Analysis of the combined use of wind and pumped

storage systems in autonomous Greek islands

G. Caralis and A. Zervos

Abstract: In autonomous islands, the wind penetration is restricted due to technical reasons related
with the safe operation of the electrical systems. The combined use of wind power with pumped
storage (WPS) systems is considered as a means to exploit the abundant wind potential, increase
the wind installed capacity and substitute conventional peak supply. An approach for the simulation
of the autonomous electrical systems is proposed and applied in three islands. The simulation is
based on the non-dynamic analysis of the electrical system, in order to calculate the energy contri-
bution of the different power units. The aim is to analyse the prospects of WPS systems to decrease
the electrical system’s cost. The results show that the integration of WPS in autonomous islands
may decrease the system’s electricity production cost.

List of symbols EW,Total Cumulative annual energy production of wind

farms
a Upper part of the demand to be covered by the EW!A Wind energy absorbed by the grid directly from
hydro-turbine the wind farms out of the WPS
d Allowed instantaneous wind penetration EW!R Curtailment of surplus wind by the wind farms
BEW Benefit from the wind energy absorbed directly out of the WPS
by the grid EW,h!A Wind energy absorbed by the grid directly from
CEC Cost of conventional energy used for pumping the wind farms in the WPS
D Electricity demand EW,h!R Curtailment of surplus wind by the wind farms
in the WPS
DP Diameter of the penstock
FC Fuel cost (per kWh produced)
EC Electricity production of conventional unit
g Gravity
EPC Electricity Production Cost
HFO Heavy Fuel Oil
EPCC EPC of conventional units
i Discount rate (i ¼ 5%)
EPCS EPC of whole electrical system
LFCU Average load factor of the committed conven-
EPCW EPC of wind power tional units
EP:available Total energy available for pumping (conven- LA Predefined level of demand upon which the
tional and wind energy surplus) hydro-turbine is to be set into operation
EP:final Cumulative final energy used for pumping LH Level of the water in the higher reservoir in the
EP:C Conventional energy used for pumping beginning of the simulations’ step
EP:W,h Wind energy surplus used for pumping LP Length of the penstock (distance between the
ESO Electricity system operator higher and the lower reservoir)
ESP_T Desirable hydro-turbine production based on L Level of the water in the lower reservoir in the
the set-points beginning of the simulations’ step
ESP_P Annual conventional electricity available for LH,END Level of the water in the higher reservoir in the
pumping end of the simulations’ step
ET Electricity production of hydro-turbine LL,END Level of the water in the lower reservoir in the
EW Annual energy production of wind farms out of end of the simulations’ step
the WPS MAC Maximum conventional ability of the com-
EW,h Annual energy production of wind farms in the mitted conventional units
WPS N Lifetime of the investment (N ¼ 20 years for
the wind farms and N ¼ 30 years for the
conventional units)
NC Number of conventional units committed
# The Institution of Engineering and Technology 2007
NP Number of pumps installed
doi:10.1049/iet-rpg:20060010
Paper first received 28th November 2006 and in revised form 11th January 2007
OMCC Operation and Maintenance cost of the conven-
tional units
The authors are with the National Technical University of Athens (NTUA),
PO Box 64514, Zografou Campus, 15704 Athens , Greece OMCWPS Operation and maintenance cost of the WPS
E-mail: [email protected]

IET Renew Power Gener., 2007, 1, (1), pp. 49– 60 49

PAbsorbed Grid ability to directly absorb wind power for wind applications. Given the current infrastructure
PC Conventional unit production and the technical constraints, the prospects of wind power
Peak Last 24-hours peak demand to decrease both the system’s electricity production cost
(EPCS) and the dependence on the oil are limited [1]. The
pm Market price for the conventional electricity
reason is that wind farms operating in autonomous
used for pumping
systems are subject to output power limitations, related
pW Fixed market price for the wind power directly with technical constraints of the conventional generating
absorbed by the grid units, namely the minimum loading levels of the thermal
PP:Avail Cumulative available power for pumping units (technical minimum) and a dynamic penetration
PP:C Conventional power finally used for pumping limit, applied for stability purposes [1 – 4]. Analysis of
different island systems has shown that the increased wind
PP:Final Cumulative power finally used for pumping penetration without storage could reduce the EPCS only
PP,R Rated power of each pump up to a maximum of 7% (unpublished).
PP:W Wind power finally used for pumping Conventional units and existing wind farms comprise the
PSU Pumped Storage Unit current infrastructure of the autonomous electrical systems
to meet the demand. A wind and pumped storage (WPS)
PT,R Rated power of the hydro turbine system, comprising new wind farms, two reservoirs for
PT,Produced Turbine power production the recycling of water, hydro turbines, pumps and pen-
PW,R Wind installed capacity out of the WPS stocks, is proposed as a mean to increase the wind installed
capacity, substitute expensive fuel oil and reduce the
PW,h,R Wind installed capacity in the WPS
required conventional installed capacity. The later is poss-
PW Wind power production of the wind farms out ible because the variable output of wind power is
of the WPS managed and transformed into a guaranteed power supply.
PW,h Wind power production of the wind farms in the Recently, the idea of pumped storage units (PSU) has
WPS been analysed by the scientific community for various
PW!A Wind power absorbed directly from the wind autonomous islands [5– 10]. Inconsistency on the assump-
farms out of the WPS tions always makes difficult the comparison between
different approaches and islands.
PW!R Curtailment of surplus wind power by the wind In the current paper, the prospects of the WPS to
farms out of the WPS decrease the EPCS , to increase the renewable energy
PW,h!A Wind power absorbed directly from the wind supply and the renewable capacity credit, are examined
farms in the WPS comparatively in three islands different in size and charac-
PW,h!R Curtailment of surplus wind power by the wind teristics, using the same assumptions and approach. Three
farms in the WPS representative Greek islands have been selected as case
R Annuity factor studies.
RAE Regulatory Authority for Energy † Crete is the largest Greek island and represents high
SRC Rest capacity of the committed conventional current EPCS due to the extensive use of gas turbines to
units cover medium loads. The electrical system consists of 6
SRT Rest capacity of the hydro-turbine steam units with 107 MW installed capacity, 1 combined
cycle with 36 MW, 13 gas turbines with 403 MW and 8
SPC Set-point of the conventional units diesel with 344 MW.
SPP Set-point for the available conventional power † Lesvos is a medium-sized island with relatively low
for pumping current EPCS due to the use of heavy fuel oil and the
SPT Set-point of the turbine smooth demand distribution. The electrical system consists
TICC Total investment cost of the conventional units of 15 Heavy Fuel Oil (HFO) units with 106 MW installed
capacity.
TICW Total investment cost of wind applications † Serifos is a typical small island with very high EPCS
TMC Technical minimums of conventional units and misdistribution of the demand between winter and
TMT Technical minimums of the hydro turbine summer due to tourism. The electrical system consists of
seven diesel units with 5.8 MW installed capacity.
U Wind velocity
VH Capacity of the higher reservoir The simulation is based on the convolution of the demand
VH!L Water flow from the higher to the lower and the wind time series. For the application in each island,
reservoir the hourly data of the electricity demand, the conventional
VL Capacity of the lower reservoir units and their features (maximum ability, technical mini-
mums, order of commitment, fuel consumption, installation
VL!H Water flow from the lower to the higher cost, operation and maintenance cost, fuel cost) and the
reservoir wind hourly data are required.
WPS Wind and Pumped Storage System
WPSO WPS operator 2 Operation and architecture of the WPS

