Citation
Elements of induction machine application in hydroelectric power generation

Material Information

Title:
Elements of induction machine application in hydroelectric power generation
Creator:
Zanjani, Hamid
Place of Publication:
Denver, CO
Publisher:
University of Colorado Denver
Publication Date:
Language:
English
Physical Description:
107 leaves : ; 28 cm

Subjects

Subjects / Keywords:
Hydroelectric generators -- Design and construction ( lcsh )
Electric machinery, Induction ( lcsh )
Electric generators -- Alternating current ( lcsh )
Electric motors, Synchronous ( lcsh )
Electric generators -- Alternating current ( fast )
Electric machinery, Induction ( fast )
Electric motors, Synchronous ( fast )
Hydroelectric generators -- Design and construction ( fast )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Bibliography:
Includes bibliographical references (leaves 106-107).
Thesis:
Electrical engineering
Bibliography:
Department of Electrical Engineering
Statement of Responsibility:
by Hamid Zanjani.

Record Information

Source Institution:
|University of Colorado Denver
Holding Location:
|Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
40326500 ( OCLC )
ocm40326500
Classification:
LD1190.E54 1998m .Z36 ( lcc )

Downloads

This item has the following downloads:


Full Text
ELEMENTS OF INDUCTION MACHINE APPLICATION
IN HYDROELECTRIC POWER GENERATION
by
Hamid Zanjani
B.S., University of Colorado, 1984
A thesis submitted to the
University of Colorado at Denver
in partial fulfillment
of the requirements for the degree of
Master of Science
Electrical Engineering
1998


This thesis for the Master of Science
degree by
Hamid Zanjani
has been approved
by
5 fe -98
Date


Zanjani, Hamid (M.S., Electrical Engineering)
Elements of Induction Machine Application In Hydroelectric Power Generation
Thesis directed by Professor Pankaj K. Sen
ABSTRACT
This study identifies and investigates major aspects of utilizing induction
machines in hydroelectric power generation as a viable alternative to the use of
conventional synchronous machines. The study begins by first introducing the
elementary operating principles and features of induction machines in clear and easy
to understand terms. Once a solid technical foundation is built, the author introduces
and explores major technical engineering topics that are of concern in a successful
assessment and application of induction machines in micro- (<100 KW) and mini-
mi 00 KW and <5000 KW) size power plants driven by hydroelectric turbines.
Some of the major topics reviewed in this study include: voltage and
frequency regulation of induction machine, stand-alone versus grid connected
machines, comparison of various capacitive var compensation techniques, electronic
load governors versus turbine governors, monitoring and protection of induction
machines, and utility requirements for induction machines connected to the power
grid.
This abstract accurately represents the content of the candidates thesis. I recommend
its publication.
Sign
in


ACKNOWLEDGMENTS
I am finally getting out of the strain of doing and into the peace of being done.
It seemed like a long journey from the time when I took my first graduate class to this
point where I can finally see the light at the end of the tunnel with my endeavor. I
must confess that my journey had its share of passing through construction zones,
detours, and stopping at scenic sites along the way. Nevertheless, this journey has
been an exciting, educational, and fruitful experience for me. The valuable lessons I
learned as a student were that: 1) The purpose of an education is not to learn, but to
learn how to learn and 2) learning should be viewed as a life long goal extending
beyond college years.
I am grateful to all my professors at the University of Colorado at Denver for
making this education possible for me. I am especially indebted to Dr. Pankaj Sen
and Dr. Bill Roemish for their efforts, inspiration, patience and for the trust that they
invested in me throughout this educational program.
IV


CONTENTS
Chapter
1. Introduction.............................................................1
1.1 Operating Principles of Squirrel-Cage Induction Machines................5
1.2 Design and Performance of Induction Machine............................10
1.2.1 Design..........................7.....................................10
1.2.2 Performance..........................................................11
1.3 Equivalent Circuit of Induction Machine...............................16
1.3.1 Development of Equivalent Circuit............i........................17
1.3.2 Analysis of the Equivalent Circuit...................................23
2. Application of Induction Generators in Micro- and Mini-Hydro Stations....30
2.1 Introduction..........................................................30
2.2 The Choice of Turbine and Governor for a Micro- and Mini-Hydro Station.32
2.2.1 Turbine..............................................................32
2.2.2 Standard Governor Versus Load Governor...............................38
2.3 Induction Generators in Micro-Hydro Stations as Stand-Alone Units......51
2.3.1 De-Rating of Induction Machine for Unbalanced Operation...............51
2.3.2 Optimizing Capacitor Size for a Stand-Alone Induction Generator.......61
2.4 Induction Generators in Mini-Hydro Stations...........................75
v


3. Protection of Induction Generators
77
3.1 Circuit Breaker (ANSI device no. 52).....................................78
3.2 Instantaneous, Overcurrent and Differential Protection Devices (ANSI device nos.
50 and 87).....................................................................79
3.3 Thermal and Overload Protection Devices (ANSI device nos. 49 and 51).....81
3.4 Ground Fault Protection Device (ANSI device no. 64, 50G, 51G)............82
3.5 Rotor Overheating Protection (ANSI device nos. 46 and 47).................83
3.6 Bearing Protective Device (ANSI device no. 38)............................84
3.7 Vibration Detection (ANSI device no. 39).................................84
3.8 Mechanical Overspeed Protection (ANSI device no. 12).....................84
3.9 Reverse Power Protection (ANSI device no. 32).............................85
3.10 Overvoltage and Overexcitation Protection................................86
3.11 Typical Overall Protection Scheme........................................87
3.12 Utility Interface Protection Requirements................................93
3.12.1 Lockable Disconnect Switch..............................................95
3.12.2 Detection and Clearing a Fault on the Utility System...................95
3.12.3 Islanding..............................................................98
3.12.4 Energizing a Dead Circuit.............................................100
3.12.5 No Manual Synchronization.............................................100
3.12.6 Modification Costs to the Utility.....................................102
Conclusions...................................................................103
vi


Appendix
A. Constants C and D for equations (2.21) and (2.22) defined.............105
References...............................................................106
vii


FIGURES
Figure
Figure 1.1 Relative motion of stator and rotor magnetic fields.................7
Figure 1.2 Induction machine torque-slip curve showing braking, motor, and
generator regions...........................................................12
Figure 1.3 - Typical short circuit waveform of a synchronous generator...........15
Figure 1.4 - Typical short circuit waveform of an induction generator............15
Figure 1.5 Stator equivalent circuit of an induction motor with rotor circuit open-
circuited....................................................................17
Figure 1.6 - Rotor equivalent circuit at stator frequency........................20
Figure 1.7 - Equivalent circuit of an induction machine (motor or generator).....22
Figure 1.8 - Direction of power flow in a motor and a generator operation........22
Figure 1.9- Typical phasor diagrams for an induction motor and a generator......22
Figure 1.10 - Thevenin equivalent circuit of an induction machine...............26
Figure 2.1 Circuit configuration of the impedance controller with rectifier and
chopper.....................................................................43
Figure 2.2 Rectifier line current for phase A with a=30 without the chopper...46
Figure 2.3 Impedance controller line current for phase A with a=30, five-pulse
chopping and PW=0.5.........................................................46
Figure 2.4 Rectifier line current for phase A with a=90 without the chopper...47
viii


Figure 2.5 Impedance controller line current for phase A with a=90, five-pulse
chopping and PW=0.5...........................................................47
Figure 2.6 Area of current control offered by the impedance controller for Rdc=l .0
pu at rated voltage...........................................................49
Figure 2.7 Sequence network circuits of an induction machine.....................53
Figure 2.8 Direction of power flow in a self excited induction generator.........55
Figure 2.9 Derating of Induction machine due to unbalanced voltages..............60
Figure 2.10 Derating of induction machine due to unbalanced voltages.............60
Figure 2.11 Block diagram of a short shunt self-excited induction generator......63
Figure 2.12 Block diagram of a long shunt self-excited induction generator.......63
Figure 2.13 Equivalent circuit of a short shunt self-excited induction generator.64
Figure 2.14 Variation of Xm versus E/F...........................................65
Figure 2.15 Variation of terminal voltage with shunt capacitor Csh...............68
Figure 2.16 Effect of Cse variation on load voltage regulation...................69
Figure 2.17 Variation of load voltage VLwith load................................70
Figure 2.18 Variation of air-gap voltage E with load.............................70
Figure 2.19 Variation of stator current Is with load.............................71
Figure 2.20 Overall characteristics of a short shunt SEIG........................72
Figure 2.21 Effects of changing shunt capacitor Cshon load voltage...............73
Figure 3.1 Induction generator connections with instantaneous and differential
overcurrent protections....................................................... 80
IX


Figure 3.2 Typical protection scheme for a large-size, induction generator.
88
Figure 3.3 Typical protection and metering scheme for an induction generator less
than 10 KW...................................................................89
Figure 3.4 Typical protection and metering scheme for an induction generatorlO
KW to less than 100 KW.......................................................90
Figure 3.5 Typical protection and metering scheme for an induction generator 100
KW to 1 MW...................................................................91
Figure 3.6 Typical protection and metering scheme for an induction generator 1 MW
to less than 10 MW.......................................................... 92
Figure 3.7 High voltage ground detection using comer of delta voltage relay..98
x


1. Introduction
Rising capital cost of building large power plants, lack of interest from large
financial institutions, increased scrutiny by the environmental agencies, changes in
federal and legislative laws, in addition to recent deregulation of power utilities have
encouraged design and development of smaller power plants. The design objectives
of these smaller power plants are simplicity, reliability, and especially cost
effectiveness.
In meeting these objectives, the lower initial cost, low maintenance, simplicity
and ruggedness associated with induction generators (in comparison with
synchronous generators) have advanced the widespread application of induction units
in smaller power plants. Most induction generators in these applications are driven
by either hydraulic or wind turbines in a stand-alone or grid-connected configuration.
The most common size of turbine/induction generator set found in commercial use
ranges from 50 KW to an upper limit of 1500-2000 KW. Within this range, power
plants built with induction generators by far outnumber similar plants built with
synchronous generators [1].
An induction generator is simpler and cheaper to build since its rotor cage is a
symmetrical structure built using an automated fabrication process. Since the rotor
winding does not need an excitation system, control equipment requirements for the
generator are greatly reduced. Furthermore, an induction generator does not need a
separate power source for its rotor winding.
The induction generator, like a synchronous generator, requires excitation in
order to produce voltage and become a source of electrical power. While a
synchronous generator derives its excitation from a separate source (e.g., an on-shaft
1


rotating excitation system), the induction generator must draw its magnetizing current
from the intertied utility system or from capacitors located near the unit. Without
excitation from some source, the induction generator cannot sustain a terminal voltage
with any load connected to it.
In order to familiarize the reader with overall characteristics of induction
generators and synchronous generators, the following table makes a condensed
comparison of the two different types of generators. Some of these characteristics
will be discussed in more detail in the forthcoming sections.
2


INDUCTION GENERATORS SYNCHRONOUS GENERATORS
Rugged and simple construction Sensitive and sophisticated construction
Negligible harmonics Harmonics are inherent
Excitation var must come from the grid or capacitor banks Needs a separate dc source for excitation
Does not contribute sustained fault-current under short-circuit conditions Contributes sustained fault-current under short-circuit conditions
Capital cost is less because system requires less design and construction efforts and control devices and machine costs are lower Capital cost is more because of higher design and construction efforts and complicated control devices
Operating cost is less because these plants can be unattended and maintenance cost is low Operating cost is high because these plants need operators and maintenance rate is high
Under load rejection conditions, self- excitation condition may result in overspeed and over voltage Overspeed and overvoltage conditions are under control as a result of governor and automatic voltage regulator actions
Control equipment and governor are simple. No AVR+ and field breakers are required Control equipment and governor are sophisticated. AVR and field breakers are required
Starting is comparatively simple. Synchronization usually not required Synchronization must be precise
No problem in system stability. The system disturbance will generally not affect machine connected to the grid In case of system disturbance, the machine may lose synchronism from the system
Table 1 Comparison of induction versus synchronous generators
There are two types of induction generators available in the market. One is
the wound-rotor induction generator and the other is the squirrel-cage-rotor induction
generator. The stators of these two types of generators are very similar in design and
construction. The major difference between the two types of generators is in the way
their rotors are constructed.
* AYR is an acronym for Automatic Voltage Regulator.
3


A wound rotor in an induction generator is similar in shape to the rotor in a
synchronous generator. The leads of the wound rotor are brought out through slip
rings mounted on the rotor shaft for adding external impedance or changing the
number of poles and hence the speed of induction generator. These features enhance
the performance and flexibility of a wound rotor induction generator in certain
applications where speed, torque, power and efficiency controls are important. The
disadvantages of a wound rotor induction generator are in its higher initial capital cost
(which compares to that of a synchronous generator) and the extra maintenance
required for the slip rings. So, in meeting the power plants feasibility and cost
effectiveness objectives mentioned earlier, a wound rotor induction generator would
not be a good candidate.
A squirrel-cage rotor in an induction generator consists of parallel bars that are
located on the periphery of the rotor shaft and form a cylinder that resembles a
squirrel cage. There is a ring that goes over the top and another ring that goes over
the bottom of these parallel bars which keep the bars in place and also serve to
provide a path for the current from the bars to flow through. The construction of this
type of rotor is relatively simple and cheap. Furthermore, this type of rotor is rugged
and has low maintenance compared to a wound rotor. The lower cost of this type of
induction generator makes it attractive and economically viable in small power plant
applications when compared to a synchronous or a wound rotor induction generator.
For this reason, the scope of this thesis is focused on the squirrel cage induction
generator and very little time is dedicated to exploring the wound rotor induction
generator.
4


