Applied Thermal Engineering 105 (2016) 1021–1029

Contents lists available at ScienceDirect

Applied Thermal Engineering

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

Research Paper

Annual performance variation of an EGS power plant using

an ORC with NDDCT cooling q
Sam Duniam ⇑, Hal Gurgenci
School of Mechanical & Mining Engineering, The University of Queensland, St. Lucia, Queensland 4072, Australia

h i g h l i g h t s

Analysis of ORC powered by enhanced geothermal system with NDDCT cooling.

Annual performance variation analysis using coupled NDDCT and cycle model.

Typical daily range of net power is ±11%.

Typical change in mean daily net power for consecutive days is 3%, can be up to 20%.

a r t i c l e i n f o a b s t r a c t

Article history: The purpose of this paper is to model an Enhanced Geothermal System (EGS) power plant using an
Received 19 December 2015 Organic Rankine Cycle (ORC) cooled by a Natural Draft Dry Cooling Tower (NDDCT) and to investigate
Accepted 7 April 2016 the influence of the variation of performance of the NDDCT due to changing ambient temperature on
Available online 8 April 2016
cycle performance. The ORC used in this work is the supercritical butene recuperated ORC. The EGS heat
source conditions used are those found at the Habanero 1 MW pilot plant in South Australia, with
Keywords: geothermal brine inlet temperature of 220 °C, minimum brine temperature of 80 °C, and brine mass flow
Enhanced geothermal systems
rate of 35 kg/s. A one dimensional NDDCT model was developed and integrated into the cycle model,
Natural draft dry cooling tower
Organic Rankine cycle
enabling a novel method of coupled analysis of ORC and NDDCT interdependence, which allows analysis
of plant performance for varying ambient temperature. The analysis finds that annual average W _ net is
2.82 MWe, the typical daily range of W _ net;mean for consec-
_ net is 0.62 MWe (±11%), the typical change in W
utive days is 0.07 MWe (3%), and the largest is 0.5 MWe (20%). The maximum range at any given time
throughout the year, based on historical temperature data extremes is ±31%, but the typical expected
range ±10%.
Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction EGS resources typically occur in arid areas with no consumable

cooling water available. This necessitates the use of air cooling for
Population growth, continued industrialisation and the result- heat rejection in this application. Mechanical draft cooling systems
ing growth in energy consumption, combined with the environ- incur a significant parasitic power consumption, and this leads to
mental impacts and depletion of fossil fuel resources and their interest in Natural Draft Dry Cooling Towers (NDDCT). NDDCTs
environmental impacts are the drivers of a global search for renew- are of interest to improve efficiency and net power generation of
able and clean energy sources. Enhanced Geothermal Systems geothermal power plants by reducing parasitic power consump-
(EGS) are a potentially viable source of renewable energy and the tion for condensing the ORC fluid [1].
Organic Rankine Cycle (ORC) is recognised as the most promising Due to EGS resources typically being several km below the sur-
method of energy conversion for low to moderate temperature face of the Earth, the drilling and heat source subsystem cost forms
heat sources such as EGS. a significant proportion of total plant cost, reported at up to 70% for
resources of 2.5–5 km depth, with power cycle cost at just 15% of
total investment cost [2]. This motivates the pursuit of high power
q
conversion system efficiency, incentivising the use of options such
17th IAHR International Conference on Cooling Tower and Heat Exchanger,
7–11 September 2015, Gold Coast, Queensland, Australia.
as NDDCTs, which may have slightly higher capital cost than forced
⇑ Corresponding author. draft cooling options, but reduce parasitic power consumption of
E-mail address: s.duniam@uq.edu.au (S. Duniam). the plant.

http://dx.doi.org/10.1016/j.applthermaleng.2016.04.025
1359-4311/Ó 2016 Elsevier Ltd. All rights reserved.
1022 S. Duniam, H. Gurgenci / Applied Thermal Engineering 105 (2016) 1021–1029

