Power Electronics (By Ned Mohan) 2
Power Electronics (By Ned Mohan) 2
Power Electronics (By Ned Mohan) 2
", I w,
D_Pwr~
" "
11 lO-M>~SIJl4J76.1il9 '~. "'-I "' i!
'm "1
'<X
_A ~
~~'" ln
"""
i
,~
IC~
I.V\~) "' -i"
, ""
'"
~
""
."
"b
I ,.. -
,
~
O. 17l1nf
c,
Figure 6-8 PSpice simulation diagram (the load is decreased at 100 ms).
"w.------.------.-------.-----~------,_------r_----_.------,_----_.------,
,~+-----~~~--1-------~----_+------4_----~~----~------+_----_4------~
~L
"w~~~~-------L------~-----L------~----~-------L------~----~------~
c V(Vou~) .
/\ f\ /\ /\
1/\ 1\ fi fi 1\
1\
f\ A f\ f\ !\ !\ r !\ !\
~
o~
CIo
(\/
I (LI)
20- .... ,... .... \I \i \1 \1 \1 ~ U
Figure 6-9 Simulation results: output voltage and inductor current.
The input voltage is fed forward as shown in Fig, 6-10. ln a system with a PFC interface,
the output is nearly constant, independent of the changes in the nllS value of the input
voltage from the utility. Therefore, an increase in the utility voltage V, causes a decrease
in i L , and vice versa. To avoid propagating the input voltage disturbance though the
PFC feedback loops, the input voltage peak is fed forward , as shown in Fig. 6-10, in
deterrnining i; .
6-10
v'd
,,,
,,
L ___________________________ _ ..J
REFERENCES
1. "UC3854 Controlled Power factor Correction Circuit Design," Philip C. Todd,
Unitrode Application Note U-1 34.
2. "High power Factor Switching Pre-Regulator Design Optimization," Lyod Dixon,
Unitrode Design Application M anual.
3. N. Mohan, T . M. Undeland, and W.P. Robbins, Power Electronics: Converters,
Applications and D esign, 3'" Edition, Wiley & Sons, New York, 2003 .
PROBLEMS
6-11
Chapter 7
The purpose of this chapter is to revlew some of the basic concepts associated with
magnetic circuits and to develop an understanding of inductors and transformers, which
are needed in power electronics.
7-1 AMPERE-TURNS AND FLUX
Let us consider a simple magnetic structure ofFig. 7-1 consisting ofan N -tum coil with a
current i, on a magnetic core made up of iron. This coil applies Ni ampere-tums to the
core. We will assume the magnetic field intensity H m in the core to be uniform along
the mean path length em' The magnetic field intensity in the air gap is denoted as Hg.
From Ampere's Law, the closed line integral of the magnetic field intensity along the
mean path within the core and in the air gap is equal to the applied ampere-tums:
(7-1)
/ /
=il
-
Im = I. =I
j
~
.&( tI:
"./ p.l d
- / w
/
Figure 7-1 Magnetic structure with air gap.
ln the core and in the air gap, the flux densities corresponding to H m and Hg are as
follows:
(7-2)
(7-3)
B B
-.!!!..e m
+ -Le g = Ni (7-4)
11m Jlo
7-1
Since flux lines form closed paths, the flux crossing any perpendicular cross-sectional
area in the core is the sarne as that crossing the air gap. Therefore,
(7-5)
B
m
=..'L
A and (7-6)
m
t/J(
cm
C
+ -'-) = Ni (7-7)
.4.&
'R.
:!C-
'R.
ln Eq. 7-7, the two terms within the parenthesis equal the reluctance ~m of the core and
the reluctance ~g of the air gap, respectively. Therefore, the effective reluctance ~ of
the whole structure in the path ofthe flux lines is the sum ofthe two reluctances:
(7-8)
(7-9)
Eq. 7-9 allows the flux t/J to be calculated for the applied ampere-tums, and hence Bm
and Bg can be calculated from Eq. 7-6.
7-2 INDUCTANCE L
At any instant of time in the coil of Fig. 7-2a, the flux linkage ofthe coil m' due to flux
lines entirely in the core, is equal to the flux t/Jm times the number of turns N that are
Iinked. This flux linkage is related to the current i by a parameter defined as the
inductance Lm :
m =N'"
'I'm =Lm i (7-10)
x(N)
",
'-~-- ---~)
(a) (b)
Figure 7-2 Co illnductance.
7-2
where the inductance L m (= ,m / i) is constant if the core material is in its linear operating
region. The coil inductance in the linear magnetic region can be ca1culated by
multiplying ali the factors shown in F ig. 7-2b, which are based on earlier equations:
(7-11)
Eq. 7-11 indicates that the inductance Lm is strictly a property of the magnetic circuit
(i.e., the core material, the geometry, and the number ofturns), provided the operation is
in the linear range of the magnetic material, where the slope of its B-H characteristic can
be represented by a constant f1m .
Energy in an inductor is stored in its magnetic field . From the study of electric circuits,
we know that at anytime, with a current i , the energy stored in the inductor is
(7-12)
where [./], for Joules, is a unit of e nergy. Initially assuming a structure without an air
gap, such as in Fig. 7-2a, we can express the energy storage in terms of flux density, by
substituting into Eq. 7-12 the inductance from Eq. 7-11 , and the current from the
Ampere' s Law in Eq. 7-1 :
[J] (7-13)
where Am im = volume, and in the linear region Bm = f1mH m . Therefore, from Eq. 7-13,
the energy density in the core is
(7-14)
Sirnilarly, the energy density in the air gap depends on f10 and the flux density ln it.
Therefore, from Eq. 7-14, the energy density in any medium can be expressed as
7-3
2
1B 3
w = - - [J/m 1 (7-15)
2 f.J
ln inductors, the energy is primarily stored in the air gap purposely introduced in lhe palh
of tlux lines.
ln lransforrners, there is no air gap in the path of the tlux lines. Therefore, the energy
stored in the core of an ideal transforrner is zero, where the core permeability is assumed
infinite, and hence Hm is zero for a finite tlux density. ln a real transforrner, the core
perrneability is finite, resulting in some energy storage in the core.
d d
e(t) = -,(t) = N-rp(t) (7-16)
dt dt
This assumes that ali tlux lines link ali N-tums such lhat , = Nrp. The polarity of the
emf e(t) and the direction of rp(t) in the above equation are yet to be justified.
The relationship in Eq. 7-16 is valid, no matter what is causing the flux to change. One
possibility is that a second coil is placed on the same core. When the second coil is
supplied by a time-varying current, mutual coupling causes the flux rp through the coil to
change with time. The other possibility is that a voltage e(t) is applied across the coil in
Fig. 7-3, causing the change in tlux, which can be calculated by integrating both sides of
Eq. 7-16 with respect to time:
7-4
1 '
~I) = ~O)+- Je(T).dT (7-17)
No
Recalling the Ohm's law, v = Ri, the current direction through a resistor is into the
terminal at the positive polarity. This is the passive sign convention. Similarly, in the
coil of Fig. 7-3, we can establish the voltage polarity and the flux direction in order to
apply Faraday's law, given by Eqs. 7-16 and 7-17. Ifthe flux direction is given, we can
establish the voltage polarity as follows: first determine the direction of a hypothetical
current that will produce flux in the same direction as given. Then, the positive polarity
for the voltage is at the terminal, which this hypothetical current is entering. Conversely,
if the voltage polarity is given, imagine a hypothetical current entering the positive-
polarity terminal. This current, based on how the coil is wound, for example in Fig. 7-3,
determines the flux direction for use in Eqs. 7-16 and 7-17. Following these roles to
determine the voltage polarity and the flux direction is easier than applying Lenz's law
(not discussed here).
The voltage is induced due to drl> l dI, regardless of whether any current flows in that
coil.
Just as conductors guide currents in electric circuits, magnetic cores guide jlux in
magnelic circuits. But there is an important difference. ln electric circuits, the
conductivity of copper is approximately 10 20 times higher than that of air, allowing
leakage currents to be neglected at dc or at low frequencies such as 60 Hz. ln magnetic
circuits, however, the permeabilities of magnetic materiais are, at best, only 104 times
greater than that of air. Because of this relatively low ratio, the core window in the
structure of Fig. 7-4a has "Ieakage" flux lines, which do not reach their intended
destination that may be another winding, for example in a transformer, or an air gap in an
inductor. Note that the coil shown in Fig. 7-4a is drawn schematically. ln practice, the
coil consists of multi pie layers and the core is designed to fit as snugly to the coil as
possible, thus minimizing the unused "window" area.
The leakage effect makes accurate analysis of magnetic circuits more difficult that
requires sophisticated numerical methods, such as finite element analysis. However, we
can account for the effect of leakage fluxes by making certain approximations. We can
divide the total flux ri> into two parts:
7-5
1. The magnetic flux </>m' which is completely confined to the core and links a li N
tums, and
2. The leakage flux, which is partially or entirely in air and is represented by an
"equivalent" leakage flux </>" which a lso links a li N tums of the coi l but does not
follow the e ntire magnetic path, as shown in Fig. 7-4b.
.pm
~
/'
t\
"
+
e .p,
-=-- \ /
(a) (b)
Figure 7-4 (a) Magnetic and leakage fluxes; (b) equivalent representation ofmagnetic and
leakage fluxes .
ln Fig. 7-4b, </> = </>m + </>" where </> is the equivalent flux which links ali N tums.
Therefore, the total flux linkage of the coil is
The total inductance (called the self-inductance) can be obtained by dividing both sides
ofEq. 7-18 by the current i:
. . .,
- = -m + - (7-19)
i
'-v" i i
___
~
L-'F L,.. ['(
(7-20)
where Lm is often ca lled the magnetizing inductance due to </>m in lhe magnetic core, and
L , is called the leakage inductance due to the leakage flux </>, . From Eqs . 7-19 and 7-20,
the total flux linkage ofthe coil in Eq. 7- 18 can be written as
. = (Lm + L, )i (7-21)
di di
e(t)=Lm-+L, - (7-22)
dI
-----
dI
e.. (l)
7-6
This results in the electrical circuit of Fig. 7-5a. ln Fig. 7-5b, the voltage drop due to the
leakage inductance can be shown separately so that the voltage induced in the coil is
solely due to the magnetizing f1ux . The coil resistance R can then be added in series to
complete the rcprescntation ofthe coi l.
;m
1(1) L di
-
+
+ I dI
L,
-
+ +
v (I)
l
+
e(1)
L,
-i(l)
+
em(t)
e(/) em(t) Lm
./
(a) (b)
Figure 7-5 (a) Circuit representation;
(b) leakage inductance separated from the core.
7-4-1 Mutuallnductances
Most magnetic circuits, such as those encountered in inductors and transformers consist
of multiple coils. ln such circuits, the f1ux established by the current in one coil partially
links the other coil or coi Is. This phenomenon can be described mathematically by
means of mutual inductances, as examined in circuit theory courses. However, we will
use simpler and more intuitive means to analyze mutually coupled coils, as in a Flyback
converter discussed in Chapter 8 dealing with transformer-isolated dc-dc converters.
7-5 TRANSFORMERS
ln power electronics, high-frequency transformers are essential to switch-mode dc power
supplies. Such transformers ofien consist of two or more tightly coupled windings where
almost ali of the f1ux produced by one winding links the other windings. lncluding the
leakage f1ux in detail makes the analysis very complicated and not very useful for our
purposes here. Therefore, we will include only the magnetizing f1ux t/Jm that links ali the
windings, ignoring the leakage f1ux whose consequences will be acknowledged
separately.
To understand the operating principies of transformers, we will consider a three-winding
transformer shown in Fig. 7-6 such that this analysis can be extended to any number of
windings. ln this transformer, ali windings are linked by the sarne f1ux t/Jm' Therefore,
from the Faraday's law, the induced voltages at the dotted terminais with respect to their
undotted terminais are as follows:
7-7
(7-23)
dtPm-
e2 =N2 - (7-24)
dI
e = N dtPm (7-25)
, ' dI
+
e3
i2
+ e2 -
dtPm
-- = -=-2 e,
=-
e e,
(7-26)
dI N, N2 N,
ln accordance to the Arnpere's law g iven in Eq. 7-9, the flux tPm at any instant oftirne is
supported by the net arnpere-tums applied to the core in Fig. 7-6,
(7-27)
ln Eq. 7-27, the core ofFig. 7-6 offers reluctance mmin the flux path and the currents are
defined positive into the dotted terminaIs of each winding such as to produce flux lines in
the sarne direction. Eqs. 7-26 and 7-27 are the key to understanding transforrners: at any
instant of time, applied voltage (equal to the induced voltage if the winding resistance
and the leakage flux are ignored) to one of the windings dictates the flux rate-of-change
and hence the induced voltage-per-tum in other windings. The instantaneous flux tPm IS
obtained by expressing Eq. 7-26 in its integral form below (with proper integral lirnits)
(7-28)
7-8
that requires corresponding net ampere-tums given by Eq . 7-27 to sustain this flux,
overcoming the core reluctance. It is important to note that it is immaterial to the core the
apportionment ofthese winding currents.
The analysis above is based on neglecting the leakage flux, assuming that the flux
produced by a winding links ali the other windings. ln a simplified analysis, the leakage
flux of a winding can be assumed to result in a leakage inductance, which can be added ,
along with the winding resistance, in series with the induced voltage e(t) in the winding
in the electrical circuit representation.
REFERENCE
1. N. Mohan, T. M. Undeland, and W.P. Robbins, Power Electronics: Converters,
Applications and Design, 3'" Edition, Wiley & Sons, New York, 2003.
PROBLEMS
Inductors
A magnetic core has the following properties: the core area Am = 0.93 1cm 2 , the magnetic
path length of Rm = 3. 76cm, and the relative permeability of the material is
J.1, = J.1m I J.1o = 5000 .
7-1 Calculate the reluctance 9i ofthis core.
7-2 Calculate the reluctance ofan air gap oflength Rg = lmm, ifit is introduced in
the core of Problem 7-1.
7-3 A coil with N = 30 tums IS wound on a core with an air gap described in
Problem 7-2. Calculate the inductance ofthis coi!.
7-4 If the flux density in the core in Problem 7-3 is not to exceed 0.2 T , what is the
maximum current that can be allowed to flow through this inductor coil?
7-5 At the maximum current calculated in Problem 7-4, calculate the energy stored in
the magnetic core and the air gap, and compare the two.
Transformers
7-9
7-6 Calculate and draw the magnetizing current waveform, along with the applied
voltage waveform, and the waveforms of the voltages induced in the open
windings 2 and 3.
7-7 Calculate the self-inductances of each winding in Problem 7-6.
7-8 Calculate the peak flux density in Problem 7-6.
7-9 A load resistance of 100 is connected to winding 2. Calculate and draw the
currents in windings I and 2, along with the applied voltage waveform.
7-10 Assuming that winding 1 is not applied a voltage, what is the peak amplitude of
the square-wave voltage at 100 kHz that can be applied to winding 3 of this
transformer, if the peak fl ux density calculated in Problem 7-8 is not to be
exceeded?
7-11 ln Problem 7-6, what is the peak amplitude of the voltage that can be applied to
winding I without exceeding the peak flux density calculated in Problem 7-8, if
the frequency of the square wave voltage is 200kHz? What is the peak value of
the magnetizing current, as compared to the one at 100 kHz?
7-10
Chapter 8
4
I I
V;n Ou !pu!
6)-
* *
~ ~
;; ~:
I
ac ; " II r- ~ ~
I
V.
I J''L. I
- I I
I I I HF transformerl I
______________________ J
....----'-
topology to convert I Feedback
de to de with isolation I controller V
Figure 8-1 Block diagram of switch-mode de power supplies.
8-2 NEED FOR ELECTRICAL ISOLATION
Electrical isolation by means oftransformers is needed in switch-mode dc power supplies
for three reasons:
I. Safety. It is necessary for the low-voltage dc output to be isolated from the
utility supply to avoid the shock hazard .
2. Different Reference Potentials. The de supply may have to operate at a different
potential, for example, the dc supply to the gate drive for the upper MOSFET in
the power-pole is refereneed to its Souree.
8-1
3. Voltage matehing. Ifthe de-de eonversion is large, then to avoid requiring large
voltage and eurrent ratings of semieonduetor deviees, it may be eeonomieal and
operationally more suitable to use an eleetrieal transfonner for conversion of
voltage leveis.
8-3 CLASSIFICATION OF TRANSFORMER-ISOLATED DC-DC
CONVERTERS
ln the block diagram of Fig. 8-1 , there are following three categories of transforrner-
isolated dc-de eonverters, ali ofwhich are discussed in detail in this chapter:
Flyback converters are very commonly used in applieations at low power leveis below 50
W. These are derived from the Buek-Boost converter redrawn in Fig 8-2a, where the
inductor is drawn descriptively on a low perrneability core.
+
v;,
(a) (b)
(c)
Figure 8-2 Buek-Boost and the Flyback converters.
The Flyback converter in Fig. 8-2b consists oftwo mutually-coupled coi Is, where the coil
orientations are such that at the instant when the transistor is tumed-off, the eurrent
switches to the second coil to mainta in the sarne flux in the core. Therefore, the dots on
8-2
coils are as shown in Fig. 8-2b where the current into the dot of either coil produces core
flux in the sarne direction. Commonly, the circuit of Fig. 8-2b is redrawn as in Fig. 8-2c.
We will consider the steady state in the incomplete demagnetization mode where the
energy is never completely depleted from the magnetic core. This corresponds to the
continuous conduction mode (CCM) in Buck-Boost converters. We will assume ideal
devices and components, the output voltage vo(t) = Vo ' and the leakage inductances to be
zero.
Tuming on the transistor at t = O in the circuit in Fig. 8-2c applies the input voltage ~.
across coil 1, and the core magnetizing flux tPm increases linearly from its initial value
tPm(O), as shown in the waveforms ofFig. 8-3. Ouring the transistor on-interval Dr;, the
increase in flux can be calculated from the Faraday's law as
'" Ao v,.
= N DTS (8-1)
'rp-p
,
lout
0'-----'----"---'-----
t
Figure 8-3 Flyback converter waveforms.
Oue to increasing tPm' the induced voltage (N2 / N,) v,. across coil 2 adds to the output
voltage Vo to reverse bias the diode, resulting in io", = O. Corresponding to the core flux,
the current im can be calculated using the relationship tP = Ni / 9'l, where 9'l is the core
reluctance in the flux path. Therefore, using Eq. 8-1, the increase in the input current
during the on-interval , from its initial value 1,.(0) can be calculated using Eq. 8-1 as
8-3
I:l.i;n = 912 ~nD~ (8-2)
N,
During the on-interval DT" the output load is entirely supplied by the energy stored in
the output capacitor, and the core magnetizing flux and the input current reach their peak
values at the end ofthis interval:
- 9{
1,. = Ii. (O) + - , V,.DT, (8-4)
N,
After the on-interval, turning off the transistor forces the input current in Fig. 8-2c to
zero. The magnetic energy stored in the magnetic core due to the flux (Ilm cannot change
instantaneously, and hence the ampere-turns applied to the core must be the same at the
instant immediately before and after tuming the transistor off. Therefore, the current iau'
in coil 2 through the diode suddenly jumps to its peak value such that
NI I
201ll _
1",_ 0
=N1 11/n - ---.r.
I"",=v
(8-5)
- N-
I 0111 = -N ' Jln (8-6)
,
With the diode conducting, the output voltage Va appears across coil 2 with a negative
polarity. Hence, during the off-interval (1-D)T" the core flux declines linearly, as
plotted in Fig. 8-3 , by tJ.(Ilp_p, where
D)T, (8-7)
tJ.(Ilp_p = ;
, (1 -
Using Eqs. 8-1 and 8-7,
(8-8)
The change in the current im"ct) can be calculated in a manner similar to Eq. 8-2, and this
current is plotted in Fig. 8-3.