1.1 Operating Principles of Squirrel-Cage Induction Machines
An induction machine can either be a motor or a generator. For clarity and
ease of understanding, the discussion here will first concentrate on an induction motor
and then the concepts developed will be expanded to cover an induction generator.
At first, it seems puzzling how a squirrel-cage induction motor is capable of
producing rotation and torque. In comparison with other types of motor that we are
familiar with (such as dc and synchronous motors) it seems as though there is
something missing from the rotor circuit before it can start to tum-either a permanent
magnet that can interact with the stators magnetic circuit, or an electrical winding
connection in the rotor that can serve as a magnetic circuit.
A magnetic circuit produces a magnetomotive-force (mmf or/) based on the
number of turns in the circuit and the current flowing through it (/= NI). According
to magnetic field theory, the presence of two such mmfs acting upon one another and
separated by an angle between them produces rotational torque. This torque is zero
when the rotor mmf and the stator mmf are in phase with each other and maximum
when these mmfs are in quadrature.
But, since there is not a permanent magnet nor an electrical winding circuit
that we are aware of in the rotor, then how is torque produced in a squirrel-cage
induction motor? The answer lies in Induction. An electromotive-force voltage
(emf or speed voltage*) is induced in the rotor bars when voltage is applied to the
+ There are two ways that voltage can be induced. One is through transformation
which is what takes place in the windings of a transformer and another method is
through emf or speed voltage which takes place in rotating machinery. Based on
Faradays law a voltage e is induced if a conductor of length / moves with linear
velocity v in a non-time varying magnetic field of flux density B by cutting-of-flux
equation: e=Blv according to the right-hand rule. The same hold true if the
conductor is held stationary and the magnetic field moves past the conductor. The
5


stator circuit. Many books and technical papers [2, 3] describe this action to be
similar to what takes place in a short-circuited secondary winding of a transformer
when voltage is impressed on its primary winding; except, in an induction motor the
magnetic core is not continuous and has a small air-gap that separates the rotor from
the stator. Furthermore, because the rotor is moving with respect to the
magnetic field frequency of the stator, the voltage induced in the rotor is of a
different frequency than that in the stator. Based on these phenomena, the induction
machine may also be compared to a transformer and a frequency changer combined.
The concept of slip (S) is important in the operation and analysis of an
induction motor. Slip is defined as the difference between the synchronous speed
(Ns) and rotor speed (Nr) expressed as a fraction of the synchronous speed.
Slip (S) nJL^ni (1.
Ns
Slip multiplied by stators synchronous frequency (F) gives the frequency
induced in the rotors electrical or magnetic circuit, and (1-S) multiplied by
synchronous speed (Ns) gives the rotors rotational speed (Nr).
So as an example, if it is given that in a four-pole motor the synchronous
frequency (F) is 60 Hertz, slip (S) is 0.03, and corresponding synchronous speed (Ns)
is 1800 rpm, then the electrical or magnetic frequency of the rotor circuit is
0.03x60 = 1.8 Hertz.
Furthermore, the mechanical speed of the rotor is:
(l-0.03)xl800 = 0.97x1800= 1746 rpm
1746 rpm = r^>m = 29.1 revolutions or cycles per second mechanical
60 sec/ min
relative motion of the conductor and the magnetic field causes the flux lines to be cut
and hence the voltage in the conductor is induced.
6


In a four-pole motor, one complete mechanical cycle corresponds to two
complete electrical cycles. Therefore:
29.1 mech. cycles per second = 29.1x2 = 58.2 elec, cycles per second or Hertz
As can be noted from the above example, the sum of the rotors magnetic field
frequency and the rotors speed (when expressed in terms of electrical frequency)
always add up to equal the synchronous frequency of the stator (58.2+1.8=60).
With a little bit of imagination (refer to Figure 1.1), it can be seen that the
rotors magnetic field frequency is riding over the rotors mechanical speed, both
traveling in the same direction as the stators magnetic field frequency. Magnetic
fields of the ro(or and the stator have the same frequency and are stationary in space
with respect to each other at any given motor speed. (This is analogous to stating that
if a train is traveling at 58.2 mph and a person standing on top of the train is walking
toward the front of the train at 1.8 mph, then a car which is traveling alongside the
train at 60 mph would appear stationary in space to the person on top of the train.)
With two magnetic fields stationary with respect to each other, the torque produced
will be constant and no torque pulsation will exist in an induction motor. This
constant torque is what turns the rotor at a steady speed.
Rotor speed expressed in Hertz plus
Rotor magnetic field frequency in Hertz
Equals
Stator magnetic field frequency in Hertz
Figure 1.1- Relative motion of stator and rotor magnetic fields.
7


For a moment, lets reiterate how an induction motor functions in a slightly
different way for further clarity. An induction motor develops useful torque since its
rotor speed is slightly less than the synchronous speed of the magnetic field present in
the stator winding. That relative motion means that the rotor bars are cutting the
stators magnetic flux lines. The emf voltage induced in the rotor bars will cause bar
currents to flow since rotor bars are short-circuited together at their ends. In turn, this
current flowing in the rotor produces its own magnetic field. Finally the interaction
of rotors magnetic field with stators magnetic field produces torque.
Contrary to common belief, the design and operation of an induction generator
is no more complicated than that of an induction motor. In most cases, an induction
generator is nothing more than a standard squirrel-cage induction motor driven above
its synchronous speed by a prime mover connected to its rotor shaft (note: machines
specifically designed to serve as a generator are slightly different as will be explained
in the next section).
In an induction generator, the prime mover drives the rotor slightly faster than
the synchronous frequency in the stator winding. The relative motion is still present,
but now the direction is reversed. The rotor is moving faster, not slower than the
stator magnetic field. The direction of the resultant magnetic field torque is nearly
reversed. The rotor field leads the stator field instead of trailing it. The machine,
instead of producing mechanical output torque from electrical input power, now
produces electrical output power from mechanical input torque supplied by the prime
mover.
For the induction generator to supply Kilowatts, the rotor must be driven
slightly faster than synchronous speed. This differs from a synchronous generator,
which when paralleled with the utility, maintains a constant speed fixed by the utility.
The faster the induction generator is driven above its synchronous speed rating, the
8


more power it will produce. Typical induction generators operate at speeds in the
range of 1 to 2 percent above synchronous speed. Induction motors, by contrast,
operate fully loaded at 1 to 2 percent below synchronous speed.
Another consideration in an induction generator operation is the excitation
source. The excitation current which is used to magnetize the core must still come
from an external source since there is no other source of field excitation available. To
meet this excitation current requirement, lagging var must be supplied to the
generator.
In a case where the induction generator is connected to the utility grid, the
lagging var is supplied by the utilitys overexcited synchronous machines or capacitor
banks. Since induction generators magnetizing current or lagging var come from the
utility grid, the grid controls both voltage and frequency of the generator. There is
nothing in the simple induction generators design, control, and operation that can
control either voltage or frequency at all. In an induction generator, whose frequency
and voltage are fixed by the utility, generator output power is the only thing that can
be controlled by the speed of the prime mover.
In a case where the induction generator is operating in a stand-alone or
isolated mode, such lagging var must be furnished by capacitors (properly-sized for
the optimal operating condition) connected across generator terminals. Since
capacitors furnish a fixed amount of var, as the generator Kilowatt output demand
changes so does the generator terminal voltage. In order to minimize such terminal
voltage variations, different schemes have been developed and implemented to
change the amount of capacitors connected to the generator terminals by means of
mechanical and/or electronic switching devices.
9


As already mentioned, induction generators can deliver only Kilowatts by
consuming kilovars from the system. For this reason, induction generators are only
rated in terms of Kilowatt output and not in terms of Kilovars.
1.2 Design and Performance of Induction Machine
1.2.1 Design
There are no major differences in electrical or mechanical design of an
induction motor versus an induction generator. In many cases, a standard induction
motor can very satisfactorily serve as an induction generator. Nevertheless, an
induction machine built specifically to serve as an induction generator may have a
few features specifically tailored for improving its operation and efficiency.
Since the generator does not have to accelerate as a motor, it can be designed
with an especially low rotor cage resistance which results in higher efficiency at the
sacrifice of normal starting torque. Furthermore, deep rotor bars or double-cage
rotors which provide good starting torque and high full load efficiency in an induction
motor do not provide an advantage in an induction generator and therefore are not
needed.
Another area where the induction generator may differ from an induction
motor is in its voltage rating. In an induction generator, the internal air-gap voltage is
always higher than the terminal or bus voltage whereas in an induction motor the
reverse is true. Therefore, induction generators must be specified to have a higher
voltage rating than the bus voltage. For example for a 480-Volt system, the rating of
the generator voltage should be specified as 500 Volts and that for a motor should be
specified as 460 Volts.
10


One last design detail that may be incorporated into an induction generator is
its higher overspeed rating. In many cases an induction generator may experience
severe overspeed conditions resulting from loss of electrical load or separation from
the utility system. Such overspeed conditions may reach twice the rated speed of the
generator. The presence of capacitor banks at the generator terminals does make
further contributions to this overspeed condition. Some generator manufacturers
build a stiffer rotor and support structure to handle the mechanical forces involved
under overspeed conditions. Others provide centrifugal switches on the rotor shaft to
trip the generator at a preset overspeed condition.
1.2.2 Performance
A typical torque-speed or torque-slip characteristic for an induction machine
which is connected to the utility line with constant frequency and constant voltage is
shown in Figure 1.2. The normal operating region of the machine as a motor or as a
generator is between motor breakdown and generator pushover limits at
approximately +0.1 slip. The torque-speed characteristics of an induction motor and
an induction generator are almost duplicate of one another; except, one is an upside-
down mirror image of the other one.
11


Figure 1.2 Induction machine torque-slip curve showing braking, motor, and
generator regions.
In an induction motor, as the mechanical load connected to the motor shaft
increases, speed drops (slip increases) and torque goes up, until the point of
breakdown torque on the curve is reached. Further load increase past this point will
quickly reduce speed down to a stall while line current rises to its locked rotor value.
In an induction generator, as the electrical load connected to the generator
terminals suddenly drops, speed increases (slip become more negative) and net
accelerating torque increases, until the point of pushover torque on the curve is
reached. If speed increases past this point, the generator resisting torque suddenly
drops away rapidly, and the shaft speed increases until prime mover can no longer
keep up, or until the generator self destructs from over-speeding. Once an induction
generator passes the pushover speed, it can reach twice its rated speed in less than a
few seconds which is far beyond what most induction machines are designed to
endure. -
12


What takes place in an induction generator may be best described by an
analogy. Consider that someone is gradually pushing a large sphere uphill toward a
cliff. At any point before reaching the cliff, if the person lets go of the sphere, it will
roll backward to lower ground and stability. However, once the sphere is pushed over
the cliff, it continues to free fall until it hits the ground and self destructs.
Short-circuit performance of an induction generator is significantly different
from that of a synchronous generator. Under fault conditions, the current drawn by
the induction generator to provide its excitation is removed or severely reduced. For
a three-phase fault, current cannot be transferred past the fault point to the induction
generator, and therefore the unit looses its excitation. If capacitors are provided for
the unit, a three-phase short would quickly cause these capacitors to be discharged,
again resulting in removal of excitation from the generator. Once the induction
generator has lost its excitation source, the output current will quickly decay. The
speed of the current decay under fault conditions is a function of the machines ability
to store magnetic energy. After all of the stored energy is used, the output will go to
zero. The maximum and/or initial current magnitude for a short-circuit on the
terminals of an induction generator is determined by the impedance of the device.
This current magnitude will be of the same order as that seen when initially
energizing the unit (five to six times full-load current).
All this differs from the performance of a synchronous generator due in most
part to the separate source of excitation. Figures 1.3 and 1.4 provide graphical
representations of the performances of an induction and a synchronous generator each
connected to a three-phase fault. The phase with the highest magnitude of current is
shown and dc offset is neglected. From these Figures it can be observed that the
decay of short-circuit current for an induction generator is so rapid, compared to that
13


of a synchronous unit, that it is not possible to use this current as an indicator of
system fault conditions.
14


AAAAAAA/V/WM
Figure 1.3 Typical short-circuit waveform of a synchronous generator.
15