Nomenclature

A area (m2) ct cooling tower

d diameter (m) ctc cooling tower contraction
F correction factor (–) cte cooling tower expansion
f friction factor (–) cw cooling water
h heat transfer coefficient (W/(m2 K)) e effective
H height (m) f fin
K loss coefficient (–) fr frontal
k thermal conductivity (W/(m K)) ft fin tip
L length (m) gb geothermal brine
m_ mass flow rate (kg/s) he heat exchanger
Nu Nusselt number (–) i inner
Ny characteristic heat transfer parameter (m1) in inlet
n number (–) net net
p pressure (Pa) o outside
P pitch (m) out outlet
Pr Prandtl number (–) pp pinch point
Q_ heat flow rate (W) T temperature
Re Reynolds number (–) t tube
Ry characteristic flow parameter (m1) to tower outlet
T temperature (°C) tr transverse
U overall heat transfer coefficient (W/(m2 K)) ts tower supports
W_ work flow rate-power (W)
Abbreviations
Greek symbols BHE Brine Heat Exchanger
DT temperature difference (°C) EGS Enhanced Geothermal Systems
e surface roughness (m) HPCP High Pressure Cycle Pump
l dynamic viscosity (kg/(m s)) HPT High Pressure Turbine
q density (kg/m3) LMTD Log Mean Temperature Difference
LPCP Low Pressure Cycle Pump
Subscripts LPT Low Pressure Turbine
a area MDACT Mechanical Draft Dry Cooling Tower
amb ambient NDDCT Natural Draft Dry Cooling Tower
b tube bundle ORC Organic Rankine Cycle
cf cycle fluid

Low to moderate temperature power applications, such as EGS, degradation in cross-wind conditions. A number of studies have
generally have a significantly higher heat rejection ratio (heat found that wind break walls can effectively minimise this issue
rejection load/net power generation) as compared to high temper- for large scale NDDCTs [5–7] and several other studies have been
ature applications such as coal-fired thermal power plants due to performed considering small scale NDDCTs and have shown that
the lower cycle efficiency [3]. Due to this higher heat rejection small towers with windbreak walls can be used effectively in such
ratio, the efficiency of the condensing process has a significantly conditions [8–10].
higher impact on the overall system performance. This further In the literature, ORC studies commonly assume a fixed
motivates giving special consideration to the design of the con- condensing temperature by specifying water cooling such as
densing system for ORC. [11–13]. Vetter et al. [14] investigated the effect of the condensing
Mechanical Draft Air Cooled Towers (MDACT) are commonly temperature on cycle performance whilst neglecting the power
used as the condenser in ORCs as they allow for more direct control consumption associated with heat rejection.
of the air flow rate through the heat exchanger, via control of the A number of studies have analysed cycle performance with
fan speed. However, this comes at the cost of a significant amount MDACT, such as Ghasemi et al. [15] who compared simulations
of the power generated at the turbine being used by the fans in the accounting for the impact of the MDACTs on the cycle performance
MDACT. The parasitic power consumption in forced draft condens- against measured data. Augustine et al. [16] compared sub- and
ing systems can account for 10–12% of gross power, under ideal supercritical ORCs by optimising a number of fluids for each whilst
conditions, and as much as 40–50% for ambient temperatures accounting for MDACT power consumption. Jung and Krumdieck
approaching condensing temperature [4]. Whereas, NDDCTs have [17] modelled an ORC system using an MDACT at the design point.
no direct parasitic power consumption, only the relatively minor Sun and Li [18] determined how ORC thermal efficiency is affected
additional pumping power required to circulate the cooling water by the MDACT air mass flow rate for a fixed ambient temperature
through the NDDCT. NDDCTs have been widely used in large scale with R134a as the cycle fluid.
commercial power plants, generally on the hundreds of megawatts Walraven et al. [19] performed an economic system optimisa-
scale; whereas ORCs are generally applied in the hundreds of kW tion of an MDACT cooled ORC powered by low temperature
to several MW scale such as in decentralised power generation geothermal heat sources, and in a separate study investigated
applications. One of the primary concerns associated with the wet cooling MDACTs versus dry cooling using MDACTs in the
use of small scale NDDCTs is the susceptibility to performance context of minimising the levelised cost of electricity from low
 S. Duniam, H. Gurgenci / Applied Thermal Engineering 105 (2016) 1021–1029 1023

Table 1 consists of the NDDCT, the condensing heat exchanger, and the
EGS conditions – those found at the Habanero site in the Cooper Basin in South cooling water pump.
Australia [21].

