00:02

we continue our discussion of beneficial effects of shunt capacitive compensation. In this lecture, we
look at the effectiveness of shunt compensation for voltage support at transmission, sub-transmission
and distribution levels. Shunt capacitive compensation is used for controlling the voltage of a bus.

00:32

at these three transmission levels. I mean at the transmission level, at the sub-transmission and the
distribution.

00:45

The application context is shown in this diagram. The target point of common coupling, PCC, or the bus
at which the SVC is going to be applied is shown here. Now we keep this bus alone and, of course, the
loads of that bus, the local load of that bus, which is PL plus JQL.

01:14

If QL is positive, then the local load is demanding lagging reactive power. PL is positive, local load is
demanding positive active power, or the local load is consuming energy. If PL is negative, local load is
delivering energy. If QL is negative, local load is demanding leading reactive power, or delivering lagging
reactive power.

01:42

We keep only that bus, the target bus.

01:47

Svc is applied at that bus in order to control the bus voltage in some manner. May be to keep it constant
or may be to make it vary as per a regulation rule. Rest of the power system is abstracted by doing a
Thevenin's equivalent. Rest of the power system contains all the generating stations, all the
transformers, all the transmission lines.

02:13
all the local loads at other substations etc. The entire power network is reduced to a Thevenin's
equivalent of the power system at the SVC connection bus. For simplicity, we ignore the loss producing
components in the transmission lines and transformers and generators etc. So we use a simple
equivalent circuit which takes care of only the reactive portion.

02:41

the resistance portion is usually negligible, but it is present. If the resistance portion is not present, then
the power system would have been 100% efficient. There wouldn't have been any transmission loss, etc.
So when you are paying attention to the efficiency of transmission, efficiency of a power system, etc.,
you have to include all the resistive losses properly. But then for this kind of studies, a simple equivalent
circuit, the evidence equivalent circuit neglecting the resistances is enough. So...

03:10

VG angle and we take the local bus itself as the reference. So angle 0 is assigned for the local bus voltage.
It is VL angle 0. VL is the magnitude and that is the reference sine wave. And since active power is being
delivered by the source to this bus, we expect the angle of this internal Tevinens open circuit voltage to
be leading by angle delta. So the open circuit voltage is...

03:39

VG angle delta. I mean if there is nothing connected at the bus then the voltage will be VG angle delta.
Right. And...

03:53

this the the Pevenance Equivalent Impedance is Jx it is assumed to be pure reactive how do you find x?
actually the short circuit MVA capacity at the bus all the major buses in the power system will have the
short circuit MVA calculated so that will be available in the system data so from the short circuit MVA
you can calculate the value of x

04:22

it is approximately x is equal to 1 by the short circuit capacity in Pu. So the Thevenin's equivalent of
power system at SPC connection bus is assumed to be known. Now the bus is loaded by a local load Pl
plus Jql. This load is taken to be a constant power kind of load, the kind of load you use in the load flow
studies.

04:51

substation load can be modeled by a constant power load or not? that is something to be discussed my
answer would be no there is no load which is constant power load that's a no part of the answer the S
part of the answer is yes for most of the power system studies yes they can be modeled as constant
power loads

05:19

In practice, the load behavior is something between constant power load and constant impedance load.
It's neither constant power load nor constant impedance load. But as you know, it is customary to model
load as a constant power load in this kind of power system studies. Constant power load means the load
will take a stated amount of power irrespective of the bus voltage.

05:47

If the bus voltage increases, then the load will take lesser amount of current, but it will take only the
stated amount of PL and QL. If the bus voltage decreases, then the load will draw higher amount of
current, thereby taking the same amount of power. That's the nature of a constant power load. We
assume that the load connected at the bus is constant power load kind.

06:16

we plan to put a switch on an SVC there.

06:21

we plan to put Svc of a certain rating.

06:27

the rating of SVC, all quantities are in PU. Not only in this discussion, but in all discussions in this course,
we deal with per unit quantities. So the SVC rating is given as a number QC. QC. That is a...
06:52

QCPU That's the rating of the equipment

06:59

rating of an equipment is usually given assuming nominal voltage across it and what is nominal voltage in
a power system 1 pu. So if the voltage is 1 pu then the equipment in full capacity will deliver QCPU of
reactive power or take QCPU of reactive power depends upon how you term it. If it is capacitive
compensator then it will take

07:28

QCPU of leading reactive power from 1PU voltage. That's the meaning of rating. So the power rating
usually assumes that the voltage across the equipment is nominal or 1PU.

07:48

Now this equipment when connected to the bus will draw a certain capacitive current IC which will be
leading by pi by 2. This bus is at angle 0. Therefore a capacitor will take current at pi by 2. The value of
current is designated by IC.

08:13

Since the rating is given as QCPU and the rating assumes that the voltage is 1PU, IC taken at full rated
condition must obviously be numerically equal to QC. So the current in PU, IC, is numerically same as QC,
provided the voltage is 1PU. Okay.

08:43

So the current taken when the voltage is 1 pu, that current is called IC. And that current is numerically
equal to QC, because QC is the reactive power that will be taken when the voltage is 1 pu. So
numerically, IC and QC are the same. Meaning-wise, QC is reactive power. IC is reactive current. OK. Now,
PL plus JQL is the local load.
09:12

and it is modeled as constant power load. Therefore, active component of current here, I mean I L, must
be

09:26

the active component of power divided by magnitude of voltage because magnitude of voltage into
magnitude of active component of current is going to be the active power therefore PL by VL in per unit
in per unit three phase single phase confusions etc will not come so PL in PU by VL in PU will give you the
active current in PU

09:55

active current. Similarly, the reactive current, reactive current in Pu will be given by QL by VL. Now it's a
question of science. Is it minus J or plus J? Well, you know that positive QL means it's a lagging reactive
power and lagging current will be minus J. So it has to be put as minus J.