1.3 Equivalent Circuit of Induction Machine
The steady state equivalent circuit of an induction machine will be developed
in this section. The extent of this development will be limited to the point where the
reader will feel comfortable understanding the equivalent circuit and be able to make
use of it in machine performance calculations. Some useful equations may be
presented without actually being derived. The reader will be spared from the
rudimentary development of the equivalent circuit which starts off with Faradays law
and rotating magnetic field theory. Discussions on these topics can readily be found
in most classic machinery textbooks.
For simplicity, only machines with a symmetrical-three-phase winding excited
by a symmetrical-three-phase voltage are considered in this section. On a per-phase
basis it may be easier to consider a Y connected machine so that winding current is
the same as the line current and the phase voltage is the same as the line-to-neutral
voltage.
The analysis of induction machines is based on the following standard
simplifying assumptions:
1. All three phases in the rotor and stator are assumed to be balanced.
2. It is assumed that the coefficient of mutual inductance between any stator winding
and any rotor winding is a cosinusoidal function of the electrical angle between
the axes of the two windings.
3. It is further assumed that the rotor is smooth and that the self-inductances of all
the windings are independent of the rotor position.
4. The effects of hysteresis and eddy currents are neglected [4],
The values of resistances and reactances on the stator and rotor sides of the
induction machine can be determined from a combination of field tests and
calculations. Once these values are assigned to the corresponding impedances of the
16


equivalent circuit, the stator and rotor currents can then be calculated for any desired
speed or load condition. The copper losses associated with the stator and rotor
circuits can readily be calculated from the equivalent circuit. When data is available
on other types of losses (i.e., core loss, stray load loss, friction and windage loss),
then all other performance characteristics such as speed-torque, speed-current,
efficiency, power factor, and so on, can be calculated. It is important to note that
stator leakage reactance Xs, rotor leakage reactance Xr, magnetizing reactance Xm,
and rotor resistance Rr, are not constants and will vary somewhat with saturation
caused by a change in load current, frequency, and rotor speed.
1.3.1 Development of Equivalent Circuit
First assume that the rotor winding of a three-phase induction motor is open-
circuited and the bus voltage is applied to the stator winding terminals. Then the only
magnetomotive-forces acting are those produced by the alternating current flowing in
the stator winding. Figure 1.5 shows the equivalent circuit that represents this
condition.
Rs Xs
11 IsZo>""
a
Figure 1.5 Stator equivalent circuit of an induction motor with rotor circuit open-
circuited.
17


Where: Vs = stator terminal voltage per phase
Ej = E = counter emf generated by resultant air-gap flux
Is = stator current
Rs = stator effective resistance
Xs = stator leakage reactance
Xm = magnetizing reactance
From the above circuit, stator current Is can be calculated from the stator
terminal voltage and impedance as follows
1=_______Ys_____
S Rs + j(Xs + Xm)
amps
(1.2)
For the condition of an open-circuited rotor winding, it is evident that stator
current (Is) equals the magnetizing current (Im). Magnetizing current (Im) sustains
air-gap flux wave ( stator, and induces the air-gap emf voltage Ej in the winding. From the viewpoint of
equivalent circuit, the magnitude of the induced voltage Ej is equal to the difference
of stator terminal voltage and the voltage drop across the stator impedance (Rs+jXs).
Assume that the rotor winding of the machine is now closed and that bus
voltage is applied to the terminals of the stator winding. The rotating flux wave (<|)m)
in the stator, produces an induced voltage E2 in the rotor winding which in turn
(since rotor winding is now closed) causes a current to flow in the rotor. The current
flowing in the rotor produces its own magnetomotive force which when it interacts
with the stator magnetomotive-force causes the rotor to turn. The actual voltage
18


induced in the rotor winding E2 is equal to Ej at blocked-rotor conditions. In general,
E2 is equal to Ej multiplied by rotor slip (S) as defined in equation (1.3).
E2 = SE i Volts/phase (1.3)
The resulting current flowing in the rotor circuit therefore is
E2 SEi . , -i ..
Ir=-------^--------=------;k------ amps/phase (1.4)
Rr + jXr(@slip) Rr + jXr(@slip)
Where Rr = resistance of rotor winding
Xr(@ slip)= leakage reactance of rotor winding at slip frequency
As mentioned before, rotor slip (S) is defined as the difference in speed
between synchronous speed and rotor speed, expressed in percent or in per unit of
synchronous speed. The actual rotor frequency (Fr) is equal to
Ft = SFs (1.5)
Where :Fr = rotor frequency
Fs = stator frequency which is the same as synchronous frequency (F)
Accordingly, if Xr is the rotor winding leakage reactance at 60 Hertz, then the
rotor winding leakage reactance at any other rotor frequency (or slip) is given by
X
xr(@ slip)= SXr or Xr= r(@shp). ohms/phase (1.6)
Equation (1.7) for rotor current may then be written as
1= ---------= . amps/phase (1.7)
Rr+jSXr ^(Rr)2+(SXr)2
Dividing the numerator and denominator of the above equation by slip (S)
yields
19


amps/phase
(1.8)
V(Rr/S)2 + (Xr)2
The rotor impedance as seen from the stator side at synchronous frequency
would then be
In terms of stator emf voltage (Ej), and stator load current (same as rotor
current Ir), the impedance as seen across stator air-gap is (Rr/1S')+jXr The
corresponding rotor equivalent circuit as seen from the stator at synchronous
frequency is shown in Figure 1.6. Resistive quantity (R/S) represents the combined
effects of rotor copper loss and the real power consumed by the load connected to the
shaft. An equivalent way of showing the same rotor circuit would be to divide
resistance (R/S) into two parts. Resistance Rr would then account for the copper loss
in the rotor and variable resistance Rr[(l-S)/S] would account for the power
consumed by load connected to the shaft which varies with slip.
= = + jXr ohms/phase
(1.9)


r
r
E
Figure 1.6 Rotor equivalent circuit at stator frequency.
20


Combining equivalent circuit of the rotor with equivalent circuit of the stator
results in the overall equivalent circuit for the induction motor as shown in Figure 1.7.
All rotor quantities are referred to the stator side and the designation Ej is replaced by
plain E from here on to refer to the air-gap voltage. The current flowing into the rotor
equivalent circuit (Ir) is the same as the load current flowing out of the stator
equivalent circuit. Except that, when this current is viewed from the stator side, it
appears to have the same synchronous frequency present in the stator winding. The
voltage across the impedance in the rotor equivalent circuit is the same as the voltage
across the magnetizing reactance (Xm) in the stator equivalent circuit. It should be
noted again that when rotor currents and voltages are reflected into the stator, their
frequency is also changed to stator frequency. All rotor electrical phenomena when
observed from the stator side have the stator frequency because what the stator
winding sees is just magnetomotive-force and flux waves traveling at synchronous
speed.
In Figure 1.7, shunt resistance Rc (across magnetizing reactance Xm) which
represents the core loss effect is omitted for simplicity. In most calculations this
small loss is deducted from the machine at the same time when friction, windage, and
stray load losses are subtracted from the machine. This equivalent circuit is the same
for a motor and a generator. For a motor, slip is positive and for a generator, it is
negative. Thereby, a negative resistance in the equivalent circuit indicates a source of
generation. In Figures 1.7 and 1.8 directions of currents and power flow for a motor
and a generator are indicated and in Figure 1.9 their corresponding phasor diagrams
are drawn.
21


Figure 1.9 Typical phasor diagrams for an induction motor and a generator.
22


1.3.2 Analysis of the Equivalent Circuit
In equations (1.1) through (1.9) important relationships related to currents,
voltages, and impedances in the rotor and stator circuits of an induction machine were
provided. From the equivalent circuit presented at Figure 1.7, additional important
equations regarding the steady-state performance characteristics of an induction
machine can be derived. These equations can be used to calculate speed, losses as
load/torque changes, starting torque, maximum torque, etc., in an induction machine.
The following equations are presented for an induction motor, consistent with
the conventions used in the equivalent circuit of Figure 1.7. The same equations can
be used for an induction generator by using a negative value of S for slip. In turn, a
negative value of S would result in a negative resistor value in the equivalent circuit;
a negative resistor value results in negative power which is an indication of generator
operation instead of motor operation. In an induction motor, copper losses, core
losses, friction and windage losses, stray load losses, etc., are subtracted from the
input electrical power to arrive at the output mechanical shaft power. Whereas, in an
induction generator, all above losses are added to the output electrical power in order
to arrive at the input mechanical shaft power.
In a three-phase induction motor, total power (Pg) transferred from stator
circuit to rotor circuit across the air-gap is
PB = 3(Ir)^ Watts (1.10)
s S
2
A portion of this power transferred across the air-gap is consumed as I R loss
or copper loss in the rotor circuit
Rotor I2R loss = 3(1 r)Rr Watts (1.11)
23


The difference of the two equations is called developed or internal mechanical
power Pd delivered to the rotor shaft
Pd = Pg rotor I2R loss = 3(Ir)2 -3(Ir)2Rr Watts
s
= 3(Ir)2Rr =(l-S)Pg Watts (1.12)
S
Output power Pout is the net power available to do work
Pout = Pd (friction, windage, core, and stray load losses) (1.13)
In an induction motor, a fraction of the air-gap power (SPg) is dissipated as
copper loss in the rotor circuit and the rest of it (l-S)Pg is converted to mechanical
power delivered to the motor shaft. It can be seen that when a motor operates at high
slip values there is more copper loss in the rotor circuit that reduces the efficiency of
the motor.
Similarly, motor torque (T) in Newton-meters at various points can be
calculated by dividing power (P) in Watts by angular velocity (co ) in radians per
second.
T = Newton-meters
CO
T = = - 3 (I r )2 N e wton-meter s
8 s S
Td= =--------------(1-S)P= 3(Ir)2 =T Newton-meters
d CO r C0S(1 S)V g C0S S g
_ Pd (friction, windage, core, and stray load losses)
Where: Tg = torque developed at the air-gap where co is at synchronous speed
(1.14)
(1.15)
(1.16)
(1.17)
24


Td = torque input to the shaft where co is at (1-S) times synchronous speed
Tout= output torque of the shaft where co is at (1-S) times synchronous speed
CO = synchronous angular velocity =----
poles
GO = rotor angular velocity = (1-S)-------
poles
Upcoming equations (1.20) through (1.23) will be better understood if the
subject of equivalent Thevenin circuit is first introduced. The general application of
Thevenin theorem permits replacing a network of linear circuit elements and constant
phasor voltage sources with only one single equivalent voltage source in series with
one single equivalent impedance as seen from arbitrary terminal points a and b of
the original network circuit. The equivalent voltage source is the voltage appearing
across terminals a and b of the original circuit when these terminals are open-
circuited. The equivalent impedance is what appears across the same terminals
looking into the original circuit with all the voltage sources short-circuited.
If Thevenin theorem is applied to the circuit of Figure 1.7 with terminals a
and b as marked and the network circuit of interest being everything to the left of
terminals a and b (which includes the stator circuit and magnetizing branch
reactance), then Thevenin equivalent voltage source and equivalent impedance can be
calculated by the following two equations
(1.18)
Ze = Re+jXe = (Rs+jXs) in parallel with (jXm)
(1.19)
Where: yla = Thevenin equivalent voltage as a phasor quantity
Vs = supply or bus voltage as a phasor quantity
25


Ze = Thevenin equivalent impedance
The equivalent circuit of an induction motor as simplified by the use of
Thevenin theorem is presented in Figure 1.10. Notice that now there is only one
current, namely rotor current (Ir), flowing in the entire circuit. This circuit
configuration makes it easier to calculate maximum torque, slip at maximum torque,
and so on as will be shown in the following equations.
R e X g Rp X r
cMAV-W----- o--------------------------a
b
Figure 1.10 Thevenin equivalent circuit of an induction machine.
The following useful equations presented here will not be derived step by
step; however, their use will be explained. Most machinery textbooks do a good
enough job of providing a detailed account of their derivations so that there is no need
in re-inventing the wheel here. Whats important is knowing that these equations
exist and being able to use them effectively in this thesis or on any another
assignment.
Internal torque of an induction motor in terms of Thevenin equivalent circuit
is
26


1 3V?a(Rr/S)
s '(Re+Rr/S)2+(Xe+Xr)2'
(1.20)
This equation is equivalent to torque equation (1.15) presented earlier with
V?,/
(Re+Rr/S)2 + (Xe+Xr)2
replacing (Ir). From equation (1.20), the torque-
slip curve of Figure 1.2 can be drawn.
Internal torque is maximum when the power delivered to the resistive
component of the rotor circuit (R/S) is a maximum. By use of the principle of
impedance matching from circuit theory, power delivered to (R/S) will be maximum
when its magnitude equals the magnitude of impedance in the rest of the Thevenin
equivalent circuit of Figure 1.10.
= V Ri + (Xe+Xr)2 (1.21)
Smax T
Where: Smax T refers to the value of slip at the point of maximum torque
From the above equation the value of slip at maximum torque is
5max T
Rr
Vr 2 + (Xe+Xr)2
(1.22)
The value of maximum torque Tmax can now be calculated by substituting
equation (1.22) into equation (1.20) for torque as follows
T
1 max
1 1'5V?a
(1.23)
s ^Ri + (Xe+Xr)2
If torque, current, and slip are known for a particular operating point (i.e.,
starting maximum, full load), the following equations may be used to calculate the
same quantities at a different operating point.
27