Parameter Value
3. System modelling
Brine production well head temperature 220 °C
Minimum brine temperature 80 °C
Brine mass flow rate 35 kg/s 3.1. Cycle modelling
Brine production wellhead pressure 35 MPa
Brine reinjection pressure 45 MPa
The system modelling in this work was performed using IPSE-
pro, a process simulation software, which uses fluid properties
from REFPROP 9.1. The standard turbine, pump and heat exchanger
temperature geothermal heat sources [20]. These studies did not models from the Enginomix LTP Library were used. A one dimen-
consider the effect of ambient temperature on cycle performance. sional NDDCT model was developed in IPSEpro, based on a widely
No studies have been found which investigate the performance used method by Kröger [22]. Further details of the NDDCT model
of ORC or geothermal plants in general when cooled by NDDCTs, are presented in the following section. The novel aspect of this
and this is the gap in the literature that the current work seeks work is the coupled modelling method of the NDDCT and the
to explore. ORC to account for the interdependence of the performance
The EGS conditions used in this study are those found at the between the two. The assumptions for component performances
Habanero 1 MW pilot plant, at Innamincka, South Australia. These used in this analysis are summarised in Table 2.
conditions are given in Table 1. The minimum brine reinjection The cycle and component parameters were optimised at the
temperature is constrained by the temperature below which the design point using a genetic algorithm optimisation, further detail
geothermal brine will start causing fouling problems in the pipes is available in [23]. The optimised values for the design point
and heat exchanger. The high brine reinjection pressure is required parameters which are used in this work are presented in Table 3.
in order to achieve the desired geothermal brine flow rate through
the EGS reservoir.
The objective of this study is to investigate the NDDCT perfor- 3.2. NDDCT modelling
mance in varying ambient temperatures and how this affects the
net power generation of an EGS ORC power plant on a diurnal A one-dimensional NDDCT model was developed in IPSEpro
and annual basis. The cycle is first analysed at a range of ambient based on the method presented by Kröger [22]. The model balances
temperatures to map the performance versus ambient tempera- the buoyancy force generated due to the heat transfer from the
ture. This data is used to investigate the variation of plant perfor- cooling water loop to the air flow through the tower, against
mance due to changing ambient temperature on a diurnal, the resistances to the air flow through the tower. A schematic of
seasonal and annual basis. the NDDCT is shown in Fig. 2, and the key model equations are
given in Appendix A.

2. System description Table 2

Values used for cycle component performance.

The recuperated ORC for EGS (Fig. 1) consists of a Brine Heat Variable Value
Exchanger (BHE), turbine, cycle pump, recuperator, and the cooling Turbine isentropic efficiency 85%
system. The BHE transfers heat from the geothermal brine to the Pump efficiency 80%
cycle, the outlet cycle fluid is supercritical fluid, which is expanded Generator efficiency 97%
in the turbine. Most of the sensible heat remaining at the turbine Motor efficiency 96%
Brine heat exchanger pinch point temperature difference P5 °C
exhaust is utilised in the recuperator to preheat the liquid from
Condensing heat exchanger pinch point temperature difference P3 °C
point 2 to point 3, prior to inlet to the BHE. The cooling system

Fig. 1. Schematic (left) and T–s diagram (right) of the recuperated ORC.
1024 S. Duniam, H. Gurgenci / Applied Thermal Engineering 105 (2016) 1021–1029