Different options related with the design and the operation

1 Introduction of the WPS have been analysed, compared and evaluated
[7]. As a result the features that are used in this paper
The autonomous electrical systems in Greek islands are include connection of the Wind farms with the pumping
based almost entirely on oil. They are characterised by station through the central grid, peak demand supply of
high wind potential and there is a high investor’s interest the hydro-turbine, consideration of the hydro-turbine as a
50 IET Renew Power Gener., Vol. 1, No. 1, March 2007
spinning reserve to increase the direct wind power The reliability of the WPS is measured by the hydro-
absorbed, double penstock and complementary pumping turbine’s ability to correspond at the set-points. The
using conventional power given the amount of conventional required rated power of the turbine is defined by the a
units’ spinning reserve. and the annual peak demand.

2.1 Connection of the WPS with the electrical 2.3 Conventional power for pumping
system
Conventional power is proposed for complementary
The connection of the wind farms with the PSU is proposed pumping in order to provide the unfailing weekday oper-
through the central grid, under the condition that the ation of the hydro-turbine and avoid the over-dimensioning
pumping loads are considered as deferrable loads. This of the reservoir. The use of conventional power for pumping
means that in case of wind loss or other stability problem, is going to increase the total demand and the required
pumps are disconnected. conventional production, so the available conventional
Additionally, wind power from the WPS can be directly power for pumping should be defined by the ESO. For
absorbed by the electrical system, according to the technical this purpose, the assignment of a set-point by the ESO is
constraints imposed by the electrical system operator suggested. The use of the available conventional electricity
(ESO). The amount of wind power that can be absorbed for pumping is then decided by the WPSO considering the
directly by the grid is dependent on the wind installed level of the water in the higher reservoir and the availability
capacity outside of the WPS, and on the allowed instan- of the surplus wind power.
taneous wind power penetration ‘d’. As a result, the wind The preferable definition of the set-point for conventional
installed capacity outside of the WPS should be defined, pumping takes into consideration the spinning reserve of the
before going on with the analysis of the WPS. The wind committed conventional units [7]. This definition permits
power absorbed in priority from these wind farms is the use of conventional power for pumping only in case
defined by the technical constraints described by the of conventional power surplus, so further conventional
Regulatory Authority for Energy [3] and the WPS inte- units are not committed to cover pumping loads. The advan-
gration will not effect their operation. tage of this proposal is that the conventional units operation
The pumping station should also be directly connected to is improved and the available conventional power for
the main grid, in order to use surplus wind power in priority pumping is widely distributed during the day. (Conventional
and sometimes conventional power for complementary units operate more efficiently in full load.)
pumping.
In Fig. 1, the proposed structure and the interconnections
of the whole electrical system after the WPS integration is 2.4 Wind penetration
presented.
The ability of the conventional power stations to balance out
both the variability of the demand and the wind power,
2.2 Operational target of the hydro-turbine defines the wind power to be directly absorbed by the
grid. Wind power is absorbed in priority by the wind
The objective of the hydro-turbine is to provide peak
farms outside of the WPS system and then, if there is a
demand supply. Since there are seasonal and daily ups
margin for further wind power absorption, by the wind
and downs of the demand, and such an operational cycle
farms in the WPS system.
of the hydro-turbine is needed to provide the financial feasi-
Before the simulation of the electrical system, the wind
bility, the daily (and not the weekly or the seasonal) peak
penetration which is allowed should be defined. A stable
demand supply is used. So, the hydro-turbine is setting
maximum instantaneous wind penetration, as a percentage
into operation when the demand exceeds a predefined
of the load demand is used (i.e. d ¼ 30%), while the wind
level which is not stable during the year, but follows the
penetration is further increased considering the hydro-
daily peak of the demand
turbine as spinning reserve. This operation presupposes
LA ¼ ð1
aÞ  Peak ð1Þ the two-sided communication between ESO and WPSO.
Specifically, the ESO should know the amount of the hydro-
where Peak is the last 24-h peak and a is the upper part of turbine’s spinning reserve, which is dependent both on the
the demand to be covered by the hydro-turbine. The ESO, turbine’s capacity and on the water availability in the
who monitors the electricity demand, is proposed to be higher reservoir, in order to allow equal increase of wind
responsible for the assignment of the turbine’s set-point penetration. Finally, the turbine should be committed,
to the WPS’s operator (WPSO), when the demand D otherwise the required time between the possible wind
exceeds the LA . By this way the set-point of the turbine is loss and the hydro-turbine response may cause a stability
defined as problem in the autonomous electrical system.