Eq. 8-8 shows that in a Flyback converter, the dependence of the voltage-ratio on the
duty-ratio D is identical to that in the Buck-Boost converter, and it also depends on the
8-4
coils tums-ratio N 2 / N I Flyback converters require minimum number of components by
integrating the inductor (needed for a Buck-Boost operation) with the transformer that
provides electrical isolation and matching of the voltage leveIs. These converters are
very commonly used in low power applications in the complete demagnetization mode
(corresponding to the discontinuous-conduction mode in Buck-Boost), which makes their
control easier. A disadvantage of the Flyback converter is the need for snubbers to
prevent voltage spikes across the transistor and diode due to leakage inductances
associated with the two coi Is.
+ D,
D, ___._iL
V;" +
+ +
V. N, N,
A
VO
D,
q(l)
q(t)
(a) (b)
Initially, assuming an ideal transformer in the Forward converter of Fig. 8-4b, the third
winding and the diode D3 can be removed and DI can be replaced by a short circuito ln
such an ideal case, the Forward converter operation is identical to that of the Buck
converter, as shown by the waveform in Fig. 8-5, except for the presence of the
transformer tums-ratio N, / N I Therefore, in the continuous conduction mode,
(8-9)
8-5
(N, I N , )V,"
-
( N, I N, )DV,"
o ~~---U~)'-~---L--------------~-----
. ' T,
Figure 8-5 Forward converter operation.
ln the case with a real transformer, the core must be completely demagnetized during the
off-interval of the transistor, and hence the need for the third winding and the diodes DI
and D" as shown in Fig. 8-4b. Tuming on the transistor causes the magnetizing flux in
the core to build up as shown in Fig. 8-6. During this on-interval DT" D, gets reverse
biased, thus preventing the current from flowing through the tertiary winding. The diode
D 2 also gets reversed biased and the output inductor current flows through DI .
When the transistor is tumed off, the magnetic energy stored in the transformer core
forces a current to flow into the dotted temlinal of the tertiary winding, since the current
into the dotted terminal of lhe secondary winding cannot flow due to D I ' which results in
v'n to be applied negatively across the tertiary winding, and the core flux to decline, as
shown in Fig. 8-6. (The output inductor current freewheels through D 2 .) Afier an
interval ~,mag' the core flux comes to zero and stays zero during the remaining interval,
until the next cycle begins.
1J1~ T.-..,
T,
To avoid the core from saturating, T d<mag must be less lhan the off-interval (1- D)T, of
lhe transistor. Typically, windings I and 3 are wound bifilar to provide a very tight
mutual coupling between the two, and hence, N , = NI Therefore, to the core is applied
an equal magnitude but opposite polarity per-tum voltage during DI', and Tdemag ,
respectively. At the upper limit, Td<mag equals (\ - D)T" and equating it to the on-
interval DT, of the transistor yields the upper limit on the duty-ratio, Dmax' to be 0.5,
with N, = N 3
8-6
S-5-1 Two-Switch Forward Converters
Single-switch Forward converters are used in power ratings up to a few hundred watts.
However, Two-Switch Forward converters discussed below eliminate the need for a
separate demagnetizing winding and are used in much higher power ratings of a kW and
even higher.
Fig. 8-7 shows the topology ofthe Two-Switch Forward converter, where both transistors
are gated on and off simultaneously with a duty-ratio D ~ 0.5. During the on-interval
DT, when both transistors are on, diodes DI and D 2 get reverse biased and the output
inductor current iL tlows through Do, similar to that in a single-switch Forward
converter. During the off interval when both transistors are tumed off, the magnetizing
current in the transformer core tlows through the two primary-side diodes into V;., thus
applying V;. negatively to the core and causing it to demagnetize. Application of -V;. to
the primary winding causes Do to get reversed biased and the output inductor current iL
freewheels through DF .
q(t)---'------'
8-7
accounted for. We will consider steady state converter operation in the continuous
conduction mode where the output inductor current iL f10ws continuously. As with
previous converters, we will assume ideal devices and components, and vo(t) = Vo '
,1
>--
-t>: +
VA
T ..
'L
~
+
- ::r:
-'I
>-
'"
~.,
D 2
ln the Full-Bridge converter of Fig. 8-8, the voltage v, applied to the primary winding
altemates without a dc component. The waveform of this voltage is shown in Fig. 8-9,
where v, = v,. when transistors I; and T, are on dUTing Dr" and v, = -v,. when 1", and
T.. are on for an interval ofthe sarne duration. This waveform applies equal positive and
negative volt-seconds to the transformer primary. The switch duty-ratio D 0.5) is
controlled to achieve the output voltage regulation by means of zero intervals between
the positive and the negative applied voltages.
V, v,.
o I
(-V..)~
f-DT, - '-DT,-
T,
Figure 8-9 Full-Bridge converter waveforms.
The way the voltage across the primary winding, hence the secondary winding, is forced
to be zero classifies Full-Bridge converters into the following two categories:
8-8
conduction losses on the primary side of the transformer. Therefore, the PWM control
results in lower conduction losses, and it is the control method discussed in this chapter.
PSM Control. ln phase-shift modulated control, the two transistors of each power-pole
are operated at nearly 50% duty-ratio, with D = 0.5. The output of each power-pole
pulsates between v'n and O with a duty-ratio of nearly 50%. The length of the zero
intervals is controlled by phase-shifting the two power-pole outputs with respect to each
other, as the name of this control implies. During zero intervals, either both transistors at
the top, or both transistors at the botlom are on, creating a short circuit (through one of
the anti-parallel diodes, depending on the direction of the current) across the primary
winding, resulting in VI = O. During this short-circuited condition, the output inductor
current is reflected to the primary winding and circulates through the primary-side
semiconductor devices, causing additional conduction losses. However, increased
conduction losses can be offset by the reduction in switching los ses by this means of
control, as we will discuss in detail in Chapter 10 on soft-switching.
8-6-1 PWM Control
As shown by the block diagram of Fig. 8-lOa, the PWM-IC for Fu!l-Bridge converters
provides gate signals to the transistor pairs (T,,1', and T" r.) during altemate cycles of
the ramp voltage in Fig. 8-10b.
~,and T,
(a)
(h)
Figure 8-10 PWM-IC and control signals for transistors.
Corresponding to these PWM switching signals, the resulting sub-circuits are shown in
Fig. 8-11 for one-half switching cycle, where the other half-cycle is symmetric .
D, iL
iI
+ '"
y
'-+
J." +
+~ >- .!:!J..
N[
Vi" : v A
...!..._
,,
+
;~
v2 Q 2t< ...I.
v A =.0
I +
: v.
VI~ . er: ~
~, (0.2
L____ J 1 'N
_2
8-9
lnterval DT, with transistors Y; ,T, in their on state. Tuming on Y;,T, applies positive
voltage V;. to the primary winding, causing D 2 to become reverse biased and iL IS
carried through D" as shown in Fig. 8-1 la. During this interval, vA = (N2 / N , ) V;" as
plotted in Fig. 8-12.
Interval (I/2-D)T, with ali Transistors off. When ali the transistors are turned off,
there is no current in the primary winding, and the output inductor current divides equally
(assuming an ideal transformer) between the two output diodes as shown in the sub-
eircuit of Fig. 8-11 b. This ensures that the total ampere-turns acting on the transformer
core equal zero beeause of iL / 2 coming out of the dotted terminal and iL / 2 going into
the dotted terminal. Applying the Kirchhoff's voltage law in the loop consisting of the
two seeondaries in Fig. 8-1 1b shows that v2 + v; = O. Sinee v, = v;, the two voltages
must be individually zero, and henee also the primary voltage VI :
(8-10)
v,
I - - - - - - T, - - - - - - 1I
v
Aif - - - - , N,' v:
N
-. V~V
/,01 <>
N
~ 2=
N DV
'
ln
o-----I---f----j---f----t.: t
(8-1 I)
8-10
8-7 Balf-Bridge and Push-Pull Converters
Variations ofFull-Bridge converters are shown in Fig. 8-13. The Half-Bridge converter
in Fig. 8-13a consists of only two tra nsistors but requires two split capacitors to form a dc
input mid-point. It is sometirnes used at slightly lower power leveis compared to the
Full-Bridge converter. The Push-Pull converter in Fig. 8-13b has the advantage of
having both transistors gates referenced to the low-side of the input voltage. The penalty
is in the transformer where during the power transfer interval, only one half of the
prirnary winding and one half ofthe secondary winding are utilized.
+
-
+
vA
.
iL
J!;n +
V;n
T,
V;n
2
(a) (b)
REFERENCE
I. N . Mohan, T . M. Undeland, and W.P. Robbins, Power Electronics: Converters,
Applications and Des ign, 3'" Edition, Wiley & Sons, New York, 2003.
PROBLEMS
Flyback Converters:
ln a Flyback converter, V;n = 30V, N, = 30 tums, and N , = IS turns. The self-
inductance of winding I is 50 flH , and ;; = 200kHz. The output voltage is regulated at
Vo = 9V.
8-1 Calculate and draw the waveforms shown in Fig 8-3 along with the ripple current
in the output capacitor, ifthe load is 27W .
8-2 For the sarne duty-ratio as in Problem 8-1, calculate the criticai power which
makes this converter ope rate at the border of incomplete and complete
demagnetization modes.
8-11
8-3 Draw the waveforms, similar to those in Problem 8-1, in Problem 8-2.
8-4 If a Flyback converter is operating in a complete de-magnetization mode, derive
the voltage transfer ratio in terms of the load resistance R , the switching
frequency J" the self-inductance L., of winding I, and the duty-ratio D.
Forward converter:
ln a Forward converter, v'n = 30V , N, = lO tums, N, = S tums, and N, = lO tums. The
8-S Calculate and draw the waveforms for vA' iL , iin' and iD, in Fig. 8-4b.
8-6 Ifthe maximurn duty ratio needs to be increased to 0.7, calculate N, / N,?
8-7 Why is diode D, necessary in Fig 8-4b?
J, = 200kHz. The output voltage is regulated such that Vo = SV. The self-inductance of
winding I is ISO pH , and the output filter inductance is SO pH .
8-12
Chapter 9
9-1 INTRODUCTION
As discussed Chapter 8, inductors and transformers are needed in switch-mode dc power
supplies, where switching frequencies are in excess of 100 kHz. High-frequency
inductors and transformers are generally not available off-the-shelf, and must be designed
based on the application specifications. A detailed design discussion is presented in
Reference [Il. ln this chapter, a simple and a commonly used approach called the Area-
Product method is presented, where the thermal considerations are ignored. This implies
that the magnetic component built on the design basis presented here should be evaluated
for its temperature rise and efficiency, and the core and the conductor sizes should be
adjusted accordingly.
9-2 BASICS OF MAGNETIC DESIGN
ln designing high frequency inductors and transformers, a designer is faced with
countless choices. These include choice of core materiais, core shapes (some offer better
thermal conduction whereas others offer better shielding to stray flux), cooling methods
(natural convection versus forced cooling), and losses (Iower losses offer higher
efficiency at the expense of higher size and weight) to name a few. However, ali
magnetic design-optimization programs calculate two basic quantities from given
electrical specifications:
The peak flux density Bm .. in the magnetic core to limit core losses, and
The peak current density J mox in the winding conductors to limit conduction
losses
The design procedure presented in this chapter assumes values for these two quantities
based on the intended applications of inductors and transformers. However, they may be
far from optimum in certain situations.
9-3 INDUCTOR AND TRANSFORMER CONSTRUCTION
Figs. 9-1 a and b represent the cross-section of an inductor and a transformer wound on
toroidal cores. ln Fig. 9-la for an inductor, the same current i passes through ali N
tums of a winding. ln the transformer of Fig. 9-1 b, there are two windings where the
9-1
current i, in winding I, with N, bigger cross-section conductors, is in opposite direction
to that of i 2 in winding 2 with N 2 smaller cross-section conductors. ln each winding,
the conductor cross-section is chosen such that the peak current density J m~ is not
exceeded at the maximum specified current in that winding. The core area A=~ in Figs.
9-1 a and b a llows the flow of flux lines without exceeding the maximum flux density
Bmu in the core.
~indow
(a) (b)
Figure 9-1 Cross-sections.
The area-product method, based on pre-selected values of the peak flux density Bmu in
the core and the peak current density J mu in the conductors, allows an appropriate core
size to be chosen, as described below.
The windows of the toroidal cores ln Figs. 9-la and b accommodate the winding
conductors, where the conductor cross-sectional area A="" depends on the maxirnum rms
current for which the winding is designed. ln the expression for the window area below,
the window fill-factor k w in a range from 0.3 to 0.6 accounts for the fact that the entire
area of the window cannot be filled , and the subscript y designates a winding, where in
general there may be more than one, like in a transformer
A,.indow = -}-
w
L (N yA=nd.Y )
y
(9-1 )
ln Eq. 9-1, the conductor cross-sectional area in winding y depends on its maxirnum rms
current and the maximum allowed current density J m~ that is generally chosen to be the
sarne for ali windings:
9-2
I
A cond ,y -~
- J (9-2)
m~
(9-3)
which shows that the window area is linearly proportional to the number of tums chosen
by the designer.
The core cross-sectional area in F igs. 9-la and b depends on the peak flux r/J and the
choice ofthe maximum allowed flux density Bm~ to limit core losses:
(9-4)
'
r/J=- (inductor) (9-5)
N
ln a transformer, based on the Faraday's law, the flux depends linearly on the applied
volt-seconds and inversely on the number of tums. This is shown in Fig. 9-2 for a
Forward converter transformer with N , = N , , and the duty-ratio D, which is limited to
0.5. Therefore, we Can express the peak flux in Fig. 9-2 as
(9-6)
where the factor k ron" equals D in Forward converter, and typically has a maximum
vaI ue of 0.5. The factor k conv can be derived for transformers in other converter
topologies based on the specified operating conditions, for example, it equals D / 2 in a
Full-8ridge converter. ln general, the peak flux can be expressed in terms of any one of
the windings, y for example, as
(transformer) (9-7)
9-3
v ,
v:" -
o
t
- DT, - T,
(- v," )
(transformer ) (9-9)
Eqs. 9-8 and 9-9 show that in both cases, the core cross-sectional area is inversely
proportional to the number of tums chosen by the designer.
The core area-product is obtained by multiplying the core cross-sectiona l area AOON with
its window area AwmdOl" :
(9-10)
(inductor) (9- I I)
(transformer) (9-12)
Eqs. 9-11 and 9-12 shaw that the area-product that represents the overall size of lhe
device is independent (as it ought lO be) of the number of tums. Afier alI, lhe core and
the overalI co mponent size should depend on the electrical specifications and the
assumed values of Em'" and J m~ , and nol on the number of tums which is an intemal
design variable.
9-4
9-4-4 Design Procedure Based on Area-Product Ap
Once we pick the appropriate material and the shape for a core, the cores by various
manufacturers are cataloged based on the area-product Ap. Having calculated the value
of A p above, we can select the appropriate core. It should be noted that there are infinite
combinations ofthe core cross-sectional area A=,. and the window area A,."ndow that yield
a desired area-product Ap. However, manufacturers take pains in producing cores s"ch
that for a given Ap' a core has Aco" and A..mdow that are individually optimized for power
density. Once we select a core, it has specific A,ore and A..,ndow ' which allow the number
of tums to be calculated as follows:
LI
N = --- (inductor ; from Eq. 9-8) (9-13)
ln an inductor, to ensure that it has the spec ified inductance, an air gap of an appropriate
length Rg is introduced in the path of flux lines. Assuming the chosen core material to
have very high permeability, the co re inductance is primarily dictated by the reluctance
9l g ofthe air gap, such that
(9-15)
where,
(9-16)
Using Eqs. 9-15 and 9-16, the air gap length Rg can be calculated as
(9-17)
The above equations are approximate because they ignore the effects of finite core
permeability and the fringing flu x, which can be substantial. Core manufacturers
generalJy specity measured inductance as a function of the number of turn for various
values of the air gap length. ln this section, we used a toroidal core for descriptive
purposes in which it will be difficult to introd uce an air gap. If a toroidal core must be
used, it can be picked with a distributed air gap such that it has the effective air gap
9-5
length as calculated above. The above procedure explained for toro ida I cores is equally
valid for other types of cores. The actual design described in the next section illustrates
tbe introduction of air gap in a pot core.
L-------~------------------------------~t
The peak value of the inductor current from Fig. 9-3 is j = I + M = 5.375A. The rms
2
value of the current for the waveform shown in Fig. 9-3 can be calculated as
~
2 1 2
I~, = I +--M = 5.0A (the derivation is left as a homework problem).
12
From the Magnetics, Inc. catalog [2], we will select a P-type material, which has the
saturation flux density of 0.5 T and is quite suitable for use at the switching fTequency of
100kHz. A pot core 26x16, which is shown in Fig. 9-4 for a laboratory experiment, has
the core Area Aco~ = 93.1 mm 2 and the window Area A.",d~ = 39 mm 2 Therefore, we
will select this core, which has an Area-Product Ap = 93.1 x 39 = 363 I mm'. From Eq. 9-
13,
9-6
N 100jl x 5.375 = 23 Tums
0.25 x93.1 x 10-6
2
Winding wire cross sectional area A=nd = 1,= / J = x = 5.0 / 6.0 = 0.83mm We will use
tive strands of American Wire Gauge A WG 25 wires [3], each with a cross-sectional area
of 0.16mm 2 , in parallel. From Eq. 9-17, the air gap length can be calculated as
100jl
= 63.9 mm 2 , ~indow
2
For the pot core 22x 13 [2], A=re = 29.2 mm , and therefore
Ap = 1866 mm 4 For this core, the winding wire cross-sectional area is obtained as
We will use three strands of A WG 25 wues [3] , each with a cross-sectional area of
0.16mm 2 , in parallel foreach winding. From Eq. 9-14,
N = 0.5x30 =10
, (63.9 X 10-6 ) x (100 X lO' )x 0.25 .
Hence,
N, =N2 = N, = 10.
9-7
9-7 THERMAL CONSIDERATIONS
Designs presented here do not include eddy current losses in lhe windings, which can be
very substantial due to proximity effects in inductors. These effects are carefully
considered in [I]. Therefore, the area-product melhod is a good starting point, but the
designs must be evaluated for temperature rise based on lhermal considerations.
REFERENCES
Inductor Design
9-1 Derive the expression for lhe rms current for lhe current waveform in Fig. 9-3.
9-2 ln the design example of lhe inductor in this chapter, the core has an area
Acon = 93.lmm', lhe magnetic path R. = 37.6mm, and lhe relative permeability of
the core material is 1', = I'm/1'0 = 5000. Calculate the inductance with 31 tums
if lhe air gap is not introduced in lhis core in the flux path.
9-3 ln lhe inductor design presented in lhis chapter, what is the reluctance offered by
the magnetic core as compared to lhat offered by the air gap?
9-4 ln Problem 9-2, what is lhe maximum current lhat will cause the peak flux
density to reach 0.25 T ?
9-5 ln the inductor designed 10 lhis chapter, what wil be the inductance and the
maximum current lhat can be passed without exceeding the Bm~ specified, if the
air gap introduced by mistake is only one-half of the required value R. g = 1mm .
What is the stored energy with R. g = 0.5 mm compared to that at R. g = I mm ?
Transformer Design
9-6 ln lhe design example of the transformer in lhis chapter, lhe core has an area
Acon = 63.9 mm' , the magnetic path R. = 3 1.2 mm , and the relative permeability of
lhe core material is 1', = I'm / 1'0 = 5000. CaJculate the peak magnetizing current
at duty-ratio ofO.5.
9-7 What is the tertiary winding conductor diameter needed for the magnetizing
current calculated in Problem 9-6?