^max
Tfi
0.5[1 +
(c
smax T
^ Sf.l. ;
]
Tc
start _
Tfi
Ir
Smax T
< Sf.l. >
2
(1.24)
start
T
start
T,
max
!r f.l
2 x ^max T
1 + Smax T
Sf.l.
Tf.l. 2 x Smax T x Sf.l.
^max Smax T + ^f_i_
(1.25)
(1.26)
(1.27)
Note: Slip values in equation (1.27) must be entered as a percentage number (i.e.: a
slip value of 0.035 or 3.5% must be entered as 3.5) in order for this equation to work.
T 2
(1.28)
= /
Xmax

''Smax TV
+
3 max T
V

Note: Above equation is only valid for small values of slip (S).
Rr

(1.29)
Rr + Rr external ^2
Note: Above equation is only valid for small values of slip (S). This equation is only
useful for wound rotor induction machine where slip (hence speed) can be changed
from Sj to S2 by adding external resistors Rr externalt0 the rotor circuit.
/c N2
5 max T
^r start _
!r f.l.
1 +
v Sf.l. )
J1 + Sn
xS
max T
(1.30)
max T
28


Note: Slip values in equation (1.30) must be entered as a percentage number (i.e.: a
slip value of 0.035 or 3.5% must be entered as 3.5) in order for this equation to work.
Now that the basics of an induction machine have been covered, it is time to
explore some of the important engineering topics required for a successful assessment
and application of induction machines in micro- and mini-sized power plants driven
by hydroelectric turbines.
29


2. Application of Induction Generators in Micro- and Mini-
Hydro Stations
2.1 Introduction
The discussion in the preceding chapter outlined the design and operational
characteristics of an induction generator. Its relative simplicity, reliability, and low
cost compared to a synchronous generator are major factors responsible for the
induction generators popularity among small power producers.
Despite these advantages, application of an induction generator in a micro-
and mini-hydroelectric project is not without its own share of obstacles that require
careful considerations. Some of the more generic problems encountered in
developing micro- and mini-hydro stations include: economic feasibility issues, high
cost of fulfilling numerous requirements of regulatory agencies, difficulty in
obtaining permits, the need for hydrological information to determine power
potential, water rights, lack of standardized equipment that could reduce costs, and
difficulties in transmitting the generated power to the utility or consumers in a
relatively small quantity. There are also specific problems such as frequency and
voltage regulation associated with using an induction generator in a stand-alone
configuration that must be addressed.
The majority of the above problems are related to the construction and
licensing parts of the project which are unique for each site. In the limited scope,
time and resources available for this thesis, these problems cannot all be discussed.
Nevertheless, a selected few of the issues directly associated with an induction
30


generator in a hydro setting will be addressed and solutions and/or alternative options
will be proposed.
A major portion of the total cost of the project is the electrical and mechanical
equipment cost for the powerhouse. In an effort to make such small micro- and mini-
hydro projects economically feasible, it is sometimes necessary to consider low cost
alternatives for power production equipment. In this process, it is necessary to look
for energy conversion equipment which is easily available at a low cost, is reliable
and can be easily operated and maintained. In stating low cost, the benchmark is the
cost of conventional equipment. Although the conventional equipment may not be
easily available nor most reliable, still, it offers high efficiency at a premium price.
With some of the alternatives proposed here, high efficiency may be compromised in
order to gain lower equipment cost and higher reliability. This is especially important
in a micro-hydro applications where the main concern is economic feasibility and
reliability instead of high efficiency. In order to make the proper selection, an
economic feasibility study must be performed to take into account cost of equipment,
payback period, life of the plant, tariff rates, etc., for each set of options available.
In searching for easily available equipment, the criterion is to look for
equipment from proven and well-established lines of products commonly available in
the market with well-defined operating characteristics. In utilizing such commonly
available products, the intention is to use them within their capability limits but not
necessarily in the same manner or convention as that for which they were intended.
For example, an induction machine which is built to run as a motor may serve as a
generator and similarly a pump may be substituted for a turbine. By doing so, the
nominal operating mode of operation has been changed; however, the capability
limits are not exceeded. A proper selection of such equipment could have a large
impact on the viability of a small project.
31


The conventional drive mechanism for a hydroelectric unit is a turbine
governor set which is responsible for converting hydraulic energy to mechanical
energy and at the same time regulating shaft speed (i.e., real power output of the
generator). As will be covered in the following section, alternative low cost
equipment that can perform the same function as a turbine is a suitable centrifugal
pump operating in reverse mode. Furthermore, instead of having a slow acting and
expensive mechanical governor to control shaft speed (by controlling the mechanical
input power), it is possible to control speed by controlling the electrical load on the
generator. Electrical devices suited for this application come with various designs
and features. The most common names given to such devices are an electronic load
governor or an impedance controller which have the function of compensating for the
fluctuations in the consumer load so that the total load seen by the generator remains
constant and hence speed and frequency are kept the same.
2.2 The Choice of Turbine and Governor for a Micro- and Mini-Hydro Station
2.2.1 Turbine
A hydraulic turbine consists primarily of a runner connected to a shaft which
converts hydraulic power into mechanical power from the energy stored in water.
The function of a governor is to control the turbines operation, and mechanical speed
and the real power produced by the generator. The hydraulic turbine is the most
important element in a hydroelectric power plant and its proper selection is crucial.
There are two general groups of hydraulic turbines [5], One is the impulse
type where water enters the turbine with high kinetic energy (in the form of velocity)
and a relatively low value of potential energy (in the form of pressure); and, the other
one is the reaction type where the water enters the turbine with high potential energy
32


and a lesser amount of kinetic energy. The reaction-type turbine can be further
subdivided into Francis type and propeller-type turbines. Finally, the propeller type
turbine has the following sub-categories:
1. Fixed blade propeller turbine.
2. Adjustable-blade propeller (Kaplan) turbine.
3. Axial-flow propeller turbine (tube, pit, or bulb).
4. Diagonal-flow turbine.
Simplistically speaking, in the aerated housing of an impulse (Pelton) turbine,
one or more water jets hit the bowl-shaped turbine runner buckets with high kinetic
energy (high velocity) and lose most of their kinetic energy (low velocity). The
forces on the buckets are a result of the impulse or momentum change of the water as
its absolute velocity is reduced to near zero in the buckets. The impulse turbine thus
utilizes the kinetic energy of the fluid entering the turbine to generate power. Impulse
turbines are used for high heads ranging from 1000 to 5000 feet.
In a reaction-type turbine, water enters the intake side of turbine housing with
high potential energy (high pressure) and relatively low kinetic energy (low velocity).
The water then leaves the outlet side of the turbine, which is much larger in diameter
than the intake side, at a lower pressure. The pressure difference across the top and
bottom of the turbine runner exerts a force that causes the runner to rotate and deliver
mechanical energy to the shaft.
In a Francis turbine, water enters the spiral case (curled around the turbine
shaft) from the intake or penstock. Water then passes through the stationary stay rings
(which guide the water flow) and through the adjustable wicket gates (which control
the water flow to the runner and thus power output of the turbine). Finally, water
passes through the runner before it enters the draft tube and into the tailrace. A
Francis turbine runner looks somewhat like a squirrel cage rotor with deep and
33


skewed bars that are tapered in diameter towards the top. Francis turbines are
normally used for medium heads ranging from 100 to 1500 feet.
The runner of a propeller-type turbine has blades similar to a ships propeller.
The number of blades varies from 3 to 10, and they could be either fixed or adjustable
(variable pitch) as their names imply. In most propeller-type turbines the axis of each
individual blade is at a right angle to the turbine shaft; except, in a diagonal-flow
propeller turbine where the axis of each blade makes a 45 angle with the turbine
shaft. Fixed-blade and adjustable-blade vertical propeller turbines are normally used
for medium heads ranging up to 150 and 200 feet respectively. The operating range
of a propeller turbine partly overlaps the operating range of a Francis turbine.
In an axial-flow propeller turbine, the turbine shaft is either slightly inclined
(as in a tube turbine) or horizontal (as in a bulb or a pit turbine). A horizontal or
nearly horizontal turbine shaft allows for a straight-through or nearly straight-through
water passageway from intake to draft-tube discharge. Turbines like these are often
used in tidal and other low head hydro plants and can be designed to operate as a
pump or as a turbine.
The shaft speed of a smaller tube turbine is very slow. In order to reduce the
size and cost of a generator for a tube turbine, a gearbox for increasing speed is
mounted between the turbine shaft and the generator shaft. Since the shaft of a tube
turbine is slightly inclined, it permits the generator be located outside the water
passageway. Whereas in a pit or a bulb turbine the generator is housed inside a
submerged and watertight enclosure in the water passageway. In general, axial flow
turbines are used for low heads ranging from the lowest head practical to 75 feet.
The main advantages of a horizontal bulb turbine in comparison with a
vertical- shaft-propeller turbine (with the generator mounted overhead) are as follows:
34


1. Because of the straight-through water passageway, a bulb turbine is more
efficient than a vertical-shaft-propeller turbine of the same size and
ratings. In a bulb turbine water is not subjected to directional changes in
intake, semispiral casing, and draft tube.
2. The length, width, and especially height of the powerhouse structure can
be kept to a minimum.
3. The excavation and construction costs are lower.
Economics and efficiency dictate that the speed of a turbine should be selected
as high as practical. At higher speeds, the overall size of the turbine and its cost are
reduced. Furthermore, since a hydraulic turbine is usually directly coupled to a
generator, higher speed means higher generator efficiency and lower generator cost
(due to smaller number of poles).
The number of generating units in a power station should be kept to a
minimum. A couple of larger units are more efficient, require less auxiliary
equipment and maintenance compared to several smaller units. Although other
factors such as flexibility of operation, higher-efficiency during low-load demands,
and minimum loss of capacity during shutdown for repair and maintenance may argue
the need for multiple smaller units. The limiting factor in the size of a turbine
generator unit is based on the largest size turbine runner that can be shipped via
ground, railroad, or sea transportation. Still, this is seldom a problem for the turbine
runner of the mini- and micro-hydro units discussed in this thesis.
2.2.1.1 Reverse Mode Centrifugal Pumps as Turbines
A hydraulic machine converts hydraulic energy to mechanical energy or vice
versa. A machine may be designed to perform one of these tasks with maximum
efficiency, this duty being defined as the nominal operating condition. If the same
35


machine is operated with a reversed direction of energy flow, it may be said to be in a
reverse mode operation. It is certain that in reverse mode operation the performance
of the machine will change significantly from that of its nominal operating mode.
For example, hydro turbines are used as pumps in pumped storage projects
and reverse mode centrifugal pumps are used as turbines in small hydro projects. In
the first case, the hydraulic machine is designed to operate most efficiently in the
turbine mode, and in the second case in the pump mode. Despite these departures
from optimum performance and efficiency, reverse mode operation may still be
economically viable since the capital cost involved in providing a separate pump
motor set for the pumped storage facility will be much greater than the savings
involved through the use of an efficient machine. Likewise in a small hydro project,
where the price of a conventional turbine is about five times that of a reversible
centrifugal pump of essentially similar rating, the additional cost for a regular turbine
will be too large to offset the advantage gained from the energy savings involved. For
these economic reasons, a reverse mode centrifugal pump may be used in certain
micro-hydro schemes as an alternative to the conventional turbine. However, a study
of the complete characteristics of the pump is necessary to determine its suitability to
function as a turbine.
A radial-flow centrifugal pump is generally used to convert mechanical energy
into fluid energy. In reverse mode operation, the pump works as a turbine where the
fluid enters the discharge port at a high pressure, drives the impeller (runner), and
leaves the suction port at a low pressure. The fluid energy is thereby converted into
mechanical energy which can be utilized to drive a generator.
Pump manufacturers generally provide characteristics only for pump mode
operation, which relate head, speed, efficiency, power input and flow rate. However,
these characteristics are different in the reverse mode operation. Still, the general
36


characteristics of a pump operating in reverse mode can be obtained from well
documented papers like Buses [6], A few of the generalized salient features of
operating a pump in reverse mode (turbine mode) are summarized below.
1. The turbines peak efficiency point occurs at a lower flow rate than where
the pumps peak efficiency point occurs.
2. The turbines Best Efficiency Point (BEP) takes place at a higher head and
flow rate than the pumps BEP.
3. The turbines power output is higher than the pumps input power at BEP
of each.
4. The turbines maximum efficiency seems to occur over a wider range of
head and flow rate than the pumps maximum efficiency.
5. At a certain low head, water may be passing through the turbine, but no
power will be produced.
6. If the turbines performance curve is known at one speed, its performance
curve at other speeds may be calculated by the use of affinity
relationships. However, the usual degree of accuracy should not be
expected.
7. The runaway speed of the turbine may be calculated from the performance
curves and the use of affinity relationships.
8. The turbines mechanical operation is smooth and quiet like the pumps
operation.
To select a pump for a particular reverse mode duty, it is necessary to specify
the head, flow rate, speed and power output for the turbine duty. Usually the
manufacturers data sheet provides data on the pumps characteristics only. Still,
there are couple of options available in determining the pumps reverse mode
operating characteristics. One option is the use of conversion factors such as those
37