SPT ¼ D
LA ð2Þ
2.5 Double penstock
Grid
Double penstock is used providing operational flexibility
and the direct quick response of the turbine when it is
WPS (Hybrid needed. The same time that the stochastic and variable
system)
wind power production is curtailed due to the technical
Conventional Wind Farms out Wind Pumped Storage/ Load constraints imposed for the safe operation of the electrical
Units of the WPS Farms Hydro turbine
system and pumps should be set into operation, the
system may need the uninterrupted and scheduled
Fig. 1 Structure of the electrical system after the WPS production of the hydro-turbine. This can be achieved
integration only with the two penstocks.
IET Renew Power Gener., Vol. 1, No. 1, March 2007 51
3 Simulation The main steps of the calculations are:

For the simulation, the design parameters of the plant † For every time-step (i.e. 1 h) the water level in the higher
should be known: hydraulic head H, length LP and diameter LH and lower reservoir LL , the electricity demand D and the
DP of the penstock, capacity of the higher VH and lower wind velocity U are known. The wind power production of
reservoir VL , rated power of the wind farms PW,R , PW,h,R the wind farms PW,h and PW is calculated given the wind
(with the index h the wind farms inside the WPS are distin- installed capacity PW,h,R and PW,R. For the first step of
guished from the wind farms outside the WPS), rated power the simulation, an initial water level in the two reservoirs
of the turbine PT,R , number of pumps NP and rated power of is needed. The independence of the results is proved using
each pump PP,R and the part of the peak to be supplied by different initial water levels.
the turbine a. An overview of the basic calculations is pre- † The hydro-turbine’s set-point SPT [(1) and (2)] and the
sented in the Table 1. number of conventional units committed NC are calculated

Table 1: Overview of the basic calculations

Input Calculation/assumptions Output

1 U a common power curve for the wind farms is used. PW,h , PW

2 D Equations (1) and (2) SPT
constraint: the SPT cannot be lower than the technical minimums of the turbine,
i.e. TMT ¼ 30%PT,R . So if D 2 LA , TMT then SPT ¼ 0.


PT;R
SPT ; if SPT . 0
SPT , LH , LH , VH , VL SRT ¼ ð3Þ SRT
0; if SPT ¼ 0

constraints considered: -water availability in the higher reservoir / level of the

water in the lower reservoir
3 D, SPT a strict order of commitment for the conventional units is considered. NC , TMC , MAC
MAC  D
SPT ð4Þ

4 D, TMC , d PAbsorbed ¼ SRT þ minfd  D; TMC g ð5Þ PW,h!A , PW!A ,


PAbsorbed ; if PAbsorbed , PW PW,h!C , PW!C
PW!A ¼ ð6aÞ
PW ; if PAbsorbed  PW
PW!C ¼ PW
PW!A ð6bÞ
PW;h!A
8( ) 9
>
< PAbsorbed
PW!A ; if PAbsorbed
PW!A , PW;h ; if P >
=
Absorbed . PW
¼ PW;h ; if PAbsorbed
PW!A  PW;h ð7aÞ
>
: >
;
0; if PAbsorbed , PW
PW;h!C ¼ PW;h
PW;h!A ð7bÞ

5 D, PW,h!A , PW!A , SPT SPC ¼ D
SPT
PW!A
PW;h!A ð8Þ SPCU

6 PW,h!C , SPC , MAC , SPP ¼ MAC
SPC ð9Þ PP:Avail

PP:Avail ¼ PW;h!C þ SPP ð10Þ

7 LH , LL , VH , VL , PP,R , NP constraints considered: number of pumps and rated power of the pumping PP:Final
station; level of the water in the higher reservoir/water availability in the
lower reservoir

)
8 PP:Final PP:W ¼ PW;h!R ; PP:C ¼ PP:Final
PW;h!R ; if ðPP:Final  PW;h!R Þ PP:W, PP:C
ð11Þ
PP:W ¼ PP:Fianl ; PP:C ¼ 0; if ðPP:Final , PW;h!R Þ

9 PP:Final operational curves of the pumps and the penstock. VL!H

10 SPT , LH , LL , VH , VL constraints considered: technical minimum of the hydro-turbine; water PT,Produced

availability in the higher reservoir/level of the water in the lower reservoir
11 PT,Produced operational curves of the turbine and penstock. VH!L
12 LH , LL , VL!H , VH!L LH;END ¼ LH þ VL!H
VH!L ð12Þ LH,END , LL,END
LL;END ¼ LL
VL!H þ VH!L ð13Þ

Note: In the case that SPT þ PW,h!A þ PW!A . D, which could happen when a þ d . 1, an iterative is used before the required SPT and
the NCU are defined

