AckermannTh. (ed) Wind Power in Power Systems
Подождите немного. Документ загружается.
//INTEGRAS/KCG/P AGIN ATION/ WILEY /WPS /FINALS_1 4-12- 04/0470855088_ 16_CHA15 .3D – 331 – [331–348/18]
20.12.2004 7:47PM
15
Wind Farms in Weak Power
Networks in India
Poul Sørensen
15.1 Introduction
Several countries support wind energy in developing countries in order to ensure
sustainable development that can meet the rapidly growing demand for power in these
countries. India is the developing country that has experienced by far the fastest
expansion of wind energy. It is therefore selected as a case study in this chapter. India
is also an interesting and typical case, as the networks are weak and in the process of
being built. The present chapter is based on a study of wind farms connected to weak
grids in India (Sørensen et al., 2001).
India has experienced a fast increase in wind energy, especially in the mid-1990s, when
subsidies and tax allowances for private investors in wind farms promoted the installa-
tion of wind farms. The State of Tamil Nadu in the southern part of India was leading
this development, but the State of Gujarat also installed a substantial capacity of wind
energy. Figure 15.1 shows the centres of this development: the M uppandal area in Tamil
Nadu and the Lamba area in Gujarat.
The total wind energy capacity installed in Tamil Nadu during the 1990s amounted to
approximately 750 MW (Ponnappapillai and Venugopal, 1 999), 20 MW of which were
funded by the government as demonstration projects. The remaining 730 MW were
funded by private investors. According to the load dispatch centre of Tamil Nadu
Electricity Board (TNEB), the total conventional power supply capacity in Tamil Nadu
amounts to 7804 MW and comprises thermal, nuclear and hydro power stations. That
means that in Tamil Nadu wind power capacity corresponds to 8.8 % of total power
Wind Power in Power Systems Edited by T. Ackermann
Ó 2005 John Wiley & Sons, Ltd ISBN: 0-470-85508-8 (HB)
//INTEGRAS/KCG/P AGIN ATION/ WILEY /WPS /FINALS_1 4-12- 04/0470855088_ 16_CHA15 .3D – 332 – [331–348/18]
20.12.2004 7:47PM
supply capacity. In 1999–2000, the wind power production in Tamil Nadu was
1157 GWh, supplying 3.1 % of total consumption (37 159 GWh).
In Gujarat, the utilisation of wind energy started with government demonstration
projects in 1985. The total installed wind energy capacity is 166 MW: 16 MW have been
funded as government demonstration projects, and the remaining 150 MW have been
funded by private investors. In 2000, the capacity of conventional power supply in
Gujarat amounted to 8093 MW, including private sector and national grid allotment.
Consequently, wind power capacity corresponded to 2.0 % of the total power supply
capacity in Gujarat.
In comparison, the Western part of Denmark had approximately 1900 MW of
installed wind power by early 2001, according to Eltra (http://www.eltra.dk) which is
the transmission system operator (TSO) in charge of this region. Other power plants had
a capacity of 4936 MW. Wind energy capacity corresponded to 28 % of total installed
power capacit y. In 2000, the wind power production in the Eltra area was 3398 GWh,
corresponding to 16.5 % of the total consumption (20 604 GWh).
Table 15.1 summarises the above-mentioned numbers. Even though the periods the
figures refer to do not exactly coincide, a comparison is still reasonable. The figures
indicate that the wind energy penetration level in Tamil Nadu is high, though not
exceptionally high, whereas in Gujarat, the wind energy penetration is moderate.
MEGHALAYA
INDIA
2001
ARUNACHAL
PRADESH
JAMMU &
KASHMIR
PUNJAB
HIMACHAL
PRADESH
UTTARANCHAL
SIKKIM
HARIANA
DELHI
UTTAR
PRADESH
RAJASTHAN
GUJARAT
BIHAR
ASSAM
NAGALAND
MANIPUR
MIZORAM
TRIPURA
W. BENGAL
JHARKHAND
MADHYA
PRADESH
Lamba
MAHARASHTRA
ORISSA
ANDHRA
PRADESH
KARNATAKA
GOA
DAMAN & DIU
D & N HAVELI
TAMIL
NADU
PONDICHERRY
KERALA
Muppandal
A & N ISLANDS
C. CHANDIGARH
P. PONDICHERRY
P
P
P
CHHATISGARH
LAKSHADWEEP
Figure 15.1 Map of India, 2001, showing Muppandal (State of Tamil Nadu), and Lamba (State
of Gujarat)
332 Wind Farms in Weak Power Networks in India
//INTEGRAS/KCG/P AGIN ATION/ WILEY /WPS /FINALS_1 4-12- 04/0470855088_ 16_CHA15 .3D – 333 – [331–348/18]
20.12.2004 7:47PM
Evaluating the penetration levels, it has to be taken into account that the power
systems in Tamil Nadu and Gujarat are both parts of larger power systems, similar to
the power system in the western part of Denmark, which is part of the strong central
European power system. The power system in India is divided into four subsystems.