9-8 Derive kco~ for a transformer in a Full-Bridge converter.
9-8
Chapter 10
10-1 INTRODUCTION
ln converters so far, we have discussed hard switching in the switching power-pole, as
described in Chapter 2. ln this chapter, we will look at the problems associated with hard
switching, and some of the practical circuits where this problem can be minimized with
soft-switching.
10-2 HARD-SWITCIllNG lN SWITCIllNG POWER-POLES
ln the switching power-pole repeated in Fig. 10-1 a, the hard-switching waveforms are as
shown in Fig. 10-lb, which were di scussed in Chapter 2. Because ofthe simultaneously
high voltage and current associated with the transistor during the switching transition, the
switching power losses in the transistor increase linearly proportional to the switching
frequency and the times t c(OO) and t c( ojf) shown in Fig. 10-lb, assuming an ideal diode
(10-1)
v,o v,o
'x'DS lo V ns /
o -----"rL--y--..;.I! ~\ ,/['S,
---s: : r"
1---,
,
1--
,"-.)
. -
. ",
, . '
~ I I
- ,'C"-',
o
~.
' 0- '
--I
~
o --L---------~--~_?~5~--~--------~~--~----~-
~--lc.on-->1
~
..._ _t".o~-i
(a) (b)
Figure 10-1 Hard switching in a power-pole.
ln hard-switching converters, in addition to the switching power losses decreasing the
energy efficiency, the other problems are device stresses, thermal management of power
losses, and electromagnetic interference resulting from high di/dt and dv/dt due to fast
10-1
transitions in the converter voltages and currents. The above problems are exacerbated
by the presence of the stray capacitances and leakage inductances associated with the
converter layout and the components.
ln order to reduce the overall converter size, while maintaining high energy efficiency,
the trend is to design dc-dc converters operating at as high a switching frequency as
possible (typically IDO-200 kHz in small power ratings), using fast-switching MOSFETs.
At high switching frequency, the switching power losses become unacceptable if hard-
switching is used, and hence soft-switching is often employed, as briefly described in this
chapter.
The problems described above, associated with hard-switching, can be minimized by
means of the following:
Circuit layout to reduce stray capacitances and inductances
Snubbers to reduce di/dt and dv/dt
Gate-drive control to reduce di/dt and dv/dt
Soft-switching
It is always recommended to have a layout to reduce stray capacitances and inductances.
Snubbers, as describe in the Appendix on the accompanying CD, consist of passive
elements (R, C and possibly a di ode) to reduce di/dt and dv/dt during the switching
transient, by shaping the switching trajectory. The trend in modem power electronics is
to use snubbers only in transformer-isolated dc-dc converters, where the leakage
inductance associated with the high-frequency transformer can be substantial, in spite of
a good circuit layout. Generally, snubbers do not redu ce the overall losses; rather they
shift some of the switching losses in the transistor to the snubber resistor.
By controlling the gate voltage of MOSFETs and IGBTs, it is possible to s low down the
tum-on and tum-off speed, thereby resulting in reduced di/dt and dv/dt, at the expense of
higher switching losses in the transistor. The above techniques, at best, resuIt in a partial
solution to the problems ofhard-switching.
However, there are certain topologies and control, as described in the next section, that
allow soft-switching that essentially eliminate the drawbacks of hard-switching without
creating new problems.
10-3 SOFT-SWITCIDNG lN SWITCIDNG POWER-POLES
There are many such circuits and controI techniques proposed in the literature, most of
which may make the problem of EM! and the overall losses worse due to large
conduction losses in the switches and other passive components. Avoiding these
topologies, only a few soft-switching circuits are practical.
10-2
The goal in soft-switching is that the switching transition in the power-pole occurs under
very favorable conditions, that is, the switching transistor has a zero voltage and/or zero
current associated with it. Based on these conditions, the soft-switching circuits can be
c1assified as follows:
ZVS (zero voltage switching), and
ZCS (zero current switching)
We will consider only the converter circuits using MOSFETs, which result in ZVS (zero
voltage switching). The reason is that, based on Eq. 10-1, soft-switching is of interest at
high switching frequencies where MOSFETs are used. ln MOSFETs operating at high
switching frequencies, a significant reduction of switching loss is achieved by not
dissipating the charge associated with the junction capacitance inside the MOSFET each
time it tums on. This implies that for a meaningful soft-switching, the MOSFETs should
be tumed on under a zero voltage switching (ZVS) condition. As we will see shortly, the
tum-off also occurs at ZVS.
10-3-1 Zero VoItage Switching (ZVS)
To illustrate the ZVS principIe, the intrinsic anti-parallel diode of the MOSFET is shown
as being distinct in Figs. 10-2a and b, where a capacitor is used in parallel. This
MOSFET is connected in a circuit such that before applying the gate voltage to tum the
MOSFET on, the switch voltage is brought to zero and the anti-parallel diode is
conducting as shown in Fig. 10-2a. This results in an ideal loss-Iess tum on at ZVS. At
tum-off, as shown in Fig. 10-2b, the capacitor across the switch results in an essentially
ZVS tum-off where the current through the MOSFET channel is removed while the
voltage across the device remains small (essentially zero) due to the parallel capacitor.
This ZVS principIe is illustrated by moditying the synchronous-rectified Buck dc-dc
converter, as discussed below.
A
I
'- -
----j
(a) (b)
10-3
complimentary gate signals q+ and q- . The waveforms associated with this
synchronous Buck converter are shown in Fig. 10-3b, where the inductor is large such
that the ioductor-current ripple is smalI (shown by the solid curve in Fig. 10-3b), and the
inductor current remains positive in the direction shown in Fig. 10-3a in the continuous
conduction mode.
(a)
q'
T'
C
+ q
,-1
ir. 0----+----
f--_...r-.'Y''--.,--,+ l- O
~R
T-
q
-1 Vo : ----t-it-Ir~~~~_-'~
~ '--
' dd ay
(a) (b)
Figure 10-4 Synchronous-rectified Buck converter with ZVS.
We wilI consider the transition at lhe time I = O labeled in Fig. 10-3b and Fig. 10-4b,
when the inductor current is at its peak iL in Fig. 10-4a. The gate signal q + of the
transistor T+, which is initialIy conducting II. ' goes to zero, while q - remains zero, as
shown in Fig. 10-4b. During the transition time during which the current transfers from
T+ to T - is very short, and it is reasonable to assume for discussion purposes that the
inductor current remains constant at iL as shown in Fig. 10-5a.
10-4
T'
_J ~
+ q' = o '~I---L---...J
r
i, _---1 i,
q
(a) (b)
Fig. 10-5 Transition in synchronous-rectified Buck converter with ZVS.
ln the circuit of Fig. 10-5a with both q and q - equal to zero, initially vc' (O) = O and
v (O) = v'n' where from Kirchhoff's law these two voltages must add to the input
voltage
(10-2)
As the current through T + declines, equal and opposite currents flow through the two
capacitors in Fig. 10-5a, which can be derived from Eq. 10-2 as follows, assuming equal
capacitances C : differentiating both s ides of Eq. 10-2 and multiplying both sides by C
d d
C-vc ' +C-v =0 (10-3)
di di
(10-4)
(10-5)
ln Fig. 10-5a, during the tum-offtransition of T +, v c' rises to v'n and v declines to o.
The voltage v cannot beco me negative because of the diode D - (assuming an ideal
diode with zero forward voltage drop), which begins to conduct the entire iL in Fig. 10-
5b, marking the ZVS tum-off of T + .
10-5
Once D - begins to conduct, the voltage is zero across T - , which is applied a gate signal
q- to tum-on, as shown in Fig. 10-4b, thus resulting in the ZVS tum-on of Y-.
Subsequently, the entire inductor current begins to flow through the channel of T - in
Fig. 10-5b. The important item to note here is that the gate signal to T - is appropriately
delayed by an interval T"'lay shown in Fig. 1O-4b, making sure that q - is applied after
D - begins to conduct.
The next half-cycle in this converter is similar with the ZVS tum-off of T - , followed by
the ZVS tum-on of T + , facilitated by the negative peak ofthe inductor current.
Although this Buck converter results in ZVS tum-on and tum-off of both transistors, the
inductor current has a large ripple, which will also make the size of the filter capacitor
large since it has to carry the inductor current ripple. ln order to make this circuit
practical, the overall ripple that the output capacitor has to carry can be made much
smaller (similar ripple reduction occurs in the current drawn from the input source) by
interleaving oftwo or more such converters, as discussed in section 3-11 of Chapter 3.
10-3-3 Phase-Shift Modulated (pSM) DC-DC Converter
Another practical soft-switching technology is the phase-shift modulated (PSM) dc-dc
converter shown in Fig. 10-6a.
/
VB1
L-. --'-/_------'\~--L[_
VAB
O c--\
ILJ
iAB lfl-l_ _\-\--_-+-
I _-,-
( a)
~ ''-_---'T t
(b)
10-6
and a superior hybrid topology shown in Fig. 10-7, patented by the University of
Minnesota [3-5], are fully described in the Appendix to this chapter on the accompanying
CD.
Ql Q5
+
J J J
V in +
Vo
Q2
J J
Figure 10-7 A superior hybrid topology to achieve ZVS down to no load [3-5].
10-4 INVERTERS FOR INDUCTION HEATING AND COMPACT
FLUORESCENT LAMPS
It is possible to achieve soft-switching in converters for induction heating and compact
fluorescent lamps. These converters are discussed in the Appendix to this chapter on the
accompanying CD.
REFERENCES
I. N. Mohan, T. M. Undeland, and W.P. Robbins, Power Electronics: Converters.
Applications and Design, 3,d Edition, Wiley & Sons, New York, 2003.
2. R. Ayyanar, N. Mohan, and E . Persson, " Soft-Switching in DC-DC Converters:
Principies, Practical Topologies, Design Techniques, Latest Developments," Tutorial
at IEEE-APEC 2002.
3. R. Ayyanar and N. Mohan, US Patent 6,310,785, University ofMinnesota, 2001.
4. R. Ayyanar and N. Mohan, "Novel soft-switching dc-dc converter with full ZVS-
range and reduced filter requirement - Part I: Regulated output applications", IEEE
Transactions on Power Electronics, vol. 16 (2001), March 2001, p. 184-192.
5. R. Ayyanar and N. Mohan, ''Novel soft-switching dc-dc converter with full ZVS-
range and reduced filter requirement - Part II: Constant-input, variable output
applications", IEEE Transactions on Power E lectronics, vol. 16 (2001), March 2001,
p. 193-200.
PROBLEMS
10-1 ln a synchronous-rectified Buck converter with ZVS shown in Fig. 1O-4a,
v'n = 12 V, v;, = 5 V, f, = 100 kHz, and the maximum load is 20 W. Calculate
the filter inductance such that the negative peak current is at least 1.5Amps.
10-7
10-2 ln Problem 10-1 , calculate tbe capacitances across the MOSFETs if the
charge/discharge time is to be no more than 0.5 JlS .
10-8
Chapter 11
11-1 INTRODUCTION
Motor drives (ac and dc) fonn an extremely important application area of power
electronics with market value oftens ofbillions dollars annually in applications described
in Chapter 1. Figure 11-1 shows the block diagram of an electric-motor drive, or for
short, an electric drive. ln response to an input command, electric drives efficiently
control the speed and/or the position of the mechanical load. The controller, by
comparing the input command for speed and/or position with the actual values measured
through sensors, provides appropriate control signals to the power-processing unit (PPU)
consisting of power semiconductor devices.
Electric Drive
1-------,
I Power
Processing Load
fixed Unit
Electric Source fonn
(utility) I
I
I
I
_ _ _ -1
input command
(speed I position)
Figure 11-1 810ck diagram of an electric drive system.
As Fig. 11-1 shows, the power-processing unit gets its power from the utility source with
single-phase or three-phase sinusoidal voltages of a fixed frequency and constant
amplitude. The power-processing unit, in response to the control inputs, efficiently
converts these fixed-form input voltages into an output of the appropriate form (in
frequency, amplitude, and the number of phases) that is optimally suited for operating the
motor. The input command to the electric drive in Fig. 11-1 may come from a process
computer, which considers the objectives ofthe overall process and issues a command to
control the mechanical load. However, in general-purpose applications, electric drives
operate in an open-Ioop manner without any feedback.
Prior to discussing the need for power electronics in electric drives for speed and position
control of mechanical systems, we will brief1y examine the requirements of mechanical
11-1
systems, and various types of electric machines in terms of their terminal characteristics
in steady state in order to determine the voltage and current ratings in designing the
power electronics interface.
11-2 MECHANICAL SYSTEM REQUIREMENTS
Electric drives must satis1'y the requirements of torque and speed imposed by mechanical
loads connected to them. Most electric motors are of rotating type.
11-2-1 Rotational Motor-Load Systems
To understand rotating systems, consider a lever, pivoted and free to move. When an
external force f is applied in a perpendicular direction at a radius r from the pivot, then
the torque acting on the lever is
T =f r
[Nm] [N] [m]
(II-I)
determined by its moment-of-inertia J. The net torque TJ acting on the rotating body of
inertia J causes it to accelerate. Similar to systems with linear motion, Newton's Law in
rotational systems becomes
T; =Ja (11-2)
dllJ TJ
a = -- m = - (11-3)
dt J
ln systems such as the one shown in Fig. 11-2a, the motor produces an electromagnetic
torque T.m. The bearing friction and wind resistance (drag) can be combined with the
load torque TL opposing the rotation. ln most systems, we can assume that the rotating
part of the motor with inertia J M is rigidly coupled (without tlexing) to the load inertia
JL The net torque, the difference between the electromagnetic torque developed by the
motor and the load torque opposing it, causes the combined inertias ofthe motor and the
load to accelerate in accordance with Eq. 11-3 :
11-2
Figure 11-2 Motor and 10ad torque interaction with a rigid coupling.
d
-lV = -
0 (11-4)
dI m J,q
where the net torque 0 = T,m - TL and the equivalent combined inertiaJ,q = J M + J L
Eq. 11-4 shows that the net torque is the quantity that causes acceleration, which in turn
leads to changes in speed and position. Integrating the acceleration a(t) with respect to
time,
,
Speed lVm(t) = lVm(O) + Ja(T)dT (I 1-5)
o
where lVm(O) is the speed at t=0 and T is a variable of integration. Further integrating
lVm (t) in Eq. 11-5 with respect to time yields
,
Om(/) = Om(O) + JlVm(T)dT (11-6)
o
where 0m (O) is the position at I = O, and T is again a variable of integration. Eqs. 11-4
through 11-6 indicate that torque is the fundamental variable for controlling speed and
position. Eqs. 11-4 through 11-6 can be represented in a block-diagram form, as shown
in Fig. 11-2b.
(11-7)
If this differential rotation takes place in a differential time dI, the power can be
expressed as
dW =T __
p= __ dOm =TlV
(I 1-8)
dI dI m
11-3
Figure 11-3 Torque, work and power.
An analogy with electrical circuits can be very useful when analyzing mechanical
systems. A commonly used analogy, though not a unique one, is to relate mechanical and
electrical quantities as shown in Table II-I.
For the mechanical system shown in Fig. 11-2a, Fig. 11-4 shows the electrical analogy,
where torques are represented by current sources. lnertias are represented by capacitors,
from its node to a reference (ground) node, and the two capacitors representing the two
inertias are combined to result in a single equivalent capacitor.
Jeq=JM+J L
Electric machines, as motors, convert electrical power input into mechanical output, as
shown in Fig. II-S. Machines may be operated solely as generators, but they also enter
the generating mode when slowing down (during regenerative braking) where the power
flow is reversed. ln this section, we will briefly look at the basic structure and the
fundamental principIes of the electromagnetic interactions that govern their operation.
11-4
Electrical MechanicaI
Electrica1
System Machine System
Consider a conductor of length R in Fig. 11-6a. The conductor is carrying a current i and
is subjected to an externally-estab/ished magnetic field of a uniforrn flux-density B
perpendicular to the conductor length. A force f.m is exerted on the conductor due to the
electromagnetic interaction between the external magnetic field and the conductor
current. The magnitude ofthis force / ,m is given as
f.m=Bi
~ y
R
.........
(I 1-9)
lN) ITJ[A}[m)
As shown in Fig. 11-6a, the direction of the force is perpendicular to the directions of
both i and B. To obtain the direction of this force, we will superimpose the flux lines
produced by the conductor current, which are shown in Fig. 11-6a. The flux lines add up
on the right side of the conductor and subtract on the left side, as shown in Fig. I 1-6b.
Therefore, the force f.m acts from the higher concentration of jlux /ines 10 the lower
concentralion, that is, from right to left in this case.
externa! B field
v "
f- f,.
~) )
f.m f
\ r-
" /
resuhant
(b)
11-5
11-3-2 Induced EMF
ln Fig. 11-7a, a conductor of length I! is moving to the right at a speed u. The B-field is
uniform and is perpendicularly directed into the paper plane. The magnitude of the
induced emf at any instant oftime is then given by
e =B / u
................... (11-10)
[VI [T][mllm ls )
The polarity of the induced emf can be established as follows: due to the conductor
motion, the force on a charge q (positive, or negative in the case of an electron) within
the conductor can be written as
where the speed and the tlux density are shown by bold letters to imply that these are
vectors and their cross product determines the force. Since u and B are orthogonal to
each other, as shown in Fig. 11-7b, the force on a positive charge is upward. Similarly,
the force on an electron will be downwards. Thus, the upper end will have a positive
potential with respect to the lower end. This induced emf across the conductor is
independent ofthe current that would tlow ifa c10sed path were to be available (as would
normally be the case). With the current tlowing, the voltage across the conductor will be
the induced-emf e(t) 10 Eq. 11-10 minus the voltage drops across the conductor
resistance and inductance.
B(into paper)
+
X X X X f q+
6
q+
X X t
X X B(into paper)
"
X
X
X
X 19
(a)
f. -
X
X
X
X
/
(b)
U
11-6
r=:::::~~il gap
o
o
(a) (b)
11-7
stator
~~~ ~imagnets
rotor
"",-..----winding
,.
11-8
Ali conductors under a stator pole have currents in the sarne direction but the current
direction reverses when the conduc tors reach the other pole. This is needed so that the
forces produced on each conductor are in the sarne direction and add up to yield the total
torque. Similarly, ali conductors under apoie have induced emfs of the sarne polarity,
which changes in polarity every ha lf revolution of the rotor when the conductors reach
the other pole.
The above description shows that the magnitude and the direction of the electromagnetic
torque depend on the armature current ia' Therefore, in a permanent-magnet dc machine,
the electromagnetic torque produced by the machine is linearly related to a machine
torque-constant k T , which is unique to a given machine and is specified in its data sheets
(J 1-12)
Similarly, the induced emf ea depends only on the rotational speed "'m' and can be
related to it by a voltage-constant k E , which is unique to a given machine and specified
in its data sheets
(J 1-13)
Equating the mechanical power ("'m T,m ) to the electrical power ( eaia )' the torque-
constant k T and the voltage-constant k E are exactly the sarne numerically in MKS units
(11-14)
Reversing the Torgue Direction. The direction of the armature current ia determines the
direction of currents through the conductors. Therefore, the direction of the
electromagnetic torque produced by the machine ais o depends on the direction of ia '
This explains, how a dc machine wh ile rotating in a forward or reverse direction ean be
made to go from motoring mode (where the speed and torque are in the sarne direction) to
its generator mode (where the speed and torque are in the opposite direetion) by reversing
the direetion of ia .
Four-Ouadrant Operation. The above diseussion shows that a de maehine ean easily be
made to opera te as a motor or as a generator in forward or reverse direction of rotation .
