Optimizing the Use of Breaker Switched Capacitors in Ceb Power Grids
-
Upload
kusum-shnathi -
Category
Documents
-
view
241 -
download
0
Transcript of Optimizing the Use of Breaker Switched Capacitors in Ceb Power Grids
-
8/3/2019 Optimizing the Use of Breaker Switched Capacitors in Ceb Power Grids
1/11
OOPPTTIIMMIIZZIINNGG TTHHEE UUSSEE OOFF BBRREEAAKKEERR
SSWWIITTCCHHEEDD CCAAPPAACCIITTOORRSS IINN CCEEBB PPOOWWEERR GGRRIIDDSSUpul Dompege1 , J.P.Karunadasa2, Kusum Shanthi3
1Transmission Projects Branch, Ceylon Electricity Board, Sri Lanka
2Department of Electrical Engineering, University of Moratuwa, Sri Lanka
3Transmission O & MS Branch, Ceylon Electricity Board, Sri Lanka
Abstract: Ceylon Electricity Board (CEB) as many
other utilities uses breaker switched capacitor (BSC)banks for voltage support and reactive power
compensation in grid substations. At present it has a320Mvar installed capacity in 33kV level and according
to CEB transmission plan 70Mvar more to be added innext few years. The main intentions of the use ofcapacitor banks is to give voltage support at the
substation level, reduction of losses in powertransformers and transmission lines, and to release the
capacity constraints in transformers and lines. CEB uses
power factor regulation for switching these capacitor
banks for above purposes but no studies have been doneto evaluate its suitability. It is learnt that the switchingbased on this criteria does not fully match with the
system requirement and therefore sometimes necessaryto manually switch on them overriding the autocontrollers or vice versa. Optimizing the use such an
economical reactive power source to the specific
intention for which they have been installed is a key
issue to be addressed. The paper describes the workscarried out to evaluate the present switching criteria of
BSC banks in CEB and the proposal of a economicalway of switching with considering all technical aspects
related to capacitor bank operations in medium voltage
level including simulation and real time monitoring..
I INTRODUCTION
The 33kV capacitor banks in the CEB network areconnected to the 33kV load bus at Grid sub stations.
However at Pannipitiya the capacitors are connected tothe 33kV tertiary winding of the 220 / 132 / 33 kV inter
bus transformer. At all locations, the switching ON
criterion is based on power factor at 33kV transformerincoming feeder. Switching off is based on leading
reactive power limit or leading power factor. If voltage
support is necessary, the banks should be switchedconsidering the voltage at the point of connection. If the
capacity constraints or loss minimization is concerned,
then they shall be fully utilized to minimize drawing varfrom remote generation. Under these considerations,why CEB controls them in an indirect way like power
factor is a question. It should be checked whether the
requirements are best met with or the available resourcesare fully utilized with the present switching criteria [1].
As observed, there are situations where some of the
33kV capacitor banks at the grid substation are keptunused, while having an acute problem of heavy reactive
power requirement in transmission system. This happens
mostly when power across the companys transmission
system does not coincide with load conditions inlocations where the capacitor banks are fixed. In some
situations, the power factor may be within acceptablelimits but the voltages are below the nominal or on load
tap changer is forced on higher taps to take care of thevoltage. The substation level capacitor bank can directlyserve for voltage support or var support, without
depending on power factor regulation which is anindirect measure of voltage or var requirement.
The objective of this paper is to fill this void by
presenting the work carried out in following areas.
verify the applicability of present switching criteria
check and ensure the possibility of connectingmaximum capacitor banks installed withoutviolating technical constraints
review and optimizing the present switchingparameters, if the present switching criteria is the
optimal solution for the CEB.
and to design and propose a suitable switchingcriteria for the capacitors by means of network
simulation and practical implementation with
continuous monitoring.
II SITE SELECTION
Precise data at the substations is beneficial for such ananalysis but studying the total system is practically
impossible in a live system. However, a case study is a
sufficient and satisfactory solution for a research like
this. The duration of data measurements shall cover a
substantial duration to represent the actual systemvariations. The general practice of such a study is to
have one week duration. Sub station at Panadura wasselected as a pilot station and the research was based on
the findings for this sub station. The load curves both
real and reactive were compared with the systembehaviour and found satisfactorily matching and
representing the system as a whole. Details of thesubstation are as follows.
Sub station capacity 2 x 31.5 transformers
Incoming feeders T connection to PannipitiyaMatugama line / Double
circuit
No of feeders 6
No of capacitor banks 4 x 5 Mvar
Maximum average night peak 46MW +27Mvar
Minimum average load 19MW +12Mvar
-
8/3/2019 Optimizing the Use of Breaker Switched Capacitors in Ceb Power Grids
2/11
LEM Qwave Premium power quality analyzer and Ellite4 Pholyphase power meter was used for data
measurement. MW and Mvar, 33kV bus voltage, power
factor at 33kV incomer TF 1, tap position of on load tap
changer TF 1, load side harmonics were measured at33kV side. MW & Mvar, power Factor at 132kV busbar and 132kV bus voltage were recorded at the 132kV
level.
III SWITCHING CRITERIA OF CEB
There are two types of switching methods in the CEB
system. In both types the criterion for switching on thebanks is the lagging power factor. The controller
evaluates the power factor of the 33kV transformerincomer feeder using voltage and current analogue
signals and switches the first filter bank when the power
factor is below a certain specified limit. Generally, thislimit is 0.9800. The next banks are switched on as per
the same condition considering the calculated powerfactor. In one type of controllers, switching off is based
on leading power factor. In the second type, controllercompares the reactive power calculated using measured
power factor and measured the real power with thereactive power calculated using the set power factor andmeasured real power.
If the difference is greater than a multiple of minimumstep of the banks, then the banks are switched off
gradually. This multiple is calculated as (1+Hysterisis)
where the hysteresis setting is generally about 10% [2].