These are interconnected only through DC-links and consequently do not operate
synchronously (Hammons et al., 2000). Tamil Nadu is part of the power system in the
southern part of India, together with the three neighbouring States of Kerala, Karnataka
and Andhra Pradesh. Gujarat, however, is included in the power system of western
India together with the two neighbouring States of Maharastra and Madhya Pradesh.
In India, the power systems are publicly owned. Each State has its own electricity
board, [e.g. the Tamil Nadu Electricity Board (TNEB) and the Gujarat Electricity
Board (GEB)]. Each State also has its own development agency for planning wind
energy development and for handling the grid connection [e.g. the Gujarat Energy
Development Agency (GEDA)]. In Gujarat, GEDA owns and operates dedicated wind
farm substations, whereas in Tamil Nadu, the substations are owned and operated by
TNEB and are not co mpletely dedicated to the wind farms.
In India, as in other countries in eastern Asia, industrial and economic development is
progressing very quickly and, at the same time, the population is growing very quickly
too. This results in fast growing demands on power syst ems. The electricity boards have
not been able to develop the necessary power system capacities for supply, transmission
and distribution accordingly. As a consequence, the power systems have in several ways
been weakened over the years.
The major part of the wind energy installations is concentrated in a few rural areas,
where the existing distribution and transmission grids are particularly weak. As will be
shown next, the power quality at the connection points of the wind turbines is mostly
poor, which affects the operation of the wind turbines. To some extent, the connection of
wind turbines to these grids has even worsened the power quality, because grid reinforce-
ments have not been sufficient to integrate the wind energy into the power systems.
In this chapter I describe the interrelationship between wind turbines and power
quality in the Indian power systems. First, I will characterise the grids in the areas
Table 15.1 Penetrations in Gujarat, Tamil Nadu and the Eltra area
(Denmark) in 2000
Area
Tamil Nadu Gujarat Eltra
Wind capacity (MW) 750 166 1 900
Conventional capacity (MW) 7 804 8 093 4 936
Total capacity (MW) 8 554 8 259 6 836
Wind capacity penetration (%) 8.8 2.0 27.8
Wind production (GWh) 1 157 — 3 398
Consumption (GWh) 37 159 — 20 604
Wind production penetration (%) 3.1 — 16.5
— Production and consumption data for Gujarat are not available
Wind Power in Power Systems 333
//INTEGRAS/KCG/P AGIN ATION/ WILEY /WPS /FINALS_1 4-12- 04/0470855088_ 16_CHA15 .3D – 334 – [331–348/18]
20.12.2004 7:47PM
where the wind turbines are connected and describe the overall electrical design of the
grid connections of the wind turbines. Referring to the characteristics of grids and wind
turbines, I will describe the impact of the wind turbi nes on power quality, on the one
hand, and the effect that the power quality has on the wind turbines, on the other hand.
15.2 Network Characteristics
This section summarises the main characteristics of the Indian grids, which are important
with regard to the interrelationship between wind turbines and grids. I will focus on the grid
in the Muppandal region, where more than half of the wind turbines in Tamil Nadu are
connected. In addition, the grid in the coastal region around Lamba in Gujarat is described.
15.2.1 Transmission capacity
One of the most critical issues for the development of wind energy in India has been the
transmission capacity of the grid in the areas where the wind farms were built. Similar to
many other countries, wind farms are concentrated in rural areas, where the existing
transmission grids are very weak. In addition, in India the new wind farms were built
during a comparatively short period of time and were restricted to a few areas, and the
reinforcement of the transmission system in these areas has lagged behind the fast
development of wind energy.
