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

[email protected]

[email protected]

[email protected]

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


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


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


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


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


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


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


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)



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)



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


8/11

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


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


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)


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.

Optimizing the Use of Breaker Switched Capacitors in Ceb Power Grids

Documents

Transcript of Optimizing the Use of Breaker Switched Capacitors in Ceb Power Grids