In the CEB system, if more than one controller is usedfor set of banks on each bus section, these works as
independent controllers when the bus section is openand in master slave mode if the bus section is in ON
position. In independent operation, the controllerswitches the banks assigned to it, typically two. First is
always the filter bank and compensator bank later. In
the master slave mode, the master will control all thebanks if the communication between the controllers is
established
IV PRESENT SWITCHING PATTERN
The figure 1a below shows the behaviour of the
capacitor banks over the full range of measurements (9days) with the present switching criteria. It is more
elaborated in figures 1b and 1c in a days window fortwo selected days. The figures 2a and 2b indicate HV
side voltage and reactive power requirement at 33 bus,for same two days.
Utilization of Cap Banks under present sytem
0
1
2
3
4
5
6
20.01.2009
21:30:00
21.01.2009
17:30:00
22.01.2009
13:30:00
23.01.2009
09:30:00
24.01.2009
05:30:00
25.01.2009
01:30:00
25.01.2009
21:30:00
26.01.2009
17:30:00
27.01.2009
13:30:00
28.01.2009
09:30:00
Time of Day
Noofcapbanks
No of Cap Banks at Panadura GSS at present schemeNo of Max Cap Banks
If only pF control is used
Utilization of Cap Banks (22.02.09)
-45.00
-40.00
-35.00
-30.00
-25.00
-20.00
-15.00
-10.00
-5.00
0.00
5.00
10.00
15.00
20.00
25.00
30.00
35.00
40.00
45.00
22.01.2009
00:00:00
22.01.2009
03:20:00
22.01.2009
06:40:00
22.01.2009
10:00:00
22.01.2009
13:20:00
22.01.2009
16:40:00
22.01.2009
20:00:00
22.01.2009
23:20:00
Time of Day
Phaseangle
0
1
2
3
4
5
6
NoofCapBanks
P ha se A ng le a t 3 3 bu s ( no C ap s) P has e An gle wit h C ap B ank s ( ca lc ula te d)Ph. angle (Nominal) No of Cap BanksNo of Max Cap Banks
Utilization of Cap Banks (24.02.09)
-45
-40
-35
-30
-25
-20
-15
-10
-5
0
510
15
20
25
30
35
40
45
24.01.2009
00:00:00
24.01.2009
03:20:00
24.01.2009
06:40:00
24.01.2009
10:00:00
24.01.2009
13:20:00
24.01.2009
16:40:00
24.01.2009
20:00:00
24.01.2009
23:20:00
Time of Day
Phaseangle
0
1
2
3
4
5
6
NoofCapBank
s
Pha se Angl e at 33 bus (no Cap s) Phase Angle wi th Cap Banks (c alcula ted)Ph. angle (Nominal) No of Cap BanksNo of Max Cap Banks
Fig.1a Switching pattern of capacitor banks over full measurement period
Fig.1b Switching pattern of capacitor banks 22nd
Jan 2009 Fig.1c Switching pattern of capacitor banks 24th Jan 2009
-
8/3/2019 Optimizing the Use of Breaker Switched Capacitors in Ceb Power Grids
3/11
Reactive power and HV side voltage (22.01.09)
-10.00
-5.00
0.00
5.00
10.00
15.00
20.00
22.01.2009
00:00:00
22.01.2009
03:20:00
22.01.2009
06:40:00
22.01.2009
10:00:00
22.01.2009
13:20:00
22.01.2009
16:40:00
22.01.2009
20:00:00
22.01.2009
23:20:00
Time of Day
Mvar
60.00
62.00
64.00
66.00
68.00
70.00
72.00
74.00
76.00
78.00
80.00
Voltage
Reacti ve powe r 132kV s ide volt age Nom inal vol tage
Reactive power and HV side voltage (24.01.09)
-10.00
-5.00
0.00
5.00
10.00
15.00
20.00
24.01.2009
00:00:00
24.01.2009
03:20:00
24.01.2009
06:40:00
24.01.2009
10:00:00
24.01.2009
13:20:00
24.01.2009
16:40:00
24.01.2009
20:00:00
24.01.2009
23:20:00
Time of Day
Mvar
60.00
62.00
64.00
66.00
68.00
70.00
72.00
74.00
76.00
78.00
Voltage
Reactive power 132kV si de vol tage Nomi nal vol tage
The figures show that the present switching pattern doesnot fully utilize the installed capacitor banks with thepresent power factor regulation switching criteria.
Comparing figures 1b with 2a and 1c with 2b, we can
see that at the start of green arrow, the load phase angle
becomes close to the setting value but next step of thebanks is not switched on, since the phase angle is
marginally above the set point. However, during thisperiod, the voltage goes down and the reactive power
consumption is high. Therefore, for both days, the 4th
bank could be switched on at around 10.00 hrs. As for
those two days, the situation is generally common
through out and therefore the present switching criteriais not a suitable solution.
V OBSERVATIONS FROM MEASUREMENTS
Behaviour of the substation load is cyclic and has twodistinct peak points. There is a load peak in the early
morning hours and highest peak is in the night.
MW / Mvar Curve - Panaduara GSS (20th to 28th Jan 2009)
Measured at 33kV side
0
10
20
30
40
50
60
20.01.2009
21:30:00
21.01.2009
17:30:00
22.01.2009
13:30:00
23.01.2009
09:30:00
24.01.2009
05:30:00
25.01.2009
01:30:00
25.01.2009
21:30:00
26.01.2009
17:30:00
27.01.2009
13:30:00
28.01.2009
09:30:00
Time of the day
MW
/Mvar
MW Mvar
Day time load is considerably flat and has a drop at
lunch time and at the close of office hours. The morningand night peaks are generally due to lighting loads and
day time industrial and commercial load is naturallyinductive. This is shown in thefigure 3.
Behaviour of the system power factor at the mediumvoltage bus and high voltage bus describes the
composition of the load. During the said morning peakand night peak the PF is comparatively high.