Table 3 implementation of off-design models where necessary for compo-

Cycle design point parameters used in this work. nents. The turbine(s) and pumps are assumed to perform at con-
Component/stream Variable Value stant isentropic efficiency, as it is assumed that the changes
Turbine inlet Temperature (°C) 192.4 resulting from varying condensing temperature will have relatively
Pressure (MPa) 6.43 minor influence on the turbine and pump operating point, and they
Condenser inlet Pressure (MPa) 0.51 will remain at near design point operation. A geometrically depen-
Condensing temperature (°C) 49.5 dent shell and tube heat exchanger model is used, which calculates
Cooling water Mass flow rate (kg/s) 732 the outlet conditions based on the calculated heat transfer coeffi-
cients and pressure drops at off-design conditions for the heat
Brine heat exchanger Heat transfer area (m2) 875
Tube length (m) 11.5 exchanger geometry which has been fixed at the design point.
Tube inner diameter (mm) 5 Similarly, the NDDCT design is fixed at the design point and so
Tube pitch (do) 1.15 off-design performance is calculated based on the NDDCT
Recuperator Heat transfer area (m2) 250 geometry.
Tube length (m) 3.3
Tube inner diameter (mm) 8
Tube pitch (do) 1.25
4. Results and discussion
Low pressure turbine inlet Temperature (°C) –
Pressure (MPa) –
The cycle was evaluated at intervals of 5 °C across a range of
ambient temperatures, and this data was used to generate the
cycle performance map shown in Fig. 3. The plant thermal effi-
The heat exchanger bundle used in the model is as per [22]; the ciency accounts for the brine reinjection pump power consump-
air side heat transfer and pressure drop are calculated via experi- tion, whereas the cycle thermal efficiency does not. The power
mentally determined characteristic correlations. The NDDCT generation, power consumption and turbine back pressure versus
geometry used is as per [23], and is summarised in Table 4. ambient temperature are shown in Fig. 4.
The cycle net power generation exhibits an approximately
3.3. Cycle off-design modelling linear dependence on ambient temperature, Tamb, according to
Fig. 3, whereas the cycle thermal efficiency, gth, shows a change
This study aims to investigate the performance variation of the in gradient at Tamb = 40 °C. This coincides with the point at which
plant across a range of ambient temperatures. This necessitates the the recuperator becomes ineffective due to the increasing

Fig. 2. NDDCT schematic and summary of specified and calculated variables in NDDCT model.

Table 4
Specifications of the cooling tower and heat exchanger.

Parameter Value Parameter Value

Tower height, H5 52 m Effective frontal area per bundle, Afrb 32.573 m2
Base diameter, d3 37 m Tube material ASTM A214 mild steel
Outlet height, d5 26 m Fin material ASTM 6063 aluminium
Inlet height, H3 5.8 m Tube thermal conductivity, kt 50 W/m K
Number of tower supports, nts 27 Fin thermal conductivity, kf 204 W/m K
Length of tower supports, Lts 6.7 Fin diameter, df 57.2 mm
Number of heat exchanger bundles, nb 22 Tube inner diameter, di 21.6 mm
Heat exchanger arrangement Horizontal Tube outer diameter, do 25.4 mm
Tube arrangement Staggered Fin root diameter, dr 27.6 mm
Number of tube rows, nrows 4 Fin tip thickness, tft 0.25 mm
Number of passes, np 2 Fin thickness (mean), tf 0.5 mm
Number of tubes per bundle, ntb 154 Fin root thickness, tfr 0.75 mm
Transversal tube pitch, Ptr 58 mm Fin pitch, Pf 2.80 mm
Longitudinal tube pitch, Pl 50.22 mm Fin surface roughness, ef <1 lm
Length of finned tube, Lt 15 m Relative tube surface roughness, e/d 5.24  104 m2
Effective length of finned tube, Lte 14.4 m Ratio of minimum to free stream flow area, r 0.433
 S. Duniam, H. Gurgenci / Applied Thermal Engineering 105 (2016) 1021–1029 1025

20 25

3.5 18
Net Power Generaon (MWe)

16 20

Cycle thermal efficiency (%)

3 14

Heat Transfer (MWth)

12
15
10
2.5 Q_in
8 Q_out
6 10
Q_in - Q_out
2 Net Power Generaon
4 Q_recup
Cycle thermal efficiency
plant thermal efficiency 2 5
1.5 0
0 10 20 30 40 50
Ambient Temperature (oC) 0
0 10 20 30 40 50
Fig. 3. Net power generation and thermal efficiency vs. ambient temperature. Ambient Temperature (oC)

Fig. 6. Heat transfer versus ambient temperature.

5 1.2
4.5 condensing temperature approaching the design point geothermal
1 brine outlet temperature, Tgb,BHE,out, shown in Figs. 5 and 6. The
4
Turbine back pressure (MPa)

recuperator is able to maintain the design point cycle fluid BHE

3.5
0.8 inlet temperature, Tcf,BHE,in, from 25 °C to 40 °C, below 25 °C Tcf,BHE,in
3 decreases due to recuperator performance limitations, with the
result that the turbine inlet temperature increases.
Power (MPa)