52 IET Renew Power Gener., Vol. 1, No. 1, March 2007

by the ESO (4). The cumulative maximum ability of the ESO in order to accept this price. The autonomous Greek
committed conventional units MAC and the hydro-turbine islands are excluded from the market liberalization and
should be able to meet the demand even if all the wind the system operator remains the owner of the local power
power is lost, providing the safe operation of the electrical stations.
system. The EPC of the turbine EPCT is defined under the
† Given the turbine’s set-point, the amount of the hydro- assumption that the whole investment is considered as a
turbine’s spinning reserve SRT (3) is taken into consider- mean to provide guaranteed electricity supply during
ation to increase the wind penetration. The SRT is calculated peak demand, so the wind energy sold in a fixed price is
given the hydro-turbine’s rated capacity PT,R , the turbine’s considered as inflow
production SPT and the available water in the higher reser-
voir LH . The ability of the electrical system to absorb wind TICWPS  R þ OMCWPS þ CEC
BEW
EPCT ¼ ð14Þ
power is increased by the SRT (5). ET
† The wind power absorbed (PW,h!A , PW!A) and
curtailed (PW,h!C , PW!C) are calculated [(6a), (6b), (7a), where TICWPS is the total investment cost, OMCWPS the
and (7b)]. operation and maintenance cost of the WPS, CEC the cost
† The conventional units set-point SPC (8) and the avail- of conventional energy used for pumping and BEW the
able conventional power for pumping SPP (9) are calcu- benefit from the wind energy directly absorbed by the
lated. So, the total available power for pumping PP:Avail is grid. If the market price is pm , then
derived (10).
† The final power used for pumping PP:Final is constrained CEC ¼ pm  EP:C ð15Þ
by the rated power of pumps (PP,R , NP) the capacity of
the reservoirs (VH , VL) and the available water. The part and if the fixed price for wind power is pw , then
of the PP:Final which is derived from the wind farms or the
grid (PP:W , PP:C) is calculated (11). Then, the water BEW ¼ pW  EW;h!A ð16Þ
flow from the lower to the higher reservoir VL!H is
calculated given the operational curves of the pumps and The degree of accomplishment of hydro-turbine’s set-point
the penstock. DA_SPT is the index to measure the reliability of the WPS.
† The desirable hydro-turbine production is defined by the This index is highly dependent on the design of the WPS;
set-points sent by the ESO. The final turbine’s power pro- namely the wind installed capacity, the capacity of the
duction PT,Produced depends on the availability of the water reservoir and the operational target of the turbine. It is
in the upper reservoir. Then the water flow from the defined as the ratio of the actual energy production ET to
higher to the lower reservoir VH!L is calculated given the the desirable production defined by the hydro-turbine’s set-
operational curves of the turbine and penstock. points (ESP_T)
† Finally, at the end of each step, the level of the higher
and lower reservoirs (LH,END , LL,END) are calculated [(12) ET
DA SPT ¼ ð17Þ
and (13)]. ESP T

When all the above calculations for all the time-steps The EPCC is defined as
(8760 h) are completed, then the following annual energy
amounts are derived: TICC  R þ OMCC
EPCC ¼ ð18Þ
EC
† desirable hydro-turbine’s production ESP_T;
† final hydro-turbine’s production ET; where TICC is the total investment cost of the essential
† production of wind farms (EW , EW,h , EW,Total); conventional units, OMCC the operation and maintenance
† wind energy absorbed by the grid directly (EW!A , cost and EC the conventional energy production. The
EW,h!A); OMCC has a fixed cost part, a variable cost part and
† wind energy curtailed (EW!C , EW,h!C); the fuel cost.
† conventional energy available for pumping ESP_P;
† total energy available for pumping (conventional and OMCC ¼ FixedCost þ VariableCost þ FuelCost ð19Þ
curtailed wind energy) EP:available;
† energy used for pumping: cumulative, conventional and The EPCS is calculated, assuming that the redundant units
wind energy (EP:final , EP:C and EP:W,h). are uninstalled

TIC  R þ OMC
EPCS ¼ ð20Þ
4 Evaluation criteria ETotal

The contribution of the WPS, together with economical and where TIC includes the cumulative investment cost of all the
reliability indexes, are used to describe the performance of power plants (essential conventional units, WPS and wind
the electrical system after the WPS integration. The conven- farms outside the WPS); the OMC includes the fixed cost,
tional units’ EPC EPCC , the EPCS and the EPCT are used to the variable cost and the fuel cost for the operation and main-
describe the economic impact of the WPS to the electrical tenance of the system; and ETotal is the total electricity demand.
system. The most critical is the EPCS , when it is compared The annuity factor R is defined as
with the current cost, the resulting benefit, if any, from the
WPS integration is defined. The EPCT is important for the i
R¼ ð21Þ
private investor, indicating a first estimation of the required 1
ð1 þ iÞ
N
price for the turbine’s electricity production which provides
the feasibility of the investment. Finally, the modification where i is the discount rate and N the lifetime of the
of the EPCC due to the WPS integration is critical for the investment.
IET Renew Power Gener., Vol. 1, No. 1, March 2007 53
5 Dimensioning procedure 160000000 153
214
140000000 275
In fact, given the wind potential, the wind installed capacity

Reservoir's volume (m^3)

336

is the parameter which defines the available for exploitation 120000000 397
458
energy. The part of this energy which can be directly 100000000 519
580
absorbed by the grid, depends on the technical constraints 80000000 641
of the electrical system. The part of the surplus wind 703
60000000 764
energy which can be finally exploited depends on the archi- 825
tecture the design and the size of the WPS. For a given wind 40000000 886
947
installed capacity, there is a direct relation between the size 20000000 1008
of the reservoirs, and the maximum operational target that 1069
0 1130
can be covered by the hydro-turbine with reliability reach- 0% 20% 40% 60% 80% 1191
ing the 100%. The capacity of the hydro-turbine is then 1252
WPS share of the peak 1314
defined by this target.
a
The aim of the proposed approach is to take comparable
results for the three islands, so the basic parameters in issue 0.20 153

are introduced dimensionless. The wind installed capacity is 214

275
defined as percentage of the mean annual load demand, 0.18 336
which is a fair way to levelize wind installed capacity in 397

Turbine's EPC (€/kWh)

458
different in size systems [1]. Twenty different values for 0.16 519
wind installed capacity are examined from 50 to 430% of 580
641
the mean annual load demand by step 20% (20 cases). 0.14
703
The capacity of the reservoir is defined from 10 to 150 by 764

step 20, times the maximum hourly water pumping 0.12 825
886
ability (8 cases), namely as a function of the wind installed 947
0.10 1008
capacity. 1069
For each wind installed capacity and reservoirs capacity, 0.08
1130
1191
the maximum operational target that can be achieved 0% 20% 40% 60% 80% 1252
by 100% is calculated using an iterative procedure. WPS share of the peak 1314