Power factor - Panaduara GSS (20th to 28th Jan 2009)
Measured at 33kV bus and 132 kV bus
0.75
0.80
0.85
0.90
0.95
1.00
20.01.2009
21:30:00
21.01.2009
17:30:00
22.01.2009
13:30:00
23.01.2009
09:30:00
24.01.2009
05:30:00
25.01.2009
01:30:00
25.01.2009
21:30:00
26.01.2009
17:30:00
27.01.2009
13:30:00
28.01.2009
09:30:00
Time of the day
Power
Factor
PF at 33 bus PF at 132 bus
Power factor - Panaduara GSS (21st Jan 2009)
Measured at 33kV bus and 132 kV bus
0.75
0.80
0.85
0.90
0.95
1.00
0.00 1.40 3.20 5.00 6.40 8.20 10.00 11.40 13.20 15.00 16.40 18.20 20.00 21.40 23.20
Time
PowerFactor
PF at 33 bus PF at 132 bus
During day time the power factor is low due to highly
inductive industrial load. However it takes a somewhatflat profile showing that both real and reactive loads
increase in same proportion. Figures 4a and 4b illustrate
this pattern.
The intention of CEB in using the capacitor banks is a
key factor in the analysis. It is quite clear that CEBs
intention is to give a voltage support at the mediumvoltage bus which drops due to increasing load.Dropping the system voltage at the load centres is a
critical problem especially in locations where there is no
Fig.2b HV side voltage and reactive power on 24.01.09Fig.2a HV side voltage and reactive power on 22.01.09
Fig.3 Daily load pattern
Fig.4a Power factor over the full measurement period
Fig.4b Power factor in a days window
-
8/3/2019 Optimizing the Use of Breaker Switched Capacitors in Ceb Power Grids
4/11
close by generation for reactive power compensation.Release of sub station capacities and reduction of losses
are secondary expectations in CEBs point of view
although they too are very important. As said earlier,
voltage decreases due to large resistive loads at nightand morning peaks and due to heavy industrial andcommercial loads at day time. Voltage improves in mid
night till early morning with decreasing loads butconsiderable base reactive load exists through out.
Comparison of voltage at high voltage bus and powerfactor measured at medium voltage bus is shown in
figure 5a.
Comparison of 132kV Voltage & Phase angle measured at 33 bus
-45.00
-40.00
-35.00
-30.00
-25.00
-20.00
-15.00
-10.00
-5.00
0.00
5.00
10.00
15.00
20.00
21.01.2009
00:00:00
21.01.2009
20:00:00
22.01.2009
16:00:00
23.01.2009
12:00:00
24.01.2009
08:00:00
25.01.2009
04:00:00
26.01.2009
00:00:00
26.01.2009
20:00:00
27.01.2009
16:00:00
Time of Day
Phaseangle
65.00
67.00
69.00
71.00
73.00
75.00
77.00
79.00
Voltage(kV)
PF a t 33 b us ( no Ca ps ) 13 2k V Vo ltag e w ith ou t c ap s No min al 13 2 Vo ltag e
Comparison of 132kV Voltage & Phase angle measured at 33 bus on 21st J an
2009
-45.00
-40.00
-35.00
-30.00
-25.00
-20.00
-15.00
-10.00
-5.00
0.00
5.00
10.00
15.00
20.00
0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00 20.00 22.00
Time of Day
Phaseangle
65.00
67.00
69.00
71.00
73.00
75.00
77.00
79.00
Voltage(kV)
P F a t 33 bu s (no Ca ps ) 1 32 kV Vo ltage w itho ut c ap s N om ina l1 32 V ol ta ge
Power factor goes high during night and morning peaks
causing tendency to switch off the capacitor banks but
bus voltage goes down. Due to this, we may deliberatelyignore a possibility of improving bus voltage due to
gradual disconnecting of capacitor banks during nightpeak or delay in picking up the banks in the morning. Inother words, either it is possible to keep some capacitor
banks for extended time or some banks can be
connected bit earlier. Therefore, voltage and power
factor behaves contradictorily.
In the other case, again the voltage decreases during daytime with high inductive loads and power factor
becomes low. It comes to an approximate flat profilelater. The figures 5a and 5b show this clearly. This
means that the increase of real and reactive power is insame proportion. The point that has to be considered is
that if the first come banks correct the power factor thenthe others will not come even if there is a possibility of
compensating more reactive power or increasing the bus
voltage. As explained earlier, this is clearly observed in
figures 1b, 1c, 2a and 2b. The possibility of stepping tothe 4
thbank is still there at around 9.00 hrs on both days
despite the phase angle is just above the setting value.
Analysis of measured data shows a considerableuncompensated reactive power with the present
switching criteria. It is further explained in figures 6aand 6b for two other days with in measurement period
and indicates unsuitability of the present switching
criteria.Uncompensated Var (26.01.09)
-10.00
-5.00
0.00
5.00
10.00
15.00
26.01.2009
00:00:00
26.01.2009
03:20:00
26.01.2009
06:40:00
26.01.2009
10:00:00
26.01.2009
13:20:00
26.01.2009
16:40:00
26.01.2009
20:00:00
26.01.2009
23:20:00
Time of Day
Mvar
1
2
3
4
5
6
7
8
9
10
NoofCapBanks
Un served V ar No of Ca p Ba nk s N o o f Ma x Ca p Ba nk s
Uncompensated Var (27.01.09)
-10.00
-5.00
0.00
5.00
10.00
15.00
27.01.2009
00:00:00
27.01.2009
03:20:00
27.01.2009
06:40:00
27.01.2009
10:00:00
27.01.2009
13:20:00
27.01.2009
16:40:00
27.01.2009
20:00:00
27.01.2009
23:20:00
Time of Day
Mvar
1
2
3
4
5
6
7
8
9
10
NoofCapBanks
Uns erve d Va r N o o f C ap Ba nk s No of Max Ca p Ba nk s
On Load Tap Changer (OLTC) and the AVR in such
scale utility substations is also available to adjust the LV
bus voltage. Use of capacitor banks in voltage support isthe most economical since it reduces apparent power
drawn from the system hence reducing losses. Tapchanger improves the voltage by changing the tapposition and reduces only small amount of reactive
power and overall effect on reactive power due to
increased voltage is an increase. Therefore, full
utilization of already installed capacitor banks is better
in adjusting the bus voltage than adjusting the taps.Figure 7 which contain the pattern of the tap position
with no capacitor banks shows that the system operatesat higher tap positions during day time with decreasing
voltages.