In Tamil Nadu, almost 600 MW of wind energy were inst alled over the three- year
period from 1994 to 1997. Approximately 400 MW of the wind energy capacity are
located in the Muppandal area. Figure 15.2 shows the grid connection diagram for that
area. The 110 kV Arumuganeri–Kodayar ring mains was constructed to supply the
Muppandal area with power. The capacity of the ring mains itself is only 200 MVA.
To improve transmission capacity, an additional 110 kV line to Melak Kalloor and two
230 kV lines from the S. R. Pudur substation to the Tutucorin 230/110 kV substation
and the Kayathar 230/110/11 kV substation, respectively, have now been installed as a
supplement to the ring mains.
Figure 15.2 indicates the sho rt-circuit levels at selected points of the 110 kV and
230 kV grid. The short-circuit power at the wind farm substations of the 110 kV ring
mains is about 800 MVA. This is very low considering that approximately 200 MVA of
wind power are connected directly to the ring.
In Gujarat, there are 60 MW of wind power capacity in the coastal area around
Lamba. The wind farms are connected to three dedicated substations, as shown in
Figure 15.3. The power from the wind farms is evacuated through a 66 kV line with a
capacity of 30 MVA, and a 132 kV double line with a total capacity of 60 MVA.
In June 1998, the capacity installed at these substations amounted to 75 MW. However,
in June 1998, a capacity of approximately 40 MW was damaged by a cyclone, which also
damaged the overhead lines that were used for the transmission of the power from the
wind farms. Before the cyclone, the only transmission line was a 66 kV line with a capacity
of 30 MVA. This configuration led to grid outages as a result of load shedding and grid
breakdowns after a 66 kV conductor snapped because of age and overloading. The process
of ageing of the line has been accelerated by the saline atmosphere close to the sea.
334 Wind Farms in Weak Power Networks in India
//INTEGRAS/KCG/P AGIN ATION/ WILEY /WPS /FINALS_1 4-12- 04/0470855088_ 16_CHA15 .3D – 335 – [331–348/18]
20.12.2004 7:47PM
15.2.2 Steady-state voltage and outages
The insufficient capacity of the power syst ems causes large variations in the steady-state
voltage, and outages on the grids. Figure 15.4 shows measurements of the steady-state
voltage on the primary side (110 kV) of the wind farm substation of Radhapuram in
1999 (Sørensen et al., 2000). Some of the measurements are from the windy season in
May and August, others are from the less windy season in October. The Radhapuram
Pechipparai
110
/11 kV
Veeyanoor
110
/11 kV
Kuzhithurai
110/11 kV
Munchirai
110
/11 kV
Thuckalay
110
/11 kV
Nagercoil
110
/33/11 kV
Manavala
Kuruchi
33
/11 kV
Kenyakumari
33/11 kV
Villankumara
33/11 kV
Aralvai
110/33 /11 kV
Muppandal
110/11 kV
Panagudi
110/11 kV
S.R.Pudur
230
/110 kV
Karungulam
110
/11 kV
Veppilankulam
110/11 kV
Radhapuram
110/11 kV
Perungudi
110/33 /11 kV
Koodankulam
110/11 kV
Kodayar
PH2
Kodayar
PH1
Veerava
Nalloor
110
/11 kV
Melak
Kalloor
110
/11 kV
Palaya
Pettai
110
/11 kV
Talayuthu
110/33 /11 kV
Karantha
Neri
110/11 kV
Katai
Karunkulam
110
/33/11 kV
Sathankulam
110/33 /11 kV
Moolaikarai
33/11 kV
Srivaikuntam
33/11 kV
Arumuganeri
110/11 kV
Tutucorin
230
/110 kV
Kayathar
230/110 /11 kV
TTPS
5×210 MW
Substation
Wind farm substation
230 kV line
110 kV line
33 kV line
2006 MVA
1109 MVA
809 MVA
806 MVA
712 MVA
Pazhavoor
110/11 kV
Short-circuit power
Figure 15.2 Muppandal grid with the original 110 kV Arumuganeri–Kodayar ring mains and
230 kV line in Muppandal, Tamil Nadu, based on drawings by an engineer of the Muppandal
office of the Tamil Nadu Electricity Board
Bhogat
132
/66/11 kV
23 MW
Bhatia
132
/66 kV
Navadara
132/66 /11 kV
12 MW
Lamba
132
/66/11 kV
25 MW
66 kV
30 MVA
132 kV
60 MVA
Substation
Wind farm substation
66 kV line
132 kV line
Figure 15.3 Wind farms and substations on the costal lines of Porbandar, Gujarat: dedicated
lines for transmission and no tapping for rural distribution
Wind Power in Power Systems 335
//INTEGRAS/KCG/P AGIN ATION/ WILEY /WPS /FINALS_1 4-12- 04/0470855088_ 16_CHA15 .3D – 336 – [331–348/18]
20.12.2004 7:47PM
substation is one of the eight wind farm substations connected to the 110 kV ring mains
in Muppandal (Figur e 15.2).