2.5 0.6 The heat input to the cycle, Q_ in , remains constant at Tamb of less
2 than 40 °C, due to the minimum brine reinjection temperature
0.4 constraint of 80 °C, whilst heat rejection from the cycle, Q_ out , drops
1.5 W_net and the heat recuperated, Q_ recup increases with decreasing ambient
W_turb
1 temperature, leading to increased W _ net .
W_in,cfpump 0.2
0.5 W_in,gbpump The NDDCT performance is dependent on both the ambient
Turbine back pressure conditions and the cooling water conditions at the NDDCT inlet,
0 0 which is constrained by the minimum pinch point constraint in
0 10 20 30 40 50
the condenser. The objective of the cooling system is not to reject
Ambient Temperature (oC) as much heat as possible, but rather to reject heat at the rate
required by the cycle in order to condense the cycle fluid at the
Fig. 4. Power generation, power consumption, and turbine back pressure vs.
ambient temperature. lowest possible temperature in order to give the highest W _ net
possible. This is because, as shown in Fig. 7, whilst increasing the
250 cooling water mass flow rate, m _ cw , increases Q_ out , the cooling water
outlet temperature, Tcw, also increases, with the result that the
cycle condensing temperature increases, and W _ net decreases. So
200 _ cw should be selected to achieve a suitable trade-off between high
m

T_cf,BHE,in 19 45
Temperature (oC)

150 T_cf,turb,in
18.5 Q_out 44
Heat Rejecon Rate (MWth)

T_cf,turb,out
T_cf,sat,cd 18 T_cw,out 43
Temperature (oC)

100 T_gb,BHE,out 17.5

42
17
41
16.5
50
40
16
15.5 39
0
15 38
0 10 20 30 40 50 0 500 1000 1500
Ambient Temperature (oC) Cooling water mass flow rate (kg/s)

Fig. 5. Temperature profiles at points of interest for each of the cycles for varying Fig. 7. NDDCT heat rejection rate, Q_ out , and cooling water outlet temperature,
ambient temperatures. Tcw,out, versus cooling water mass flow rate for Tamb = 25 °C and Tcw,in = 46.5 °C.
1026 S. Duniam, H. Gurgenci / Applied Thermal Engineering 105 (2016) 1021–1029

1200 1.1 80 10

Normalised Value (by 25oC operang point)

70 9
1000
8

Temperature Difference ( oC)

1.05 60
Heat Transfer (MWth)

Temperature (oC)
800 7
50
6
600 1 40
5
30
400 4
T_a,in
0.95 20 T_cw,in
m_cw 3
m_air T_a,out
200 10
Q_trans - Normalised T_cw,out 2
K_he - Normalised LMTD
0 0.9 0 1
0 10 20 30 40 50 0 10 20 30 40 50
Ambient Temperature (oC) Ambient Temperature (oC)

_ cw and constant ITD.

Fig. 8. NDDCT performance versus ambient temperature with constant m

1200 80

70
1000
60
Mass flow rate (kg/s)

800 m_a
50 m_cw
Temperature (oC)
T_a,in
600 40 T_a,out
T_cw,in
T_cw,out
30
400 T_cf,sat,cd
ITD
20
200
10

0 0
0 10 20 30 40 50
Ambient Temperature (oC)

Fig. 9. NDDCT and cycle parameters versus ambient temperature.

Q_ out and low Tcw,out, in order to be able to achieve a low cycle fluid of typical temperatures for each season were used. Temperature
condensing temperature and high W _ net . data from BOM for 2012 was used and the sample temperatures
An additional effect to account for is that for constant m _ cw and were taken from the third week of January, April, July and October
inlet temperature difference, ITD, which is defined as ITD = to represent the seasons. The diurnal and seasonal variation is
illustrated in Fig. 10.
Tcw,in  Ta,in, there is a decrease in Q_ out for increasing Tamb, as is
shown in Fig. 8. This is due to a decrease in m_ a which results from
an increase in the heat exchanger air side loss coefficient, Khe, due
4
to changing air properties. The implication of this is that for
increasing Tamb it becomes necessary to achieve the desired m _a
Net Power Generaon (MWe)

3.5
via increasing the ITD rather than through increasing m _ cw . This is
the control strategy used in this work, and is illustrated in Fig. 9. Summer
3
These two effects result in the non-linear m_ cw curve in Fig. 9. At Autumn
higher ambient temperatures m _ cw decreases, in order to allow a Winter
2.5
slight increase in ITD, by increasing Tcw,in. This allows a linear Spring
change in condensing temperature and the linear W _ net trend
2
shown in Fig. 3.
1.5
4.1. Diurnal and seasonal variation 0 1 2 3 4 5 6 7
Day
To investigate the influence of seasonal ambient temperature Fig. 10. Diurnal and seasonal plant performance variation for the sample daily
_ net of the cycle on a diurnal basis, a sample set
variation on the W temperatures of 2012.
 S. Duniam, H. Gurgenci / Applied Thermal Engineering 105 (2016) 1021–1029 1027