Fig.5a Comparison of HV bus voltage and PF
Fig.5b Comparison of HV bus voltage and PF in a days
window
Fig.6a Uncompensated reactive power under present
switching criteria 24.01.09
Fig.6b Uncompensated reactive power under present
switching criteria 27.01.09
-
8/3/2019 Optimizing the Use of Breaker Switched Capacitors in Ceb Power Grids
5/11
132kV voltage pattern, tap position & no of cap banks (25th to 27th)
65
67
69
71
73
75
77
79
81
83
85
25.01.2009
00:00:00
25.01.2009
08:20:00
25.01.2009
16:40:00
26.01.2009
01:00:00
26.01.2009
09:20:00
26.01.2009
17:40:00
27.01.2009
02:00:00
27.01.2009
10:20:00
27.01.2009
18:40:00
Time of Day
132kVbusvoltage(ph
_E)kV
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
N
oofCapBanks/Tapposition
132kV bus voltage Nominal 132kV voltageMax. Continuous 132 voltage No of Cap BanksNo of Max Cap Banks Tap position
The voltage rise obtained by raising one tap position up,
is 1.5 % of the voltage at the point of measuring. This isas per the specifications of the OLTC. At 33kV voltage
this rise is about 0.495 kV. The approximatedpercentage voltage rise given by switching one 5Mvar
capacitor bank is given as (kvar / kva) * Xt Where
kvar = addition of reactive load, kva = transformerrating and Xt = transformer reactance in % [3].
When two transformers are in parallel, this value
becomes 0.79% and the voltage rise is 0.260kV at33kV. As these figures suggests, the effect of rise in one
tap step is same as adding two 5Mvar capacitor banks
when two transformers are paralleled or one 5Mvarbanks when one transformer is connected. Considering
the above, the system could be operated at least withtwo taps below if the capacitor banks are connected.
The table1, an abstract of the results from network
simulation shows that if start from point (A) with LVbus voltage 31.98kV, the tap position changes from 10
to 12 until it adjust the bus voltage to 32.98kV (PointB). Switching of 3 capacitor banks of 5Mvar can keepthe bus voltage 32.87kV while retaining at the same tap
(point C) but reduce the HV side current in one
transformer by about 14A.
VI SYSTEM MODELLING FOR SIMULATIONS
Following effects due to the switching of capacitor
banks to the system was studied by modelling thenetwork with PSCAD which is widely used simulationsoftware for network simulations [4].
Maximum voltage rise due addition of capacitorbanks at the bus bar
The capability of transformer OLTC and AVR tohandle those voltage variations by changing tapposition, when necessary.
The capability of OLTC to handle the currentthrough it without exceeding its current switching
capacity during back feeding reactive power into
the system
The effect of resonance when adding morecapacitor banks under various load conditions and
system harmonic levels
Effects on voltage distortion caused by loadharmonics at 33kV bus, when adding more
capacitor banks
Cost analysis considering the reduction of losses
due to power factor improvement, release of systemcomponent capacities etc. and many others.
VII VOLTAGE RISE
The substation model was adjusted to have same
measurement condition as measured without capacitorbanks. The changes in parameters when simulating the
switching of the capacitor banks as per present criteria,for maximum var compensation and for maximum
capacitor banks were recorded next. The maximum
voltage rise which may occur at maximum source
voltage and minimum load with all capacitors was also
simulated and recorded.
Multiple RunOutput File
All CapsLoad -17.2 MW 9.6Mvar
Run # Tap HV Volt.(kV-phE) LV Volt(kV) HV_Current (pk-A)
1 8 83.08 36.86 .105
2 7 83.07 36.31 .102
3 6 83.06 35.78 .099
4 5 83.05 35.26 .096
5 4 83.04 34.76 .094
6 3 83.03 34.28 .091
7 2 83.03 33.80 .089
8 1 83.02 33.34 .086
Multiple RunOutput File
No CapsLoad -17.2 MW 9.6Mvar
Run # Tap HV Volt.(kV-phE) LV Volt(kV) HV_Current (pk-A)
1 8 82.55 35.49 .102
2 7 82.56 34.97 .099
3 6 82.56 34.46 .097
4 5 82.57 33.97 .094
5 4 82.56 33.49 .091
6 3 82.58 33.03 .089
7 2 82.59 32.58 .086
8 1 82.59 32.14 .084
Under such a worst case, LV bus voltage with no
capacitor banks is 33.494 kV at tap position to 4. Whenall banks are connected at this stage, AVR & tap
changer is capable to maintain the bus voltage at
Run # Tap HV LV Ph TF_HV_Position Voltage Volt age Ang_LV LV_MW LV_MVar Current (pk)
1 13 74.78 33.50 -23.51 47.12 20.50 .17
2 12 74.79 32.98 -23.51 45.66 19.87 .16-------(B)
3 11 74.80 32.47 -23.51 44.27 19.26 .16
4 10 74.82 31.98 -23.51 42.95 18.68 .15 ------(A)
5 9 74.83 31.50 -23.51 41.68 18.13 .15
Multiple Run Output File 3 cap banks
Run # Tap HV LV Ph TF_HV_Position Voltage Volt age Ang_LV LV_MW LV_MVar Current (pk)
1 1 3 75.17 34.45 -6.17 49.83 5.38 .16
2 1 2 75.17 33.91 -6.17 48.28 5.22 .15
3 1 1 75.17 33.38 -6.17 46.80 5.06 .15
4 1 0 75.18 32.87 -6.17 45.38 4.90 .14-----( C)
5 9 75.18 32.38 -6.17 44.03 4.76 .14
Table 1 An abstract of the results from network
Fig.7 Behaviour of tap position with no capacitor banks
Table 2 Results from simulations for worst case analysis
-
8/3/2019 Optimizing the Use of Breaker Switched Capacitors in Ceb Power Grids
6/11
33.344kV at tap changes from 4 to 1. Practically this isnot a desired condition but such a worst case will not be
allowed by the system operator. Table 2 shows the
results from the simulations.