10:24

if you don't like that explanation then you remember that complex power is given by V i conjugate. In
other words

10:36

This number PL plus JQL, which is the complex power, must be equal to VL angle zero and IL phasor
conjugate. Now you conjugate the entire thing. Conjugate the entire thing. Then this will become

11:05

minus and this conjugate will go and VL angle 0 has to be conjugated but it is the same. Therefore IL
phasor is going to be PL minus JQL by VL. So PL by VL minus JQL by VL. So

11:30

I L A is this number and I L R is this number if I write with a negative sign. Okay, so current is lagging that
is understandable. But the mathematical explanation for that is this physical explanation is positive Q L
means positive reactive power which is lagging reactive power in power system.

11:50

lagging reactive power means current must be lagging. If current is to lag it must be minus J. So, in the
current it will be minus J and the amount will be Q L by V L. That is the way another way to explain it.

12:06

So the current taken by the load is given by this. When the switch is open here, the current delivered by
the Thevenin's Equal and also will be the same. So the current which is flowing here is this.

12:25

But with the switch open, the load bus voltage has a magnitude of VL. When we close the switch, the
SPC will take leading reactive current. That leading reactive current will partially cancel this lagging
reactive current. And therefore, the current here will be different from what is written here. Of course,
current here will be same, same as before.

12:54

but this negative component will be cancelled by partially by this positive component. Partially, fully or
overcompensation depends upon what is the magnitude of IC, but usually it is partial. So here also, then
some negative component will flow in the reactive part, but the magnitude of that negative component
will come down significantly by this amount. So the reactive component of current flowing in the
impedance Jx will go down.

13:24

As a result the drop in Jx will go down and as a result the terminal voltage VL will become a new voltage
VL prime. Of course the angle delta will also undergo a small change but the change in delta will be small
because we understand power decoupling principle.

13:46
by power decoupling principle active power is decided mostly by delta reactive power is decided by the
magnitude difference we are manipulating only the reactive power and therefore we hope we will
manipulate only the magnitude voltage difference and since this magnitude is constant we expect this
magnitude will respond but this power decoupling

14:16

Active power is to a small extent governed by voltage differences and reactive power is to a small extent
decided by angle difference.

14:26

due to that small influence or due to the small error in power decoupling principle we expect delta will
also change a little but by a small quantity only when we close the switch whereas we expect this voltage
to change meaningfully when we close the switch because we are manipulating the reactive flow, right.

14:49

So the voltage will take up a new value VL prime and the angle will take up a new value delta prime. We
expect that delta prime will be very close to delta and we don't look at that matter any further. If
required, we can calculate it, but that's not our focus. We want to find out what is the new voltage VL
prime.

15:11

The new voltage VL prime can be written as the old voltage VL plus some delta VL. We expect delta VL to
be positive because we expect the bus voltage to increase and improve. We want to find out what is the
amount of improvement.

15:30

How is the amount dependent upon the various quantities? Does the amount of increase depend on the
load? If it depends on the load, does it depend on the active component of the load? Or does it depend
only on the reactive component or does it depend on both?

15:53
the change in voltage or improvement in voltage does it depend on the value of this current or the
capacity Qc obviously it should now delta VL the improvement does it depend on X the answer is yes but
how does it depend on it and delta VL does it depend on VG yes it does does it depend on the initial
value VL

16:24

Well, it depends on all these. We want to develop an explanation for that. So for that, first let us find out
the pre-switching on voltage. I mean VL. When the switch is open, what is the bus voltage? Then when
the switch is closed, what is the bus voltage? And then we will find the difference, and that is
improvement, et cetera. So first let us find out VL. The voltage that will be there at the bus with the SVC
switched off.

16:52

the bus delivering a constant power load PL plus JQL through a Thevenin's impedance Jx from a source
VG angle delta. So we draw a phasor diagram for that. The phasor reference is VL angle zero and the
active component of current is PL by VL. The reactive component of current which is lagging is QL by VL
and therefore the current IL is this parallel

17:22

So this is I L. The angle between V L and I L which I haven't shown is the power factor angle experienced
by the Thevenin's equivalent. Right now this current is flowing through an impedance J X. Therefore the
voltage drop across the impedance will be

17:44

current into impedance x rotated by 90 degree in the anti-clockwise direction leading voltage. Voltage
across inductance will lead the current through the inductance. So going by components this component
of current will produce this component of voltage drop in the inductance rotate by 90 degree anti-
clockwise multiplied by x. So the reactive component of current produces

18:13

qL by VL into x. Notice that that voltage drop is in line with the terminal voltage. This is why it's called
inline drop. So who produces the inline drop? The reactive component of current produces the inline
drop. And what is produced by the active component of current?

18:38

Active component of current is this, multiply by x and rotate it up by 90 degree. That is this. That will be
PL by VLX. So that produces quadrature component of voltage drop. Now you close this triangle. This
length is the voltage drop across the impedance. So the terminal voltage plus the voltage drop across the
impedance will be the source voltage VG.