Fig. 10 shows there can be significant variation on a daily basis, The annual average Wnet value is 2.82 MWe, which corresponds
as well as across the year, due to seasonal changes in ambient to a cycle thermal efficiency of 15.8%. However, significant fluctu-
temperature. Also shown in Fig. 10 is the potential for different ations of the daily mean net power generation W_ net;mean are seen on
_ net for each season, for winter in this
degrees of variability of W both a monthly basis and over the space of a few days. The highest
sample data, W _ net is consistently within 3.0–3.5 MWe, whereas change in W_ net;mean over a short period is by approximately
for summer the range is 2.3–2.95 MWe at the beginning of the 0.8 MWe, over a week. The change from day to day in W _ net;mean is
sample week and has dropped to 1.95–2.5 MWe by the end of shown in Fig. 12.
the sample week. For the sample data used the variability in winter _ net;mean for consecutive days, as shown
The largest increase in W
is lower, at roughly 10%, whereas the variability of summer and
in Fig. 12, is 0.49 MWe (20%), the largest drop is 0.27 MWe (10%)
spring is greater than 20% for the temperature data considered.
and the mean magnitude of consecutive day change is 0.07 MWe
_ net , shown in Fig. 13, varies considerably
(3%). The range of daily W
4.2. Annual performance variation across the year and does not show a strong seasonal dependence.
From Fig. 13, the mean daily range of W _ net is 0.62 MWe, the
The annual plant performance was calculated using daily
highest range is 1.08 MWe, and the lowest daily range is
temperature data for 2012 and is shown in Fig. 11.

3.6
3.4
Net Power Generaon (MWe)

3.2
3
2.8
2.6
2.4
2.2
2 Actual net power generaon
Mean daily net power generaon
1.8

Fig. 11. Annual net power generation using the daily temperature data for 2012, from the Bureau of Meteorology [24].
for consecuve days (MWe)

0.5
0.4
0.3
0.2
0.1
0
-0.1
-0.2
-0.3

20
(%)

15

10

-5

-10

-15

_ net;mean for consecutive days, absolute value (top), and as a percentage of W

Fig. 12. Change in W _ net;mean of that day (bottom).
1028 S. Duniam, H. Gurgenci / Applied Thermal Engineering 105 (2016) 1021–1029

1.2

(MWe) 1

0.8

0.6

0.4

0.2

1.3

1.2

1.1

0.9

W_net,min / W_net,mean
0.8
W_net,max / W_net,mean
0.7

_ net absolute value (top), and as a ratio to W

Fig. 13. Range of daily W _ net;mean for that day (bottom) for 2012 temperature data from BOM.

0.05 MWe. This corresponds to a mean of approximately ±11%, a 5. Conclusions

minimum of ±0.8% and a maximum of approximately ±21%.
The above data is for a specific year so the variation is an exam- An annual performance analysis of an EGS power plant using
ple only and not necessarily representative of the maximum range NDDCT cooling is performed in this paper. A one dimensional
of possible cycle performances. The historical summary tempera- NDDCT model was developed and integrated into the ORC model
ture data for the period of 1972–1999 was used to plot the maxi- and this was used to investigate the diurnal, seasonal and annual
mum and minimum W _ net that can be expected (Fig. 14) as well performance variability of the system.
as the more typical range of expected power plant performance For the temperature data of 2012, the annual average Wnet value
using the mean maximum and mean minimum temperature data is 2.82 MWe, which corresponds to a cycle thermal efficiency of
for this period. 15.8%. There is a significant degree of variation of net power gen-
This data shows that there is a much higher potential for plant eration for the recuperated ORC powered EGS power plants using
variation, in the most extreme case W _ net may vary by over NDDCTs, both across the year and on a daily basis. The mean, max-
1.5 MWe in a short period. In summer this could be about ±25% to imum and minimum daily W _ net ranges for the sample year consid-
30%, and in winter the range is approximately ±15% to ±20%. The ered are 0.62 MWe (±11%), 1.08 MWe (±21%), and 0.05 MWe
mean maximum and mean minimum temperatures would be a more _ net;mean for consecu-
(±0.8%) respectively. The average change in W
representative range for a typical year, and the mean expected plant tive days is 0.07 MWe (3%), however the largest change is
performance range for these conditions is 0.5–0.6 MWe, or ±8% to 0.5 MWe (20%).
±12% of Wnet,mean across the year.