With present configuration, the maximum effectivereactive power injection through a transformer when
either transformers in parallel, or transformers areindependent, is 10Mvar (since each transformer is
connected with two banks). Addition of 20 Mvar gives arise of about 1.04kV at 33kV bus voltage. Rise of 1 tap
position changes the voltage by 0.015 pu and this is
about 0.495kV at 33kV and therefore the effect of risein voltage over the nominal value due to addition of
maximum capacitor banks can be handled with two tappositions.
The simulation results were studied on voltage at theHV bus with the no banks, maximum banks, when the
banks are switched to give optimum var compensationand when banks are switched with the present scheme.
As per the above results, switching the maximumcapacitor banks under any real time condition is
obviously possible as far as the voltage rise at bus bar isconcerned.
Real time data measurement of 132 kV voltages with all
4 capacitor banks in ON condition was done and
compared with the simulation results.
Variation of tap position - 21st & 22nd January 2009
0
2
4
6
8
10
12
14
16
18
21.01.200900:00:00
21.01.200909:00:00
21.01.200918:00:00
22.01.200903:00:00
22.01.200912:00:00
22.01.200921:00:00
24.01.200906:00:00
24.01.200915:00:00
Time of day
Numberoftaps
Tap - actual measurement with no cap banks Tap - Simulated results - under present scheme
Tap - Simulated results with maximum cap banks
VIII VOLTAGE CONTROL BY OLTC & AVR
Behaviour of tap changer position in response to rise of
voltages beyond the nominal values due to capacitorbanks were simulated and Figure 9a shows the
comparison. The figure 9b shows the tap positionrecorded in real time monitoring with all capacitorbanks connected. Tap position behaves within theacceptable range.The current through the OLTC with
maximum capacitor banks does not exceed rated current
under any condition and therefore, it is not a decisive
factor.
IX RESONNANCE AND VOLTAGE DISTORTION
The effects of resonance due to switching on thecapacitor banks were studied using the same sub station
Tap - real measurements 18th Feb to 21st Feb 2009
4
5
6
7
8
9
10
11
12
13
14
15
16
18.02.2009
12:00:00
18.02.2009
19:30:00
19.02.2009
03:00:00
19.02.2009
10:30:00
19.02.2009
18:00:00
20.02.2009
01:30:00
20.02.2009
09:00:00
20.02.2009
16:30:00
21.02.2009
00:00:00
21.02.2009
07:30:00
Time of day
Tapposition
Tap - real measurements
model with slight modifications to add harmonic currentsource, distortion level measurement and frequency scan
module blocks. The simulations were done for differentload combinations and for different substation
configurations as well. The simulation results are showninfigures 10a to 10e. Followings are the observations
With only filter banks, two resonance pointsare observed, one with highest impedance andone with lowest.
When filter banks are mixed with normalbanks, one additional high impedance point is
observed where the value is higher than thefirst one.
Minimum resonance point (close to 250Hz /tuned to 5
thharmonic) and the first high
impedance point(190Hz-inter harmonic or
close to 4th
harmonic) is same under all
conditions. Therefore impacts are negligible. The second high impedance point varies withthe configuration.
Higher the load lower the magnitude of theimpedances
For only normal banks, the minimumresonance point not seen (detuned banks)
The effect of voltage harmonic distortionmainly due to second highest impedance pointis the fact to be considered
Resonance Characteristics (Single TF/1 filter bank - for different loads)
0
10
20
30
40
50
60
70
80
90
100
110
120
130
50 150 250 350 450 550 650 750 850 950 1050 1150 1250
Frequency (Hz)
Impedenace(ohm)
M ini mu m t f lo ad Max da y lo ad sh are fo r1 tf M ax nig ht pe ak sh ar e f or 1 t f Ma x t f l oa d
Fig.9a Tap position variation to give constant LV busvoltage (simulated results)
Fig.9b Tap position variation with all cap banks (actualmeasurements)
Fig.10a Frequency scan single t/f & one filter bank
-
8/3/2019 Optimizing the Use of Breaker Switched Capacitors in Ceb Power Grids
7/11
Resonance Characteristics (Single TF/2 CAP banks - for different loads)
0
10
20
30
40
50
60
70
80
90
100
110
120
130
50 150 250 350 450 550 650 750 850 950 1050 1150 1250
Frequency (Hz)
Impedenace(ohm)
M in imum tf lo ad Ma x da y l oa d s ha re fo r 1 t f M ax nig ht pe ak sh ar e f or 1 t f Ma x tf lo ad
Resonance Characteristics (2 TF/1 filter bank - for different loads)
0
10
20
30
40
50
60
70
80
90
100
110
120
130
50 150 250 350 450 550 650 750 850 950 1050 1150 1250
Frequency (Hz)
Impedenace(ohm)
Minimum tf load Max day loadf Max night peak Max tf load
Resonance Characteristics (2 TF/2 filter bank - for different loads)
0
10
20
30
40
50
60
70
80
90
100
110
120
130
50 150 250 350 450 550 650 750 850 950 1050 1150 1250
Frequency (Hz)
Impedenace(ohm)
Minimum tf load Max day loadf Max night peak Max tf load
Resonance Characteristics (2 TF/3 filter bank - for different loads)
0
10
20
30
40
50
60
70
80
90
100
110
120
130
50 150 250 350 450 550 650 750 850 950 1050 1150 1250
Frequency (Hz)
Impedenace(ohm)
Minimum tf load Max day loadf Max night peak Max tf load
Resonance Characteristics (2 TF/4 filter bank - for different loads)
0
10
20
30
40
50
60
70
80
90
100
110
120
130
50 150 250 350 450 550 650 750 850 950 1050 1150 1250
Frequency (Hz)
Impedenace(ohm)
Mini mum tf load Max day loadf Max night pe ak Max tf load
Under all configurations, the voltage distortion at 33kV
bus level is below 7.2% level (6.5% is the planningvalue as per IEC 61000-3-6 and 8% tolerance value as
per EN 50160) [5] . High distortion is resulted when all
capacitor banks are connected. Therefore the impact toallowable voltage distortion levels by maximum use of
capacitor banks is under acceptable levels.