That is to say, that the maximum turbine size which is b

justified by the wind installed capacity and the reservoirs
capacity should be defined. It is obvious that a bigger Fig. 2 Example of the dimensioning procedure (Crete)
target could be set, but it would be achieved in less than a Reservoir’s capacity
100% of the year. The target of the dimensioning is to b Turbine’s EPC
guarantee the reliability of the turbine operation (achieve
DA_SPT ¼ 100%), and conventional units are supposed to
be removed. So, the conventional installed capacity is (unpublished). The WPS is introduced as a means to
redefined by the peak demand, the supply of the turbine further increase the development of wind energy.
and a safe margin of conventional installed capacity back † The diameter of the two penstocks is calculated given the
up (20%). rated capacity of the pumping station and the hydro-turbine
For each different value of wind installed capacity, there (Section 10).
is an optimum value for the reservoir which provides the † Ten centrifugal pumps connected in parallel are used to
lower EPCT . By the investor’s point of view the target is provide operational flexibility [7]. Given the available
the maximum contribution with the least cost. So, from power for pumping the numbed of committed pumps is
160 cases resulted – all of them provide the reliability of calculated (Section 10).
the turbine operation – the 20 cases with the maximum † The efficiency of the pumping station and hydro
target achievement and the least cost comprise the generation are analytically calculated for the different
optimum solution (lower envelope curve of the EPCT’s points of operation (annex 2). The results show that the
curves) for the current island and scenario. An example is whole efficiency of the plants (the ratio of the annual
given for the island of Crete in Fig. 2, where the 20 different energy produced by the hydro-turbine compared to the
curves represent different wind installed capacity inside the annual energy consumed for pumping) in the examined
WPS (153 – 1314 MW). cases varies between 55 and 69%.
† For the estimation of the WPS investment cost, empirical
6 Assumptions formulas have been used [7]. For the wind farms an
investment cost of 1000 E/kW has been considered [11].
The following assumptions are considered. An overview of the WPS’s cost estimation is presented in
Table 3.
† Hydraulic head between the two reservoirs H ¼ 300 m, † The calculation of the conventional EPC is based on the
and distance between them LP ¼ 3000 m. assumptions (Table 2) related with the installation cost, the
† Annual mean wind velocity: 8.1 m/s at the hub-height. fixed OMC, the variable OMC and the fuel consumption in
(The same wind time-series are used in the three islands). the various operational points.
† Financial evaluation without any subsidy. † The discount rate is considered i ¼ 5% and the lifetime N
† Reference price for the oil 54$/b, which is the annual of the investment 20 years for the wind applications and 30
mean value for 2005. years for the conventional units.
† Wind installed capacity out of the WPS is assumed to be † Market price for the conventional electricity used for
70% of the mean annual load in Crete, 40% in Lesvos and pumping pm is 0.035 E/kWh (low night tariff) or 0.09 E/
60% in Serifos. This is the wind installed capacity which kWh (regular tariff).
provides the minimization of the EPCS without storage † Fixed price for the wind production pw is 0.08 E/kWh.
54 IET Renew Power Gener., Vol. 1, No. 1, March 2007
 Table 2: Overview of the formulas and assumptions for the WPS cost estimation

Equipment, cost symbol Data/formula for cost estimation, E

Wind farms, CW 1000 /kW

Pumps, CP CP ¼ NP . C0,P . (PP,rated/H 0.3
P )
0.82
, C0,P ¼ 1814
Hydro-turbine, CT CT ¼ C0,T . (PT,rated/H 0.3
T ) 0.82
, C 0,T ¼ 4687

Reservoir, CR CR ¼ 420 . V 0.7

82 39
>
> ðWM  pDI  eI  LÞ  CM þ ðppI  LÞ  CI þ >
Penstock (CPenstock) > ffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl} |fflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflffl} 7>
>6 |fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl >
>
>6 7>
X> >
Material Cost Insulation Cost
<6 ! 7=
1:25  6
6 pDI2 7
7
>6 1:5 
I >
 L  CE >
7>
>
> 4 4 5>
>
>
> |fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl} >
>
: ;
Excavation Cost

Grid connection, CGC 4% (CP þ CT þ CR þ CPenstock)

Control system, CCS 1.6% (CP þ CT þ CR þ CPenstock)
Transportation of equipment, CT 2.4% (CP þ CT þ CR þ CPenstock)
Personal, CP 30% (CP þ CT þ CR þ CPenstock)
Others, CO 2% (CP þ CT þ CR þ CPenstock)
Operation and maintenance, OMCWPS 2% (CP þ CT þ CR þ CPenstock þ CW)

7 Results The EPCT is expected higher in Serifos than in the other

islands due to the small size of the system. Far from that, the
7.1 Comparison between islands current EPCS is higher in Serifos, providing higher competi-
tiveness for the WPS. In Crete and Serifos, the system’s
The earlier described indexes are presented for the three EPCS is decreased, thanks to the introduction of the WPS
representative islands (Crete, Serifos and Lesvos) in Fig. 3. up to 40 and 20%, respectively, whereas in Lesvos it is
increased. In parallel, the EPCC is increased in Serifos
† The Turbine’s EPC in Fig. 3a. and Lesvos, whereas in Crete is decreased or almost
† The Conventional units’ EPC in Fig. 3b. remains the same. This is justified because in Crete the
† The hybrid’s energy supply (%) in Fig. 3c. number of conventional units is larger than in the other
† The electrical system’s EPC (EPCC - E/kWh) in Fig. 3d. islands and so the system is more flexible to meet the
demand more efficiently and to remove the redundant
Additionally, the EPCC in Figs. 3a and b and the EPCS in units. Even in cases where the EPCC is expected to be
Fig. 3d before the WPS integration are presented with higher, the public utility is going to earn the difference
straight lines. between its current cost (presented with straight lines in

0.50 90%
0.45 80%
Hybrid's energy supply (%)

0.40
Turbine's EPC (€/kWh)

Serifos 70%
0.35
60%
0.30
50%
0.25
Crete 40%
0.20
Crete 30% Crete
0.15
Lesvos Lesvos Lesvos
0.10 20%
Serifos Serifos
0.05 10%
0.00 0%
0% 20% 40% 60% 80% 100% 0% 20% 40% 60% 80% 100%
WPS share of the peak WPS share of the peak
a c

0.50 0.50
Conventional units' EPC (€/kWh)