proposed by Buse. These conversion factors relate the turbine BEP performance with
pump BEP performance. A second option is to test a potentially suitable pump in a
test laboratory or at a factory. The test results obtained will be much more accurate
than from any mathematical conversion factors available. Availability of hydraulic
pumps (new or used) far exceeds those of hydraulic turbines and the cost savings
involved could very well make a micro-hydro station economically feasible.
2.2.2 Standard Governor Versus Load Governor
In the past, many studies have been done to regulate the voltage and frequency
of a self-excited induction generator under varying consumer loads. In most such
studies, the frequency regulation is achieved by regulating the speed of the prime
mover by using a mechanical governor. The control of supply voltage is usually
performed by controlling a variable reactive power source. However, changes in
prime mover speed do not result in a linear change in frequency under varying loads
due to changes in slip speed of the machine. Furthermore, difficulties in building a
smoothly variable reactive power source at a low cost have restricted the performance
of voltage regulators. Moreover, the transient response of the generator system to a
load change is limited by the large time constant of the mechanical governor.
Based on several studies [7, 8], regulation of both voltage and frequency can
be achieved by controlling the current drawn from the generator while operating with
an unregulated turbine and a constant excitation capacitance. Regulation of the load
current is accomplished with the aid of an impedance controller connected at the
generator terminals which is capable of drawing a controllable lagging current.
Unlike in large hydro power stations, the conservation of water in the
reservoir is not usually a constraint in micro- and mini-hydro stations. Therefore, full
power is generated with a fully opened gate valve, regardless of the demand on
38


consumer load. The unnecessary power is dissipated as heat in the impedance
controller. If required, the conservation of water in the reservoir is possible by
closing the gate valve slowly during long periods wich low power demands. The
dissipated heat may be used for other purposes such as water heating and steam
making. Thus, the mechanical speed governor responsible for slow transient response
and a significant percentage of total cost of a small power station are eliminated.
A regulator is designed to control the impedance controller such that the load
current of the generator is always at a value which results in rated voltage and rated
frequency under all operating conditions. When the current drawn by the consumer
load drops below the calculated value, the difference in currents is drawn by the
impedance controller connected in parallel to the consumer load. An impedance
controller which provides fast response to control commands can be implemented
with fast switching power electronic devices. As a consequence, the proposed load
current regulator is capable of improving the transient performance of the generator
system significantly.
The quality of an ac power supply is primarily measured by its ability to
supply real and reactive power demand of the consumer load at the rated voltage and
frequency under all operating conditions. The main cause of difficulty in maintaining
the voltage and the frequency at their rated values is the ever-changing nature of
consumer demand. In order to maintain constant voltage and the frequency, the
supply and the demand of real and reactive power should be exactly balanced.
2.2.2.1 Control of Terminal Impedance
A different method to maintain the balance of real and reactive power of the
system with varying consumer loads would be to keep both the power supply and the
demand of power from the consumers constant. The supply of real power is kept
39


constant by using a constant power source such as a hydro turbine operated with a
fixed gate valve position at a constant head. The reactive power supply is maintained
constant by using a constant bank of excitation capacitors. The demand of real and
reactive power from the generator can be kept constant despite the variations of
consumer load by controlling an additional load at the generator terminals, connected
in parallel with the consumer load. This auxiliary load is controlled to consume the
difference between the supply and the demand of both real and reactive powers. In
another words, the impedance seen by the generator at its terminals is maintained
constant. Therefore, the auxiliary load at the generator terminals can be identified as
the impedance controller of the generator system.
There are several advantages in using this control strategy. First, the
regulation of both voltage and frequency of the system is performed by a single
controller instead of two controllers required in the standard method. This results in
reduced complexity of control. Secondly, a mechanical governor regulating the speed
of the turbine is not required for this method of generator control. Thus the cost of
the generating station drops significantly. Thirdly, the slow response of the
mechanical governor system is replaced by the fast response of an electronically
switched impedance controller. As a consequence, the terminal impedance controller
has the potential to provide excellent transient performance which cannot be achieved
by a conventional mechanical governor.
2.2.2.1.1 Proposed Regulator
Due to the availability of power semiconductor devices capable of switching
at high frequencies, an impedance controller with desired properties can be realized in
practically many different configurations. Also, the control of such an impedance
controller is now easier than ever before, mainly due to the development of powerful
40


micro-controllers at lower costs. Despite the availability of powerful micro-
controllers, the control strategies described in this thesis are fairly simple and do not
fully exploit the capabilities of such powerful devices.
Considering a small hydro power station operating with an unregulated
turbine and a constant excitation capacitance bank, the types of disturbances
experienced can be divided into two main categoriesfast and slow disturbances. The
dominant cause of fast disturbances is sudden changes in consumer load an its
impedance. The quality of the power supplied mainly depends on the ability of the
regulator to act against such fast disturbances. In addition, slow disturbances are also
to be expected due to many causes such as fluctuations of the water level of the
reservoir and drifts in system parameters with aging and changes in ambient
conditions.
A regulator with a feed-forward control loop which would monitor the rate of
change in load current would provide an effective control strategy for regulation
against fast disturbances. Similarly, addition of a proportional-integral feedback
control loop acting on the steady-state errors in voltage and frequency is necessary to
regulate the induction generator through the impedance controller against slow
disturbances.
2.2.2.1.2 Impedance Controller
For the successful implementation of the proposed regulator, the impedance
controller must fulfill a number of requirements. First, the controller must be able to
control the flow of real and reactive currents independently in order to be able to
compensate for arbitrary changes in real and reactive power demands from the
consumer load. This can only be achieved when the impedance controller has at least
two control variables. Second, it must provide a fast response to a given control
41


command so it can have high performance during load changes. With feed-forward
regulation, the delay in acting against changes in load current depends only on the
delay in calculating the desired control commands and the delay in responding to
these commands. The calculation of the proper values of the control commands is
mainly determined by the speed of the microprocessor. Since present day
microprocessors can perform this task within a few hundreds of a microsecond, the
regulator performance is primarily limited by the response time of the impedance
controller.
Third, a continuous control of real and reactive currents drawn from the
impedance controller minimizes any jumps in voltage and frequency due to actions of
the regulator. Therefore, an impedance controller with high control resolution is
important.
Other important features of an impedance controller include: a simple and
rugged circuit with high reliability, minimum number of circuit components to keep
the cost down, and the ability to minimize harmonic distortion from the fast switching
of power electronic devices.
There are several different circuits that could satisfy the basic requirements of
an impedance controller. In most cases, solid-state switching devices such as phase-
controlled thyristors are used in some form of a static var circuit configuration to
control and vary the effective amount of resistive, capacitive, and inductive loads on
the generator. After a careful examination of all the available options, the circuit
configuration that best suits our current application and meets the above requirements
is the one which uses three fixed capacitors, a controlled bridge rectifier, and a single-
quadrant chopper (operates in the first quadrant only) and a resistor on the dc side of
the rectifier. This circuit configuration is shown in Figure 2.1.
42


Figure 2.1 Circuit configuration of the impedance controller with rectifier and
chopper.
The resistively loaded bridge rectifier draws both real and reactive currents
determined by the rectifier delay angle. The function of the chopper is to change the
magnitude of the current drawn by varying the conduction period as is done in pulse
width modulation. This configuration requires only one resistor, seven power
semiconductor switching devices, and three fixed excitation capacitors. The
thyristors of the bridge rectifier are line commutated and the single-quadrant chopper
can be implemented with a switching device such as GTO or BJT which have gate
turn-off capabilities. The reduction of harmonic distortion on the ac supply voltage is
accomplished by switching the chopper at high frequencies. The delay angle of the
rectifier and the pulse width of the chopper can be continuously and independently
controlled to achieve high resolution. Furthermore, with high frequency switching of
the chopper, this configuration provides the fastest response of all alternatives under
consideration.
43


In implementing the selected impedance controller, the first element to be
evaluated is the controller resistance R current ratings of all power semiconductor devices in the circuit. The selection of Rdc
is made such that the impedance controller provides the desired range of control for
the consumer load.
This impedance controller draws both real and imaginary current components
through the bridge rectifier and chopper. Both direction and magnitude of the current
drawn could be varied as changes are made to the delay angle of the rectifiers a and
pulse width of the chopper PW. As a result, a and PW are used as the two control
commands to control of power factor and magnitude of the current drawn from the
circuit.
One of the common problems with power electronic devices is the harmonic
distortion injected into the supply voltage waveform. Similar to a bridge rectifier, the
impedance controller under consideration introduces 5, 7, 11,13,..., etc., harmonics
into the ac voltage supply. The excitation capacitors at the input terminals of the
rectifier act as a low pass filter to reduce some the distortion of the ac voltage
waveform caused by the impedance controller. Further reduction of the harmonics
can be achieved by increasing the switching frequency of the chopper analogous to
the pulse width modulation of an inverter. The high switching frequency of the
chopper reduces the magnitudes of lower order harmonics such as 5, 7, 11 at the
expense of increased higher order harmonics. Since the ac system acts like a low pass
filter, the net result is a reduction in distortion in the voltage waveform.
Although as the switching frequency of the chopper is increased, the overall
voltage distortion gets reduced; this benefit diminishes as the switching frequency is
increased beyond a certain limit.
44


Furthermore, practical problems such as the maximum switching frequency of
power semiconductor devices such as GTOs limit the number of switchings of the
chopper that can take place during a 60 period that each of the six rectifiers is
conducting. The results of a study performed by Hoops [9] show that five switchings
of the chopper per each conduction period of the rectifier (every 60) to be the
optimum switching frequency.
For easier understanding, Figures 2.2 and 2.4 show the current waveforms of
the bridge rectifier alone without the presence of the chopper circuit for delay angles
of 30 and 90 respectively. Figures 2.3 and 2.5 show the complete operation of the
bridge rectifier and the chopper combined with the same delay angles, five pulse
chopping per each conduction period, and a chopper pulse width of 0.5 (50% on, 50%
off). With five-pulse chopping there are 30 choppings per each cycle of the ac supply.
This amounts to chopper frequency of 1800 Hertz for a fundamental frequency of 60
Hertz on the ac supply.
45


Figure 2.2 Rectifier line current for phase A with a=30 without the chopper.
chopping and PW=0.5.
46


Figure 2.4 Rectifier line current for phase A with a=90 without the chopper.
chopping and PW=0.5.
47


In a case where there is a strict requirement for reducing the harmonic
distortion to a minimum, a harmonic filter can designed by using some or all of the
excitation capacitors and extra inductors to form an LC filter. Typically, the
harmonic distortion of voltage without any filtering is only about five percent for
five-pulse chopping. The contribution of an LC filter is fairly limited. In reported
studies, it was able to reduce the total harmonic distortion from about five percent to
about 1.4 percent with added cost and complexity to the project. Therefore, the use of
such LC filters is not recommended for this application.
In the interest of brevity, most of the mathematical analysis, computer
simulation studies, and intermediate experimental results associated with performance
and characteristics of the impedance controller are excluded from this thesis. Instead,
a summary is presented on the effects of varying delay angle (a) and pulse width
(PW) on real and imaginary currents drawn from the impedance controller and how
well the closed loop control regulator performs in conjunction with the impedance
controller.
The primary effect of the rectifier delay angle (a) is to change the power
factor of load current and the primary effect of the chopper pulse width (PW) is to
change the amplitude of the current. The pulse width of the chopper has almost no
effect on the power factor of the current drawn. As a result, power factor of the
impedance controller is almost completely determined by the rectifier delay angle (a)
and the amplitude of current is almost directly proportional to the per unit pulse width
(PW). In short, after the power factor has been determined by a, PW determines the
amplitude of the current. Consequently, a change in the amplitude of the load current
can be corrected much faster than a change in the phase angle of the load since
chopper frequency is five times faster than the rectifier frequency.
48


Figure 2.6 shows the control range of the impedance controller as delay angle
(a) and pulse width (PW) are varied for terminal voltage (V) and impedance
controller resistance (Rdc) set to one per unit each.
Figure 2.6 Area of current control offered by the impedance controller for
Rdc=l .0 pu at rated voltage.
Furthermore, experimental results using a thyristor bridge rectifier and a GTO
chopper showed close correspondences among theoretical, computer simulation, and
laboratory experimental results.
The feed-forward regulator displayed a superb performance against load
changes. When the generator system is regulated only by the feed-forward regulator,
the steady-state errors of voltage and frequency are 0.04 pu and 0.01 pu respectively.
However, when the same load change is applied to the generator system regulated
with both feedback and feed-forward regulators, the errors of voltage and frequency
49


are slowly brought back to zero. Thus the small errors resulting from the use of
approximately calculated control commands can be eliminated by the action of the
feedback regulator.
The results of computer simulation studies demonstrate the superior
performance of the chosen regulator. The action of the feed-forward regulator results
in excellent transient performance for changes in consumer load. Except for the small
steady-state errors resulting from the use of approximately calculated control
commands, the transients in voltage and frequency are over within two cycles. The
remaining steady-state errors are corrected by the slow-acting feedback regulator.
The feedback regulator guarantees errorless operation in steady-state even when the
operating parameters change from their normal values.
According to the proposed strategy, the generator is operated with an
unregulated prime mover eliminating expensive mechanical speed governors. The
regulation of voltage and frequency is achieved by drawing the correct amount of load
current resulting in rated voltage and rated frequency under any changes in operating
conditions. An impedance controller connected at the terminals of a generator draws
the difference between the generator current and the consumer load current. The fast
disturbances resulting from changes in consumer load impedances are quickly
compensated by the feed-forward regulator.
Compensation for slow-acting disturbances and steady-state errors resulting
from feed-forward regulation is made by a feedback regulator. Thus the regulator
draws the correct amount of current even when operating parameters slowly drift
from their normal values. Therefore, the described regulator and impedance
controller have the potential to improve the performance of the induction generator in
great proportions.
50