19:07

But now you remember the Pythagoras theorem. You can see that the length of the source voltage is
practically equal to the length of the terminal voltage plus this length. Actually, it is VL plus QL by VL into
x whole square plus PL by VL x whole square square root. But then, even when this length is 10

19:35

20% of this total length. Squaring, adding and taking root, you will see that the hypotenuse is nearly very
close to the base. So quadrature component is relatively insignificant or ineffective in lengthening the
hypotenuse. Especially when the perpendicular quadrature distance

20:06

So for a right angle triangle with base dominant and altitude less than 20% of base, hypotenuse is
practically equal to base within 1 or 2% error.

20:26

If you draw an arc from here, the excess distance here is practically equal to this distance because the
inclination angle is small. That's another way to look at it geometrically. We expect delta to be small,
typically 5 to 10 degrees only. So with all this reasoning, it is clear that you can forget about the vertical
component in calculating the difference between VL and VG.

20:55

the difference between VL and VG is practically equal to QL by VL into x. And therefore you can write
magnitude of source voltage which is VG is approximately equal to VL plus QLx by VL. This approximation
rests upon Pythagoras theorem.

21:21

hypotenuse is approximately base if altitude is less than 20 percent of base. And we could use the
Pythagoras theorem approximation because the reactive component of current flowing in an inductance
will produce inline voltage drop and the active component of current will produce the quadrature
voltage drop. And now it turns out that inline voltage drop is the

21:48

most significant voltage drop. That voltage drop decides the matter. That voltage drop decides the
magnitude difference between the bus voltage and the internal source voltage. And this is why in-line
voltage drop, in-line, I N L I N E, in-line voltage drop is something you should always pay attention to.

22:16

virtually forget about the quadrature voltage drop. And inline voltage drop is produced by reactive
component of current. And that is why reactive component turns out to be very, very significant in
voltage magnitude difference. And from there you reach power decoupling principle. That is why power
decoupling principle is more or less correct, right? And if you want delta,

22:44

Delta can be written as sine inverse of this height by this. Sorry, this by this. Or tan inverse of this height
by this height. So it is tan inverse of this number by this plus this. But tan inverse of delta is
approximately equal to delta. Sorry, tan delta is equal to delta for small delta in radians. Therefore, delta
in radians is given by this by this.

23:13

So delta N radians is approximately PLx by VL by VL plus QLx by VL. But then our aim was to find out the
bus voltage. So this equation has to be solved for bus voltage. Bus voltage is unknown quantity. We know
QLx and Vg. Therefore forming the quadratic, it is a quadratic equation. And solving for VL, you get this.
23:42

Up to that it is exact. But in the next step I put approximation. What is the approximation I performed? I
have written root of 1 minus 4 QLx by VG square as 1 minus 2 QLx by VG square. I am using the following
approximation.

24:09

It is one of those approximate formula that come from binomial approximation when x is very small.

24:19

compared to 1. So, root of 1 minus x is 1 minus x by 2 when x is very small compared to 1. You may verify
suppose you square 1 minus x by 2 you will get 1 plus x sorry 1 minus x plus x square by 4. x square by 4
is negligible if x is very small compared to 1. So, it will be 1 minus x. So, this is the approximation I am
using.

24:53

and therefore I write.

24:57

this square root as this and from there I get the terminal voltage VL is VG minus QLX by VG. I am taking
the addition here not subtraction. Please justify that think it out. In fact that subtraction also has some
meaning but for our purpose what we need is the positive sign.

25:27

VL is going to be VG minus QLX by VG.

25:33

the terminal voltage.

Now simplified versions of this formula are possible. Because here it involves VG and there's a by VG
here. So if VL can be, so I'm writing a note here, if VL can be assumed to be very close to 1PU, what is the
basis for that assumption? In a normal steady state running power system, all voltages, bus voltages will
be close to 1PU.

26:04

If they are not very close to 1pu, then the system is in an abnormal condition. By close to 1pu, I mean
something between 1.05 to 0.95. That's a usual situation. Therefore, for the first cut solution or first try
solution or for mental calculation, you may safely assume that all the voltages are close to 1pu. And
wherever the voltages are required in the formula, you may substitute 1pu and get a first order estimate
of the answer you are looking for.

26:33

in that case if it is very close to 1 pu then ILR is close to QL what is ILR the reactive component of current
the reactive component of current if VL is close to 1 pu it is QL itself and that current is flowing through
impedance X so it will be QLX and therefore VL plus QLX will be VG and therefore VL will be VG minus
QLX.

27:00

you don't have to divide by VG, VG is close to 1, you can take it that way. So...

27:08

See the reasoning here is the reactive component of current is the reactive power taken by the load
divided by the bus voltage. But that division by bus voltage may be omitted if the bus voltage is known to
be close to 1. Therefore, the reactive component of current is the reactive power in P u itself. Reactive
component of current will produce inline drop. That inline drop will be reactive component of current
into inductive impedance.

27:38

and therefore it is ql into x quadrature drop is pl into x but quadrature drop we are simply forgetting we
are using Pythagorean approximation therefore vl plus ql into x is vg and from there vl is vg minus xql
that will be easy to calculate for mental calculation or approximate calculation we may use that.

28:06

That was a note. You can use this kind of approximation when you solve problems. And when the answer
is sufficiently accurate, of course, VL has to be close to 1 pu. So suppose VL, finally, it is expected to be
1.03, and VG is 1.04, 0.05, et cetera. You can use this approximation. But if VL goes down to 0.85 and all,
then I mean.

28:34

Suppose you calculate it using this formula and then you see that VL is 0.85. Now you cannot use that
approximation. Now you realize that VL is low. So we cannot ignore the division by VG. So we'll have to
go for this. OK. Now with S closed, the leading reactive current drawn by the SVC will depend on
whether it is

29:02

This is the important point here. Why we are resolving into three cases. With S closed, we take case by
case. What are the three cases? The three possibilities are maybe our static work compensators are fixed
capacitor. Or maybe it is a static work compensator using variable impedance type. Or maybe it is a stat-
com. Maybe we are using a switched converter.