X NEW SWITCHING CRITERIA
Utilization of the capacitor banks under present criteria,calculated on daily average and with reference tomaximum utilization is about 75 % for power factor /
var control switching criteria and 70.03 % for purepower factor control. (Utilization = (Mvar1*t1 + Mvar2*t2 + ---------+ Mvarn*tn) / Maximum Mvar * 24 where Mvarn =
switched capacitor rating at time slot tn and tn is taken as 10min interval.
Although these values are comparatively high, it does
not indicate the optimality of the use. The present
scheme contains unnecessary utilization at certain time
periods and periods of partial utilization of capacitor
banks even the opportunity is there to fully use them. Inreal situation, sometimes the network operators
manually switch off the banks to avoid high leadingpower factor and bus voltage rises or switch on the
banks which are already in off position due to improved
power factor. Therefore, the high utilization factor is not
the mere deciding factor for the optimal usage. Loss
minimization, voltage support, releasing capacityconstraints etc., are the factors to be considered.
As discussed in the previous chapters, the possibility of
connecting the maximum number of capacitor banksinto the LV bus under any system conditions is obvious.The analysis shows that the harmful effects can be
maintained with marginally affecting the regulationsand not violating the technical limitations. Therefore,
following conclusions can be made.
For the selected substation, it is possible to connectall four capacitor banks under any systemcondition.
Therefore, any other combinational arrangement, tosuit the local requirements is also possible.
Fig.10b Frequency scan single t/f & 1filter & 1 normalbank
Fig.10c Frequency scan 2 t/f & 1filter bank
Fig.10e Frequency scan 2 t/f & 2filter 1normal banksFig.10d Frequency scan 2 t/f & 2filter banks
Fig.10fFrequency scan 2 t/f & 2filter 2normal banks
-
8/3/2019 Optimizing the Use of Breaker Switched Capacitors in Ceb Power Grids
8/11
The first point can be considered in the system point ofview when capacitor banks in substations are kept
unused while transmission system needs reactive power
for other locations. This happens mostly when power
across the companys transmission system does notcoincide with load conditions in locations where thecapacitor banks are fixed. This situation can be mostly
experienced in substations which are heavilyinterconnected. Under those conditions, keeping a
definite economical reactive energy source underutilizedor unutilized depending on local requirements, while
generation or some other means producing and
transmitting them in the system, is not justifiable. CEBhas to take advantages of ON Demand Control to use
the already installed capacitor banks in this manner. Ifthe transmission system needs var as explained and if a
centralized network control center monitor the load flow
in its transmission system, then switching of unusedcapacitor banks at such a time can be used to inject
reactive power. This needs a comprehensive load flowstudy, fully pledged SCADA system and sometimes
remote station control facility etc., to implement theabove schemes. Interestingly, those are already in touch
with the CEB transmission network. Therefore, ifnecessary CEB can use its maximum installed capacitorbanks without any difficulty.
Second option is to meet local requirements in each
substation. As explained voltage or var control or acombination of both seems to be better compared topower factor. It is always an indirect measure of
reactive power. And also it has no concerns over theeffects beyond the substation, such as voltage rises due
to predominant line capacitance during very light loaded
conditions. In such cases, considerable lagging reactive
load at load centres is beneficial. If the substationreactive power requirement is fully compensated during
these periods, the voltage rise at receiving ends will be a
problem. In such cases capacitor bank switching basedon voltage control may have more benefits.
XI SWITCHING CRITERIA BASED ON
REACTIVE POWER
Switching based on reactive power requirements is amore flexible and natural means of capacitor control
concepts. It adds a fixed amount of lagging reactive
power into the system regardless of most otherconditions. Losses and the capacity release are directly
proportional to the apparent power. Injecting the
reactive power reduces apparent power andconsequently losses.
In var control based switching, due consideration has tobe given to avoid hunting or PUMPING of the banks.
Hysteresis or restraint control is suggested to avoid sucha problem. In general, switching ON based on about2/3 of a step and switching off based on slightly beyond
1/3 of the step in leading direction. To avoid respondingto sudden reactive power changes, restraint control or
integration of inputs over certain time period can be
used. These are available in most of the capacitor bank
controllers.
Considering the above basis, parameters for reactive
power control switching for master slave control modewas suggested as follows. The minimum step setting
depends on controller, based on master slave mode or
independent mode. Considering the results obtained by
simulations, following points can be considered in a
reactive power control based switching criteria for CEB.
When transformers are paralleled, one controllerfeels only a half of the capacity of a switched bank.
Step size of a bank is 5Mvar.
Switching ON when lagging reactive powerexceeds 2.5 *2/3 =1.6Mvar (lag)
Switching OFF when leading reactive power
exceeds (2.5 *1/3) *1.4 1.2Mvar(lead)
Switching points were selected from simulation resultswith approximated AVR control and shown infigure 11
The switching points based on lowest reactive power
drawn from system and power factor close to unity(optimum compared to losses) was also show in the
diagram.