11-9
11-4-3 DC-Machiue Equivaleut Circuit
It is convenient to discuss a dc machine in tenns of its equivalent circuit ofFig. 11-11,
which sbows conversion between electrical and mechanical power. The annature current
ia produces the electromagnetic torque 1',m(= kr ia ), represented by a dependent current-
source necessary to rotate the mechanical load at a speed Jm ' Across the annature
terminaIs, the rotation at the speed of Jm induces a voltage, called the back-emf
ea (= kEJm ) , represented by a dependent voltage-source.
On the electrical side, the applied voltage v a overcomes the back-emf ea and causes the
current ia to flow. Recognizing that there is a voltage drop across both the annature
winding resistance Ra (which includes the voltage drop across the carbon brushes) and
the annature winding inductance La' we can write the equation of the electrical side as
(11-15)
On the mechanical side, the electromagnetic torque produced by the motor overcomes the
mechanical-Ioad torque TL to produce acceleration:
where J ,q is the total etfective value of the combined inertia of the dc machine and the
mechanical load.
Note tbat the electric and the mechanical systems are coupled. The torque 1',m in the
mechanical system (Eq. 11-12) depends on the electrical current ia' The back-emf ea in
the electrical system (Eq. 11-13) depends on the mechanical spced Jm The electrical
11-10
power absorbed from the electrical source by the motor is converted into mechanical
power and vice versa. ln a dc steady state, with a voltage Va applied to the armature
terminais, and a load-torque TL be ing supplied to the load, the equivalent circuit is as
shown in Fig. 11-l2a, where the inductance La (not shown) in the electrical portion and
the capacitance (shown dotted) representing J ,q in the mechanical portion of the circuit
do not play a role in dc steady state. Hence, in Fig. 11-12
(11-17)
(11-18)
Based on Eqs. 11-17 and 11-18, the steady state torque-speed characteristics for various
values of Va are plotted in Fig. 11-12b.
J,,---
+ + 1 "'m Vai> Va2 > Va3 > Va4
I Wm,rated Va/=rated :
W. I Va 2 :
PPU
v. ..,-
-'-
I J ..
Va 3 -+-----
,
Va4
I T, 1'------_,1.:. ___ _
,
E=kwT
.. =~ 1
Em~m"TQ
(a) (b)
Figure 11-12 DC motor in the dc steady state.
11-11
ia Sinuso idal
Power
@- Processing h
PMAC
~
W
Load fm={ Position
sensor
I
ic
u tility Unit motor
We will consider 2-pole machines, like the one shown schematically in Fig. 11-14a. This
analysis can be generalized to p-pole machines where p > 2. The stator contains three-
phase, wye-connected windings shown in the cross-section ofFig. 11-14a, each ofwhich
produce a sinusoidally-distributed f1ux-density distribution in the air gap, when supplied
by a current.
11-5-2 PrincipIe of Operation
The permanent-magnet pole pieces mounted on the rotor surface are shaped to ideally
produce a sinusoidally-distributed f1ux density in the air gap. The rotor f1ux-density
distribution (represented by a vector ,) peaks at an axis that is at an ang le em(I) with
respect to the a -axis of phase winding a, as shown in Fig. 11-14b. As the rotor tums, the
entire rotor-produced f1ux density distribution in the air gap rotates with it and "cuts" the
stator-winding conductors and produces emf in phase windings that are sinusoidal
functions of time. ln the ac steady state, these voltages can be represented by phasors.
Considering phase-a as the reference in Fig. 11-14b, the induced voltage in it due to the
rotor f1ux cutting it can be expressed as
.~
b -axlS '. ,
,,
,
Br(l)
---------
:C-axIS
,
a-axis
Ca) Cb)
Figure 11-14 Two-pole PMAC mach ine.
11-12
(11-19)
It should be noted that the induced voltage in phase-a peaks when the rotor-flux peak in
Fig. 11-14 is pointing downward, 90 before it reaches the phase-a magnetic axis. Since
the flux produced by the permanent magnets on the rotor is constant, the rms magnitude
of the induced voltage in each stator phase is linearly related to the rotational speed mm
by a per-phase voltage-constant k E . pha"
(11-20)
An important characteristic of the machines under consideration is that they are supplied
through a current-regulated power-processing unit shown in Fig. 11-13, which controls
the currents ia ' ib , and ic supplied to the stator at any instant oftime. To optimize such
that the maximum torque-per-ampere is produced, eaeh phase eurrent in ae steady state is
eontrolled in phase with the indueed voltage. Therefore, with the voltage expressed in
Eq. 11-19, the eurrent in ae steady state is
(11-21)
ln ae steady state, aeeounting ali three-phases, the input eleetrie power, supplied by the
eurrent in opposition to the induced baek-emf, equals the meehanieal output power.
Using Eqs. 11-19 through 11-21
ln Eq. 11-23, the eonstant that relates the rms eurrent to the per-phase torque is the per-
phase torque-constant k". ph~ . Therefore, similar to that in de maehines, in MKS units
At this point, we should note that PMAC drives eonstitute a elass, whieh we will eall self-
synchronous motor drives, where the term "self' is added to distinguish these maehines
from the eonventional synehronous maehines. The reason for this is as follows: in
PMAC drives, the stator phase eurrents are synehronized to the meehanieal position of
the rotor sueh that, for example, the eurrent into phase-a, in order to be in phase with the
induced-emf, peaks when the rotor-flux peak is pointing downward, 90 before it reaehes
11-13
the phase-a magnetic axlS. This explains the necessity for the rotor-position sensor,
unless the rotor position is mathematically computed by sensed voltages and currents.
11-5-3 Mechanical System ofPMAC Drives
The electromagnetic torque acts on the mechanical system connected to the rotor, and the
resulting speed mm can be obtained from the equation below:
m = _T,,
_d_m_ 'm,-::-_T"-
L (11-25)
dt J ,q
where J ,q is the combined motor-Ioad inertia and TL is the load torque, which may
include friction . The rotor position Om (t) IS
(11-26)
Figure 11-15 Equivalent circuit diagram and the phasor diagram ofPMAC (2 pole).
On the electrical s ide, the applied voltage V. in Fig. I 1-15a overcomes the back-emf Em.
and causes the current 7. to f1ow. The frequency of the phasors in Hz equals mm /(2n)
11-14
in a 2-pole PMAC machines. There is a voltage drop across both the per-phase stator
winding resistance R, (neglected here) and a per-phase inductance L" which is the sum
of the leakage inductance caused by the leakage f1ux of the stator winding, and the effect
of the combined flux produced by the currents f10wing in the stator phases. On the
mechanical side in Fig. ll-15a in steady state, the capacitance representing J ,q in the
mechanical portion ofthe circuit is of no effect, and T.m = TL
Note that the electric and the mechanical systems are coupled. ln the electrical system,
the back-emf Emo magnitude (Eq. 11-20) depends on the mechanical speed aJm ln the
mechanical system, the torque T.m (Eq. 11-23) depends on the magnitude ofthe electrical
current lo, which depends on the load torque being demanded of the machine. The
electrical power absorbed from the electrical source by the motor is converted into
mechanical power and vice versa. The phasor diagram, neglecting R" is shown in Fig.
11-15b.
11-5-5 PMAC Torque-Speed Characteristics
ln PMAC machines, the speed is independent ofthe electromagnetic torque developed by
the machine, as shown in Fig. 11-16a, and depends on the frequency of voltages and
currents applied to the stator phases of the machine.
J;>fz>J, > f.
f - - - - - - - - J; Vo
f--------fz
f - - - - - - - - J,
1-------1.
/
o '----------'l~ 0'-'-------1-
T.m f= aJm
(a) 2n (b)
Figure I I - I 6 Torque-speed characteristics and the voltage versus frequency in PMAC.
ln the per-phase equivalent circuit of Fig. 1 1-15a and the phasor diagram of Fig. I I - I 5b,
the back-emf is proportional to the speed aJm Similarly, since the electrical frequency is
linearly related to aJm , the voltage drop across L, is also proportional to aJm Therefore,
neglecting the per-phase stator winding resistance R" in the phasor diagram of Fig. 1 1-
15b, the voltage phasors are ali proportional to aJm , requiring that the per-phase voltage
magnitude that the power-processing unit needs to supply is proportional to the speed
aJm , as plotted in Fig. 11-16b. This relationship between the voltage and speed shown in
11-15
above that indicated by the dotted line, is needed at very low speeds, where the voltage
drop across the stator winding resistance R, becomes substantial at higher torque
loading, and hence cannot be neglected.
11-5-6 The Controller and the Power-Processing Unit (pPU)
As shown in the block diagram ofFig. 11-13, the task ofthe controller is to dictate the
switching in the power-processing unit, such that the desired currents are supplied to the
PMAC motors. The reference torque signal is generated /Tom the outer speed and
position loops discussed in Chapter 13. The rotor position Bm is measured by the
position sensor connected to the shaft. Knowing the torque constant kT . pha" allows us to
calculate the rms value ofthe reference current /Tom Eq. 11-23. Knowing the current rms
vai ue and Bm allows the reference currents for the three phases to be calculated, which
the PPU delivers at any instant oftime.
11-6 INDUCTION MACIDNES
lnduction motors with squirrel-cage rotors are the workhorses of industry because of their
low cost and rugged construction. When operated directly from line voltages (a 50- or
60-Hz utility input at essentially a constant voltage), induction motors operate at a nearly
constant speed. However, by means of power electronic converters, it is possible to vary
their speed efficiently.
11-6-1 Strnctnre of Induction Machines
The stator of an induction motor cons ists of three-phase windings, distributed in the stator
slots. These three windings are displaced by 120 0 in space with respect to each other, as
shown by their axes in Fig. 11-1 7a.
The rotor, consisting of a stack of insulated laminations, has electrically conducting bars
of copper or aluminum inserted (molded) through it, close to the periphery in the axial
b-ms
(a) (b)
Figure 11-17 (a) Three-phase stator; (b) squirrel-cage rolor,
11-16
direction. These bars are electrically shorted at each end of the rotor by electrically
conducting end-rings, thus producing a conducting cage-like structure, as shown in Fig.
11-17b. Such a rotor, called a squirrel-cage rotor, has a low cost, and rugged nature.
11-6-2 Principies oflnduction M o tor Operation
Figure 11-18a shows the stator windings whose voltages are shown in the phasor diagrarn
ofFig. 11-18b, where the frequency ofthe applied line-voltages to the motor is j in Hz,
and V~, is the magnitude in rms
v
- 'V,}-'-+---'
- n-
- --- 1<
v.
v;. I ma
(a) (b)
Figure 11-18 Induction mach ine: applied voltages and magnetizing currents.
By Faraday' s law, the applied stator voltages given in Eq. 11-27 establish a rotating flux-
density distribution in the air gap by drawing magnetizing currents of rms value 1m'
which are shown in Fig. ll-18b:
(11-28)
As these currents vary sinusoidaUy with time, the combined flux-density distribution in
the air gap produced by these currents rotates at a synchronous speed CO,yn' where
The rotor in an induction motor tums (due to the electromagnetic torque developed, as
will be discussed shortly) at a speed com in the sarne direction as the rotation of the air
gap flux-density distribution, such that com < CO" n. Therefore, there is a relative speed
between the flux-density distribution rotating at CO,yn and the rotor conductors at COm .
This relative speed, that is the speed at which the rotor is "slipping" with respect to the
rotating flux-density distribution, is called the slip speed:
11-17
By Faraday's Law (e = B l u) , voltages are induced in the rotor bars at the slip frequency
due to the relative motion between the flux-density distribution and the rotor, where the
slip frequency hUp in terms ofthe frequency f ofthe stator voltages and currents is
Since the rotor bars are shorted at both ends, these induced bar voltages cause slip-
frequency currents to flow in the rotor bars. The rotor-bar currents interact with the flux-
density distribution established by the stator-applied voltages, and the result is a net
electromagnetic torque T.m in the sarne direction as the rotor's rotation.
The sequence of events in an induction machine to meet the load torque demand IS as
follows: At essentially no load, an induction machine operates nearly at the synchronous
speed that depends on the frequency of applied stator voltages (Eq. 11-29). As the load
toque increases, the motor slows down, resulting in a higher value of slip speed. Higher
slip speed results in higher voltages induced in the rotor bars and hence higher rotor-bar
currents. Higher rotor-bar currents result in a higher electromagnetic torque to satisfy the
increased load-torque demando
Neglecting second-order effects, the air gap flux is totally determined by the applied
stator voltages. Hence, the air gap flux produced by the rotor-bar currents is nullified by
the additional currents drawn by the stator windings, which are in addition to the
magnetizing currents in Eq. 11-28.
11-6-3 Per-Phase Equivalent Circuit ofInduction Machines
ln this balanced three-phase sinusoidal steady state analysis, we will neglect second-order
effects such as the stator winding resistance and leakage inductance, and the rotor circuit
leakage inductance. As shown in Fig. 11-19a for phase-a in this per-phase circuit, ti;,
and so on are applied at a frequency f, which results in magnetizing currents to
establish the air gap flux-density distribution in the air gap, represented as flowing
through a magnetizing inductance Lm' The voltage magnitude and frequency are such
that the magnitude of the flux-density distribution in the air gap is at its rated value and
this distribution rotates counter-clockwise at the desired },yn (Eq. 11-29) to induce a
back-emf in the stationary stator phase windings such that
(11-32)
11-18
Next, we will consider the effect ofthe rotor-bar currents. The rotor-bar currents result in
additional stator currents (in addition to the magnetizing current), which can be
represented on a per-phase basis by a current 7;." in-phase with the applied voltage (since
the rotor-circuit leakage inductance is ignored), as shown in Fig. 11-19a
+
!. I
I
PPU t .,-
-'-
+
I Jq
T.. T,
(b)
I' DC--j
v:. = Ema = kE . ,.,.,.,~(j}syr> LO 'E; = kE.pIttuJomLO r:", = 3kr .phsuJ:
(a)
Figure 1 1-19 Induction motor equivalent circuit and phasor diagram.
The flux-density distribution (although produced differently than in PMAC machines)
interacts with these stator currents, just like in PMAC machines, and hence the per-phase
torque-constant equals the per-phase voltage constant
(11-35)
Pem,phase =wT.
m em,p ase =wk
m E., p 1'ra a~'e
(11-36)
ln Eq. 11-36, (WmkE.p.~, ) can be considered the back-emf E~ ,just like in the DC and the
PMAC machines, as shown in Fig. 1 [-[9a
(1 [-37)
ln the rotor-circuit, the voltages induced depend on the slip-speed W' lip' and overcome the
IR voltage drop in the rotor bar resistances. The rotor-bar resistances, and the voltage
drop and the power losses in them, are represented in the per-phase equivalent circuit of
11-19
Fig. l1-19a by a voltage drop across an equivalent resistance R;. Using the Kirchhoff's
voltage law in Fig. 11-19a,
(11-38)
Hence,
(11-39)
Using Eqs. 11-34, 11-36 and 11-39, the combined torque ofall three phases is
(11-40)
where k r ",
. ~"P in the above equation is a machine constant that shows that the torque
produced is linear1y proportional to the slip speed. Using Eqs. 11-30 and 11-40,
8ased on Eq. 11-41, the torque-speed characteristics are shown in Fig. 11-20a for various
applied frequencies to the stator. The stator voltage as function of frequency is shown in
Fig. 11-20b to maintain the peak of the flux-density distribution at its rated vaI ue. The
relationship between the voltage and the synchronous speed or f shown in Fig. 11-20b
is approximate. It needs a substantially higher voltage than indicated (shown dotted) at
very low frequencies where the voltage drop across the stator winding resistance R,
becomes substantial at higher torque loading, and hence cannot be neglected.
OJm f..>/,>/,>f.
Va
f..
f,
/,
f.
O O
I;m f = msyn
(a) (b) 2"
Figure 11-20 lnduction motors: Torque-speed characteristics and voltage vs. frequency.
11-20
11-7 SUMMARY
The discussion above shows the requirements of three common types of machines
supplied through power-electronics based power processing units in steady state.
REFERENCE
I. N. Mohan, Electric Drives: An Integrative Approach, year 2003 Edition, MNPERE,
Minneapolis, www.MNPERE.com.
2. N. Mohan, Advanced Electric Drives: Analysis, Modeling and Simulation using
Simulink, year 2001 Edition, MNPERE, Minneapolis, www.MNPERE.com.
PROBLEMS
Mechanical Systems
I l-I A constant torque of 5 Nm is applied to an unloaded motor at rest at time t = O.
The motor reaches a speed of 1800 rpm in 4 s. Assuming the damping to be
negligible, ca1culate the motor inertia.
11-2 ln an electric-motor drive similar to that shown in Fig. 11-2a, the combined
inertia is J eq= 5xIO-3kg.m2. The load torque is T =0.05kgm 2 . Draw the
L
electrical equivalent circuit and plot the electromagnetic torque required /Tom the
motor to bring the system linearly /Tom rest to a speed of 100 rad/s in 4 s, and
then to maintain that speed.
11-3 Plot the power required in Problem I 1-2.
DCMotors
11-21
11-7 ln a 2-pole, three-phase (PMAC) brushless-dc motor drive, the torque constant
kr and the voltage constant k E are 0.5 in MKS units. The synchronous
inductance is 15 mH (neglect the winding resistance). This motor is supplying a
torque of 3 Nm at a speed of 3,000 rpm in a balanced sinusoidal steady state.
Calculate the per-phase voltage across the power-processing unit as it supplies
controlled currents to this motor.
Induction Motors
11-8 Consider an induction machine that has 2 poles and is supplied by a rated voltage
of 208 V (line-to-line, rms) at the frequency of 60 Hz. It is operating in steady
state and is loaded to its rated torque. Neglect the stator leakage impedance and
the rotor leakage fluxo The per-phase magnetizing current is 4.0 A (rms). The
current drawn per-phase is 10 A (rms) and is at an angle of 23.56 degrees
(lagging). Calculate the per-phase current if the mechanical load decreases 50
that the slip speed is one-halfthat ofthe rated case.
11-9 ln Problem 11-8, the rated speed (while the motor supplies its rated torque) is
3475 rpm. Calculate the slip speed OJ'''p' and the slip frequency f."p of the
currents and voltages in the rotor circuito
II-lO ln Problem 11-9, the rated torque supplied by the motor is 8 Nm. Calculate the
torque constant, which linearly relates the torque developed by the motor to the
slip speed.
II-II A three-phase, 60-Hz, 4-pole, 440- V (Iine-line, rms) induction-motor drive has a
full-Ioad (rated) speed of 1746 rpm. The rated torque is 40 Nm. Keeping the air
gap flux-density peak constant at its rated value, (a) plot the torque-speed
characteristics (the linear portion) for the following values of the frequency f:
60 Hz, 45 Hz, 30 Hz, and 15 Hz. (h) This motor is supplying a load whose torque
demand increases linearly with speed, such that it equals the rated torque of the
motor at the rated motor speed. Calculate the speeds of operation at the four
values offrequency in part (a).
11-12 ln the motor drive of Problem II-II , the induction motor is such that while
applied the rated voltages and loaded to the rated torque, it draws 10.39 A (rms)
per-phase at a power factor of 0.866 (Iagging). R, = 1.50. Calculate the voltages
11-22
Chapter 12
12-1 INTRODUCTION
Motor drives (ac and dc) are important application areas ofpower electronics with market
value of tens of billions dollars annually, as described in Chapter 1. Uninterruptible
power supplies (UPS) are a special case of ac motor drives in discussing the role of
power electronics, and will be briefly discussed in section 12-6 in this chapter.
ln motor drive applications, the voltage-link structure of Fig. 1-16, repeated in Fig. 12-1 ,
is used, where our emphasis will be to discuss how the load-side converter with the dc
voltage as an input synthesizes dc or low-frequency sinusoidal voltage outputs.