29:30

The behavior of the static-wire compensation unit when the bus voltage is not equal to 1 pu is different
in these three cases. At least it is different between first two cases and the third case. That is because
consider a fixed capacitor. What is the reactive power taken by a fixed capacitor dependent on? It
depends on its reactive impedance and the bus voltage.

30:01

The fixed capacitor will have a certain reactive impedance Xc which is equal to 1 by omega c and then
working from VL it will take VL by Xc.
30:16

and then the reactive power will be V L into I c. So, it will be V L squared by x c.

30:29

So a fixed impedance reactive power taken by a fixed capacitor is directly proportional to square of the
voltage. Therefore, if the voltage is low, it will take only lower amount of reactive power, not the rated
amount of reactive power. It will take only lower amount of reactive current, not the rated amount of
reactive current. That has to be accounted. But if it is a stat-com...

30:59

Inside a STATCOM what you have is a generator, inverter. The rating of inverter is decided by heating in
the inverter, not by the terminal voltage. Based on the heating, the designer has given you a certain
current rating. Of course he won't give the rating in amperes, rather he will give it as MVA or KVA. So a
VA rating is given. But that rating is based on the heating effect.

31:31

that VA rating assumes that the terminal voltage is standard voltage or nominal voltage.

31:39

So from the rating given for the STATCOM, you may divide by the nominal voltage, then you will get its
current rating. And that much current it can deliver always. Even against a short circuit, it can deliver that
current. Because current is the capability.

32:00

It's not a fixed impedance. So the inverter is capable of delivering a certain amount of amperes. If you
exceed that many amperes, inverter will overheat. If you go below that many amperes, inverter will run
cool. So the rating of inverter and such equipment will be current based because it is heating based. How
does it matter whether the bus voltage is low or high?

32:29
inverter doesn't bother, it is able to deliver its rated current. So STACCOM has to be treated as a current
source capable of delivering a certain amount of current, certain maximum amount of current
irrespective of whatever is the terminal voltage. Whereas a fixed capacitor has to be treated as an
impedance and therefore if the bus voltage is very low, capacitor can take only low current.

32:56

and therefore reactive power delivered by the capacitor will also come down. This is the major
difference between a fixed capacitor and a STATCOM. And coming to SVC, what is inside an SVC? Variable
impedance type static power generator. So SVC is basically passive element only. It's only that you are
controlling the value of the passive element. So SVC usually will have a capacitor whose value is

33:24

controlled by a parallel inductor or something like that. And therefore at the extreme, that is at the full
rated condition, SVC will be a capacitor only. Because to get the full rating you will probably switch off
the parallel inductance and keep the capacitor alone. So there also it behaves like a fixed capacitor.

33:53

or switched capacitor or variable impedance type static wire generator, all these, the current taken by
them will go down with bus voltage and go up with bus voltage. Therefore, the reactive power delivered
also will depend on the bus voltage, okay, in square manner. That is, the reactive power delivered will be
proportional to square of the bus voltage.

34:22

or the reactive power that can be taken by a fixed capacitor is proportional to square of the bus voltage.
And the maximum reactive power that can be taken by a SVC is also proportional to square of the
voltage, because SVC under maximum capacity condition is a fixed impedance only. But in the case of
STATCOM, under the maximum delivery capability, under the maximum delivering condition or taking
condition, it takes a constant current.

34:52

whatever is the best voltage. Therefore, the reactive power taken or delivered by a STATCOM under the
extreme condition or full rated condition will be proportional to terminal voltage not proportional to
square of terminal.

35:09

Okay. This is the major difference between these two kinds of units. That is why we are isolating different
cases. So in case one, I consider a fixed capacitor of value C such that it consumes QC per unit leading
reactive power from a nominal voltage of 1 pu. Remember, in power systems, capacitors are not rated in
microfarads or anything. They are rated in kilowatts or megawatts.

35:39

but that reactive power rating is a calculated one. How is it calculated? Terminal voltage is assumed to be
nominal, and then the current taken by that particular capacitance is calculated, and then terminal
voltage into capacitance gives you the reactive power rating. So QCPU leading reactive power it takes
from nominal voltage of one.

36:04

and therefore the current taken by capacitor when the bus voltage is 1 pu let's say it is ic that is
numerically equal to qc because 1 pu is the voltage qc pu is the reactive power so reactive power by
voltage will give you the current taken and therefore current taken which is ic numerically it is same as
qc in pu and therefore the reactance of the capacitance is

36:34

by Ic that is same as 1pu voltage by Qcpu of current because current is same as reactive power rating in
pu and therefore it is 1 by Qc this is xc that is what remains constant so the value of xc remains constant
therefore when the voltage is not 1pu but some other voltage then the current drawn will be that some
other voltage divided by xc so it becomes multiplication by qc.

37:03

Therefore the reactive current drawn by this fixed capacitor when the bus voltage is VL prime is QC into
VL prime That's a reactive current drawn Okay, and what is the reactive power drawn? You'll have to
multiply by one more VL prime because current into voltage Reactive power drawn will be QC into VL
prime square Reactive current drawn will be QC into VL prime
37:31

where QC is the reactive power that will be drawn if the voltage is 1 pu. And that is the rating.