The curves in Figure 15.4 show distinct diurnal variations in the voltages, with typical
low voltages during the peak load period in the evening and high voltages at night.
During two of the da ys, the steady-state voltage is about 15 % below the rated voltage
for approximately 2 h. This is below the tolerance of 12:5 % specified in the Indian
Standard (IS) 123600 (Indian Standard, 1988). The voltage variations on the 110 kV
mains indirectly affect the voltages at the wind turbine terminals. The transformers in
the substations in Muppandal have no automatic voltage regulation to compensat e for
this. Remote voltage regulation is carried out manually from the substation by observa-
tion of the voltages in the substation.
Figure 15.4 also shows four outages with a total duration of approximately 1 h,
during approximately 200 h of measurements. Since the measurements are taken on
the primary side of the substation, these outages only include outages on the 110 kV
level or higher. The number of outages is expected to be higher on the low voltage side,
where the wind turbine terminals are connected to the grid. According to TNEB, the
grid availability of the dedicated wind farm feeders during peak production is in the
range of 95–96 %, with an annual average of 98 %.
The main reason for the voltage variations in Figure 15.4 is the load variation, which
can be observed by the regular diurnal variation. The wind farms will, however, also
affect voltage, both as a result of power production (which increases voltage) and as a
result of reactive power consumption (which decreases voltage).
In Gujarat, the interruption time on the line to Lamba was extremely long, before the
cyclone and the reinforcement of the 66 kV line with a 132 kV line. Rajsekhar, Hulle and
Gupta (1998) concluded that the total operational time of the wind turbines in Lamba
was reduced by 4–5 % as GEDA disconnected 11 kV feeders in order to avoid the 66 kV
0.80
0.85
0.90
0.95
1.00
1.05
1.10
1.15
1.20
0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00 0:00
Time (hh:mm)
Voltage (p.u.)
6–7 May
27–29 August
12–21 October
IS 12360
Figure 15.4 Steady-state voltage measured in Radhapuram substation. Note: IS 12360 ¼ Indian
Standard 12360 (Indian Standard 1988)
336 Wind Farms in Weak Power Networks in India
//INTEGRAS/KCG/P AGIN ATION/ WILEY /WPS /FINALS_1 4-12- 04/0470855088_ 16_CHA15 .3D – 337 – [331–348/18]
20.12.2004 7:47PM
line tripping, an d by another 474 h per year (5 %) in 66 kV outages due to faults in the
GEB system. After the installation of the 132 kV line, GEB has reduced the number of
interruptions to two per month, with a duration of 6–8 h per month.
15.2.3 Frequency
In the Indian power systems, the frequency varies significantly compared with European
systems. Figure 15.5 shows the frequency measured simultaneously with the voltage.
The curves in Figure 15.5 have diurnal variations in frequency which are similar to the
voltage variations shown in Figure 15.4, with frequencies of about 51 Hz at night and down
to 48 Hz during the day. According to IS 12360, the frequency should be 50 Hz 3 % (i.e.
in the interval of 48.5 Hz to 51.5 Hz). It appears from the curves that the frequency is
floating until a lower limit of 48 Hz is reached. Then the frequency is regulated, presumably
by load shedding, in order to minimise the gap between supply and demand.
The wind farms are not a major source of frequency variation, as the installed wind
power capacity is relat ively small compared with the total capacity of the interconnected
system in south India.
