. Binary ORC (organic Rankine cycles) power plants for the exploitation of mediumelow temperature geothermal sourcesepart B: techno-economic optimization. Energy 2014b;66:435e46. [8] Madhawa Hettiarachchi H, Golubovic M, Worek WM, Ikegami Y. Optimum design criteria for an organic Rankine cycle using low-temperature geothermal heat sources. Energy 2007;32(9):1698e706. [9] Franco A, Villani M. Optimal design of binary cycle power plants for waterdominated, medium-temperature geothermal ﬁelds. Geothermics 2009;38(4):379e91. [10] Walraven D, Laenen B, D’haeseleer W. Optimum conﬁguration of shell-andtube heat exchangers for the use in low-temperature organic Rankine cycles. Energy Convers Manag 2014;83(C):177e87. [11] Wang C-C, Chang Y-J, Hsieh Y-C, Lin Y-T. Sensible heat and friction characteristics of plate ﬁn-and-tube heat exchangers having plane ﬁns. Int J Refrig 1996;19(4):223e30. [12] Wang C-C, Lee C-J, Chang C-T, Lin S-P. Heat transfer and friction correlation for compact louvered ﬁn-and-tube heat exchangers. Int J Heat Mass Transf 1998;42(11):1945e56. [13] Wang C-C, Tao W-H, Chang C-J. An investigation of the airside performance of the slit ﬁn-and-tube heat exchangers. Int J Refrig 1999;22(8):595e603. [14] Wang C, Fu W, Chang C. Heat transfer and friction characteristics of typical wavy ﬁn-and-tube heat exchangers. Exp Therm Fluid Sci 1997;14(2):174e86. [15] Yang L, Tan H, Du X, Yang Y. Thermal-ﬂow characteristics of the new waveﬁnned ﬂat tube bundles in air-cooled condensers. Int J Therm Sci 2012;53: 166e74. [16] Chang Y-J, Wang C-C. A generalized heat transfer correlation for Iouver ﬁn geometry. Int J Heat Mass Transf 1997;40(3):533e44. [17] Chang Y-J, Hsu K-C, Lin Y-T, Wang C-C. A generalized friction correlation for louver ﬁn geometry. Int J Heat Mass Transf 2000;43(12):2237e43. [18] Manglik RM, Bergles AE. Heat transfer and pressure drop correlations for the rectangular offset strip ﬁn compact heat exchanger. Exp Therm Fluid Sci 1995;10(2):171e80. [19] Macchi E, Perdichizzi A. Efﬁciency prediction for axial-ﬂow turbines operating with nonconventional ﬂuids. J Eng Power 1981;103(4):718e24. [20] Thome JR. Engineering databook III. Wolverine Tube, Inc; 2010. [21] Petukhov B, Popov V. Theoretical calculation of heat exchange and frictional resistance in turbulent ﬂow in tubes of an incompressible ﬂuid with variable physical properties (Heat exchange and frictional resistance in turbulent ﬂow of liquids with variable physical properties through tubes). High Temp 1963;1:69e83.

D. Walraven et al. / Energy 80 (2015) 104e113 [22] Gnielinski V. New equations for heat and mass-transfer in turbulent pipe and channel ﬂow. Int Chem Eng 1976;16(2):359e68. [23] Premoli A, DiFrancesco D, Prina A. A dimensionless correlation for the determination of the density of two-phase mixtures. Termotecnica (Milan) 1971;25(1):17e26. [24] Chisholm D. Pressure gradients due to friction during the ﬂow of evaporating two-phase mixtures in smooth tubes and channels. Int J Heat Mass Transf 1973;16(2):347e58. [25] Shah MM. An improved and extended general correlation for heat transfer during condensation in plain tubes. HVAC&R Res 2009;15(5):889e913. [26] Craig H, Cox H. Performance estimation of axial ﬂow turbines. Proc Ins Mech Eng 1970;185(1):407e24. [27] D’haeseleer W. Synthesis on the economics of nuclear energy, study for the European commission. DG Energy; 2013. Available from: http://ec.europa.eu/ energy/nuclear/forum/doc/ﬁnal_report_dhaeseleer/synthesis_economics_ nuclear_20131127-0.pdf.

113

[28] IEA. Technology roadmap: geothermal heat and power. Tech. Rep. International Energy Agency; 2011. [29] Smith R. Chemical process design and integration. New York: Wiley; 2005. [30] Sinnott RK. Chemical engineering design. Butterworth-Heinemann; 1999. [31] Towler G, Sinnott R. Chemical engineering design: principles, practice and economics of plant and process design. Butterworth-Heinemann; 2008. [32] Andersson J, Åkesson J, Diehl M. CasADi e a symbolic package for automatic differentiation and optimal control. In: Forth S, Hovland P, Phipps E, Utke J, Walther A, editors. Recent advances in algorithmic differentiation. Lecture notes in computational science and engineering, vol. 87. Berlin Heidelberg: Springer; 2012. p. 297e307. [33] Büskens C, Wassel D. The ESA NLP solver WORHP. In: Modeling and optimization in space engineering. Springer; 2013. p. 85e110. [34] Lemmon E, Huber M, Mclinden M. NIST reference ﬂuid thermodynamic and transport properties REFPROP. The National Institute of Standards and Technology (NIST); 2007. version 8.0.

Copyright © 2021 COEK.INFO. All rights reserved.