Comparison of var control Vs present scheme21st &22nd January 2009
0
1
2
3
4
5
21.01.200900:00:00
21.01.200906:00:00
21.01.200912:00:00
21.01.200918:00:00
22.01.200900:00:00
22.01.200906:00:00
22.01.200912:00:00
22.01.200918:00:00
Noofbanks
5
10
15
20
25
30
Time of day
Mvar
Present scheme Optimum based minimum loss proposed Var contro l Mvar with no capacitors
The figure shows that the proposed switching policy
based on reactive power control goes neck to neck withthe loss optimized switching pattern than the present
switching criteria (Blue and red curves). No of
switching operations per day is on the safe side. Atypical capacitor bank switch can operate 6times per day
considering 50,000 no of operations and 20 years lifetime. The table 4 shows the results.
The utilization factor is 80% and better than the presentsystem and also closer to the theoretical loss optimized
Number of switching
Date Bank 1 Bank 2 Bank 3 Bank 4
21.01.09 0 0 2 2
22.01.09 0 0 2 224.01.09 0 0 1 2
Fig.11 Switching pattern with var control for 21st & 22n Jan 2009
Table 4. Noof switching operations on selected days
-
8/3/2019 Optimizing the Use of Breaker Switched Capacitors in Ceb Power Grids
9/11
pattern. Increase or decrease of energy loss compared topresent switching criteria was calculated based on the
point that the losses are directly proportional to I2. Three
days were considered and a decrease of 1.8%, 4.9% and
5.04% was observed with an average reduction of3.94% (Considering only the transformer losses). Thecapacity release of the substation was calculated as
below. The same capacity will be released from the
generation as well [3].
Where KVAs - release of substation
KVAs - Capacity of substation
KVAR - Capacity of next step of the banks
Cos and Sin- Cos and sine of power factor before adding
next step
For the selected substation, addition of 5Mvar for 2 *31.5MVA transformers at the conditions as at 8.30hrs
on 24th
January 2009, the capacity release KVAs wascalculated as,
MVAs = [ {1-(5*Cos7.13/63)2}+Sin7.13 * (5 / 63) - 1]63
= 0.425 MVA
With the simulation results as in table 5 for same timeslot on the selected day, it can be calculated as;
MVAs =(33.012
+4.072) - (32.62742+.896852)
= 0.620 MVABut this is with a tap position change as well. Thereforethe simulation results can be justified. Considering the
simulation results, total average energy released byswitching from present scheme to proposed var controlscheme is 15.64 MWh per day (calculated based on
30min sample time). The scheme maintains the tap close
to nominal tap while keeping the 33 kV voltages alsowithin the range.
Under Present criteriaDate
&Time
MW Mvar33
Volt132Volt
No ofBank
s
Phangl
e
Utilization
HV A Tap
24.01
. 09
08:30
:00
33.01 4.07 32.98 74.97 3 -7.13 7.50 76.08 10
Proposed var control scheme
Date &Time
MWMva
r33
Volt132Volt
Noof
Banks
Phangl
e
Utilizatio
n
HVA
Tap
24.01. 09
08:30:0032.62 -0.89
32.7
8
75.0
94 1.57
10.0
0
73.7
69
XII SWITCHING CRITERIA BASED ON VOLTAGE
CONTROL
The difficulty in voltage control based switching is thatit has to coordinate with the voltage regulator of the
power transformers since the latter also tries to control
the voltage. Coordination is required in such a case and
following factors have to be considered.
During switching on for decreasing bus barvoltages, the capacitors shall come first if the
reactive power load is more than a specified
percentage of the minimum step of a bankotherwise the tap changer can increase the voltage.
The purpose of this is to minimize the losses by
maintaining the power factor close as much as
possible to unity.
During switching off for increasing terminalvoltages, reactive power at the time of decision
must be considered and the algorithm shall decidethe most economical step, whether to reduce the tap
or to switch off a capacitor bank.
The purpose of above two conditions is to maintainthe bus voltage while reducing the losses to
minimum. If the only requirement is to control the
voltage, then proper dead band selection for two
controllers also can serve the purpose. Differentiate the integration time, the time period
over which the measurement is averaged, also can
be used with hysteresis control to make the control
philosophy more simple.
A voltage selection scheme based on a hysteresis
control as in the figure12 is evaluated for comparisonwith the present and proposed var control schemes.
The approximated switching points of capacitor banks
based on above voltage control scheme, was selected
using the simulation results for three selected days andfigure 13 shows the comparison.
Comparioson present scheme Vs voltage control
21st & 22nd Jan 2009
0
1
2
3
4
5
21.01.200900:00:00
21.01.200906:00:00
21.01.200912:00:00
21.01.200918:00:00
22.01.200900:00:00
22.01.200906:00:00
22.01.200912:00:00
22.01.200918:00:00
Time
No
ofbanks
70
71
72
73
74
75
76
77
78
79
80
HVv
oltage
(Ph-E)kV
Present scheme Proposed voltage control
Optimum switching based on minimum loss 132kV Voltage with no caps
Table 5. Simulation results 8.30hrs 24t
Jan 2009
32.50 kV 32.75kV 33.00kV 33.50kV 33.75kV
Nominal
Voltage
Cap bank
ON
Cap bank
OFF
AVR tap
lower
AVR tap
raise
Fig.12 Switching points for cap bank controller andAVR
Fig.13 voltage control switching compared with present
& loss optimized criteria 21st & 22nd January
-
8/3/2019 Optimizing the Use of Breaker Switched Capacitors in Ceb Power Grids
10/11
The criterion does not maximize the utilization. Gradualswitching off of banks around 17.00 to 18.00 hrs also
observed due to reduction of loads after office hours.
There is a voltage rise during this period and the load
rises after that due to lighting. Voltage control schemedoes not coincide with the curve with minimum loss,like in reactive power control.
As the data shows, following conclusions can be made.
Maximum switching operations per 3rd and 4thbanks is about 4 so that the switching does not
cause any unnecessary impact.