3.5
Net Power generaon (MWe)

W_net @ highest max T_amb

3 W_net @ mean max T_amb
W_net @ lowest max T_amb

2.5 W_net @ highest min T_amb

W_net @ mean min T_amb
W_net @ lowest min T_amb
2

1.5
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Fig. 14. Upper and lower limits of expected annual variation of net power generation calculated from monthly temperature data from BOM for 1972–1999.
 S. Duniam, H. Gurgenci / Applied Thermal Engineering 105 (2016) 1021–1029 1029

Based on the historical temperature data, in the extreme case The friction factor inside the tubes, fDt, is calculated using the
the maximum plant W _ net range possible is 1.6 MWe ±30%, but Colebrook equation
"
1:11 #2
the typical range expected across the year, based on the mean 6:9 e=d
maximum and mean minimum temperatures is between 0.5 and f Dt ¼ 0:3086 log þ ð10Þ
Recw 3:7
0.6 MWe, which corresponds to ±8% to ±12% of Wnet,mean.

Acknowledgements References

[1] K. Hooman, Dry cooling towers as condensers for geothermal power plants, Int.
This study was supported by the Queensland State Government Commun. Heat Mass Transfer 37 (9) (2010) 1215–1220.
and the University of Queensland. [2] S. Kranz, Market Study – Germany, Low-Bin Project [23/02/2015]; Available
from: <http://www.lowbin.eu/public/GFZ-LowBin_marketsituation.pdf>2009.
[3] K.H.H. Gugenci, Different Heat Exchanger Options for Natural Draft Cooling
Appendix A Towers.pdf., 2010.
[4] A. Franco, M. Villani, Optimal design of binary cycle power plants for water-
dominated, medium-temperature geothermal fields, Geothermics 38 (4) (2009)
The NDDCT model is based on the method presented in Kröger 379–391.
[22]. The balance of the heat transfer and buoyancy force is calcu- [5] Z. Zhai, S. Fu, Improving cooling efficiency of dry-cooling towers under cross-
lated using the draft equation wind conditions by using wind-break methods, Appl. Therm. Eng. 26 (10) (2006).

" 
 3:5 #
H 3 H4
pa1 f1  0:00975ðH3 þ H4 Þ=2T a1 g3:5 1  0:00975 H5   T a4  ð1  0:00975H5 =T a1 Þ3:5
2 2

2 ,

2 ,
_a
m H 3 H4
3:5
_a
m
¼ ðK ts þ K ct þ K hes þ K ctc þ K he þ K cte Þhe ð2qa34 Þ 1  0:00975 H5   =T a4 þ ð1 þ K to Þ 2qa5 ð1Þ
Afr 2 2 A5

where the heat exchanger loss coefficient, Khe, is the dominant [6] M. Goodarzi, A proposed stack configuration for dry cooling tower to improve
cooling efficiency under crosswind, J. Wind Eng. Ind. Aerodyn. 98 (12) (2010)
source loss coefficient, and for the heat exchanger bundle specified
858–863.
in Table 4 is given as [22]: [7] M. Goodarzi, R. Keimanesh, Heat rejection enhancement in natural draft
cooling tower using radiator-type windbreakers, Energy Convers. Manage. 71
K he ¼ 1383:94795 Ry0:332458 ð2Þ (2013) 120–125.
[8] Y. Lu et al., Windbreak walls reverse the negative effect of crosswind in short
The rate of heat transfer, Q_ out , is calculated using an energy natural draft dry cooling towers into a performance enhancement, Int. J. Heat
balance and the Log Mean Temperature Difference (LMTD) method, Mass Transfer 63 (2013) 162–170.
[9] Y.S. Lu et al., Experimental study of crosswind effects on the performance of
where the product of the heat transfer area and the overall heat small cylindrical natural draft dry cooling towers, Energy Convers. Manage. 91
transfer coefficient is given by: (2015) 238–248.