Electrical system's EPC (€/kWh)

0.45 0.45
0.40 Serifos 0.40
0.35 0.35 Serifos
0.30 0.30
0.25 0.25
0.20 Crete
0.20 Crete Crete
Crete
0.15 0.15
Lesvos Lesvos
Lesvos
0.10 0.10 Lesvos Serifos
Serifos
0.05 0.05
0.00 0.00
0% 20% 40% 60% 80% 100% 0% 20% 40% 60% 80% 100%
WPS share of the peak WPS share of the peak
b d

Fig. 3 Comparison of the three islands Crete, Lesvos and Serifos

a Turbine’s EPC
b Conventional units’ EPC
c Hybrid’s energy supply
d Electrical system’s EPC

IET Renew Power Gener., Vol. 1, No. 1, March 2007 55

Table 3: Assumptions of the conventional EPC’s calculation

Type of unit Installation cost, Fixed OMC, Variable OMC, Fuel consumption in the
E/kW E/kW E/kWh nominal point, g/kWh

Steam 1345 16 0.0011 221

Comb.cycle 662 22 0.0008 410
Gas 321 25 0.0007 310
Diesel 1272 9.4 0.0016 240
HFO 1345 9.4 0.0016 240
Diesel (small units) 2544 9.4 0.0016 225

Figs. 3a and b) and the price which has to pay to the WPS winds called ‘meltemi’ during summer justifies a bigger
investor, which should be at least the EPCT . This means that turbine – in relation to the rest components (wind capacity
the public utility will have the opportunity to buy cheaper and reservoir) – and higher energy and peak supply
electricity instead of producing it with higher cost. achievement occurs.
Finally, the higher contribution of the WPS can be achieved In Table 4 the required wind installed capacity, the
in the case of Serifos. dimensioning of the WPS (reservoir, turbine) and the peak
In Fig. 4, the required cost per kilo watt of guaranteed and energy supply achieved are presented. Regional
power and per kilo watt of installed capacity (wind and features – for example, the existence of a reservoir or a
hydro-turbine’s capacity) is presented. The different size favourite site of installation with bigger hydraulic head or
of the islands, the different size of the plants and the differ- higher wind potential – may further improve the feasibility
ent cost of the various subsystems as it is introduced by the of the plant. Additionally, the geographical distribution of
formulas in Table 2, give reasons for the differences wind farms in a large system may decrease the required
between the three islands. As a result the plant is expected capacity of the reservoir for the same target achievement.
more expensive in Serifos due to the small size (Figs. 4b
and 3a). Additionally, the different duration curve of the 7.2 Sensitivity on the allowed instantaneous
demand in the three islands may effect the dimensioning wind penetration
of the plant and its cost per kilo watt of guaranteed power
(Fig. 4a). For example in Serifos there is a very short In this paragraph, using the reference values for the wind
period of peak demand in summer and much lower potential (mean annual wind velocity 8.1 m/s) and for the
demand in the rest of the year. Seasonal strong north oil price (54$/b), the sensitivity of the results on the
allowed instantaneous wind penetration is examined. Two
different values of the allowed instantaneous wind pen-
7000
etration (40 and 50%) are introduced in comparison with
6000 the basic value (30%).
cost of the plant per kW of
guaranteed power (€/kW)

The effect of the allowed instantaneous wind penetration

5000
on the results is not so important, as it is in case without
4000 storage (unpublished). This means that the feasibility of
Crete
the WPS is proved even if the technical constraints of the
3000
lesvos system remain strict (Figs. 5 – 7). The reason is that the
2000 serifos consideration of the turbine as a spinning reserve allows
whether or not the increase of the directly absorbed wind
1000
power.
0
0% 20% 40% 60% 80% 100% 7.3 Sensitivity on the oil price
WPS share of the peak
a The EPCS with (straight lines) and without the WPS inte-
3000 gration are compared for three different oil prices (54, 75
and 100$/b). As it is expected, in Crete and Serifos
2500
cost of the plant per kW of

(Figs. 8 and 9) the highest the WPS penetration, the

installed capacity (€/kW)

lowest is the dependence of the EPCS on the oil price. In

2000
the case of Lesvos (Fig. 10) although with the current oil
1500 price the integration of the WPS may not decrease the
Crete system’s EPC, the situation is changed as soon as the oil
1000 lesvos price is increased. This means that even in cases, where
serifos the integration of the WPS does not provide a lower cost
500 today, it reduces the risk to deal with higher costs in the
0
near future due to a probable increase of the oil price.
0% 20% 40% 60% 80% 100% Consequently, the dependence of the EPC on the oil price
WPS share of the peak can be significantly reduced with the WPS integration.
b
7.4 Financial evaluation
Fig. 4 Comparison of the cost of the WPS plants in the three
islands Crete, Lesvos and Serifos The price for the electricity produced by the turbine is not
a Per kilowatt of guaranteed power strictly defined, so there is an uncertainty in the financial
b Per kilowatt of installed capacity (wind and hydro-turbine) evaluation of the investment. The new law 3468/2006 in
56 IET Renew Power Gener., Vol. 1, No. 1, March 2007
Table 4: Proposed dimensioning in the three case-studies islands

Island Peak, MW Wind capacity, MW Reservoir, †106 m3 Turbine, MW % Peak supply % Energy supply

Crete 563 825–1314 60 –120 310–405 55– 75% 46– 72%

Lesvos 58 84 –145 4.4–14 30 –46 57– 81% 51– 79%
Serifos 2.9 2.2– 3.6 0.2–0.3 2.3– 2.9 75– 100% 68– 83%

80% 90%

Hybrid's energy supply (%)

70% 80%
Hybrid's energy supply (%)

70%
60%
60%
50% Crete 50%
Serifos
40% 40%
30%
30% 30% 30%
40%
20% 40%
20% 50%
10% 50%
10%
0%
0% 0% 20% 40% 60% 80% 100%
0% 20% 40% 60% 80% WPS share of the peak
WPS share of the peak a
a 0.35
0.20
0.30