15.2.4 Harmonic and interharmonic distortions
With decreasing prices, the application of power electronics for power supplies and motor
drives, for instance, is steadily increasing all over the world. The increased load from
power electronics causes harmonic and interharmonic distortions in the power systems.
Standard power supplies for appliances emit only low orders of odd harmonics. Since
3rd-order harmonics are cancelled out by star–delta connected transformers, the main
47.00
47.50
48.00
48.50
49.00
49.50
50.00
50.50
51.00
51.50
52.00
0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00 0:00
Time
Frequency (Hz)
6–7 May
27–29 August
12–21 October
IS 12360
Figure 15.5 Frequency measured at Radhapuram substation. Note: IS 12360 ¼ Indian Standard
12360 (Indian Standard, 1988)
Wind Power in Power Systems 337
//INTEGRAS/KCG/P AGIN ATION/ WILEY /WPS /FINALS_1 4-12- 04/0470855088_ 16_CHA15 .3D – 338 – [331–348/18]
20.12.2004 7:47PM
emission to the medium-voltage grid from power supplies are 5th and 7th harmonics.
Modern motor drives emit higher-order harmonics and interharmonics, typically with
frequencies above 1 kHz.
Measurements at the wind farm substations have shown only moderate harmonic and
interharmonic distortions of the voltages at the grid connection points on the primary
sides (110 kV and 66 kV), except for distortions at Radhapuram substation, caused by
wind turbines with power converters.
15.2.5 Reactive power consumption
The loads on the power systems in India consume a significant amount of reactive
power, mainly as a result of the use of agricultural pumps. The reactive power con-
sumption affects the power system by:
.
increasing losses in the transmission and distribut ion grids;
.
reducing thermal capacity for the transmission of active power in the distribution and
transmission grids;
.
reducing the active power capacit y of the synchronous generators in the other power
plants in the system;
.
raising the risk of voltage instability in the power system (Taylor, 1994).
Usually, the reduction of the active power capacity of the synchronous generators is not
a critical issue. In India, however, it is important because production capacity is
insufficient to supply the demand.
15.2.6 Voltage imbalance
The voltage imbalance can be high on rural feeders that are dominated by unbalanced
single-phase loads. Experience with wind turbines connected to rural load feeders in India
indicates that the electricity boards perform load shedding on single phases of the rural
feeders, which causes severe voltage imbalances. However, most of the wind turbines in
India are connected to the substations by dedicated wind farm feeders. Therefore,
measurements on load feeders to quantify this issue have not been carried out.
Voltage imbalances have been measured on the primary sides of five win d farm
substations in Tamil Nadu and Gujarat (Sørensen, Unnikrishnan and Lakaparampil,
1999), and the results show low voltage imbalances. IS 12360 does not include limits for
voltage imbalance. The European standard for voltage quality, EN 50160 (European
Standard, 1999) sets a limit of 2 %. This limit was not exceeded by any of the measur e-
ments at the substations.
15.3 Wind Turbine Characteristics
Many of the wind turbines in India have a rated capacity of between 200 kW and
250 kW, but there is the whole range, from 55 kW to 500 kW. Most of the wind turbines
are produced in India, in cooperation with the western wind turbine industry.
338 Wind Farms in Weak Power Networks in India
//INTEGRAS/KCG/P AGIN ATION/ WILEY /WPS /FINALS_1 4-12- 04/0470855088_ 16_CHA15 .3D – 339 – [331–348/18]
20.12.2004 7:47PM
The majority of wind turbines installed in India use induction generators directly
connected to the grid (wind turbine Type A)
(1)
. Both stall control (Type A0) and pitch
control (Type A1) are used. The capacitor banks in these wind turbines are typically
designed to compensate only the induction generators’ no-load consumption of reactive
power. Owing to a substantial increment in the fees for reactive power in Tamil Nadu in
2000, some manufacturers now compensate the full reactive power consumption.
In India, another widely used grid connection concept for wind turbines is to connect
the generator through a full-scale power converter (Type D). Most of these wind
turbines use pitch control (D1). However, the wind turbines with power converters
use different types of components. Both induction generators and synchronous gen-
erators are used, and the power converters use either thyristor switches or integrated
gate bipolar transistor (IGBT) switches. Power converters using thyristor switche s also
consume reactive power, whereas power converters with IGBT switches can control the
power factor fully at the wind turbine terminals. If the power factor is controlled to
unity at the wind turbine terminals, only the transformer uses reactive power.