Functionally, this converter operates as a linear amplifier, amplifYing a control s ignal, dc
in case of dc-motor drives, and ac in case of ac-motor drives. The power flow through
this converter should be reversible.
+
utility Load
controJler
(h)
Figure 12-2 Converters for dc and ac motor drives.
12-1
12-2 SWITCIllNG POWER-POLE AS THE BUILDING BLOCK
To synthesize de or low-frequeney s inusoidal ae outputs, these eonverters eonsist of the
switehing power-pole shown in Fig. 12-3a with the bi-direetional eurrent eapability,
whieh was introdueed in Chapter 3. [t eonsists of a parallel eombination of Buek and
Boost eonverters, as shown in Fig. 12-3b, where the two transistors are eontrolled by
eomplimentary gate signals. This implementation and the eomplimentary gating of
transistors allow a eontinuous bi-direetional power and eurrent eapability as diseussed
below, and henee ideally a diseontinuous eonduetion mode does not existo The average
representation by an ideal transforrner is shown in Fig. 12-3e.
Buck .. , Boas!
.-----------
, . .---------,.
~
eonverters for de drives, and between any two of the three poles for ae drives, shown in
Fig. 12-2a and b, the de offset is neutralized as will be diseussed in the foIlowing seetions
in this ehapter.
12-2
+
Vd ,< !
,idA
r +
VAN
~
IA
T --
T
VAN
rI
1: dA
~
v c, A
(b)
ln motor drives, to generate the switching function qA' a triangular-waveform signal v,,;
is used for comparison in the PWM-IC, as shown in Fig. 12-5a, where the peak value V,ri
is kept constant and the frequency of the triangular wavefonn establishes the switching
frequency J, .
- i"
o ,
,,,
,
,
I I IT I I I
~ , ," ', "
f '- j - 1-1---5'.
q , (t) : ' "
o d J')
0 , __ 1 0 o t
I .",,(1)1: I
v_ (1) ~ r:
o f - i f - =Jfi
v_
t
i~~
t
00 w
Figure 12-5 Switching power-pole and its voltage and current waveforms.
The resulting switching-signal waveforms are shown in Fig. 12-5b. The duty-ratio d A'
obtained by comparison of these linear waveforms, can be derived by considering two
values of the control signal Vc A: V,ri results in d A to equal I, and (-V,ri) results in d A to
equal O. Linearly interpolating between these two values, the expression for d A can be
written as
12-3
(12-1 )
which shows that it consists of two terms - a dc offset of 0.5 and a term that is linearly
proportional to the control voltage V'.A(t). Therefore, the average output voltage of the
power-pole can be written as
(12-2)
where O.5Vd is the dc offset and kpo,< is the gain by which the control signal is amplified
to produce the output voltage. The average representation of the pulse-width-modulated
switching power-pole is shown Fig. 12-6.
- idA
r - - --,
+
+1 1 i iA Vd
vd-T ~
IA i>AN
I
L_ _ ...J
q A(I)
Vc A(I) Vc. A(I)
(a ) (b)
Although only their average values are of interest, we must recognize that the voltage vA
and the current idA contain harmonics in addition to their intended average values.
Fourier analysis of these waveform s shows that they contain components at frequencies
at the multiples ofthe switching frequency /" with sidebands located at the multiples of
the frequency !, being synthesized by the switching power-pole. These harmonic
frequencies can be expressed as foll ows, where k, and k, are constants, which can take
on values 1,2, 3, and 50 on:
12-4
This harmonic analysis is graphically shown in Fig. 12-7, which highlights the
importance of selecting a high switching frequency. Clearly, the harmonic components
in v A at frequencies f. given by Eq. l2-3 are undesirable but also unavoidable in a
switching power-pole. However, the minimum value in .f" depends on J" and by
selecting a high switching frequency, ali the harmonic frequencies are pushed to the
correspondingly high values. At these high values, the series inductance at the current-
port becomes very effective in ensuring that the current iA has a small ripple in spite of
the pulsating vA ' Similarly, the high-frequency components in i dA are filtered by a
relatively small capacitor across the voltage-port. For this reason in modem power
electronic systems, it is typical to use switching frequencies a few hundred kHz in dc-dc
converters at low power, and a few tens of kHz in converters for motor drives up to a few
hundred kW.
Vh
f, 3fs 4 fs
12-5
I
+ Vo
I x X
Vd
T I A B
+
I
Vo ea l
l______________ ~_J
I
x io
( a) (b)
Figure 12-8 De-motor four-quadrant operation.
ln dc-motor drives, as shown in Fig. 12-9a, the control voltage vc(t) produced by the
feedback controller is applied as vc.A (t), equal to vc(t), to produce the switching function
q A (t) for pole-A. The control voltage for pole-B is negative of the control voltage for
pole-A: vc.B (t) = -vc(t). The average representation ofthe two poles is shown in Fig. 12-
9b and the switching waveforms are shown in Fig. 12-9c. Using Eq. 12-l
(12-4)
(12-5)
( 12-6)
where kpwm is the gain by which the control voltage is amplified by this switch-mode
converter to apply the voltage at the terminais of this dc machine. This output voltage
can be controlled in a range from (-Vd ) to Vd . Thus, a four-quadrant converter IS
(12-7)
Using the values of the duty-ratios of the two poles, and recognizing the directions with
which the currents are defined,
(12-8)
12-6
Tbus, using Eq. 12-1 to aehieve d A and d B in terms of vc(t) and substituting them in
Eq. 12-7 along with the currents from Eq. 12-8:
+
A
B
N
vJt) , ,, t
,,, ,
(a)
vo(t) I
~ i !
,,
I
,
,
tt
,
:,
,
;0 IX Motor io(t)
idO
~ --=--=- --------: Or>----+--+-~f---4--_i__+_- t
+ v...ft):
, ::
,-------------" idA(t) - - - - ~
BW ~N~) n--~~-_+_:~-~f___L___ t
()--~
, -7,--i---+--~,f---+-
, - t
Vc . A (I) .
I dB
()-- ~T--
t :: ,::
--~ ~-- ~
n f1t= ~
I I I ' ,
I I , ' ,
"" I ,. '" ,
(h)
id (t)o;1 t;j + t-.
(c)
The instantaneous quantities assoeiated with this switeh-mode de-de converter are
pulsating, as shown in Fig. 12-ge. Based on the diseussion of harmonic eomponents in
the voltage at the eurrent-port and in the eurrent at the voltage-port of eaeh power-pole,
we ean eaIculate eaneellation of some harrnonies in the eombination of the two power-
poles in a de-drive converter. A detailed diseussion ofthis ean be found in referenee [1].
12-7
12-4 AC-MOTOR DRlVES
Ac-motor drives are the workhorse of industry. These three-phase motors are controlled
in speed and position by applying adjustable-magnitude, and adjustable-frequency ac
voltages by three bi-directional switching power-poles, as shown in Fig. 12- I Oa. Their
average representation is shown in Fig. 12- I Ob.
A '-=--I---I-~""-{
+
.
acmotor
(b)
The resulting voltage waveforms are shown in Fig. 12-1 I. Using Eq. 12-10 into
equations similar to Eq. 12- I for ali three phases produces the following duty-ratios in the
average representation shown in Fig. 12-10b, where the control voltages and the duty-
ratios are plotted in Fig. 12-12:
12-8
38bl_
D DDDDD DDn~ ~ ~ ~~n ~
v". i~ I-, - - - -~,-,- - - ~'-'- - - -~~
' ~ D~
D - - -~'- '- - -~
-'------'----'-
o 0-----,-,-------
0 -'O --------'-D---"--'
-L- D------'-0----,------,---
0
where the pole gain kpo'c is the sarne as in Eq. 12-2, and it amplifies the control voltage
for each switching power-pole. These voltages are plotted in Fig. 12-12. The average
output voltages of the switching power-poles have the sarne offset voltage, equal to
0.5 Vd These equal offset voltages in each phase are considered zero-sequence voltages,
which cannot cause any current to flow when applied to a balanced three-phase motor
load shown in Fig. 12-IOa or b.
12-9
0.5
OL-------------------------------------__
t
Vd
2
t
OL-------------------------------------_
Figure 12-12 Duty-ratios and the average output voltages ofthe power-pole.
Henee the de eomponent in eaeh phase voltage (terminal to neutral n) of the motor load
equals zero. As a eonsequenee in Fig. 12-10b, the average phase-neutral voltages aeross
the motor phases, with respeet to the motor neutral n, are as intended
(12-13)
(12-14)
12-10
We should note that at high switching rrequency Is relative to the fundamenta l
rrequency j, being synthesized, average voltages in Eq. 12-13 equal the fundamental-
frequency components (v Anl' VBnl ' VCnl ) in the voltages applied to the motor phases (at
the terminais with respect to the motor-neutral n).
To obtain the average currents drawn rrom the voltage-port of each switching power-
pole, we will assume the average currents drawn by the motor load in Fig. 12-10b to be
sinusoidal but lagging with respect to the average voltages in each phase by an angle tA ,
~(t) = j sin(w,t - tA)
T,;(t)=jsin(w,t-tA -120) (12-15)
I.;(t) = j sin(w,t - tA - 240)
Therefore in Fig. 12-10b, the average currents drawn from the voltage-port are
(I2-16)
The total average current drawn by the converter from the dc-side voltage source is the
sum ofthe three currents above:
(12-1 7)
ln Eq . 12-15, applying Kirchoff's law to average currents at the motor-neutral, the sum of
three average currents equals zero,
(12-18)
Therefore, substituting Eqs. 12-16 and making use ofEq. 12-18 in Eq. 12-17
_id(t) = 0.5-,!i-
V I,[Sin(w,t)Sin(liV-tA)+sin(llJlt- 1200)Sin(llJlt-tA -1200)] (12-20)
V;ri +sin(w,t-2400)sin(w,t-tA -240 )
12-11
that simplifies to a dc current
- 3 V _
iA/) = I d = --"E-I costA (12-21)
4 V,ri
(12-22)
which shows that the average power input at the voltage-port equals the total three-phase
power out ofthe current-port in this converter that is assumed loss-Iess.
ln Fig. 12-IOa, the instantaneous quantities associated with this switch-mode converter
are pulsating, as shown in Fig. 12-11. Based on the discussion of harmonic components
in the voltage at the current-port and the current at the voltage-port of each power-pole,
we can calculate cancellation of some harmonics in the combination of the three power-
poles in this ac-drive converter. A detailed discussion of this can be found in reference
[ 1].
12-4-1 Space Vector Pulse Width Modulatioo
The above method of synthesizing the three-phase sinusoidal output is called the
sinusoidal-PWM. As discussed in the appendix to this chapter on the accompanying CD,
the sinusoidal-PWM is near ideal, except it does not utilize the available dc-link voltage
to its fullest. A modified PWM technique called the space vector PWM (SV-PWM) is
described in the Appendix on the accompanying CD, which can produce from the sarne
dc-link voltage a three-phase output voltage that is higher by approximately fifteen
percent.
12-5 VOLTAGE-LINK STRUCTURE WITH BI-DIRECTIONAL POWER
FLOW
ln many applications of ac-motor dr ives, the voltage-link structure in Fig. 12-1 is such
that the power tlow through it is bi-directiooal. Normally, power tlows from the utility to
the motor, and while slowing down, the energy stored in the inertia of machine-Ioad
combination is recovered by operating the machine as a generator and feeding power
back into the utility grid. This can be accomplished by using three-phase converters
discussed in section 12-4 at both ends, as shown in Fig. 12-13a, recognizing that the
power tlow through these converters is bi-directional. ln the normal mode, converter at
the utility-end operates as a rectifier and the converter at the machine-end as an inverter.
The roles of these two are opposite when the power tlows in the reverse direction during
energy recovery.
12-12
The average representation of these converters by means of ideal transformers is shown
in Fig. 12-13b. ln this simplified representation, where los ses are ignored, the utility
source is represented by a source v ,a ' etc. in series with the internal system inductance
L, . The machine is represented by its steady state equivalent circuit, the machine
equivalent inductance L in series with the back-emf e A , etc . Under balanced three-
phase operation at both ends, the role of each converter can be analyzed by means of the
per-phase equivalent circuits shown in Fig. 12-13c. ln these per-phase equivalent
circuits, the fundamental-frequen cy voltages produced at the ac-side by the two
converters are v aI and v An l , and the ac-side currents at the two sides can be expressed in
the phasor form as
(12-23)
(12-24)
where, the phasors in Eq. 12-23 represent voltages and current at the utility-system
frequency IV, typically 60 (or 50 Hz), and the phasors in Eq. 12-24 at the fundamental-
frequency cq synthesized by the m achine-side converter.
Tal
(c)
12-13
ln Fig. 12-13a, for a given utility voltage, it is possible to control the current drawn from
the utility by controlling the voltage synthesized by the utility-side converter in
magnitude and phase. ln these circuits, if the losses are ignored, the average power
drawn from the utility source equals the power supplied to thc motor. However, the
reactive power at the utility-s ide converter can be controlled, independent1y of the
reactive power at the machine-side converter that depends on the motor load.
Energy
Storage
12-14
consists of two switching power-poles, where as shown, the inductance of the low-pass
filer establishes
the current-ports of the two power- poles. The control voltage is at the desired output
frequency to produce the desired output voltage amplitude:
+ .J "
~ ----- ---- '
y
, +
+
.. -b '
-G-j~------ f-_ .!:
V,O " :v Critical v,
'
,
o
T Y
,
,
Load 0 Criticai
Load
- I
(a) (b)
q, -
q.
This control signal equals V ,.A. For the switching power-pole B, a phase shift of 1800 IS
introduced in Eq. 12-25, that is by a factor of (-I) , to produce its control signal
(12-26)
As shown in Fig. 12-16a, these control signals are compared with a triangular waveform
at the switching frequency, typically above 20 kHz to produce the switching functions for
the transistors in the switching power-poles. The switching waveforms are shown in Fig.
12-16a.
v,
v",.. (t) o.
o
/ \
o r-------+-------~------~---
\ /
v. (t)
\ /
'--' (b)
12-15
dA(I) = 0.5 + 0.5 ~c sin li'"
V,ri (12-27)
d.(t) = 0.5 - 0.5 ~c sin w,t
V,ri
With these duty-ratios, the average outputs of the two power-poles and the average output
voltage vo(t) are as follows, plotted in Fig. 12-16b
where,
(12-29)
ln order to calculate the average current drawn from the dc source, we will assume that
the average ac-side current is sinusoidal and lagging behind the output ac voltage by an
angle tA, as sbown in Fig. 12-17:
(12-30)
.0
(12-31)
12-16
which shows that the average current drawn from the dc source has a dc component Id
that is responsible for the average power transfer to the ac side of the converter, and a
second harmonic component id 2 (at twice the frequency of the ac output), which is
undesirable. The dc-link storage in a 1-phase inverter must be sized to accommodate the
f10w of this large ac current at twice the output frequency, similar to that in PFCs
discussed in Chapter 6. Of course, we should not forget that the above discussion is in
terms of average representation of the switching power-poles. Therefore, the dc-link
storage must also accommodate the flow of switching-frequency ripple in id discussed
below.
The switching waveforms in the single-phase UPS are shown in Fig. 12-16a, which
confirm that the voltage vo(t) and id(t) are pulsating at the switching frequency. A low-
pass filter is necessary to remove the output voltage harmonic frequencies, which were
discussed earlier in a generic manner for each switching power-pole. The pulsating
current ripple in iAt) can be by-passed from being supplied by the batteries by placing a
small high quality capacitor with a very low equivalent series inductance.
REFERENCES
I. N. Mohan, T . M. Undeland, and W.P. Robbins, Power Electronics: Converters,
Applications and Design, 3 m Edition, Wiley & Sons, New York, 2003.
PROBLEMS
Switching Power-Pole
12-1 ln a switch-mode converter pole-A, Vd = 150V, V;rl = 5V, and J, = 20kHz.
Calculate the values of the control signal vc A and the pole duty-ratio dA during
which the switch is in its top position, for the following values of the average
output voltage: V AN = 125V and V AN = 50V .
12-2 ln a converter pole, the iA(t) waveform is as shown in Fig. 12-5b. lncluding the
ripple, show that the current relationship of an ideal transformer is valid.
DC-Motor Drives
12-3 A switch-mode dc-dc converter uses a PWM-controller IC which has a triangular
waveform signal at 25 kHz with V;,j = 3 V. If the input dc source voltage
Vd = 150V , calculate the gain k pWM in Eq. 12-6 ofthis switch-mode amplifier.
12-17
12-4 ln a switch-mode dc-dc converter, VC I V", = 0.8 with a switching frequency
is = 20kHz and Vd = IS0 V. Calculate and plot the ripple in the output voltage
V o (I) .
12-S A switch-mode dc-dc converter is operating at a switching frequency of 20 kHz,
and Vd = ISO V. The average current being drawn by the dc motor is 8.0 A. ln
the equivalent circuit of the dc motor, Eo = I OOV , Ro = 0.2S0 ,and Lo = 4mH .
(a) Plot the output current and calculate the peak-to-peak ripple, and (b) plot the
dc-side current.
12-6 ln Problem 12-S, the motor goes into regenerative braking mode. The average
current being supplied by the motor to the converter during braking is 8.0 A. Plot
the voltage and current waveforms on both sides of this converter at that this
instant. Calculate the average power f10w into the converter.
12-7 ln Problem 12-S, calculate i,;,., ],;0' and ],;(= l d)
12-8 Repeat Problem 12-S if the motor is rotating in the reverse direction, with the
same current draw and the sarne induced emf Eu value of the opposite polarity.
12-9 Repeat Problem 12-8 if the motor is braking while it has been rotating in the
reverse direction. It supplies the same current and produces the sarne induced
emf Eo value ofthe opposite polarity.
12-10 Repeat problem 12-S if a bi-polar voltage switching is used in the dc-dc
converter. ln such a switching scheme, the two bi-positional switches are
operated in such a manner that when switch-A is in the top position, switch-B is
in its bottom position, and vice versa. The switching signal for pole-A is derived
by comparing the control voltage (as in Problem 12-S) with the triangular
waveform.
VAn(t) = ; 0.7Ssin(/XI,t) , where Vd =3 00V and /XI, = 2tr x4S rad ls. The
12-18
ean be represented as eA(t) = 106.14sin(w,t-6.6 )V. (a) Caleulate and plat
d A(t), dB(t) , and d e (t), and (b) sketeh 7;.(t) and t.;,,(t).
12-14 ln Problem 12-13, ealeulate and plat 7;(t), whieh is the average de eurrent drawn
fram the de-side.
Single-Phase UPS
12-19
Chapter 13
13-1 INTRODUCTION
Many applications, such as robotics and factory automation, require precise control of
speed and position. ln such applications, a feedback control, as illustrated by Fig. 13-1,
is used. This feedback control system consists of a power-processing unit (PPU), a
motor, and a mechanical load. The output variables such as torque and speed are sensed
and are fed back, to be compared with the desired (reference) values. The error between
the reference and the actual values are amplified to control the power-process ing unit to
minimize or eliminate this error. A properly designed feedback controller makes the
system insensitive to disturbances and changes in the system parameters.
error
error p I I output
des ired Electri c Mec h
'0-- p I I
I I
(reference) amplifier U I Machine I Load
sigo ai
~
ElectricaJ Mechanical
Systern System
The control system in Fig. 13-1 is shown simplified in Fig. 13-2, where Gp(s) is the
Laplace-domain transfer function of the plant consisting of the power-processing unit, the
motor, and the mechanical load. G Js ) is the controller transfer function. ln response to
a desired (reference) input X'(s) , the output of the system is X(s) , which (ideally)
equals the reference input. The controller Gc(s) is designed with the following
objectives in mind:
13-1
a zero steady state error.
a good dynamic response (which implies both a fast transient response, for example
to a step-change in the input, and a small settling time with very little overshoot).