Electrical system's EPC

0.18
Electrical system's EPC

0.16 0.25 Serifos

0.14 Crete 0.20

(€/kWh)
30%
0.12
(€/kWh)

0.15 40%
0.10 30%

0.08 0.10 50%

40%
0.06 0.05 without
50%
WPS
0.04
without 0.00
0.02 WPS 0% 20% 40% 60% 80% 100%
0.00 WPS share of the peak
0% 20% 40% 60% 80%
WPS share of the peak
b
b
Fig. 7 Sensitivity of the results in Serifos on the allowed instan-
Fig. 5 Sensitivity of the results in Crete on the allowed instan- taneous wind penetration (8.1 m/s, 54$/b)
taneous wind penetration (8.1 m/s, 54$/b)
a Hybrid’s energy supply (%)
a Hybrid’s energy supply (%) b Electrical system’s EPC (E/kWh)
b Electrical system’s EPC (E/kWh)

90% Greece proposes that this price should be defined, respect-

ively, to the average variable cost of the conventional elec-
Hybrid's energy supply (%)

80%
70% tricity production in the autonomous electrical system in the
60% period of the license issue. In Figs. 11– 13, the required
50% Lesvos prices to provide an Internal Rate of Return (IRR) 8, 12
40% and 16% to the investors in the three islands are presented
30% 30%
under the following assumptions.
20% 40%
10% 50% † Own capitals: 50%.
0% † Subsidy: 0%.
0% 20% 40% 60% 80% 100% † Loan: 50% with discount rate 6% and 10 years payment
WPS share of the peak
a
period.
0.14 † Lifetime of the project 20 years and residual value 35%.
0.12 † Tax 25%.
Electrical system's EPC

Lesvos
0.10 0.30
30% 100$/b
0.08
(€/kWh)

Electrical system's EPC

0.25 Crete
40%
0.06 75$/b
0.20
50%
0.04 54$/b
(€/kWh)

54$/b
without 0.15
0.02 75$/b
WPS
0.10 100$/b
0.00
0% 20% 40% 60% 80% 100%
0.05 without
WPS share of the peak
WPS
b 0.00
0% 20% 40% 60% 80% 100%
Fig. 6 Sensitivity of the results in Lesvos on the allowed instan-
taneous wind penetration (8.1 m/s, 54$/b) WPS share of the peak

a Hybrid’s energy supply (%)

b Electrical system’s EPC (E/kWh) Fig. 8 Sensitivity of the electrical system’s EPC in Crete on the
oil price with and without WPS integration (8.1 m/s, d ¼ 30%)
IET Renew Power Gener., Vol. 1, No. 1, March 2007 57
 0.20
0.18 100$/b
Electrical system's EPC
0.16 Lesvos
0.14 75$/b
54$/b
0.12
(€/kWh)

54$/b
0.10 75$/b
0.08
100$/b
0.06
0.04 without
0.02 WPS
0.00
0% 20% 40% 60% 80% 100%
WPS share of the peak

Fig. 9 Sensitivity of the electrical system’s EPC in Lesvos on the

oil price with and without WPS integration (8.1 m/s, d ¼ 30%) Fig. 13 Required price for the electricity produced by the hydro-
turbine to provide IRR ¼ 8, 12 and 16% in Serifos (8.1 m/s,
d ¼ 30%)
0.50 100$/b
0.45 The results show that in Crete and Serifos the investments
75$/b
Electrical system's EPC

0.40 are feasible, since the required price is competitive to the

0.35 54$/b EPCS and the fuel cost of the system before the WPS inte-
Serifos
0.30
gration. The situation is different for Lesvos, where the
(€/kWh)

54$/b
0.25
75$/b feasibility of the investment is marginal.
0.20
0.15 100$/b

0.10 without
0.05
WPS 8 Conclusions
0.00
0% 20% 40% 60% 80% 100% With the introduction of the WPS, the wind penetration in
WPS share of the peak autonomous systems can be increased, simultaneously
decreasing the system’s EPC. As a consequence, the oper-
Fig. 10 Sensitivity of the electrical system’s EPC in Serifos on ation of conventional units and their required installed
the oil price with and without the WPS integration (8.1 m/s, capacity can be significantly reduced.
d ¼ 30%) The financial benefit from the introduction of the WPS
should be shared between the ESO and the investor, by
the definition of a suitable price. The main parameters
which should be taken into consideration for the definition
of the suitable price are the size of the plant, the size of
the island, the current cost of the system and the duration
curve of the demand.
Another important advantage of the WPS integration is
that the production cost is to a large extent known in
advance, contrary to the current cost which depends
strongly on the oil price. Thus the installation of WPS can
provide both financial and environmental benefits and is
strongly recommended.

9 References
Fig. 11 Required price for the electricity produced by the hydro-
turbine to provide IRR ¼ 8, 12 and 16% in Crete (8.1 m/s, 1 Caralis, G., and Zervos, A.: ‘Assessment of the wind penetration in
d ¼ 30%) Autonomous Greek islands’. 3rd National Conf. for RES, RENES,
February 2005, Athens, pp. 23–25 (in Greek)
2 Papathanassiou, S., and Boulaxis, N.: ‘Power limitations and energy
yield evaluation for wind farms operating in island systems’, Renew.
Energy, 2006, 31, (4), pp. 457– 479
3 Regulatory Authority of Energy: ‘Methodology for the assessment
of wind penetration in non-interconnected islands’ Available
at: http://www.rae.gr/k2/ape-penetration.pdf, http://www.rae.gr/K2/
deliberation-ape.html (in Greek), accessed February 2003
4 Katsaprakakis, D., and Christakis, D.: ‘The Wind power penetration in
the island of Crete’. RES & RUE for Islands Int. Conf., Cyprus, 30– 31
August 2004
5 Zervos, A., Caralis, G., Gorgoulis, M., and Zografakis, N.:
‘Implementation plan for the large scale deployment of RES in
Crete-Greece’. Altener 2000 Conf., 2000, Toulouse-France
6 Boulaxis, N.G., and Papadopoulos, M.P.: ‘Assessment of the
contribution of hybrid systems in renewable energy penetration in
islands’. ISAP 2003, Lemnos Greece, 31 September, Paper
ISAP03/141
7 Caralis, G., and Zervos, A.: ‘Prospects of wind and pumped storage
Fig. 12 Required price for the electricity produced by the hydro- systems’ integration in Greek islands’. EWEC 2006, Athens, Greece
turbine to provide IRR ¼ 8, 12 and 16% in Lesvos (8.1 m/s, Available at http://www.ewec2006proceedings.info/allfiles2/457_
d ¼ 30%) Ewec2006fullpaper.pdf