37:39

So the capacitor draws this much of reactive current, leading reactive current. Therefore, the net
reactive current that flows here in X, that is, reactive current of the load and reactive current of the
capacitor put with proper polarities. And therefore, net reactive current flowing through the impedance
X with C acting. The reactive current taken by the load is...

38:07

reactive power demanded by the load remember it's a constant power load so reactive power
demanded by the load divided by bus voltage VL prime VL prime is a new bus voltage you will get when
you close the switch S so this is the reactive current drawn by the load it is leading lagging reactive
power reactive current and leading reactive current drawn by the capacitor is this much so you subtract
the remaining so the the difference

38:36

That will be the final lag reactive power. So many PU lag.

38:44

and VL' is source voltage minus the drop, inline drop. Inline drop is the lagging reactive current into the
impedance of the reactants, the inductance. So this quantity into X. So VL' is approximately VG minus,
where is this approximation coming from? Pythagorean approximation. Right, from here you form the
quadratic and solve for VL'.

39:14

B l prime is this. Once more use the approximation that root of 1 minus x is 1 minus x by 2 provided x is
small. By the way did I explain why x is small here.
39:31

go back a bit here. I claimed that I can use root of 1 minus x is 1 minus x by 2, assuming x is small
compared to 1. Is x small?

39:44

Not really. There can be situations where it is not small. But if you suspect such a situation, then you
don't use this approximation. You'll have to solve this. And with the calculator and with numbers, you
can solve it. You don't have to use the approximation. Only for deriving expressions that I'm using that
approximation. Now, the typical value of x is something like 0.05, 0.1. And very rarely it goes above 0.1.

40:13

Typical value of VG is around 1. Typical value of QL is also around 0.5 to 1. And therefore, this number
could be some 0.2, 0.3, et cetera. Root of 1 minus 0.3 is not actually 0.85. Approximately, it is OK. So I
would say up to when this number becomes 0.3,

40:42

you start re-examining your assumption. Till then it seems to be okay. Right? So root of.7 versus.85.
Okay? Root of.... I think it will be okay. Second decimal place change will be there. So up to that it is okay.
And similarly here. Here also same argument, not only that. Here this number is less than one.

41:10

So here the approximation is better, 1 minus q c x is less than 1. So approximation is better. So use that
approximation.

41:20

and write it this way. So, here there is approximation, binomial approximation or root of 1 minus x is 1
minus x by 2, x is small compared to 1. Between this and this

41:36
there is no approximation, there is only algebraic manipulation. Okay? And now there is one more
approximation in this step. That approximation is Qc x is usually less than 1. Because x is usually 0.05,
0.1, 0.15, but hardly ever more than 0.2, unless the power system is really bad. Okay? Because you see
the... What is the short-circuit capacity?

42:05

that you can expect at the LT distribution.

42:13

Short circuit capacity at LT distribution level is typically between 5 to 10 pu. At the sub transmission level
it is around 10 to 15 and the transmission level it is much more than 20, 30 etc. And therefore the value
of X at transmission level is hardly 0.03, 0.02 etc. Value of X in the sub transmission level is like 0.07 and
primary distribution may be 0.1.

42:42

and secondary distribution level it may come to around 0.2, LT distribution. So 0.2 is an extreme value.
Okay. And QC, you hardly ever implement or use equipment which can deliver many, many PUs of power.
Power is usually somewhere close to one PU at the most and therefore this number is typically 0.2 max.

43:10

and therefore 1 by 1 minus a small quantity can be written as 1 plus that small quantity. That's another
binomial approximation. That is 1 by 1 minus alpha is 1 plus alpha if alpha is much less than 1. I am using
that approximation and then I am recognizing that this combination is exactly VL which is the pre-
switching on voltage at the bus.

43:39

this quantity is VL that has been derived here. So wherever that combination comes, I will write VL. So it
is VL plus something. And that obviously that plus something must be delta VL. Therefore, the
improvement in voltage or climb in voltage is rating of the capacitor into the Thevenin's impedance into
the generator voltage or the open circuit voltage.
44:06

QCXVG That's an expression to be remembered Now if VG and VL prime I mean that the venance open
circuit voltage and the bus voltage Are close to 1 pu as they are in a normal power system then The
current drawn by the capacitor will be the rated current which is numerically equal to QC And then the
reactive current in X will be QL minus QC by QL

44:36

because load will demand QL units of reactive power but the best voltage is close to 1 take it as 1
therefore the reactive component of load current is QL only reactive component of capacitor current is
QC only therefore the difference is the net reactive current in X net reactive current in X into X is the
inline drop source minus inline drop is terminal and therefore you can say this drop plus this one this is
the old VL

45:06

So you can say delta VL is QC into X. Okay. So you can forget this multiplication by VG and say VG is close
to one. So approximate improvement in voltage by switching on a capacitor of rating QC per unit,
immediately you can answer it is QC into X, where X is the Thevenin's impedance. That's a first level
answer. Further refinement of the answer, multiply by VG if you know it.

45:36

Sometimes you may not know what is VG value, except an assurance that it is close to 1. Further
refinement, still further refinement. If you want still further refinement, don't use here. This
approximation don't use. Okay. Solve this root and solve for VL prime. Still further refinement. Don't use
Pythagorean approximation. Go back to

46:04

the phasor diagram and solve it properly. I mean square and add this number also. That is when the
active component of load power will also have a bearing in the final value. But of course we expect that
influence to be small only. So to the question will the improvement in voltage when you switch on a
capacitor depend upon the active power component of the load? The answer is yes, but it is small
influence because it is a quadrature
46:35

A small influence, usually negligible, but if you want to be exact, then you can consider that. But then no
formula will work. You will have to come to the phasor diagram, solve the phasor diagram and find the
length VG and VLX, the relation between VG, VL and all that.