Controller Plant
X'(s)
-, ""I Ge(s)
,I Gp (s)
I
, X(s)
I
Fig 13-2 Simplified control system representation.
To keep the discussion simple, a unity feedback will be assumed. The open-Ioop transfer
function (incIuding the forward path and the unity feedback path) G OL (s) is
(13-1 )
X( s)
The cIosed-loop transfer function in a unity feedback system is
X'(s)
(13-2)
ln order to define a few necessary control terms, we will consider a generic Bode plot of
the open-Ioop transfer function G OL (s ) in terms of its magnitude and phase angle, shown
in Fig. 13-3a as a function of frequency. The frequency at which the gain equals unity
(that is IGOL (s)1 = Odb) is defined as the crossover frequency fc (angular frequency me)
At the crossover frequency, the phase delay introduced by the open-Ioop transfer function
must be less than 180 in order for the closed-Ioop feedback system to be stable.
Therefore, at fe' the phase angle ~oL IJ; ofthe open-Ioop transfer function, measured with
(13-3)
Note that ~oLlf. has a negative value. For a satisfactory dynamic response without
oscillations, the phase margin should be greater than 45, preferably cIose to 60 .
The magnitude of the closed-Ioop transfer function is plotted in Fig. 13-3b (idealized by
the asymptotes), in which the bandwidth is defined as the frequency at which the gain
drops to (- 3 dB). As a Irrst-order approximation in many practical systems,
13-2
100
50
50
100
OdB1:====~==:s~3.il~'@
I
BW
3dB
; 0 1-
150
(b)
, ""'''' """ !
O--'---........"'-'-""-t-
-200 '-.'2 -'--'--'-'-.......ll....,, -'--'-'-'-'-'=-;; """
, -'--'-...............-'J 2
10 10 10 10 10
(a) ----..
For a fast transient response by the control system, for example a response to a step-
change in the input, the bandwidth of the closed-Ioop should be high. From Eq. 13-4,
this requirement implies that the crossover frequency f c (of the open-Ioop transfer
function shown in Fig. 13-3a) should be designed to be high.
13-3 CASCADE CONTROL STRUCTURE
ln the following discussion, a cascade control structure such as that shown in Fig. 13-4 is
used. The cascade control structure is commonly used for motor drives because of its
f1exibility. It consists of distinct control loops; the innennost current (torque) loop is
followed by the speed loop. If position needs to be controlled accurately, the outennost
position loop is superimposed on the speed loop. Cascade control requires that the
bandwidth (speed of response) increase towards the inner loop, with the torque loop
being the fastest and the position loop being the slowest. The cascade control structure is
widely used in industry.
torque speed
pos; l ; o n 0 -
+ I: Pos ition
controller
- .
~~
+
-
Speed
controller
~~
+
-
Torque
controller - Electrical
System
E~
H
-
Mech
System
hiJ POS ;I; on
torque(current)
speed
position
13-3
13-4 STEPS lN DESIGNING THE FEEDBACK CONTROLLER
Motion control systems ofien must respond to large changes in the desired (reference)
values of the torque, speed, and position. They must reject large, unexpected load
disturbances. For large changes, the overall system is ofien nonlinear. This nonlinearity
comes about because the mechanical load is ofien highly nonlinear. Additional
nonlinearity is introduced by voltage and current limits imposed by the power-processing
unit and the motor. ln view of the above, the following steps for designing the contro!ler
are suggested:
I. The first step is to assume that, around the steady-state operating point,
the input reFerence changes and the load disturbances are ali small. ln
such a small-signal analysis, the overall system can be assumed to be
linear around the steady-state operating poin!, thus allowing the basic
concepts of linear control theory to be app lied.
2. Based on the linear control theory, once the controller has been designed,
the entire system can be simulated on a computer under large-signal
conditions to evaluate the adequacy of the controller. The controller
must be "adjusted" as appropriate.
13-5 SYSTEM REPRESENTATlON FOR SMALL-SIGNAL ANALYSIS
For ease of the analysis described below, lhe system in Fig. 13-4 is assumed linear and
the steady-state operating point is assumed to be zero for ali ofthe system variables. This
linear analysis can be then extended to nonlinear systems and to steady-state operating
conditions other than zero. The control system in Fig. 13-4 is designed with the highest
bandwidth (associated with the torque loop), which is one or two orders of magnitude
smaller than the switching frequency .f,. As a result, in designing the controller, the
switching-frcqucncy components in variou s quantities are of no consequence. Therefore,
we will use the average variables discussed in Chapter 3, where the switching-frequency
components were eliminated.
13-5-1 The Average Representation of the Power-Processing Unit (PPU)
For the purposes of designing the feedback controller, we will assume that the de-bus
voltage Vdc within the PPU shown in Fig. 13-5a is constant. Following the averaging
analysis in Chapter 3, the average representation of the switch-mode converter is shown
in Fig. 13-5b. ln terms ofthe dc-bus voltage V'c and the triangular-frequency waveforrn
peak v" p the average output voltage vaU) oFthe converter is linearly proportional to the
control voltage:
(k PIPA., -_ -V'c
.-
) ( 13-5)
v,,,
13-4
where k l'WM is the gain constant of the PWM converter. Therefore, in Laplace domain,
the PWM controller and the dc-dc switch-mode converter can be represented s imply by a
gain-constant k pWM , as shown in F ig. 13-5c:
(1 3-6)
where Va( s) is the Laplace transfo rm of Va(t) , and v,,(s) is the Laplace transform of
vc(t) The above representation is valid in the linear range, where -v,,, ,;:; Vc ,;:; v,,, .
+
+
A B Va
=- iB
(a)
(c)
Figure 13-5 (a) Switch-mode conve rter fo r dc motor drives; (b) average model of
the switch-mode converter; (c) linearized representation.
13-5
13-5-2 Modeling of the De Machine and the Mechanical Load
The dc motor and the mechanical load are modeled as shown by the equivalent circuit in
Fig. 13-6a, in which the speed (J)m(t) and the back-emf ea(t) are assumed not to contain
switching-frequency components. The electrical and the mechanical equations
corresponding to Fig. 13-6a are
(13-7)
and
(13-8)
where the equivalent load inertia J ,q (= J M + J L ) is the sum of the motor inertia and the
load inertia, and the damping is neglected (it could be combined with the load torque TL ).
ln the simplified procedure presented here, the controller is designed to follow the
changes in the torque, speed, and position reference values (and hence the load torque in
Eq. 13-8 is assumed to be absent). Eqs. 13-7 and 13-8 can be expressed in the Laplace
domain as
(13-9)
or
(13-10)
T
La
=- (13-11)
, R
a
(13-12)
From Eq. 13-8, assuming the load torque to be absent in the design procedure,
(13-13)
Eqs. 13-10 and 13-13 can be combined and represented in block-diagram form, as shown
in Fig. 13-6b.
13-6
7 = 1'em
kr
+ TL(s)
~
R.
VQ(s) 1.(5~ Ior
(UM 5
+ s./,
v.
6.(5)
kE
eQ _ k~ IV..
(a) (b)
Fig 13-6 De motor and mechanicalload (a) equivalent circuit; (b) block diagram.
(13-14)
where kp is the proportional-contro ller gain. ln torque and speed loops, proportional
controllers, if used alone, result in a steady-state error in response to step-change in the
input reference. Therefore, they are used in combination with the integral controller
described below.
ln the integral controller shown in F ig. 13-7, the o utput is proportional to the integral of
the error E(s), expressed in the Laplace domain as
(13-15)
where k, is the integral-controller ga in. Such a controller responds slowly because its
action is proportional to the time integral of the error. The steady-state error goes to zero
13-7
for a step-change in input because the integrator action continues for as long as the error
is not zero.
Vc, p(s)
kp
G p (.1")
x '(s)
+ '-
!L V<.I (s~~ X(s)
s
- E(s)
Vc(s)
(13-16)
La 5.2mH
J ,q 152 X 10 6 kg m '
B O
KE O.IV /( rad /s)
kT O.I N m / A
Vd 60V
V,ri 5V
/, 33kHz
13-8
We will design the torque, speed, and position feedback controllers (assuming a unity
feedback) based on the small-signal analysis, in which the load nonlinearity and the
effects ofthe limiters can be ignored.
13-7-1 The Design ofthe Torque (Current) Control Loop
As mentioned earlier, we will begin with the innermost loop in Fig. I3-8a (utilizing the
transfer function block diagram of Fig. 13-6b to represent the motor-Ioad combination,
Fig. 13-5c to represent the PPU, and Fig. 13-7 to represent the PI controller).
PI
(a)
v, (S)
1;(S~
~ 11 R" 10 (s) I
+ '-~ PI f--- k pWM
l +sTe
kT f..L- sJ,q
~
- -
Ea(s)
k EkT
I---
la(s) sJ,q
(b)
PI
V,(s )
J;(s) II R" la(s)
+ :::~
kil (1+killk J
S
S
pJ
kpWM
Va(s)
I+STe
l a(s)
(c)
Fig. 13-8 Design ofthe torque controlloop.
ln permanent-magnet dc motors in which ,pj is constant, the current and the torque are
proportional to each other, related by the torque constant k T Therefore, we will consider
the current to be the control variable because it is more convenient to use. Notice that
there is a feedback in the current loop from the output speed. This feedback dictates the
induced back-emf. Neglecting 7;., and considering the current to be the output, E a (s)
13-8a can be redrawn as shown in Fig. 13-8b. Notice that the feedback term depends
13-9
inversely on the inertia J ,q' Assuming that the inertia is sufficient1y large to justifY
neglecting the feedback effect, we can simplifY the block diagram, as shown in Fig. 13-
8c.
The current-controller in Fig. 13-8c is a proport.i onal-integral (PI) error amplifier with the
proportional gain k p , and the integral gain ki/' lts transfer function is given by Eq. 13-
16. The subscript "I" refers to the current loop. The open-Ioop transfer function
G, OL(S) ofthe simplified current loop in Fig. 13-8c is
To select the gain constants of the PI controller m the current loop, a simple design
procedure, which results in a phase margin of 90 degrees, is suggested as follows:
Select the zero (kil / k p , ) ofthe PI controller to cancel the motor pole at (I Ir, ) due to
the electrical time-constant T, of the motor. Under these conditions,
or (13-18)
Cancellation of the pole m the motor transfer function renders the open-Ioop transfer
function to be
=k,.OL (13-19a)
G IOL ()
S
. S
where
(13-19b)
ln the open-Ioop transfer function of Eq. 13-19a, the crossover frequency m" = k , OL .
We will select the crossover frequency h ,(= m" / 27t) ofthe current open-Ioop to be
approximately one to !WO orders of magnitude smaller than the switching frequency
of the power-processing unit in order to avoid interference in the control loop from
the switching-frequency noise. T herefore, at the selected crossover frequency, from
Eq.13-19b,
k - me/ R. (13-20)
i/ -k
PWM
13-10
This completes the design of the torque (current) loop, as illustrated by the example
below, where the gain constants k p, and k" can be calculated from Eqs. 13-18 and 13-
20.
~ Example 13-1 Design the current loop for the example system of Table 13-1,
assuming that the crossover frequency is selected to be 1 kHz.
3
Solution From Eq. 13-20, for OJ" = 2" x 10 rad / s ,
= 1050.0
o ,, ,, ,, ,
M,,~~~~~~~~~~~~
~40
I 1 1 11 11 11 1 11 1 11111 I 1 11 1 111
<
11 111111
11 111 111
1 1 111111
1 1 111111
I I II .
1 1 11 1111
I I 11 111 1
11 111111 ,
1 1 11 111
1 11 1111
~ -10
2O nur
I I 11 1111
, ,, ,,
-I..J. 1-01 U.III- _I -l .1-1 UIIL _ 1-1 .J.IU III. _ I ..J.1.1J1ll_1- 1-l1.J.1
,
,
~ ,
- .., TI - I - ' , I nm - r i . fTl 1 1 1 11111 1 11 11111 1 1 11 1111 11111 1 11 1111
1 1 111 111 1 11 11111 1 11 11111 1 11 1111
'2 o
1 11111 111 1 1 1111111
- 111111111 -11 flfill! -' - I T in i'
I III 1 1 1 1 11111 1 1 11 1 111
-'- ,"'T 'l lIii - 1-l llTl
~ 20 , ,, ,, ' " '"
--1.1 LJ U.III- _I...J .1-1 U IIL _ 1--1 .J. I U III. _ 1_ 1..I.1.1J _Li...JUI
f -20
_'_ I --.!I~I I I
I I I 11 111 :>
1 1 11 11 11
1 1 111 111
1 11 11111
1 1 111111
11 1111 11
11 1111i1
11 111111
11 111111
1 11 1111
1 11 1111
4O ~LU~~~Wlli~~~~Lll~~UW~ -30
10f) 10' 102 103 10 10~ 10' 10' 10' 10' 10' 10'
~ r-__________~
F~~ ~)~________~
~~Y~~ Frequency (Hz)
I 11 111111 1 1111111 I 1111 111 1 111 1 1111 I 1 1 11 111
O
, 1 1 111111 I I 11111 I I 11 1111
f _' .1 U I.!IIL _I.J LI UlIl. _,_, .i I!."!!. _' _ I .lIJ.J lll _ LI...J '1.1 I 1 1 111111 I I 111 111 I I 11 1111
;-
-89.5
-90
I I I 111111
111 1 1 1111
I I I 111111
I 1 111111 1
"""1
I I I 111 111
I ' 1 11 11' 1
I"",
I I I 111111
"""i
I 1 11 1 1111
I I I 11111
I 1 11 11 11 f -50
1 1 111111 1 11 11111
, I I ' 11 111
'1 11111
,, ,,
i i illi ,ii i i i' i i r " ,TI
~ -90. 5 I I I 1 1 1111 I I I 111111 I 1 I 111 11 1 I I I 1 1 1111 I r I 1 1 111
Ii I 111 1111 I I 111 11 1
a. -I "i 11 H-Ilj- -1 -1
I I I 111111
ri HItT-
I I 111111
- I"'" "i ' H II1"
I I 111 11 1
- I -I "i lti ltt -
I I 111111
1- 1--1 !TI
I I I 11111
~
, 11 111111
I 11 11 111 , I 11 11 11 1
I 111111 1
, 11 1 '111
, 11 11111
91 100
10' 10' 10' 10' 10' 10' 10' 10' 10'
Frequencv (HZ)
(a) (b)
Figure 13-9 Frequency response ofthe current loop (a) open loop; (b) closed
loop.
The open-Ioop transfer function is plotted in Fig. 13-9a, which shows that the crossover
frequency is 1 kHz, as assumed previously. The closed-Ioop transfer function is plotted
in Fig. 13-9b. ~
13-11
GO.OL (S) = k,o [1 + s I(k,o I kpn )] ! l:.L (13-21)
, S , currenf loop sJ'"I
PI co~u'()lJel'
------
lorque +mu1ia
'(s)
+ "-
-
O- PI I';(S , I'(S) kT i -
,
sJ. q
ilIm{S)
com(s)
This shows that the open-Ioop transfer function consists of a double pole at the origino At
low frequencies in the Bode plot, this double pole at the origin causes the magnitude to
decline at the rate of - 40 db per decade while the phase angle is at -180 . We can
select the crossover frequency w'" to be one order of magnitude smaller than that of the
current loop. Similarly, we can choose a reasonable value of the phase margin I/Jpm.o '
Therefore, Eq. 13-22 yields two equations at the crossover frequency:
=1 (13-23)
and
The two gain constants of the PI controller can be calculated by solving these two
equations, as illustrated by the following example.
'" Example 13-2 Design the speed loop controller, assuming the speed loop crossover
frequency to be one order of magnitude smaller than that of the current loop in Example
13-2; that is, f cn = 100Hz, and thus w'" = 628rad I s. The phase margin is selected to
be 60 .
and I/JPM.o = 60 at the crossover frequency, where s = )w", = )628 , we can calculate that
13-12
kpn =0.827 and k,n =299.7.
150,-;-,-'-,:c",::::",,'--;-,-'-"C:"::::,,,,'--;-,-:C"C:";;',,,,---'-,CC"",,"::::,,,---'-,-:C,,"'"::::,,,--:-,
, :C"-",,,eJ" '" r~,~,~,,~,,",,-'-,~,~,,~,,",,~,~,~"~,~,,.'-~,~,~"~"",,,-',~,~,~,,~,,'-,,-:-,~,~"~,,::l,,
~ 100 : ~I!.!':* -:~ H:~::- ~:~ :*::~ ~ -i-:H:m-:-:H:*::- TH:~:: ~ o
I t IIIIU 1 1111 111 11 1111111 ' 1 1111111 1 1 1111 11
IIIIIU I IIII II! I I I I I'IA -tllllii,-llli'illl I 11i'i1l1
' I t 11111
{l-20 '"11111 " "111 11 11111111 1 "11 11 1 11111111 11111111
:a
~
S'J
o
111111111 1 11 111111
HHII+! - 1--1 H~II -
1 111 11 11
+- HI~I l..,
,. r i ",n - ,-, ntnl,- r n mi" ..., T rrl
11I1I1i1 ""11111 11111 11. 111 111111
-+ HHIH - I .... H ,''',,- +-HI+I I .... -I-II-+IHII-i-I-+I
- rITITl1!r ,- TIni ,
I '1111" 11111111
-I +-1 ... ,
aO~'~LWW,~0~'~.w.W,rJ
11111111 1111111
'-,-~=,WO~'~~~'~0';-'-~"""'0'
"1111'" "'"111
1.,00
11111111
_-' _LUWI 2i 11111111 '11111" " 1 1 , . , , ,111111 111111111 11111111
1 11111 11 I IIIIUII I 111 1111 ~ 11111111 11111101 11111' 111111111 1111' 11 11 11111111
e -BD Tnnln- " nI1Tl1-rnm ,TrTUm-rITlnllri'Tlnl
j 1~
o. I I 1111111 ' I 1111 11 I 111111
~
......
11111 111
I I I " ' II
11 1111111
I I 1111111
1111111
I I 11111'
111 111111
I 1111111
11111111111111111
" I 111111 I I 111111
.ooL:::~::::::~"~~~~~"~'~"'~"~'~~~~'~"~"~"
"IIIUI 1111 11111 1111 1 111 I 11111111
~ ~ ~ ~ ~ ~ ~ 10' 10'
FA3QJI3IlCY (Hz) F~(Hz1
(a) (b)
Figure 13-11 Speed loop response (a) open loop; (b) closed loop.
The open- and the closed-Ioop transfer functions are plotted in Figs. 13- I I a and 13- I I b.
(13-25)
O:(S)
+\..
o--- k.
aJ~(s)
J
aJ".(s)
-
J
S
m ( s)
( 13-26)
13-13
.. Example 13-3 For the example system of Table 13-1 , design the position-Ioop
controller, assuming the position-Ioop crossover frequency to be one order of magnitude
smaller than that of the speed loop in Example 13-3 (that is, f di = 10Hz and
llJdI = 62.8 rad / s).
50
, I I I II HI I I I 111m I I I ti til ' I I 11111 o r --;-,OC":C,,;;,,,-C,C 7," ,",f',;;:::-;-;~::--;-;-cc;;c--;-,--:,--:,:7";;;],,,
iii I I 1 111111 I 1 1 1 1111 ' 1111111 '111lil! I 111111>, I 'I 1111
"- 1 1 1 111111 I "IIIU, I 111111 ...J J. Ul1l1'
t
,.
o
.5Q
1
I 1 1 111111
I 1 1 1 Hill
1111 1111
I 1 , 11 1111
I 11 111111 1 ' 1' 1"
I ' II IHI'
I 1 11 11'
1 11 11'11
11 1 ' 1,11
,
, 1 1111111 , I 111111 , 11111111 , '11 11 "
11 1111
(a) (b
Figure 13-13 Position loop response (a) open loop; (b) c10sed loop.