58 IET Renew Power Gener., Vol. 1, No. 1, March 2007

 8 Kaldellis, J., Kavvadias, K., and Vlachou, D.: ‘Electricity load 120%
management of APS using wind-hydro solution’. Proc. MedPower,
2002, Athens, Greece
100%
9 Katsaprakakis, D., Betzios, G., and Christakis, D.: ‘combined
management of water sources and wind energy in Lesvos – Is the
80%

η dimensionless
target of 100% renewable electricity supply in Lesvos possible?’.
17th Conf. Pan-Hellenic Network of Ecological Organizations,
Lesvos-Greece, 7 –9 October 2005 (in Greek) 60%
10 Theodoropoulos, P., Mantas, Z., Zervos, A., and Betzios, G.:
‘Integrated power system of Serifos Island with high RES
penetration using pump-storage’. Int. Conf., RES for Island Tourism 40%
and Water, 2003, EREC, Crete-Greece
11 EWEA. ‘wind energy the facts: an analysis of wind energy in the 20%
EU-25’ February 2004
12 Manolakos, D., Papadakis, G., Papantonis, D., and Kyritsis, S.: ‘A 0%
simulation-optimisation programme for designing hybrid energy
systems for supplying electricity and fresh water through 0% 50% 100% 150% 200%
desalination to remote areas. Case study: the Merssini village, Qdimensionless
Donoussa island, Aegean Sea, Greece’, Energy, 2001, 26,
pp. 679–704 160%

140%
10 Appendix
120%
10.1 Pumps – hydro-turbine simulation

Hdimensionless
100%
The problem of the simulation of the pumping station can be 80%
described as ‘Which is the required power to commit i
60%
pumps and which is the flow of the water’, and in the
case of the hydro-turbine’s simulation as ‘Which is the 40%
required volume of the water to produce the required
20%
power’.
The pump and turbine characteristics (Head-flow curve 0%
and efficiency-flow curve) are introduced dimensionless 0% 50% 100% 150% 200%
(Figs. 14, 15). The presented curves were assumed as repre- Qdimensionless
sentative for a wide range of centrifugal pumps and Pelton
turbines with different nominal characteristics [12]. Fig. 14 Dimensionless efficiency-flow and head-flow curve for a
The operation point (Fig. 16) of the pumping station is representative centrifugal pump
calculated as the intersection of the pumps characteristic
(H, Qi), where i is the number of pumps in operation and
the pipeline characteristic HP(Q). 120%

HP ðQÞ ¼ H þ z  Q2 ð22Þ 100%

ηdimensionless

where z is the loss coefficient of the penstock, H the hydrau- 80%

lic head and Q the water flow.
60%
The operation point of the hydro-turbine is calculated as
the intersection of the hydro-turbine characteristic (H, Q), 40%
and the pipeline characteristic HP(Q).
20%
HT ðQÞ ¼ H
z  Q2 ð23Þ
0%
0% 20% 40% 60% 80% 100% 120%
The loss coefficient is calculated Qdimensionless
L 1
z¼l ð24Þ Fig. 15 Dimensionless efficiency-flow curve for a representative
DP 2gA2 Pelton turbine
where, l (The loss coefficient and the dimensionless loss
coefficient are calculated for the different water flows.) is
the dimensionless loss coefficient, LP the length of the pen-
stock, DP the diameter of the penstock and A the area of the
penstock. For the calculation of the l, the formula of the
Colebrook and White is used
 
1 2; 51 1s
pffiffiffi ¼
2 log pffiffiffi þ ð25Þ
l Re  l 3; 71
where 1s is the relative roughness of the penstock’s wall
1s ¼ 1/D and 1 the absolute roughness.
A short iterative procedure is used to solve the above
complex formula. The Reynolds is defined as Re ¼ c . D/v,
where c is the water velocity in the penstock c ¼ Q/A
and v is the cinematic viscosity of the water
v ¼ 1.31  1026 (m2 . s21) for water temperature 108C. Fig. 16 Definition of the operation point of the pumping station
IET Renew Power Gener., Vol. 1, No. 1, March 2007 59
 Definition of the Penstock’s diameter: The diameter of flow of each pump in the nominal point of operation QP,R
the two penstocks for the turbine DP,T , and the pumping
station DP,P is calculated, assuming a maximum permitted r  g  hP  NP;R
Qmax;P ¼ NP  QP;R ¼ NP  ð28Þ
water velocity Vmax , which occur for the maximum water H
flow Qmax . The following equations are applied for the pen-
stocks of the pumping station and the hydro turbine. where hP ¼ 0.85 is the efficiency of the pump in the
nominal point.
In the case of the hydro-turbine, the maximum water flow
4Qmax
DP ¼ ð26Þ Qmax,T is defined by the hydro-turbine
ðVmax  pÞ1=2
r  g  NT;R
Vmax ¼ 0:125  ð2gHÞ0:5 ð27Þ Qmax;T ¼ 1:3  QT ;R ¼ 1:3 
hT  H
ð29Þ

In the case of the pumping station, the maximum water flow where hT ¼ 0.90 is the efficiency of the hydro-turbine in the
Qmax,P is defined by the number of pumps and the water nominal point.

60 IET Renew Power Gener., Vol. 1, No. 1, March 2007

The author has requested enhancement of the downloaded file. All in-text references underlined in blue are linked to publications on ResearchGate.