46:57

The point is, starting from this approximation, there are different levels or layers of refinement possible.
This approximation can be used for quick mental calculation. This can be used for slight further
improvement. I think that improvement will matter. Because VG could be 1.05, 1.07, so it is a 5 or 7%
improvement on this quantity. So that's the next level of refinement.

47:25

Third level of refinement is don't use this approximation, solve the root. Fourth level of approximation is
don't use Pythagorean approximation here, go to the facet bag.

47:40

Now we can express the voltage improvement delta VL in terms of the bus voltage VL before switching
on the shunt compensation equipment. You see the voltage improvement here is expressed in terms of
reactive power rating of the capacitor, the evidence impedance of the bus and the open circuit voltage
or source voltage of the bus. That you may not know sometimes. The evidence impedance can be
worked out from the MVA, short circuit MVA rating of the bus.

48:11

Short circuit MVA rating of the bus will be immediately known if you know the circuit breaker rating
which is installed in that substation Okay, or short circuit information is usually known that data will be
there in the power system data But VG may not be known Okay, if we could express VG in terms of the
terminal voltage or bus voltage before you switched on the capacitor Maybe that will be better But that
is okay

48:39

because VL and VG are related by this formula. VL and VG are related. VG is approximately equal to VL
plus. So I can substitute for VG.

48:53

So, delta VL is QCXVG instead of VG you substitute this expression. So, if you know the pre-compensation
voltage you can predict post-compensation improvement.

49:10

Okay, pre-compensation voltage will be usually known because only because you know that you are
thinking about compensating. Maybe it is low, that is where you are trying to compensate. So you will
generally know that voltage. Now, the effectiveness of compensation can be properly measured if you
define percentage improvement. The improvement expressed as a percentage over the pre-improved
value.

49:38

So delta VL by VL into 100 percent. If you do that, you will get QCX into 1 plus QLX by VL square. Of
course, you are free to assume VL is very close to 1. That will give you a cruder level of approximation in
which case it will be percentage improvement will be QC into X into 1 plus QL into X. So the
improvement depends upon what was the load reactive power. Okay.

50:10

Okay, provided VL is approximately 1 pu is justified. Now I'm showing an illustrative calculation where,
okay, this approximation is used and this better formula is used. For example, say our source voltage is
1.05 pu, load reactive power is 1 pu, the evidence impedance is 0.1 pu, then VL will be VG minus QLX by
VG, from this equation, and that is 0.9548.

50:43

I am assuming Qc 0.75 Pu. Then percentage improvement will be using this formula. Qc x, Qc x is 0.075.
1 plus.

50:59
1 and 0.1 divided by 0.9548 square into 100 is 8.32%. But if VL is not calculated and it is taken as 1 pu,
percentage improvement as per this formula is 8.25. This percentage has to be interpreted as a
percentage of this number, whereas this percentage has to be interpreted as a percentage of this
number. Therefore, delta VL in the first cases

51:28

0.0794 Pu in the second approximation it is 0.0825 Pu they are different only in the third decimal place if
you round this it is 0.08 if you round this also it is 0.08 so in the third decimal place there is difference
that is usually tolerable so I think you can use this approximation unless VL and VG are far away from

51:58

And if they are far away like suppose VG is 1.2 and VL is 0.8 or the other way, well they are simply not
permitted in a power system. Something will be done. So unless the power system is in extremely bad
condition, you may assume that all the voltages are close to 1 pu. So that was a discussion on switching
on the static power compensator and that compensator happens to be a fixed capacitor compensator.

52:27

then what are the expressions for estimating the improvement in voltage and what are the factors which
govern the improvement the answer to that question what are the predominant factors which govern
the improvement the answer is the predominant factors which govern the improvement in voltage are
the rating of the compensation equipment and the Thevenin's impedance or short circuit capacity

52:57

that means smaller the short circuit MVA, better the improvement for a given rating. So the amount of
voltage improvement is proportional to the reactive power equipment rating and it's proportional to the
Tevinin's impedance at that bus. On other dependences are only marginal. Of course it depends slightly
on the load reactive.

53:25

but not much. Next, case two. In case two, the static bar compensator is a variable impedance type SVC
with rating QCPU. With rating QCPU means, at rated condition, the capacitor will be fully active. The
paralleling inductance will be simply switched off.
53:48

So at full rating, it is simply a capacitor only. And therefore, whatever we said about a fixed capacitor is
correct here also. There it's only that you can vary that capacitance from zero to C, where C is the same
value as in the first case, because the rating is same. So since the capacitor can be varied.

54:14

The improvement in voltage also can be varied from no improvement to full improvement and the full
improvement value is QC into X into VG. So the result is very similar to that of fixed capacitor. Not much
different. Now coming to case 3, STATCOM with rating of same rating QCPU. That means at full rated
condition the STATCOM is capable of delivering this much of current.

54:42

because rating is based on one pu voltage so rated reactive power divided by rating voltage that is the
rated reactive power current or rated current so it can deliver this much of current even if the voltage is
zero it can deliver that so stat can deliver stat com can deliver this much current at any bus voltage
therefore

55:07

current delivered by the or current taken by the STATCOM delivery and taking that depends upon which
power you are talking about. If you say leading power, leading reactive power then you will have to say
the current taken leading current taken that is IC. So the leading current taken by the STATCOM will
always be QCPU in the maximum of course you can control it to be less than QCPU but maximum you
can dose QCPU always.