The open- and the c1osed-loop transfer functions are plotted in Figs. 13-l3a and 13-13b . ..
REFERENCE
1. M. Kazmierkowski and H. Tunia, "Automatic Control of Converter-Fed Drives,"
Elsevier, 1994, 559 pages.
PROBLEMS
13-1 ln a unity feedback system, the open-Ioop transfer function is of the form
k
GOL(s ) = . Calculate the bandwidth ofthe closed-Ioop transfer function.
l+s / llJp
How does the bandwidth depend on k and llJp ?
13-2 ln a feedback system, the forward path has a transfer function of the form
k
G(s) = , and the feedback path has a gain of kfb which is less than
l+s /llJp
unity . Calcu late the bandwidth of the closed-Ioop transfer function. How does
the bandwidth depend on k fb ?
13-3 ln designing the torque loop of Examp le 13-1 , include the effect ofthe back-emf,
shown in Fig. 13-8a. Desig n a PT controller for the sam e open-Ioop crossover
frequency and for a phase margin of 60 degrees. Compare your results with
those in Example 13-1.
13-14
13-4 ln designing the speed loop of Exarnple 13-2, include the torque loop by a first-
order transfer function based on the design ln Exarnple 13-1. Design a PI
controller for the sarne open-Ioop crossover frequency and the same phase
margin as in Exarnple 13-2 and compare results.
\3-5 ln designing the position loop ofExample 13-3, include the speed loop by a first-
order transfer function based on the design ln Example 13-2. Design a P-type
controller for the sarne open-Ioop crossover frequency as in Example 13-3 and
for a phase margin of 60 degrees. Compare your results with those in Exarnple
13-3.
\3-6 ln an actual system in which there are Iimits on the voltage and current that can
be supplied, why and how does the initial steady-state operating point make a
difference for large-signal d isturbances?
13-7 Obtain the time response of the system designed in Example 13-2, in terms of the
change in speed, for a step-change ofthe load-torque disturbance.
13-8 Obtain the time response of the system designed in Example 13-3, in terms of the
change in position, for a step-change of the load-torque disturbance.
13-9 ln the example system of Table \3-1, the maximum output voltage of the de-de
converter is limited to 60 V. Assume that the current is limited to 8 A in
magnitude. How do these two limits impact the response ofthe system to a large
step-change in the reference vai ue?
13-10 ln Exarnple 13-2, design the speed-Ioop controller, without the inner current
loop, for the same crossover frequency and phase margin as in Exarnple 13-2.
Compare results with the system of Example 13-2.
13-15
Chapter 14
THYRISTOR CONVERTERS
14-1 INTRODUCTION
Historically, thyristor converters were used to perform tasks that are now performed by
switch-mode converters discussed in previous chapters. Thyristor converters are now
typically used in utility applications at very high power leveIs. ln this chapter, we will
examine the operating principIes of thyristor-based converters.
14-2 Thyristors (SCRs)
Thyristors are a device that can be considered as a controlled diode. Like diodes, they
are available in very large voltage and current ratings, making them atlractive for use in
applications at very high power leveIs.
Thyristors, shown by their symbol in Fig. 14-la are sometimes referred to by their trade
name of Silicon Controlled Rectifiers (SCRs). These are 4-layer (p-n-p-n) devices as
shown in Fig. 14-1 b. When a reverse (v AK < O) voltage is applied, the flow of current is
blocked by the junctions pn I and pn3. When a forward (v AK > O) polarity voltage is
applied and the gate terminal is open, the flow of current is blocked by the junction pn2,
and the thyristor is considered to be in a forward-blocking state. ln this forward-blocking
state, applying a small positive voltage to the gate with respect to the cathode for a short
interval supplies a pulse of gate current iG that latches the thyristor in its on state, and
subsequently the gate-current pulse can be removed.
A
A p
1----l....... p.~..I...
(a)
G~ K
(b) G
N
p
....._.p..~~.
1----l ........P.V.2.
N
14-1
voltage appears across the thyristor (anode A is positive with respect to cathode K), and if
the thyristor were a diode, a current would begin to flow in this circuito This instant
wl = O in this circuit we will refer as the instant ofnatural conduction. With the thyristor
forward blocking, the start of conduction can be controlled (delayed) with respect to
wt = O by a delay angle a at which instant the gate-current pulse is applied. Once in the
conducting state, the thyristor behaves Iike a diode with a very small voltage drop of the
order of 1 to 2 volts across it (we wi\l idealize it as zero), and the R-L load voltage vd
equals v" as shown in Fig. 14-2b.
i,
(a)
(b)
14-2
----+
id
-
Ls
is
1 3 1
is +
Vs
3
+
-rv vd Id
vd
4 2
'4 2
(a) (b)
Figure 14-3 Full-Bridge, single-phase thyristor converter.
mt
mt
I , 2 3, 4 I
Figure 14-4 Single-phase thyristor converter waveforrns.
Thyristors 1 and 2, and thyristors 3 and 4 are treated as two pairs, where each thyristor
pair is supplied gate pulses delayed by an angle a with respect to its instant of natural
conduction; mt = 0 0 for tbyristors 1 and 2, and 0Jt = 1800 for tbyristors 3 and 4, as shown
in Fig. 14-4. ln tbe positive half-cycle of the input voltage, thyristors 1 and 2 are
forward-blocking until they are gated at 0Jt = a when tbey immediately begin to
conduct, and thyristors 3 and 4 become reverse-blocking. ln tbis state,
These relationships hold true until a + 1r in the negative half-cycle of the input voltage,
when thyristors 3 and 4 are gated and begin conducting. ln this state,
Vd(1) = -v,(t) and i,(t) = -1d a+1r < mt"; a+21r (14-2)
which holds true for one half-cyc1e, until the next half-cycle begins with the gating of
thyristors 1 and 2.
14-3
Assuming the input ac voltage to be sinusoidal, the average value Vd of the voltage
across the dc-side of lhe converter can be obtained by averaging lhe vAI) waveform in
Fig. 14-4 ove r one half-cycle during a < wr ,., a + 1[ :
1 a+.>r 2 "
V, = -
1[
f V, sinwrd(lt) =-v,
a
"
1[
cosa (14-3)
On the ac-side, lhe input current i, waveform is shifted by an angle a wilh respect to lhe
input voltage as shown in Fig. 14-4, and the fundamental-frequency component i' l (I) has
a peak value of
- 4
1' 1 = -1, (14-4)
1[
ln terms of voltage and current peak values, the power drawn from the ac-side is
I - -
p = -v,
2
/ ' 1cosa (14-5)
Assuming no power loss in the thyristor converter, the input power equals lhe power to
the dc-side ofthe converter. Using Eqs. 14-3 and 14-4, we can reconfirm the following
relationship:
(14-6)
Power flow can be controlled by the delay angle a; increasing it toward 90 reduces lhe
average dc-side voltage Vd while simultaneously shifting lhe input eurrent i, (/)
waveform farther away wilh respect to lhe input voltage waveform. The de voltage as a
function of a is plotted in Fig. 14-5a and the eorresponding power direction is shown in
Fig. 14-5b. The waveforms in Fig. 14-4 show Ihat for the delay angle a in a range of 0
to 90, Vd has a positive value as plotted in Figs. 14-5a and b, and the converter operates
as a reetifier, wilh power flowing from the ac-side to lhe dc-side.
Delaying the gating pulse sueh that a is greater Ihan 90 in Fig. 14-4 makes V, in Eq.
14-3 negative, and the converter operates as an inverter as shown in Fig. 14-5b, with
power flowing from the dc-side to the ac-side. ln practiee, the upper limit on a in lhe
inverter is below 180, for example 160 lo avoid a phenomenon known as the
commutation failure, where the current fails to commulate fully from the conducting
Ihyristor pair to lhe next pair, prior to the instant beyond which the conducting pair keeps
14-4
on conducting for another half-cycle. This commutation-failure phenomena is fully
described in [I].
Rectifier
P=VdId = +
,j---+---:rr:.,,-~ a
,, ,,
1500., 1/800
,
, Inverter
" P=VdId =-
(a) (b)
Fig. 14-5 E ffect ofthe delay angle a.
Previously, our assumption was that the ac-side inductance L, equals zero. ln the
presence of this inductance, the input current takes a finite amount of time to reverse its
direction, as the current "commutates" from one thyristor pair to the next.
1 3 O
, iI
v
t i3 .
is--L + " t _ ~~s /
+ '\.t- j
vL +
vd ld iG ! b
. .
j n t
1:7:
'
4
1;4
~~)
2 ti2
tId \ t
! i \ I-1d
a: u :
(a) (b)
From basic principies, we know that changing the current through the inductor L, in the
circuit of Fig. 14-6a requires a finite amount of volt-seconds. The dc-side is stiH
represented by a dc current Id' The waveforms are shown in Fig. 14-6b, where
thyristors 3 and 4 are conducting prior to mI = a, and i, = -Id
At J{ a , thyristors 1 and 2, which have been forward blocking, are gated and hence
=
they immediately begin to conduct. However, the current through them doesn ' t jump
instantaneously as in the case of L , = O where i, instantaneously changed from C-Id ) to
14-5
(+ I d ). W ith a tinite L" during a short interval called the commutation interval u, ali
thyristors conduct, applying v, across L, in Fig. 14-6a. The volt-radians needed to
change the inductor current /Tom (- I d ) to (+ I d ) can be calcu lated by integrating the
inductor voltage v L (= L, . di, / dI) from a to (a + u), as follows:
The above volt-radians are "Iost" from the integral of the dc-side voltage waveform in
Fig. 14-6b every half-cycle, as shown by the shaded area in Fig. 14-6b. Therefore,
dividing the volt-radians in Eq. 14-7 by the 1C radians each half-cycle, the voltage drop
in the dc-side voltage is
(14-8)
This voltage is lost from the dc-side average voltage in the presence of L, . Therefore,
the average voltage is smaller than that in Eq. 14-3:
2 2
Vd =-V,cosa--JL,Id (14-9)
1C 1C
We should note that the voltage drop in the presence of L, doesn't mean a power loss in
L, ; it simply means a reduction in the voltage available on the dc-side.
The converter waveforms, where the delay angle a (measured with respect the instant at
which phase voltage waveforms cross each other) is zero are similar to that in diode
rectifiers discussed in Chapter 5 (see Figs. 5-11 and 5-12). The average dc voltage is as
calculated in Eq. 5-23, where VL" is the peak value ofthe ac input voltage:
(14-10)
14-6
va ia
-
id
J
J 3 5 P
L,
4 6 2
(a) (b)
Figure 14-7 Three-phase Full-Bridge thyristor converter.
Oelaying the gate pulses to the thyristors by an angle a measured with respect to their
instants of natural conductions, the waveforms are shown in Fig. 14-8, where L, = O .
wt = o
Figure 14-8 Waveforms with L, = O .
ln the de-side output voItage waveforms, the area Aa corresponds to " volt-radians loss"
due to delaying the gate pulses by a every 7r / 3 radiano Assuming the time-origin as
shown in Fig. 14-8 at the instant at whieh the phase voltagc waveforms cross, the line-
line voltage v Qe
waveform ean be expressed as Vu sin rol. Therefore from Fig. 14-8, the
drop LI. Va in the average dc-side voltage can be calculated as
(14-11)
14-4-1 Effect of L,
As shown in the waveforms of Fig. 14-9, the average de output voltage due to the
presence of L, is reduced by the area A" every 7r / 3 radiano Ouring the eommutation
interval u , from a to a + u , the instantaneous de voItage is redueed due to the vo1tage
drop across the inductanee in series with the thyristor to whieh the current is
14-7
commutating from o to (+Id ). Using the procedure ln Eq. 14-7 for single-phase
converters,
a+u Id
Au = J
a
VI. d(ml) = mL, Jdi, = mL,Id
O
(14-12)
Substituting results from Eqs. 14-10, 14-11, and 14-13 into Eq. 14-14
(14-15)
14-8
14-5 Current-Link Systems
Thyristors are available in very large current and voltage ratings of several kAs and
several kVs that can be connected in series. ln addition. thyristor converters can block
voltages of both polarities but conduct current only in the forward direction. This
capability has led to the interface realized by thyristor-converters with a dc-current link
in the middle, as shown in Fig. 14-10. Unlike in voltage-link systems, the transfer of
power in current-link systems can be reversed in direction by reversing the voltage
polarity of the dc-link. This structure is used at very high power leveIs, at tens of
hundreds ofMWs, for example in H igh Voltage DC Transmission (HVDC).
+ +
(14-16)
(14-17)
0 0
By controlling the delay angles a I and a 2 in a range of 0 to 180 (practically, this
value is limited to approximately 150 0
) , the average voltage and the average current in
the system ofFig. 14-10 can be controlled, where current flow through the dc-link can be
expressed as
(14-18)
14-9
where the dc-Iinl< resistance Rd is generally very small. ln such a system, for the power
f10w from system I to system 2, Vd2 is made negative by controlling a, sucb that it
operates as an inverter and establishes the voltage of the de-I ink. The converter 1 is
operated as a rectifier at a delay angle aI such that it controls the current in the dc-link.
The converse is true for tbese two converters if the power is to f10w from system 2 to
system 1.
The above discussion of current-link systems shows the operating principie behind
HVDC transmission systems discussed in the next chapter, where using transformers,
six-pulse thyristor converters are connected in series on the dc-side and in parallel on the
ac-side to yield higher effective pulse number.
REFERENCE
PROBLEMS
Single-Phase Thyristor Converters
14-10
Chapter 15
15-1 INTRODUCTION
Power electronics applications in utility systems are growing very rapidly, which promise
to change the landscape of future power systems in terms of generation, operation and
control. The goal of this chapter is to briefly present an overview of these applications to
prepare students for new challenges in the deregulated utility environrnent and to
motivate them to take further courses in this field, where a separate advanced-Ievel
graduate course is designed, for example at the University of Minnesota, to discuss these
topics in detail.
ln discussing these app lications, we will observe that the power electronic converters are
the same or modifications of those that we have already discussed in earlier chapters.
Therefore, within the scope of this book, it will suffice to discuss these applications in
terms of the block diagrams of various converters. These utility applications of power
electronics can be categorized as follows:
15-1
15-2 POWER SEMICONDUCTOR DEVICES AND THEm CAPABILITlES [IJ
Fig. 15-1 a shows the commonly used symbols of power devices. The power handling
capabilities and switching speeds of these devices are indicated in Fig. 15-1 b. Ali these
devices allow current flow only in their forward direction (the intrins ic anti-parallel diode
of MOSFETs can be eXplained separately). Transistors (intrinsically or by design) can
block only the forward polarity voltage, whereas thyristors can block both forward and
reverse polarity voltages. Diodes are uncontrolled devi ces, which conduct current in the
forward direction and block a reverse voltage. At very high power leveIs, integrated-gate
controlled thyristors (IGCTs), which have evolved from the gate-tum-off thyri stors, are
used. Thyristors are semi-controlled devices that can switch-on at the desired instant in
their forward-blocking state, but cannot be switched off from their gate and hence rely on
the circuit in which they are connected to switch them off. However, thy ristors are
available in very large voltage and current ratings.
10'
~
e 106
IGBT
~
o lO'
"-
ln a very broad sense, the role of power electronics in these power system applications
can be categorized as follows :
15-3-1 Solid-State Switches
By connecting two thyristors in anti-parallel (back-to-back) as shown in Fig. 15-2, it is
possible to realize a solid-state switch which can conduct current in both directions, and
tum-on or tum-off in an ac circuit with a delay of no more than one-half the line-
frequency cycle. Such switches are needed for applications such as dual feeders , shunt-
compensation for injecting reactive power at a bus for voltage control, and series-
compensation oftransmission lines.
15-2
Figure 15-2 Back-to-back thyristors to act as a solid-state switch.
15-3-2 Converters as an Interface
Power electronic converters provide the needed interface between the electrical source,
often the utility, and the load, as shown in Fig. 15-3. The electrical source and the
electrical load can, and often do, differ in frequency, voltage amplitudes and the number
of phases. The power electronics interface allows the transfer of power from the source
to the load in the most energy efficient manner, in which it is possible for the source and
load to reverse roles. These interfaces can be classified as below.
,,, -------------,,,
, ,
}-~ Converter f+-{
Source Load
,
:_ L._-_..:_::.:_:.:_:.:_:.:_:.:_:.:_"-_-_..J_ J
Figure 15-3 Power electronics interface.
15-3-2-1 Voltage-Link Systems
The semiconductor devices such as transistors ofvarious types and diodes can only block
forward-polarity voltages. These devices with only unipolar voltage-blocking capability
have led to the structure with two converters, where the dc ports of these two converters
are connected to each other with a parallel capacitor forming a dc-link in Fig. 15-4,
across which the voltage polarity does not reverse, thus allowing unipolar voltage-
handling transistors to be used within these converters. The transfer of power can be
reversed in direction by reversing the direction of currents associated with the dc-link
system.
15-3
6>--i
ACl
ACl
-
Combi ned Heat and Power Wind Power Plants
Plant (eHP) ). Village
>Factory );o Commerc ial
>Commercial Bui lding Building
> House
>Apanrnent Building
Micro-Turbine
Fuel Cell ). Hospital
:> Smart House :> Commercial
,. Perfonnance ~ Building
Building
15-4
There is a move away from large central power plants towards distributed generation due
to environrnental and economic reasons. Renewable resources such as wind and
photovoltaic systems are growing in their popularity. There are proposals to place highly
efficient small-scale power plants, bascd on fuel cells and micro-turbines, near load
centers to sirnultaneously avoid transmission congestion and line losses. Many of
distributed generation systems are discussed below.
15-4-1 Wind-Electric Systems
Wind energy is the fastest growing energy resource in the world. ln 2002, the amount of
installed wind energy in the world reached nearly 30,000 MW, and the installed wind
energy in the U.S. alone increased by 66% in 2001. This remarkable growth is spurred
by a combination of environrnental concems, decreasing costs of wind electric systems,
and financial incentives for wind park developers.
There are many schemes for hamess ing wind energy, out of which those using doubly-
fed induction generators are enjoying increased popularity. As shown in Fig. 15-7, in this
system, the stator which produces most of the power is connected directly to the grid.
The rotor windings are accessible througb slip-rings and brushes to a power electronic
converter system, which supplies the rotor current. By injecting currents into the rotor,
the machine can be made to act like a generator at both sub-synchronous and super-
synchronous speeds. The converter rating is about 15% of the rated power, for example,
in the 750 kW doubly-fed induction machines used in the southwestem Minnesota for
wind-electric systems, only 150 kW flows through the rotor, and the rest through the
stator. Injecting current into the rotor provides control over the real and the reactive
power (Ieading or lagging).
Wound rotor
lnduction Generator
r-----------------~~~
w,~\.:-~t:"rl /~,..~_...~=DC==;
A=CC= ] /i1~rtr
Turbine V DC / AC
Generator-s lde Gnd-slde
Converter Converter
Figure 15-7 Doubly-fed induction generators for wind-electric systems.
15-4-2 PbotovoItaic (PV) Systems
Photovoltaic systems are the ultimate in distributed generation and have even a greater
potential than the wind-electric systems. ln 1997, PV systems surpassed the 500 MW
15-5
threshold in worldwide installation, and annual global sales of $1 billion. A few years
back, the Clinton administration announced the 'million solar roof initiative',
emphasizing the growing importance of PV systems.
ln PV systems, the PV arrays (typically four of them connected in series) provide a
voltage of 52-V to 90-V DC, which the power e1ectronic system, as shown in Fig. 15-8,
converts to 120V160Hz sinusoidal voltage suitable for interfacing with the single-phase
utility. PV modules are still very expensive, as much as $1,500 for only a 300 W peak
output. Such PV modules can be connected in parallel for higher output capacity.