2.3 Induction Generators in Micro-Hydro Stations as Stand-Alone Units
2.3.1 De-Rating of Induction Machine for Unbalanced Operation
2.3.1.1 Introduction
A major difficulty of power production in micro range (less than 100 KW) is
that most of these units are located in remote areas, operate as stand-alone units, and
serve only the local loads. Types of loads found in the majority of these remote
locations are either single-phase or unbalanced. Yet, the generators used for these
applications are balanced three-phase generators that require balanced three-phase
loads in order to produce their rated outputs. One way of handling this problem is
simply by de-rating the induction generator. But from an economic point of view,
this is not attractive. Another solution would be to study each load profile and
redistribute the electrical loads. The goal is to redistribute the electrical load in such a
way that for any given time period, each phase of the generator has an equal share of
the total load. Most likely this wont be a one-hundred-percent solution and a small
amount of de-rating of the induction generator still has to be done based on the worst
case of unbalance condition.
Another method suggested in reference [10] proposes the use of symmetrical
components to calculate reactance values for an unbalanced three-phase excitation
bank such that when connected to the unbalanced load bank, the combined impedance
of the two banks as seen by the generator appears balanced. This concept is also
further developed to make a single-phase load appear as a balanced three-phase load
to the induction generator.
Quite often, de-rating of an induction generator seems simpler and more
straightforward compared to the other alternatives available. So it would be prudent
51


to discuss and develop a method for handling de-rating of an induction generator in
this section. The following analytical methods are applicable for de-rating an
induction motor as well as a generator.
It is widely known that induction machines are sensitive to unbalanced
operation. The National Electrical Handbook states that if the supply voltage has an
unbalance greater than five percent, the induction motor either must be taken out of
service or properly de-rated first. A relatively small unbalance in the voltage of an
induction motor causes a rather large unbalance in line currents which result in
excessive machine temperature rise. The formulas developed in the following section
can be used to de-rate an induction machine subject to unbalanced three-phase
voltages.
2.3.1.2 Balanced Induction Machine Subjected to Unbalanced Voltages
The effects of unbalanced voltages on a balanced three-phase induction
machine can be studied by the use of symmetrical components. In a three-phase,
three wire system (no neutral wire) the unbalanced voltages can be resolved into a set
of positive voltages and another set of negative voltages. Since the induction
machine is assumed to be symmetrical, the positive sequence voltages give rise to the
positive sequence currents in the positive sequence equivalent circuit of the machine,
and similarly (yet independently) the negative sequence voltages give rise to the
negative sequence currents in the negative sequence equivalent circuit of the machine.
The following Figure shows the positive and negative sequence equivalent circuits of
an induction machine.
52


Rg Xg Xf< Rg Xg X r-
Figure 2.7 Sequence network circuits of an induction machine.
In the positive sequence circuit the direction of power flow is in for a motor
and out for a generator. In the negative sequence circuit the direction of power flow
is always inward since the value of variable resistive R,7(2-S) is always positive under
either motoring or generating operating conditions.
Z+ and Z- reflect the equivalent positive and negative sequence impedances
of the circuits. Next, their relative magnitudes at different operating points are
estimated for later use in the calculations.
At standstill (S = 1): Z+= Z~ or Z+/Z_= 1
At synchronous speed (S = 0): Z+ Z~
At no load speed (S = 0.005): Z+/Z-= 12
At full load speed (S s 0.025): Z+ /Z~ = 7
As will be demonstrated by the following example, the large ratios of
Z+ / Z~ at no load and at full load indicate that an induction machine is highly
sensitive to voltage unbalances. A small percentage of voltage unbalance will cause a
53


much larger percentage of negative sequence current to flow. For example, if the
unbalance factor (ratio of V~ / V+) in a typical motor operating at full load is 10%,
then:
= V-/Z = x = 0.1x7 = 0.7 per unit (2.1)
I+ v+/z+ V+ Z"
So for a 10% unbalance factor, the negative sequence current (I~) is about
70% of positive sequence current (I+). The positive and negative sequence currents
produce positive and negative sequence torques respectively. The positive sequence
torque is productive and does the work in a motor or the generator while the negative
sequence torque is always counterproductive. It acts to brake or slow down the
machine, causes heating problems, and consumes energy from the electrical system.
2.3.1.3 Balanced Self-Excited Induction Generator Subjected to Unbalanced
Loads
In the case of a self-excited induction generator, the capacitors connected
across generator terminals furnish excitation reactive power requirements of the
generator. The capacitors must also meet the var requirements of the load. When an
induction generator is operating in the self-excited mode, the generator frequency
adjusts itself to a value where the following two power equilibrium conditions are
met.
1. The terminal voltage assumes a value where the net reactive power flow in
the circuit is zero.
2. The slip assumes a value where the net real power flow in the circuit is
zero.
54


Figure 2.8 Direction of power flow in a self-excited induction generator.
In this stand-alone (self-excited) mode of operation, the only energy source in
the circuit is the variable negative resistance shown in the equivalent circuit above.
Positive sequence power is fed from the generator to the load. Some of this power is
absorbed by the unbalanced load and the remainder is converted to negative sequence
power and fed back to the induction generator. Some of negative sequence power fed
back to the generator is absorbed as copper loss, core loss, and the rest is converted
into parasitic negative sequence torque. Figure 2.8 shows how real and reactive
power are accounted for in a balanced self-excited induction generator feeding an
unbalanced load.
2.3.1.4 Calculation for De-rating of an Induction Machine
Calculation for de-rating of a balanced three-phase induction machine
subjected to unbalanced three-phase voltages is done by studying the positive- and
negative-sequence equivalent circuits separately. The total copper loss due to
unbalanced voltages equals the sum of copper losses produced in the positive- and
negative-sequence equivalent circuits independently.
55


In calculating an approximate expression for the copper loss, the magnitude of
rotor current (Ir) is assumed to be equal to the magnitude of the stator current (Is).
This is a fair assumption since the magnetizing current (Im) is on the order of 0.4 per
unit and it is nearly in quadrature with the rotor current ( Is =Im +Ir ).
Therefore, the magnitudes of Is and Ir are within 90% of each other. Then, the
approximate expression for the total copper loss can be written as follows.
Pr(& + I?-)(Rs+Rr) (2-2)
For an unbalance factor of 0.1 and a typical ratio of sequence impedances of
Z+ / Z_ =7 at full load, it follows that
Is- v8-/z-_vz
Is+ V/z+ vs-
2
= 0.72 = 0.49 per unit
x = 0.1x7 = 0.7 per unit
T
+
(2.3)
(2.4)
li_(Rs + Rr) = 0.49ls+(Rs + Rr) (7-5)
P"U = 0.49PJU (2.6)
This last expression shows that total copper loss will be increased
substantially even for a small value of unbalance factor. The negative sequence
copper loss may reach even higher values if the voltage unbalance in one phase is
significantly higher than the average unbalance of the three phases. Under the above
unbalance conditions, the motor cannot operate continuously at its rated output level
and it will sustain damage from overheating. Therefore, it is necessary to establish a
method for de-rating the induction machine.
Before a method can be developed, a guideline or a criterion for de-rating
must be established. One limiting criterion can be that the current in the most heavily
56


loaded phase should not exceed the rated value. This criteria is somewhat limiting
because it does not make any allowance for temperature equalization between phases
to take place.
A better criterion which is more acceptable for de-rating of smaller induction
machines states that the total positive- and negative-sequence copper losses should
not exceed those losses experienced under balanced rated conditions. Under this
criterion, even if one phase experiences a higher temperature rise than the other two
phases, temperature equalization between phases should help to bring down the
temperature of the hottest phase. Experimental test results reaffirm this assumption
and therefore this criterion will be adopted as the basis for the calculation that
follows.
l2i(Rs + Rr)-l2+(Rs + Rr) + l|_(Rs + Rr)
's+=M^-
(2.7)
(2.8)
Where: Is+
is the maximum allowable positive-sequence stator current
111
If.l.
Is_ = V- / Z~ is the negative-sequence stator current
If.l. = V/Z+ is rated stator current under balanced condition and
V is the voltage under rated balanced condition
x
v x Z~
(2.9)
Note: for small unbalance factors =
V v+
Is = If.l.
x
V+ Z"
By substituting equation (2.10) into equation (2.8) an expression for
maximum allowable output under the unbalanced voltage condition results.
(2.10)
57


Is+=If.l.
1-
V- z+
--x
V+ Z
(2.11)
Maximum Allowable Output=
. 1- v~ z+ x
1 V+ Z-_
x rated output
(2.12)
This equation for de-rating of an induction machine is plotted in Figure 2.9.
Some textbooks as well as Gleason and Elmore [11] refer to the ratio of Z /Z as
being essentially equivalent to the easily obtainable ratio of the starting current to the
full load current in an induction machine.
A better approximation results if the above equation is written in terms of
rotor current (Ir). Let X be the no load magnetizing current expressed in per unit.
Then, from approximation (Is2 = Im2 +Ir2 ), per unit value of rotor current 1+ under
rated balanced condition can be written as
(2.13)
Ir(p.u.)+=t/l
2_x2
Under unbalanced conditions the same equation in per unit transforms to
IL
if.i.
it
if.i.
i2
- X1 per unit
(2.14)
At If, =1.0 equation (2.8) can be written as it = yjl~ I2- and using equation
(2.14) the per unit rotor current can be written as
IL
if.i.
if.i.
- X2 per unit
(2.15)
58


I
+
r
If.l.
1 X 2
per unit
(2.16)
1 X2
V
---x
V +
per unit
(2.17)
Maximum allowable output =
If.l.
UFx
1 X2
per unit
(2.18)
Where UF is the unbalance factor.
This equation for de-rating of an induction machine which also takes into
account the per unit magnetizing current (X) is plotted in Figure 2.10.
59


60


One area that had been not been discussed until now is the effect of
unbalanced voltages on increasing the core loss. Fortunately, the core loss of an
induction machine under balanced operation is relatively small. Furthermore, the
unbalanced voltages will have very little effect in increasing this core loss in an
induction machine. The results of Williams study [12] show that the increase in
temperature rise due to this increase in core loss will be negligible. Therefore, in this
approximate calculation for de-rating of an induction machine due to unbalanced
voltages, the core loss will be considered unchanged and it will not have a
temperature rise effect on the induction machine.
Worth mentioning, unbalanced voltages also affect a few other performance
characteristics of an induction machine. For instance, unbalance voltages cause
locked-rotor and breakdown torques to decrease. This reduction in torque must be
taken into account before selecting an induction machine for a particular application.
Also, unbalanced voltages cause a lowering in full-load speed of an induction motor.
The motor, as well as a generator, experience higher losses which translate into a
higher slip at full load. Finally, the locked-rotor current will be unbalanced to the
same degree that the voltages are unbalanced at starting (at standstill, Z+ /Z_ = 1);
however, the current will be unbalanced at about 6 to 10 times the unbalanced
voltages at normal operating speed (at full load speed, Z+ /Z" s 7).
2.3.2 Optimizing Capacitor Size for a Stand-Alone Induction Generator
The use of capacitors as an excitation var source for stand-alone (self-excited)
induction generators has already been mentioned in previous sections of this thesis. It
should be clear that when a sufficiently large three-phase capacitor bank is connected
to the terminals of an externally driven induction generator, a small voltage tends to
be generated in the windings. This voltage is generated due to the residual magnetism
61


of the rotor or the initial charge on the capacitor bank. The resulting currents flowing
in the capacitors provide a positive feedback to increase the induced emf voltage in
the generator windings. This process is called self-excitation. The voltage
continues to increase until it reaches its steady state value. A steady state voltage is
reached when the magnetizing reactance Xm saturates to a level high enough,
determined by the excitation capacitance and the rotor speed.
One of the major application setbacks of a self-excited induction generator
(SEIG) is its poor voltage regulation even under controlled speed. In order to
maintain a constant terminal voltage while the load is being increased, the amount of
capacitive var must be increased proportionally as well. There are a number of
voltage regulating schemes that reportedly work satisfactorily. Some of these
schemes include static var compensators, switched capacitors, saturable core reactors,
or any combination of the above. Most of these schemes use closed control loop,
relays, contactors, and semiconductor switching operations that are complex,
expensive, require maintenance, and produce harmonics.
In effect, these types of voltage regulating schemes negate some of the
advantages of choosing an induction generator for use in a mini- and micro-power
generating unit in the first place.
Besides the above options and the impedance controller option already
discussed, there are two simple and inexpensive capacitive var compensating schemes
that exhibit almost a flat voltage regulation profile as the load changes. As discussed
by Shridhar in reference [13], the two schemes are referred to as short shunt and long
shunt SEIG schemes. These schemes eliminate the need for an additional voltage
regulator since the schemes have inherent self-regulating features. Figures 2.11 and
2.12 show the block diagrams for these schemes.
62