1 [10] Y.S. Lu et al., The influence of windbreak wall orientation on the cooling
1 1 performance of small natural draft dry cooling towers, Int. J. Heat Mass
UA ¼ þ ð3Þ Transfer 79 (2014) 1059–1069.
hae Aa hcw Acw [11] F. Heberle, D. Brüggemann, Exergy based fluid selection for a geothermal
Organic Rankine Cycle for combined heat and power generation, Appl. Therm.
The characteristic heat transfer parameter, Ny, is defined as: Eng. (2010).
[12] F.-Z. Zhang, P.-X. Jiang, Thermodynamic analysis of a binary power cycle for
hae Aa
Ny ¼ ð4Þ different EGS geofluid temperatures, Appl. Therm. Eng. 48 (2012) 476–485.
ka34 Afr Pr0:333
a34 [13] S. Quoilin et al., Thermo-economic optimization of waste heat recovery
Organic Rankine Cycles, Appl. Therm. Eng. 31 (14–15) (2011) 2885–2893.
Rearranging [14] C. Vetter, H.-J. Wiemer, D. Kuhn, Comparison of sub- and supercritical Organic

 Rankine Cycles for power generation from low-temperature/low-enthalpy
0:333 ntb;actual geothermal wells, considering specific net power output and efficiency, Appl.
hae Aa ¼ Ny ka34 Afr Pra34 ð5Þ Therm. Eng. 51 (1–2) (2013) 871–879.
ntb;maximum
[15] H. Ghasemia et al., Modeling and optimization of a binary geothermal power
plant, Energy (2013).
where ntb,actual/ntb,maximum is to correct for the reduced effectiveness
[16] C. Augustine et al., Modeling and analysis of sub- and supercritical binary
of the tubes at the bundle ends. Ny for normal non-isothermal flow rankine cycles for low- to mid-temperature geothermal resources, Geotherm.
through the heat exchanger bundle specified in Table 4 is given by Resour. Counc. Trans. 33 (2009) 689–694.
[17] H.C. Jung, S. Krumdieck, Modelling of organic Rankine cycle system and heat
Ny ¼ 383:617313 Ry0:523761 ð6Þ exchanger components, Int. J. Sustain. Energy 33 (3) (2014) 704–721.
[18] J. Sun, W.H. Li, Operation optimization of an organic Rankine cycle (ORC) heat
where Ry is the characteristic flow parameter for the heat exchan- recovery power plant, Appl. Therm. Eng. 31 (11–12) (2011) 2032–2041.
[19] D. Walraven, B. Laenen, W. D’haeseleer, Economic system optimization of air-
ger bundle and is defined as:
cooled organic Rankine cycles powered by low-temperature geothermal heat
_a
m sources, Energy 80 (2015) 104–113.
Ry ¼ ð7Þ
la34 Afr [20] D. Walraven, B. Laenen, W. D’haeseleer, Minimizing the levelized cost of
electricity production from low-temperature geothermal heat sources with
ORCs: water or air cooled?, Appl Energy 142 (2015) 144–153.
The heat transfer coefficient inside the tubes is calculated via [21] T. Mills, B. Humphreys, Habanero pilot project – Australia’s first EGS power
the correlation proposed by Gnielinski [25] plant, in: 2013 Australian Geothermal Energy Conference, 2013.
 0:67 [22] D.G. Kröger, Air-Cooled Heat Exchangers and Cooling Towers, PennWell Corp.,
f Dt
8
ðRecw  1000ÞPr cw 1 þ Ldtee OK, 2004.
Nu ¼  0:5 ð8Þ [23] S. Duniam, Design Optimisation of an Australian EGS Power Plant using a
Natural Draft Dry Cooling Tower, University of Queensland, 2016.
1 þ 12:7 f 8Dt ðPr 0:67
cw  1Þ [24] A.G. Bureau of Meteorology, Climate Statistics for Australian Locations –
Summary Statistics MOOMBA Available from: <http://www.bom.gov.
using the definition of the Nusselt number: au/climate/averages/tables/cw_017096.shtml>2015.
[25] V. Gnielinski, New equations for heat and mass transfer in turbulent pipe and
hde channel flow, Int. Chem. Eng. 16 (2) (1976) 359–368.
Nu ¼ ð9Þ
k