Presently there is not much economy of scale in the PV modules, making it prudent to
invest in steps.
Isolated
PWM
DC-DC
Converter
Converter
Uti lity
Icp
15-6
high-temperature batteries are being developed), in Super-conducting Magnetic Energy
Storage (SMES) coils and in the inertia of f1ywheels. Ali these systems need power
electronics interface, where the interface for the f1ywheel storage is shown in Fig. 15-9.
Machine-side Gridside
Converter Converter
Flywheel
Switchmode
Utility Converter
~
I
Rectifier
15-7
15-6-1 Dual }'eeders
The continuity of service can be enhanced by dual power feeders to the load, where one
acts as a backup to the other that is supplying the load as shown in Fig. 15-11 . Usi ng
back-to-back connected thyristors, acting as a solid-state switch, it is poss ible to switch
the load quickly, without interruptio n, from the main feeder to the backup feeder and
back.
Feeder I
)1---1
Feeder2 I
Load
L-_ _ _ _- '
Energy
Storage
+
v, Power Electronic Load
Interface
15-8
irreversible in spite of recent setbacks, is to deregulate where utilities must compete to
sell power on an open market and the customers have a choice of selecting their power
provider. ln such an environment, utilities that were vertically integrated are now forced
to split into generation companies that produce power, and transmission-line operators
that maintain the transmission and distribution network for a fee.
[n this deregulated environment, it is highly desirable to have the capability to dictate the
flow ofpower on designated power lines, avoiding overloading oftransmission lines and
excessive power losses in them. ln this section, we will look at some such options.
15-7-1 High VoItage DC (HVDC) and Medium Voltage DC (MVDC) Transmission
Direct current (DC) transmission represents the ultimate in flexibiI ity, isolating two
interconnected ac systems from the requirement of operating in synchronism or even at
the same frequency. High voltage DC (HVDC) systems using thyristor-based converters
have now been in operation for several decades. Lately, such systems at medium
voltages have been proposed. Both ofthese types of systems are discussed below.
[5-7-1-1 High Voltage DC (HVDC) Transmission
Fig. 15-14 shows the block diagram of a HVDC transmission system, where power is
transmitted over dc lines at high voltages in excess of 500 kV. First, the voltages in ac
system I at the sending end are stepped up by means of a transformer. These voltages
are rectified into dc by means of a thyristor-based converter, where the ac line voltages
provide the commutation of current within the converter, such that ac currents drawn
from system I tum into dc current on the other side of the converter. These currents are
transmitted over the dc line where additional series inductance is added to ensure that the
dc current is smooth, free of ripple as much as possible. At the receiving end, there are
thyristor-based converters, which convert the dc current into ac currents and injects them
into the ac system 2, through a step-down transformer. The roles of the sending and the
receiving ends can be reversed by reversing the voltage polarity in the dc system.
ACl AC2
Figure 15-14 HVDC system block diagram (transformers are not shown).
At very high power leveis in excess of 1,000 MW, use of thyristors, at least for now,
represents the only reasonable choice.
15-9
15-7-1-2 Medium Voltage DC (MVDC) Transmission
At lower power leveis, there are proposals to use medium voltages, in which case the
block diagram ofthe system is as shown in Fig. 15-15. ln such a system, the direction of
power f10w is reversed by changing the direction of current in the dc line.
6>--1
ACI
DC transmission systems discussed earlier are an excellent choice where large amount of
power needs to be transmitted over long distances, or if the system stability is a serious
issue. ln existing ac transmission networks, limitation on constructing new power lines
and their cost has led to other ways to increase power transmission capability without
sacrificing the stability margino These techniques may also help in directing the power
f10w to designated lines.
Power f10w on a transmission line connecting two ac systems in Fig. 15-16 is given as
E,E, . <"
P = --SInu (I 5-1)
X
where E, and E, are the magnitudes of voltages at the two ends of the transmission line,
X is the line reactance and <5 is the angle between the two bus voltages.
E,LO E,L-<5
jX
~~~__nnn__- +_ _~
ACI AC2
15-10
15-7-2-1 Shunt-Connected Devices to Control the Bus Voltage Magnitude E
The reactive power compensation is very important and may even be necessary at high
loadings to avoid voltage collapse. Shunt-connected devices can draw or supply reactive
power from a bus, thus controlling the bus voltage, albeit in a limited range, based on the
internal system reactance. Various forms of such devices are being used in different
combinations. These include Thyristor-Controlled Reactors (TCR) shown in Fig. 15-17a,
and Thyristor-Switched Capacitors (TSC) shown in Fig. 15-17b for Static Var
Compensation (SVC). The advanced Static Var Compensator (STATCOM) shown in
Fig. 15-17c can draw or supply reactive power.
1 Utility
-L
=
(a)
& (b)
STATCOM
(c)
I
l...----l
Shunt-compensation devices have the following limitations for controlling the flow of
active power:
These devices, connected in series with a transmission line, partially neutralize (or add
to) the transm ission line reactance. Therefore, they change the effective value of X in
Eq. 15-1, thus allowing the active power flow P to be controlled. Various forms ofsuch
devices have been used. These include the Thyristor-Controlled Series Capacitor (TCSC)
shown in Fig. 15-18.
15-11
Figure 15-18 Thyristor-controlled series capacitor.
15-7-2-3 Static Phase Angle Control and Unified Power Flow Controller (UPFC)
Based on Eq. 15-1 , a device connected at a bus in a substation, as shown in Fig. 15-19a,
can influence power flow in three ways by:
I. controlling the voltage magnitude E
2. changing the line reactance X, and/or
3. changing the power angle <5.
Such a devi ce called the Unified Power Flow Controller (UPFC) can affect power flow in
any combination of the ways listed above. The block diagram of a UPFC is shown in
Fig. 15-19a at one side of the transmission line. lt consists of two voltage-source switch-
mode converters. The first converter injects a voltage E, in series with the phase voltage
such that
Therefore, by controlling the magnitude and the phase of the injected voltage E, within
the circle shown in Fig. 15-19b, the magnitude and the phase of the bus voltage E2 can
be controlled. Ifa component ofthe injected voltage E, is made to be 90 degrees out-of-
phase, for example leading with respect to the current phasor J, then the transmission
line reactance X is partially compensated.
sub-station
------------------------ -- --------------,
E, I~ .E j + E2
E,
-LI E,
I (b)
Shunt Series
converter ~ ~ converter
P2.Q2 ~ .QI
~------------------------ ---- ------------
(a)
Figure 15-19 UPFC.
15-12
The second converter in a UPFC is needed for the following reason: since converter 1
injects a series voltage E" it delivers real power p" and the reactive power Q, to the
transmission line (where p, and Q, can be either positive or negative):
p, = 3 Re(E,I) (15-3)
(15-4)
Since there is no steady state energy storage capability within UPFC, the power P2 into
converter 2 must equal p, if the los ses are ignored:
(15-5)
However, the reactive power Q2 bears no relation to Q" and can be independently
controlled within the voltage and current ratings ofthe converter 2:
(15-6)
By controlling Q2 to control the magnitude of the bus voltage E" UPFC provides the
same functionality as that of an advanced static var compensator STATCOM.
REFERENCES
1. N. Mohan, T. M. Undeland, and W.P. Robbins, Power Electronics: Converters,
Applications and Design, 3"' Edition, Wiley & Sons, New York, 2003.
2. N. G. Hingorani, and L. Gyugyi, Understanding FACTS, New York, IEEE Press,
2000.
3. N. Mohan, A. Jain, P. Jose and R. Ayyanar, "Teaching Utility Applications ofPower
Electronics in a First Course on Power Systems," submitted to the IEEE Transactions
on Power Systems, Special Section on Power Engineering Education, 2003.
PROBLEMS
15-1 Show the details and the average representation of converters in Fig. 15-7 for a
wind-electric system.
15-2 Show the details and the average representation of converters in Fig. 15-8 for a
photovoltaic system.
15-3 Show the details and the average representation of converters in Fig. 15-9 for a
tlywheel storage system.
15-4 Show the details and the average representation of converters in Fig. 15-10 for an
adjustable-speed drive.
15-5 Show the details and the average representation of converters in Fig. 15-12 for a
UPS.
15-13
15-6 Show the details and the average representation of converters in Fig. 15- \3 for a
DVR.
15-7 Show the details and the average representation of converters in Fig. 15-14 for an
HVDC transmission system.
15-8 Show the details and the average representation ofconverters in Fig. IS-IS for an
MVDC transmission system .
15-9 Show the details and the average representation of converter in Fig. 15-17c for a
STATCOM.
15-10 Show the details and the average representation of converters in Fig. I 5-19a for a
UPFC system.
15-14
DETAILED CHAPTER CONTENTS
DCC-l
2-6-1 Switching Frequency is, 2-11 3-9 Worst-Case Design, 3-16
2-6-2 Selection of Transistors and 3-10 Synchronous-Rectified Buck
Diodes, 2-12 Converter for Very Low Output
2-6-3 Magnetic Components, 2-12 Voltages [3],3-16
2-6-4 Capacitor Selection, 2-12 3-11 lnterleaving ofConverters [4], 3-
2-6-5 Thermal Design, 2-12 17
2-6-6 Design Tradeoffs, 2-14 3-12 Regulation of DC-DC Converters
2-7 The PWM ControJler IC, by PWM, 3-17
2-14 3-13 Dynamic Average Representation
References, 2-1 5 of Converters in CCM, 3-18
Problems, 2-1 5 3-14 Bi-Directional Switching Power-
APPENDIX 2A Diode Reverse Pole, 3-21
Recovery and Power 3-15 Discontinuous-Conduction Mode
Losses, 2-16 (DCM),3-22
2A -I Diode F orward Loss, 2-16 References, 3-24
2A-2 Diode Reverse Recovery Problems, 3-24
Characteristic, 2-16 Appendix,3-24
2A-3 Diode Switching Losses, 2-17 3A Discontinuous-Conduction Mode
(DCM) in DC-DC Converters, 3-24
Chapter 3 3A-I Output Voltages in DCM, 3-24
Switch-Mode DC-DC Converters: 3A-I-I Buck Converters in DCM, 3-24
Switching Analysis, Topology 3A-I-2 Boost and Buck-Boost
Selection and Design Converters in DCM, 3-25
3-1 DC-DC Converters, 3-1 3A-2 Average Representation ofDC-
3-2 Switching Power-Pole in DC DC Converters in DCM, 3-26
Steady State, 3-1 3A-2-1 v. and i. for Buck and Buck-
3-3 Simplifying Assumptions, 3-4 Boost Converters in DCM, 3-26
3-4 Common Operating Principies, 3-4 3A-2-2 v. and i. for Boost
3-5 Buck Converter Switching Converters in DCM, 3-28
Analysis in DC Steady State, 3-4
3-6 Boost Converter Switching Chapter 4
Analysis in DC Steady State, 3-7 Designing Feedback Controllers in
3-7 Buck-Boost Converter Analysi s in Switch-Mode DC Power Snpplies
DC Steady State, 3-10 4-1 Introduction and Objectives of
3-7-1 Other Buck-Boost Topologies, 3- Feedback Control, 4-1
13 4-2 Review ofL inear Control Theory,
3-7-1-2 Cuk Converter, 3-14 4-2
3-8 Topology Selection [2], 3-15
DCC-2
4-2-1 Open-Loop Transfer Function 5-4 Diode-Rectifier Bridge "Front-
G,, (s),4-3 Ends", 5-10
4-2-2 Crossover Frequency .Ic(s) of 5-4-1 The Single-Phase Diode-Rectifier
Bridge, 5-13
G L (s), 4-3
4-2-3 Phase and Gain Margins, 4-3 5-4- l-I Effects of L, and Cd on the
4-3 Linearization of Various Transfer Waveforrns and the THD, 5-11
Function Blocks, 4-5 5-4-2 Three-Phase Diode-Rectifier
4-3-1 Linearizing the PWM Controller Bridge, 5- 13
IC,4-5 5-4-3 Comparison of Single-Phase and
4-3-2 Linearizing the Power Stage of Three-Phase Rectifiers, 5- I 5
DC-DC Converters in CCM, 4-6 5-5 Means to Avoid Transient lnrush
4-3-2-1 Using Computer Simulation to Currents at Starting, 5-16
Obtain "o1d , 4-9 5-6 Front-Ends with Bi-Directional
4-4 Feedback Controller Design ln Power Flow, 5-16
Voltage-Mode Control, 4-11 References, 5-16
4-5 Peak-Current Mode Control, 4-15 Problems, 5-16
References, 4-20
Problems, 4-20 Chapter 6
Power-Factor-Correction (PFC)
Chapter 5 Circuits and Designing the Feedback
Rectification ofUtility Input Using Controller
Diode Rectifiers 6-1 I ntroduction, 6-1
5-1 Introduction, 5-1 6-2 Operating Principie of Single-
5-2 Distortion and Power Factor, 5-1 Phase PFCs, 6-1
5-2-1 Rms Value ofDistorted Current 6-3 Control of PFCs, 6-3
and the Total Harrnonic 6-4 Designing the Inner Average-
Distortion (THD), 5-2 Current-Control Loop, 6-4
5-2-2 The Displacement Power Factor 6-4-1 d(s)/Vc(s) for the PWM
(DPF) and Power Factor (PF), 5-
Controller, 6-5
5
6-4-2 IL(s) 1 d(s) for lhe Boost
5-2-3 Deleterious Effects ofHarrnonic
Converter in the Power Stage, 6-5
Distortion and a Poor Power
Factor, 5-7 6-4-3 Designing the Current Controller
5-2-3-1 Harmonic Guidelines, 5-7 G,(s),6-5
5-3 Classifying the "Front-End" Of 6-5 Designing the Outer Voltage Loop,
Power Electronic Systems, 5-9 6-6
6-6 Example ofSingle-Phase PFC
Systems, 6-8
DCC-3
6-6-1 Design of the Current Loop, 6-8 8- 7 Practical Considerations ln
DCC-4
10-3-1 Zero Voltage Switching (ZVS), 11-5-3 Mechanical System ofPMAC
10-3 Drives, 11-15
10-3-2 Synchronous Buck Converter 11-5-4 PMAC Machine Equivalent
with ZVS, 10-3 Circuit, 11-15
10-3-3 Phase-Shift Modulated (PSM) 11-5-5 PMAC Torque-Speed
DC-DC Converter, 10-6 Characteristics, 11-16
10-4 Inverters for Induction Heating 11-5-6 The Controller and the Power-
and Compact Fluorescent Lamps, ProcessingUnit(PPU),11-17
10-7 11-6 lnduction Machines, 11-17
References, 10-7 11-6- I Structure of Induction
Problems, 10-7 Machines, 11-8
11-6-2 The Principies ofInduction
Chapter 11 Motor Operation, 1 l-I 8
Electric Motor Drives 11-6-3 Per-Phase Equivalent Circuit of
11-1 Introduction, 1 l-I Induction Machines, 11-20
11-2 Mechanical System References, 11-22
Requirements, 11-2 Problems, 11-22
11-2-1 Electrical Analogy, 11-4
11-3 Introduction to Electric Machines Chapter 12
and the Basic Principies of Synthesis ofDC and Low-Frequency
Operation, I 1-4 Sinusoidal AC Voltages for Motor
11-3-1 Electromagnetic Force, 11-5 Drives and UPS
11-3-2 lnduced Emf, 11-7 12-1 Introduction, 12-1
11-3-3 Basic Structure, 11-7 12-2 Switching Power-Pole as the
I 1-4 DC Motors, I 1-8 Building Block, 12-2
11-4-1 Structure ofDC Machines, 11-8 12-2-1 Pulse-Width-Modulation
11-4-2 Operating Principies of DC (PWM) ofthe Bi-Directional
Machines, 11-9 Switching Power-Pole, 12-2
11-4-3 DC-Machine Equivalent Circuit, 12-2-3 Harmonics in The PWM
11-10 Waveforms v A and idA' 12-5
11-4-4 Torque-Speed Characteristics, 12-3 DC-Motor Drives, 12-5
11-11 12-4 AC-Motor Drives, 12-9
11-5 Permanent-Magnet AC Machines, 12-4-1 Space Vector Pulse Width
11-12 Modulation, 12-13
11-5-1 The Basic Structure of 12-5 Voltage-Link Structure with Bi-
Permanent-Magnet Ac Directional Power Flow, 12-13
Machines, 11-13 12-6 Uninterruptible Power Supplies
11-5-2 Principie of Operation, 11-13 (UPS), 12-15
DCC-S
12-6-1 Single-Phase UPS, 12-16 14-3-1 The Effect of Ls on Current
References, 12-19 Commutation, 14-5
Problems, 12-19 14-4 Three-Phase, Full-Bridge
Thyristor Converters, 14-6
Cbapter 13 14-5 Current-Link Systems, 14-8
Designing Feedback Controllers for Reference, 14-10
Motor Drives Problems, 14-10
\3-1 lntroduction, 13-1
\3-2 Control Objectives, 13-1 Cbapter 15
13-3 Cascade Control Structure, 13-3 Utility Applications of Power
13-4 Steps in Designing the Feedback Electronics
ControII er, 13-4 15-1 Introduction, 15-1
13-5 System Representation for Small- 15-2 Power Semiconductor Devices
Signal Analysis, 13-4 and their Capabi lities [1], 15-2
\3-5-1 The Average Representation of 15-3 Categorizing Power E1ectronic
the Power-Processing Unit Systems, 15-2
(PPU), 13-4 15-3-1 Solid-State Switches, 15-2
13-5-2 Modeling ofthe DC Machine 15-3-2 Converters as an Interface, 15-3
and the Mechanical Load, 13-6 15-3-2-1 Voltage-Link Systems, 15-3
13-6 Controller Design, 13-7 15-3-2-2 Current-Link Systems, 15-4
13-6-1 PI Controllers, 13-7 15-4 Distributed Generation (DG)
13-7 Example of A Controller Design, Applications, 15-4
13-8 15-4-1 Wind-Electric Systems, 15-5
\3-7-1 The Design ofthe Torque 15-4-2 Photovoltaic (PV) Systems, 15-
(Current) Control Loop, 13-9 6
13-7-2 The Design ofthe Speed Loop, 15-4-3 Fue1 Cell Systems, 15-6
13-11 15-4-4 Micro-Turbines, 15-6
13-7-3 The Design of the Position 15-4-5 Energy Storage Systems, 15-6
Control Loop, 13-13 15-5 Power Electronic Loads, 15-7
References, 13-14 15-6 Power Quality Solutions, 15-8
Problems, 13-14 15-6-1 Dual Feeders, 15-8
15-6-2 Uninterruptible Power Supplies,
Cbapter 14 15-8
Tbyristor Converters 15-6-3 Dynamic Voltage Restorers, 15-
14-1 lntroduction, 14-1 8
14-2 Thyristors (SCRs), 14-1 15-7 Transmission and Distribution
14-3 Sing1e-Phase, Phase-Controlled (T &D) Applications, 15-9
Thyristor Converters, 14-2
DCC-6
15-7-1 High Voltage DC (HVDC) And 15-7-2-2 Series-Connected Devices to
Medium Voltage DC (MVDC) Control the Effective Series
Transmission, 15-9 Reactance X, 15-12
15-7-1-1 High Voltage DC (HVDC) 15-7-2-3 Static Phase Angle Control
Transmission, 15-9 and Unified Power Flow
15-7-1-2 Medium Voltage DC (MVDC) Controller (UPFC), 15-13
Transmission, 15-10 References, 15-14
15-7-2 Flexible AC Transmission Problems, 15-14
Systems (Facts), 15-10
15-7-2-1 Shunt-Connected Devices to
Control the Bus Voltage
Magnitude E, 15-12
DCC-7