Utilization when voltage control is used seems to below compared to loss optimized switching pattern.
It is 55%, 58% for the two selected days.
Due to reduced utilization, energy losses andsubstation capacity release also not be economical.However it matches with the voltage properly.
Therefore, if the need is to give voltage support,
then this kind of switching policy is very
satisfactory.
XIII A PRACTICAL APPROACH
In serving both voltage support and loss reduction
purposes, combining the two concepts, voltage controland var control will give more benefits. However, theAVR is to be inoperative while the capacitor controller
is in action but redundancy to be maintained by AVRwhen the capacitor controller is faulty, after banks are
fully utilized, during a tripping of capacitor banks or thevoltages are beyond certain extreme ends. It is learnt
from the analysis that if reactive power control is used
aiming to manually or automatically switch off the
banks at specified time intervals (from around 22.30hrsto 7.00hrs) to avoid high bus voltages during low load
conditions, the above combinational effect isachievable. . The comparison is shown infigures 14.
Voltage control Vs var control with Manual OFF at night - 21st &22nd Jan
0
1
2
3
4
5
21.01.2009
00:00:00
21.01.2009
06:00:00
21.01.2009
12:00:00
21.01.2009
18:00:00
22.01.2009
00:00:00
22.01.2009
06:00:00
22.01.2009
12:00:00
22.01.2009
18:00:00
Time of day
Noofbanks
proposed Var control from 7.00 to 22.30hrs Prposed Voltage control scheme
As we see from these figures, if reactive powercontrolled switching can be used as above, it is similar
to the voltage control scheme but less complex. The
disadvantage is the functionality of such a manual automixed control. However, if both voltage and reactive
power combined controller having multiple variable or
Boolean switching controllers can be used to switch the
banks considering voltage and var, it could be a bettersolution.
XIV ANALYSIS AND RESULTS
i. Using capacitor banks at 33kV sub distributionlevel to compensate reactive power requirement
and therein, to maintain voltage stability at same
level is economical and effective in the CEBsystem.
ii. Occasions where the capacitor banks are
switched ON and OFF manually by over-riding
the auto controller was frequently observed. Thissays that the switching criteria are not fully fit to
the requirements in CEB system. The
observations also show that present switchingcriteria at the selected substation neither
maximize nor optimize the utilization.iii. The study proves the technical feasibility of
maximum capacitor bank connections to thepoint at which they are fixed without violatingvoltage rise due to reactive power injection,
effects to voltage distortion and resonance due toharmonics with additional capacitor banks,
switching capabilities of the on-load tap changer
and the capabilities of AVR to handle voltagevariations due to reactive power injection.
iv. The maximum voltage rise under different
capacitor bank combinations (with effective Tap
control) for 21st, 22
nd& 24
thare 77.57kV, 77.8kV
& 77.17kV respectively. The maximum
percentage rise for high voltage side is .33% andthat for low voltage side is 0.95%.
v.
Effects due to resonance for the selectedsubstation are negligible.
vi. Voltage distortion levels remains marginally
below 8% hence acceptablevii. Local voltage variation due to added reactive
power can be handled by the AVR and tap
changer controls so that any combination ofbanks is feasible to connect.
viii. The current through the tap changer does notexceed its switching capacity.
ix. Reactive power controlled based switching is avery much economical method of capacitor bank
controlling as far as the utilization, loss reduction
and capacity release is concerned. Only problema utility may face is that, some times especially in
light load conditions with long transmissionlines, there may be a necessity to have some
reactive power to reduce the Ferranti effects. In
such cases, minimizing reactive power
consumption is not desired.
x. In practice, for a utility like CEB where most ofthe generation is concentrated to certain areas,maintaining bus voltages may be difficult and be
important than reducing losses using capacitorbanks. In such a, voltage control based capacitor
switching will be a good solution.
Fig.14 Voltage control Vs var control with manual off at
night. (21 and 22 Jan 09)
-
8/3/2019 Optimizing the Use of Breaker Switched Capacitors in Ceb Power Grids
11/11
XIV CONCLUSION
Considering all these factors discussed so far,
followings are the conclusions from this research study.
i. Present capacitor bank switching philosophy basedon power factor regulation does not give maximum
benefits to the CEB transmission network. Thisscheme neither maximizes nor optimises the
utilization.
ii. Considering the installed capacities and step sizes
in each substation, it is technically possible toutilize the full installed capacities in all substations
without violating the technical standards.
iii. Therefore, it is technically feasible to back feed the
excess capacitor bank capacity for reactive powercompensation in the transmission network.
iv. Use of a switching policy based on reactive power
control or voltage control is more useful as far asthe CEB system is considered. Reactive power
based switching which is simple, is useful for lossminimization and voltage based control is usefulwhen voltage stability is concerned.
v. Considering the factors discussed in 7.1 viii and ix,
for network like CEB, it is useful to consider thecontrollers with multi-parameter or Booleanswitching options. Reactive power and voltage can
be the parameters to be considered in the switchingdecisions.
XVI ACKNOWLADGEMENT
Authors wish to thank to Dr H.M. Wijekoon (CE-
planning/Distribution Region 3, CEB for his valuable
comments and Mr. L.A.S.Fernando and his staff ofOperations and Maintenance branch of CEB forfacilitation data collection and measurements.
XVII REFERENCES
[1] Kusum Shanthi K.P Benchmark the Sri Lankan
power system by power quality monitoring &
analysis Master Thesis, University of Moratuwa, 2005. Chapter 7.2 pp 51.
[2] User Manual for POCOS control reactive powercontroller and harmonic analyzer
[3] Technical paper in web [email protected]
Economics when applying shunt capacitors pp6-7[4] User Manual for PSCAD
[5] Kusum Shanthi K.P, Rangith Pererra, SarathPererra Benchmarking the Sri Lankan power
system by a power quality monitoring program2005. Section III D pp 4.