55:36

always with reference to voltage whatever is the best voltage it can do this. So reactive current in the
turbulence impedance will be the reactive current taken by the load minus the reactive power current
taken capacity of reactive current taken by the STATCOM and that is a constant number in the in the
rating limited condition it is constant I mean you are letting the STATCOM do the maximum it can.
56:05

then it will take this much of current. Now this is the reactive current in x. So multiply by x, you get the
inline drop. Add the drop to VL prime, you will get VG. So solving this equation, again it's quadratic.
Again, there's a root here. Manipulate the quantity under the root and write it as one minuses number.
And that one minus that number, again using the same argument as before and remembering that.

56:31

denominator is close to 1 etc. You can say it is a small number compared to 1. We expect the number to
be something less than 0.3. So, I can use binomial approximation root of 1 minus alpha is equal to 1
minus alpha by 2 where alpha is this number. Right. And then.

56:52

you get to this level and further approximation.

57:00

there is approximation only here. Here there is a binomial approximation. After that it's only algebra.
And then I'm carrying this forward to the next page. Okay. The denominator is written as VG into one
plus something. And then I argue that this is a small number compared to one. So one by one plus alpha
is one minus alpha binomial approximation. So I'll bring it to the numerator. And then I will write, collect
this term and this term.

57:29

and write it this way and recognize it as the pre-switching on voltage. Before SVC is switched on, what is
the best voltage? That is VL and therefore this composite quantity must be delta VL. So delta VL is QC
into X into 1 plus QLX by VG square. That's a formula that we use for STATCOM. Of course, if VG is close
to 1, okay, or VG and VL are close to 1.

57:58

then this formula will coincide with the old formula. I think it will become same, right? Okay, next. In all
the three cases, we see that the improvement that will result by using static work compensation will
directly depend upon the rating of the equipment and directly proportional to the the evidence
impedance and therefore inversely proportional to short circuit rating of the bus. Other dependences are
marginal.

58:29

It will depend on the load reactive power but marginally. And still more marginally on the active power.
We haven't actually derived it, but we expect it to be very small. So we will say that voltage
improvement at a bus is directly proportional to the rating of reactive power equipment and directly
proportional to the bus impedance or short circuit impedance at the bus or the evidence equivalent
impedance at the bus.

58:57

And therefore, this number X, for a given rating QC, the number X is very important. Suppose you have a
250 kV, a reactive pore compensator. If you connect it in the LT distribution system, you connect it in the
11 kV distribution system, you connect in the 110 kV substation level, sorry, sub-transmission system
level.

59:25

or you connect it at the 400 kV transmission system level. Same rating of the equipment. Suppose you do
that, where do you think you will get better voltage improvement? The highest voltage improvement will
be observed if you do it at the LTC side. Of course, you may ask, is it possible to connect a 250 kVA
equipment everywhere in the power system? The answer is no, because voltage levels are different.
What I'm saying is, if you design a 250 kVA equipment to work at 400 volt,

59:54

to work at 11 kV, to work at 110 kV and to work at 400 kV. Of course, four different designs but the rating
is the same. Then which unit will be most effective in improving the voltage? The one which is designed
for 400 kV, 400 volt LT distribution system because that is the point at which you will have a large value
of X.

01:00:20

Okay, and that takes us to the next topic, X versus location in a power system. What is X? X is the
Thevenin's impedance at a particular bus. That bus is kept and then rest of the power system, the entire
power system is converted to a Thevenin's equivalent and what is the impedance? What is the
Thevenin's impedance? Okay.

01:00:48

we take this kind of a visual or a representation of a power system. The core of the power system
represented by the red circle, I call it the core. And the last annular ring, I call it the last mile. The last
mile system is the LT distribution, low tension distribution. That is where our homes are located and that
is where small industries are located, etc.

01:01:18

and that is where small distribution generation units are also located like rooftop solar systems, small
wind power mills such small power distributed generation units are also interfaced to LTE distribution
voltage levels are 440 volt or 230 volt depending upon pre-phase or single phase that's a last mile system
LTE distribution system

01:01:48

primary distribution system, there also there can be distributed generation, slightly higher power
distributed generation interfacing is usually done to 11 kV primary distribution system. Primary
distribution system can run at 11 kV, 22 kV or 33 kV, sometimes 6.6 kV 2.

01:02:12

and as you move inwards you have the sub-transmission system next. There also there can be distributed
generation and small generating stations. By small I mean low power. May be a 5 megawatt or a 10
megawatt generating station. May be micro hydel, mini hydel etc. or may be a small diesel plant etc. That
will come under sub-transmission system. And power transmission will take place at

01:02:38

66 kV and 110 kV in the sub-transmission system, sometimes in 220 kV also. Whether you call 220 kV
transmission system or sub-transmission system depends upon whether you have 1200 kV lines, 750 kV
lines, etc. If you have 1200 kV lines and 700 kV lines and an HPDC link, etc. you will tend to call 220 kV as
sub-transmission.
01:03:04

But if the highest transmission voltage in your system is 400 kV only, then 220 kV is clubbed along with
400 kV and then you call both of them transmission systems. And you will call only 110 kV and 66 kV as
sub-transmission. Now the innermost is the core. In the core region you have the major generating
stations and then the high voltage and extra high voltage transmission system.

01:03:31

it's quite unlikely that you will have distributed generation in the core.

01:03:38

This is not a geometrical description. I said major generating stations are in the core. That doesn't mean
that they are located close to each other. Conceptually, element-wise, when you draw a network, right,
conceptual, or the network drawing, you will mentally tend to associate the generating stations, HV, HV
lines, et cetera, as the inner core of a power system, the central most components.