Figure 2.11 Block diagram of a short shunt self-excited induction generator.
Figure 2.12 Block diagram of a long shunt self-excited induction generator.
Shridhar [13] has demonstrated through his calculations and lab results that it
is possible to have an almost flat voltage regulation across the load with a particular
combination of capacitors connected in short shunt and long shunt configurations.
The results of his experiments conclude that the performance of a short shunt
configuration is superior to a long shunt configuration. The analytical procedures for
selecting the optimum capacitor sizes and predicting performance for the short shunt
scheme is explained in the following paragraphs.
The objective of the following procedure is to find an ideal combination of
series and shunt capacitor sizes that would maintain the load voltage within
acceptable limits from no load to full load power output. The actual study itself
involves modeling, computer simulation, analysis, and estimation of desired
parameters. Equations (2.19) through (2.29) represent the highlights of the actual
63


process involved. It will become evident that once the procedure is established and
understood the design of a self-regulating self-excited induction generator is fairly
straightforward.
The following procedure takes into account variation of magnetizing reactance
Xm with saturation as the basis of this calculation. The steady state equivalent circuit
of Figure 1.7 is modified and redrawn in Figure 2.13 to reflect the variation of
reactances with frequency. This step is necessary, because in a stand-alone induction
generator, speed (and hence frequency) is not regulated by the utility. As the
operating speed changes so will the values of the circuit reactancesespecially the
value of Xm. The rotor resistance quantity R/S" is rewritten in terms of per unit
frequency and per unit speed as Rr[F/(F-V)] where the terms inside the bracket are
equivalent to 1/S. Finally the external load resistance, series, and shunt capacitances
are added to the original equivalent circuit.
Figure 2.13 Equivalent circuit of a short shunt self-excited induction generator.
Where :Xsh = Per phase capacitive reactance corresponding to shunt capacitance, Csh
Xse = Per phase capacitive reactance corresponding to series capacitance, Cse
Rl = Per phase load resistance
64


F = Per unit (PU) frequency
V= Per unit speed
Is, Ir, IL = Per phase stator, rotor, and load currents respectively
The ratio of air-gap voltage over frequency (EIF) is an indication of flux
density in the generator air-gap. Figure 2.14 shows a typical curve demonstrating
variation of Xm versus EIF. The magnetizing characteristic of the machine which are
shown in terms of Xm here is critical in analyzing the SEIG.
Figure 2.14 Variation of Xm versus EIF.
The saturated region of the above curve can be expressed mathematically by
the following equation
E//r= KrK2Xm (2.19)
Where K1 and K2 are constants that correspond to the starting point and slope
of the above curve respectively.
For the purpose of solving the circuit unknowns in Figure 2.13, lets assume
the values of Xse and Xsh are given. Furthermore, speed of the prime mover and the
65


load Rl connected to the generator are known. Then the only unknowns remaining in
the circuit of Figure 2.13 would be the values of Xm and per unit base frequency (or
stator frequency) F. These two unknowns can be evaluated by writing a current loop
equation in the stator circuit as follows
IsZs = 0 (2.20)
Where: Zs= Zj + Z^ ZL/(Zsh+ZL) + Z2Zm/(Z2+ Zm)
Z,=Rs+jXsF Z2=Rr-E- + jXrF Zm =jXmF
F-V
Zsh = j^sh / F ZL = Rl jXse / F
When the generator is operating under load, stator current fs is greater than
zero. Therefore, from equation (2.20), the value of zs must equal zero. zs is a
complex quantity which could be broken down into its real and imaginary parts. By
setting each of the real and imaginary parts of zs to zero, now there are two
equations and two unknowns which could be solved for Xm and F. The following
two equations show the general form and order of the equations obtained.
Real Zs= (D1Xm+D2)F3+(D3Xm+D4)F2+(D5Xm+D6)F+
(D7Xm+D8) 0 (2.21)
Imaginary Zs = (CjXm+C2) F4 +(C3Xm+C4) F3 +(C5Xm+C6) F2+
(C7Xm+C8)F+(C9Xm+C! 0) = 0 (2.22)
The C and the D constants in the above equations are defined in appendix-A.
The above equations can be solved using a number of numerical methods including
the Newton Raphson or the Secant method. There are numerous papers published on
the above topic. In some of the other papers, the order of the equations derived may
be different, but the general approach is very similar to what is presented here. This
66


is mostly because of the assumptions and simplifications made in the papers. For
example, in this study for simplicity it is assumed that the value of rotor reactance Xr
and stator reactance Xs are equal and that the load is purely resistive.
Once values of Xm and F are found, then equation (2.19) can be used to
calculate E. Subsequently, once E is known, the following equations can be used to
calculate voltages and currents in the generator circuit.
Is = E/[Z1 + ZLZsh/(ZL + Zsh)] (2.23)
Ir = E / [RrF / (F V) + jFXr] (2.24)
= ^s^sh / (^L + Zsh ) (2.25)
vs = iLzL (2.26)
Vl = IlRl (2.27)
Pin = -3Ir2Rr- 10 r r F V (2.28)
P0ut-3IL R^ (2.29)
As mentioned earlier, if the values of shunt and series capacitors, prime mover
speed, and load are given, then above equations can be used to solve the remaining
circuit parameters (i.e., Xm, F, voltages, currents, and power). The next step would
be to use the above equations and perform mathematical and/or experimental trial
studies to determine what size capacitors can keep the voltage regulation within an
acceptable limit.
From the equivalent circuit of Figure 2.13, it is clear that under no load
condition, load current IL flowing through the series reactance Xse and load resistance
Rl are zero. Therefore, Xse (or series capacitor Cse ) has no effect on generator
terminal voltage under no load condition. Therefore, under no load condition, shunt
67


reactance Xsh (or shunt capacitor Csh) is solely responsible for maintaining generator
terminal voltage. The effect of varying shunt capacitor Csh on terminal voltage can be
studied to determine the range of values for a shunt capacitor that will maintain the
terminal voltage to within 6 percent of rated terminal voltage. A typical curve
which resulted from the study of a laboratory 3.7 KW, three-phase induction
generator is presented in Figure 2.15.
Figure 2.15 Variation of terminal voltage with shunt capacitor Csh.
The above graph shows that as the value of Csh is increased, so will the no
load terminal voltage Vs of the generator. As a first trial, a value of Csh equal to 16.7
[iF is chosen. This value corresponds to a no load voltage of 1.0 per unit. With the
value of Csh fixed, another mathematical and/or experimental trial study is performed
to determine regulation of load voltage VL (from no load to full load power) as the
value of series capacitor Cse is changed. Figure 2.16 shows the results of such study.
The two vertical dashed lines in Figure 2.16 represent the range of values for series
capacitor Cse that would maintain the generator terminal voltage Vs to within 6
68


percent of rated value. This Figure also shows that load voltage regulation can be
minimized to about 2 percent when the value of the series capacitor is set at 40 pF.
0 20 40 60 80 100 120
Csein micro Farads
Figure 2.16 Effect of Cse variation on load voltage regulation.
The first set of suitable values for Csh and Cse have now been computed.
Once the mathematical routine is set up to calculate one set of values (such as those
calculated above), then it is relatively simple to play a game of what if? with the
computer and change the value of series capacitor Cse. In the next step, the effects of
Cse at 20,40 and 60 pF are studied to determine impact on load voltage VL, terminal
voltage Vs, and stator current Is at different power output levels. The objective of this
study is make sure generator voltage and current ratings are not being exceeded while
the capacitors maintain load voltage regulation to within an acceptable limit. The
following three Figures graphically show the results of such study.
69


Figure 2.17 Variation of load voltage VL with load.
Figure 2.18 Variation of air-gap voltage E with load.
70


Figure 2.19 Variation of stator current Is with load.
From examining the above Figures, the following observations could be made.
When a lower value of Cse corresponding to 20 pF is used, load voltage regulation is
high (Figures 2.16 and 2.17), generator terminal voltage rises above its rating (Figure
2.8), and rating of stator current is somewhat exceeded (Figure 2.19) at rated output
power. When a higher value of Cse corresponding to 60 pF is used, load regulation is
slightly poor (Figures 2.16 and 2.17), generator terminal voltage is normal (Figure
2.18), and the stator current drops below one per unit at rated power output. So with a
higher value of Cse the generator output capability can be increased at the expense of
higher voltage regulation. There is also an extra cost involved for the purchase of
additional capacitors and the generator may experience higher overspeed conditions
in case it loses its load.
A further adjustment of the series capacitor size to Cse = 45 pF demonstrates
the best overall performance of the system. At this setting, the voltage regulation is
as low as two percent and the generator can be loaded to 107 percent without
71


exceeding the voltage and current ratings of the generator beyond manufacturers
recommendations. The low value of voltage regulation makes the SEIG attractive
when compared to the synchronous generator. The synchronous generator, in
addition to being more expensive, also has complex control systems as covered
earlier. Figure 2.20 combines all the important characteristics of the circuit in one
graph for this new setting.
Figure 2.20 Overall characteristics of a short shunt SEIG.
After going through this entire process, there may be one other aspect of the
design which may be worth investigating. It should be of no surprise that the shunt
capacitor Csh also has some effects on load voltage regulation. What would happen if
instead of setting Csh to 16.7 pF, it was changed to 15.3 and 18 pF. These values
correspond to 0.94 and 1.06 per unit of no load terminal voltage in Figure 2.15.
From the results of such a study, it is concluded that Csh set at 16.7 (corresponding to
unity no load voltage) still produces the best performance on load voltage regulation.
The results are shown in Figure 2.21. As seen from this Figure, for smaller values of
72


Csh an undesirable voltage dip appears across the load at partial generator output
levels.
Figure 2.21 Effects of changing shunt capacitor Csh on load voltage.
The last topic to be covered in this section is investigating the performance of
a long shunt SEIG (see Figure 2.12) versus a short shunt SEIG. Derivation of
equations and circuit analysis of a long shunt SEIG is fairly lengthy yet similar to
what has already been covered for a short shunt SEIG here. Therefore, the reader will
be spared from a detailed coverage of long shunt SEIG. Nevertheless, mathematical
and experimental results from the study are discussed ahead.
Analyses and experiments similar to those were conducted for a short shunt
SEIG were also repeated for the long shunt configuration. The results show that like
a short shunt SEIG, the long shunt SEIG also has self-regulating features. However,
the best voltage regulation that a long shunt SEIG can produce is about six percent
compared to two percent that was obtained from a short shunt SEIG. The size of the
series capacitor required for a long shunt configuration is about 100 pFmore than
double the size and cost of the capacitor used for the short shunt arrangement.
73


Furthermore, in the long shunt configuration, the series capacitor must have a higher
current rating to be able to handle not just the load current, but the entire generator
line current. So with minimal time spent on long shunt SEIG, it can be concluded
that performance characteristics and economics make the short shunt SEIG more
attractive than the long shunt SEIG.
In a follow up study presented in reference [14] the transient performance of
short shunt SEIG is compared to the long shunt and the simple shunt (no series
capacitor) configurations. Here, the conclusion drawn reinforces the performance
superiority of short shunt SEIG configuration over that of the other two
configurations during steady state as well as transient state. Some of the relevant and
noteworthy findings of this study are summarized below.
After a short-circuit, a simple shunt SEIG cannot build up voltage and self-
excite itself if the load remains connected to it. The residual magnetism of the
machine reduces to such a low level that an external jolt is required before re-
excitation and voltage build up is possible. A simple shunt SEIG has poor voltage
regulation, low overload capability, and problems with voltage build up after a short-
circuit on its terminals.
On the contrary, the short shunt SEIG not only can re-excite itself after a
short-circuit, it can do it while the load remains connected to the generator. The
series capacitor reinforces the voltage build up process and prevents the generator
from becoming totally de-magnetized. The short shunt SEIG has good steady state as
well as transient performance. It has a high overload capability and can withstand
switching transients of loads up to 1.6 per unit without losing self-excitation.
74