Type C wind turbines [doubly fed induction generators (DFIGs)] are not installed in
India yet. When wind energy development in India peaked during the mid-1990s, this
type of wind turbine was not very common.
15.4 Wind Turbine Influence on Grids
This section summarises the influence of wind turbines on the power quality of the weak
grids to which they are connected.
15.4.1 Steady-state voltage
In many countries, the influence on steady-state voltage is the main design criteria for
the grid connection of wind turbines to the medium-voltage distribution system. In
particular, if wind turbines are connected to existing rural load feeders, the effect of the
wind turbines on the steady-state voltage is important (Tande et al., 1996; DEFU, 1998).
According to IS 12360, the tolerance of the voltage on the 11/33 kV level is þ6/9%.
Voltage variations must be kept within these limits if wind turbines are connected to
rural feeders. If the wind turbines are connected to dedicated feeders, the voltage limits
are set by the official wind farm planning considerations (Government of India, 1998).
These specify that the voltage increase at each wind turbine must not exceed 13 % of
rated voltage.
15.4.2 Reactive power consumption
The reactive power consumption from wind farms adds to the reactive component
originating from agricultural loads. In particular, Type A wind farms consume reactive
power, but even Type D wind farms that produce at unity power factor consume
(1)
For definitions of wind turbine Types A to D see Section 4.2.3.
Wind Power in Power Systems 339
//INTEGRAS/KCG/P AGIN ATION/ WILEY /WPS /FINALS_1 4-12- 04/0470855088_ 16_CHA15 .3D – 340 – [331–348/18]
20.12.2004 7:47PM
reactive power for the transformers in wind farms and substations and in the overhead
lines. At the same time, wind farms feed active power into the grid. These two factors
together reduce the power factor at the conventional power stations.
Looking at the primary side of the wind farm substations, the wind farms affect the
reactive power of:
.
wind turbines themselves, especi ally wind turbines with directly grid connected induc-
tion generators;
.
step-up transformers between wind turbines and wind farm feeders;
.
wind farm feeders;
.
substation transformers.
Generally, these components consume reactive power, even though some wind turbines
may operate at a leading power factor and wind farm feeders produce reactive power at
low loading.
The reactive power on the primary side of five wind farm substations was measured.
Figure 15.6(a) plots the measured 10-minute-average values of react ive power against
active power at three substations in Tamil Nadu, whereas Figure 15.6(b) illustrates
similar measurements at two substations in Gujarat.
Figures 15.6(a) and Figure 15.6(b) indicate that the reactive power consumption of
the win d farm substations is lower in Tamil Nadu than in Gujarat. The main technical
reasons for the differences in reactive power consumption are that:
.
in Gujarat, approximately half of the capacitors in the wind turbines are defective,
whereas in Tamil Nadu most of the wind turbine capacitors are working;
.
the TNEB has installed capacitors on the secondary sides of the wind farm substa-
tions, which connect wind turbines that consume reactive power.
The high rate of defective wind turbine capacitors in Gujarat is mainly owing to the
fact that wind farm operators in Gujarat have no incentive to replace the defective
capacitors. In Tamil Nadu, in contrast, wind farm operators have to pay for reactive
power consumption. Another aspect could be that the environment is more aggressive in
Gujarat than it is in Tamil Nadu. However, the climate may only be a relevant aspect at
Lamba substation, which collects power from wind turbines installed on the coastline.
The wind turbin es in Dhank are situated inland, in a less aggressive environment.
The three substations in Tamil Nadu represent different technologies of connecting
wind turbines to the grid. Aralvai connects only wind turbines with directly grid
connected induction generators (Type A), Radhapuram connects only wind turbines
with IGBT-based power converters (Type D) and, finally, Karunkulam connects differ-
ent types, including wind turbines with thyristor-based power con verters (also Type D).
Figure 15.6(a) indicates that the reactive power consumption of the wind farms can be
kept low in the lower power range, independent of the wind turbine technology. There
are no measurements available regarding the upper half of the power range. At rated
power production, though, reactive power consumption is expected to increase to a
higher level in Aralvai and Karunkulam than in Radhapuram because of the different
technologies.
340 Wind Farms in Weak Power Networks in India