01:04:10

the central most components. That is, when you switch on a lamp in your house, the current taken by
the lamp has to travel. It has to travel through the LTE distribution line, then distribution transformer, get
into the 11 kV system, travel through the 11 kV lines, go to the power transformer, reach a substation,
from the sub-transmission system substation, probably go to a high voltage substation, and from there
travel

01:04:39

hundreds of kilometers or sometimes even thousands of kilometers to reach a particular generating

01:04:50

With respect to this path I am giving this diagram. Okay. Now, consider connection points anywhere in
this diagram. Suppose you have a connection point in the LTE distribution system, maybe it's a plug point
in your house. If you short that plug point and measure the current and calculate the impedance, open
circuit voltage by short circuit current, that is X.

01:05:19

that's the Thevenin's equivalent. Similarly, suppose you have a connection point here, maybe it's a 11 kV
bus at a distribution system substation. Suppose you put a short there, then the open circuit voltage by
short circuit current, that's a X value there. Similarly, X value everywhere may be calculated.

01:05:43

In fact, x value at most of the substations are known. They won't give you x value straight away, they will
give you short circuit MVA from which it can be calculated. Short circuit MVA is nothing but short circuit
current into open circuit voltage or nominal voltage and expressed in PU. Okay. So the increasing
direction of x is from core to periphery. As you travel from the core,

01:06:11

and you reach periphery, X value increases in Pu. At the innermost core, that is at the terminal of the
generating station, what is the Thevenin's impedance? Very, very small. At the sub-transmission system,
what is the Thevenin's impedance?

01:06:39

generator terminals up to the sub-transmission subsystem, up to the sub-transmission substation you

01:07:07

all the impedances between you and the core will constitute your governance impedance.

01:07:15

This is why the Tevinon impedance increases as you go towards periphery. Now, if you have a static wire
compensation unit, many of them of the same rating but designed for different voltages, you connect
the equipment in the outer periphery, you will see that it is able to produce a lot of voltage
improvement. You go inner, it doesn't produce much voltage improvement and you go to the core, it
doesn't produce anything. But then what is it doing? Of course it is...

01:07:44

taking reactive power, it is modifying the flows, maybe you get some loss advantage, maybe the losses
will go down a little here and there but you don't get much, there's no point in trying to put a reactive
power compensator straight across a generator terminal right?

01:08:03

That's my first statement, I will correct it. If the generator were ideal, there is no point in putting a
reactive power equipment across it. Just as there is no point in putting a capacitor across a DC battery. It
is as good as that capacitor not being there. What's the point in putting a capacitor straight across a DC
source?

01:08:25

Similarly here, there's no point in putting a capacitive reactive power equipment across a generator's
terminal. That's the first statement, conditional. I'll put a condition on that. The condition is that
generator should not be working in the reactive power limited mode for my statement to be correct. But
suppose that generator is unfortunately working in the reactive power limited mode.

01:08:51

then it is not a PV bus anymore, it becomes a PQ bus, you know from load flow. By the way, how does a
reactive power limit come on a generator? Why is it there? The answer is, lagging current in a
synchronous generator produces demagnetization in the flux system, field system. And therefore, the
excitation current in the synchronous generator has to go up to counter the demagnetizing effect of
lagging current.

01:09:23

As you make the current in the synchronous generator more and more lagging, that is the reactive load
on the synchronous generators, if it is increasing, then demagnetization increases. So the magnetization
has to be increased. So the excitation current has to be increased. Otherwise the terminal voltage will go
down. So excitation system control will automatically increase the field excitation current. But the field
structure has a maximum rated current.

01:09:52

you cannot increase the current above that because field will get over-heated.

01:09:58

So the excitation control system will finally raise the field current to the upper limit and that is it. It
cannot do anything more. So from then onwards field current will be kept constant. And if the armature
current is becoming more and more lagging, there will be more and more demagnetizing effect and
that's not cancelled. And therefore terminal voltage will go down. Terminal voltage has to be released
now in load flow study. So PV bus becomes PQ bus.

01:10:26

its terminal voltage will change.

01:10:31

Now suppose you put a reactive power compensation equipment straight at the generator bus.

01:10:39

So that reactive power flowing inside the generator will go down because of compensation and maybe it
will come out of the reactive power limited mode. Maybe it will become a PV bus again and that will
improve the terminal voltage and that improvement will be it will percolate down to even the
distribution level. Everywhere voltage will improve.

01:11:02

So voltage improvement by putting a reactive power equipment straight across generating stations or
close to generating stations will be possible only if the generators, generator or generators are working
in reactive power limited mode. That is sealing, excitation is at the ceiling. Otherwise, locating reactive
power equipment in the inner core doesn't lead to any much benefit.
01:11:32

some benefit will be there but not much. So reactive power compensation or static power compensation
is essentially for, okay, I mean, static power compensation for voltage improvement is essentially in the
sub-transmission system and primary distribution system and in the final LT distribution system. At the
transmission system level, if you are using reactive power compensator or fax controller like that,

01:12:02

the objective will not be improvement of voltage, it will be something else. I'm not saying that static-wire
compensators are not used in the transmission level. They are used, but for some other benefit, not
voltage improvement benefit. Okay, we will see one such application where static-wire reactive power
compensator is located in the EHV or HV line.

01:12:31

That is in the core and we will see what is the objective and how is that objective satisfied. Definitely the
objective is not to improve the voltage as such but something else.

01:12:46

that will be covered in the next lecture.