2.4 Induction Generators in Mini-Hydro Stations
In previous sections the use of induction generators in micro-hydro stations
was emphasized. Hydro power generation of capacity less than 100 KW is
categorized as micro-hydro generation. The economics as well as the technical
problems encountered in micro-hydro schemes are different from those of the mini-
hydros. The economic aspect of a micro-hydro scheme may suggest the use of a
reverse mode centrifugal pump along with an impedance controller instead of a
conventional turbine governor set. However, in mini-hydro schemes the problems are
of a different nature.
An induction generator may be cheaper, more robust, and relatively
maintenance free compared to a synchronous generator in the medium ratings (100
KW to 5000 KW), making it a better choice than a synchronous generator; but the
generation must be controllable-much more so than in a micro-hydro scheme.
Satisfactory parallel operation must be ensured by giving attention to the stability
conditions. To make the scheme economically viable the generators must work in
parallel with the grid system where there are other synchronous generators available
to supply the excitation power. These problems make the mini-hydro scheme different
from the micro-hydro schemes and suggest a different approach.
A coordination study of an induction generator in an interconnected grid
system is necessary to determine systems long term dynamic stability. The short term
transient stability aspect is kept out of the picture assuming the system is fairly
strong. Two different methods are widely used for performing such studieseither
computer simulation of the complete system or a real time simulation. A typical
study involves investigating many different systems. Each of these systems in turn
has its own sub-system. For example a thermal power generating system has a boiler,
turbine, governor, turbogenerator, voltage regulator, etc.
75


If in lieu of real time simulation study a computer simulation method is
selected, then each system is represented by its mathematical model. However, in this
approach simplifications and approximations are made in order to keep the size of the
simulation realizable on a computer. Furthermore, one has to make several trial runs,
study several computational results, and compare numerous graphs before being able
to get a feel for the characteristics of the complete generating system.
In medium range, use of induction generators in mini-hydro schemes is
contingent on economic aspects of meeting their excitation requirements. The cost of
providing the excitation capacitors offsets their economic viability as compared to
synchronous generators with larger ratings. However, in situations where induction
generators are used in a power system network in conjunction with other generators
which can provide the required excitation power, use of induction generators could be
made economically viable even in medium output range. Most induction generators
in mini-hydro application today operate in parallel with the interconnected network.
DeMello has reported [15] that induction generators are technically feasible in
interconnected power systems and where the network is strong enough from the
transient stability point of view. The var control could be implemented at the source
of excitation where it might be at some distance away from the induction generator.
76


3. Protection of Induction Generators
Protection of induction machines has been extensively covered in numerous
IEEE publications and in protective relay manufacturers literature [1]. These
protection schemes may vary from bare essentials (those required by the National
Electrical Codes) to very elaborate and expensive schemes (often implemented or
required by the larger utility companies). The choice of protection for each induction
generator should be based on sound engineering judgments and the specific
circumstances that surround each induction generator installation.
One determining factor is the size of the induction generator and its
replacement cost. The amount of protection provided for a 1000 KW generator
should not be the same as that provided for a 50 KW generator. A larger and more
expensive generator can justify a more elaborate and expensive protective scheme
than that can be afforded for a small generator.
Other important factors that should be evaluated in considering protection
include: the location of the induction generator and ease of access to replacement
parts and qualified technicians, the presence (or lack of) an operator on duty who can
respond to a problem quickly, the impact of prolonged unavailability of a generator in
lost revenue and the dependency of a remote community on its operation as its main
source of electric power.
In the forthcoming paragraphs different protective devices that make up a
typical induction generator protective scheme are discussed. Each protective devices
intended applications, benefits and limitations are underscored. Afterward, these
individual devices are consolidated together to form a typical induction generator
protective scheme.
77


Please bear in mind that the protective devices discussed in this section will be
augmented by additional operational and/or protective requirements imposed by the
utility on induction generators interconnected to the power grid. The additional
requirements and their basis will be covered in the following section under Utility
Interface Protection of Induction Generators.
3.1 Circuit Breaker (ANSI device no. 52)
Similar to all other motors and generators, there is a need for a switching
mechanism to connect and disconnect an induction generator from the rest of the
electrical circuit under normal and fault conditions. This can be accomplished in one
of two ways: either by a set of motor starter contactors (ANSI device no. 42) in series
with fuses or by a power circuit breaker.
The combination of motor starter contactors and fuses costs less than the
power circuit breaker. However, this option requires more maintenance and once a
fuse is blown due to a fault it must be replaced before generator operation can be
resumed. Furthermore, the contactors have been known to open unexpectedly
(because of their control relay dropping out) when their control voltage reaches to
about 70 percent of its normal value during voltage transients.
Power circuit breakers cost more than the above option; however, they
provide the advantage of being able to interrupt the fault current. They offer greater
flexibility and control functions versus a set of motor starter and fuses and they
should be utilized on medium or larger generator applications.
78


3.2 Instantaneous, Overcurrent and Differential Protection Devices (ANSI
device nos. 50 and 87)
It is essential to provide instant protection for an induction generator against
excessive current, or excessive rate of rise of current normally present in a fault
within the generator or its protected circuit.
Fuses, if they are used, provide this type of protection very well inherently.
However, if power circuit breakers are used, then they must be supplemented with
instantaneous overcurrent (ANSI device no. 50) or differential overcurrent relays
(ANSI device no. 87). Instantaneous overcurrent relays must be set high enough to
prevent the generator breaker from tripping because of the inrush currents during
startup. Instantaneous overcurrent relays are not as sensitive as the differential relays
and they provide only the minimum amount of protection against short circuits.
These relays are adequate for smaller induction generators where the generator neutral
leads are not externally available and a differential relay scheme can not be
implemented.
Differential type relays provide a more accurate and positive way of
monitoring fault currents within the protected zone. Their application is strongly
recommended for medium and larger size induction generators. The percentage
differential setting of these relays should be set to detect a fault from the phase leads
to neutral with a coverage of at least 80 percent of the entire length of stator winding.
Additional cost is the main drawback in applying differential relays versus
instantaneous relays. Additional costs appear in the form of the more expensive
relays, the cost of additional wiring required for differential circuit wiring, the cost
for doubling the number of current transformers, and the additional manufacturing
costs in providing induction generator neutral leads for the differential wiring circuit.
79


Most induction machine manufacturers charge extra for bringing out and
terminating the neutral leads. Instead, the neutral leads of most induction machines
are tied together internally. This is because the fault contribution period of induction
generators is relatively short and there is no need for grounding the neutrals of these
generators. Grounding of the system, if required, can take place at the neutral of the
step up transformer associated with the generator.
A more cost effective alternative is also available where the instantaneous
relays are applied in a self-balancing differential method as shown in Figure 3.1. In
this method the neutral lead as well as the phase lead passes through the same current
transformer. In the absence of an internal winding fault the currents in the two leads
cancel each other out and the instantaneous relay detects a net zero current passing
through the current transformer. The zone of protection for this scheme is less than
the percentage relay. It can also be implemented for lower cost and it is sufficiently
sensitive enough for use with medium size machines.
Figure 3.1 Induction generator connections with instantaneous and differential
overcurrent protections.
80


3.3 Thermal and Overload Protection Devices (ANSI device nos. 49 and 51)
Thermal and overload protections ensure that the generator will not operate in
excess of its thermal ratings for a prolonged period of time, thus preventing the
generator from insulation damage and loss of life.
The thermal protection can be accomplished in a number of methods. In one
method, a series combination of thermally actuated switches and generator starter
contactors is provided in the same enclosure. The load current passes through the
thermal switches as well as the generator starter contactors. The time versus current
characteristics of the bimetalic thermal switches is designed and set to prevent the
generator winding from getting too hot before the bimetalic switches open and take
the generator out of service. This method does not actually monitor the temperature
of the generator winding; instead, it estimates the winding temperature by monitoring
the load current magnitude and its duration.
In other situations where power circuit breakers are present instead of
generator starter contactors, an inverse time overcurrent set of relays can be used to
trip the breakers. The inverse time overcurrent relays (ANSI device no. 51) and
instantaneous time overcurrent relays (ANSI device no. 50) can be purchased as one
unit from most relay manufacturers. Settings of both of these relays have to be
coordinated against starting transient conditions of the generator and other
downstream relays. Time overcurrent relays do not monitor the actual winding
temperature of the generator either. Instead, they make an estimation of winding
temperature by monitoring the load current. Time overcurrent relays are most
effective when there is a generator overload condition or a persistent short-circuit
with limited fault current magnitude present. These relays would not work effectively
in case of a solid short-circuit because the decay rate of the generator fault current
magnitude is too fast for the time overcurrent relay to detect it.
81


Still, the most effective means of monitoring and protecting the generator
winding against high temperature is by use of resistive temperature detectors (RTDs)
or thermocouples. These devices are embedded in the generator stator winding slots
and provide direct and accurate readings of winding temperatures from several points
to the meters, relays, and/or microprocessor units with which they are interfaced. The
relays and the microprocessor units can be set to trip the generator either based on the
hottest temperature reading or based on an average reading from the RTDs or
thermocouples. It is a fairly standard practice for most generator manufacturers to
provide eight RTDs (two per phase plus two more for shaft bearings) for medium and
larger size machines.
3.4 Ground Fault Protection Device (ANSI device no. 64,50G, 51G)
Type and location of the ground fault protection devices depend on generating
station configuration and the type of equipment used. As was mentioned earlier, fault
current magnitude of an induction generator decays very rapidly and therefore the
generator is not considered to support or make a significant contribution to a ground
fault. Consequently, neutral leads of most induction generators are not grounded.
Nevertheless, remaining electrical equipment should be grounded in order to
stabilize and prevent overvoltages on the non-faulted phases of the system during a
fault. A grounded Y on the low voltage side of the step-up transformer or a dedicated
grounding transformer could serve this function very well.
The grounding method (i.e. high impedance, low impedance, or solid) requires
some studying and evaluation of the generating plant, the interconnecting utility
systems requirements, and the protective relaying scheme employed. In most small
and single unit generating stations a high-resistance grounding method is used to limit
the fault current to less than five amperes and minimize consequent damage to
82


generator and related equipment. High resistance grounding takes place by first
grounding the system through a dedicated grounding transformer that is connected
with grounded Y on the primary side and broken delta on the secondary side. Then a
suitable grounding resistor is connected across the broken delta on the secondary side
of the transformer. A ground protective relay is usually connected in parallel with the
grounding resistor across the broken delta terminals. In case of a ground fault, a
current passes through the grounding transformers primary side which will induce a
voltage across its broken delta secondary side terminals. This voltage causes a
current to flow through the grounding resistor that will be sensed by the ground relay
as a ground fault.
3.5 Rotor Overheating Protection (ANSI device nos. 46 and 47)
In a squirrel-cage induction generator the rotor is more robust and less
susceptible to damage from overheating than the stator winding. Nevertheless, in
earlier sections it was shown that only a few percentage of voltage or load unbalance
can cause a large magnitude of negative sequence current to flow in the rotor. This
negative sequence current produces a negative sequence (or braking) torque in the
rotor which is responsible for rotor overheating.
Rotor temperature cannot be readily measured. Instead, where economically
feasible, protective relays are installed that would measure the amount of voltage
unbalance, current unbalance, and/or negative sequence current flowing in the system.
Unbalanced voltage conditions could especially present a problem for
induction generators connected to a utility grid. Despite the utilitys best efforts to
keep the phase voltages balanced, there will always be some amount of voltage
unbalance due to daily load fluctuations. As discussed earlier, The National Electrical
83


Codes does not recommend operating an induction machine with more than five
percent of voltage unbalance on the system.
3.6 Bearing Protective Device (ANSI device no. 38)
Overheating of the generator bearings can be caused either due to lack of
lubricant or excessive mechanical wear. Bearing temperature should be monitored
closelyespecially in unattended generating stations. Then, if the temperature
reaches above a certain setpoint, the generator must be tripped. Protection against
high bearing temperature can be performed by either indicating thermometers that are
equipped with contacts to trip the generator, or RTDs that are interfaced with
auxiliary trip relays. With either method, the important thing is to place the sensors
right at the bearing surface and not at some remote location like the lube oil reservoir.
3.7 Vibration Detection (ANSI device no. 39)
High vibration caused by a mechanical imbalance can escalate and cause
serious damage to the generator in a short time. The best locations to monitor for
vibration are on the generator bearings. Here, vibration detectors can monitor any
excessive sign of shaking and quickly shut down the generator. Installation of a
vibration monitor is highly recommendedespecially in unattended generating
stations.
3.8 Mechanical Overspeed Protection (ANSI device no. 12)
Induction generators frequently experience overspeed conditions as a
consequence of load rejection or disconnection from the utility (especially when a
large capacitive bank is connected across their terminals). The monitoring of
overspeed condition is crucial in preventing generator damage.
84


The most common method in overspeed protection is to use a centrifugal
speed switch on generator shaft. At a certain speed, the contact from the switch
changes state under the centrifugal force of rotor rotation. This change in contact
state is transferred from the rotating shaft to the stationary stator through a
magnetically coupled slip ring installed on the generator shaft. The contact in turn
initiates the shutdown of the generator through a set of control relays. The generator
shutdown procedure may include application of brakes to the rotor an/or rapid
reduction of mechanical power delivered from the prime mover.
In some cases a tachometer with a contact having an adjustable setpoint for
speed may be used to monitor generator speed. Alternatively, such a contact can also
serve to shut down the generator. The tachometer senses rotor speed through a
rotating encoder mounted on the generator shaft and a stationary-electro-optic-speed
sensor mounted on the stator in close proximity to the encoder.
3.9 Reverse Power Protection (ANSI device no. 32)
Under special conditions it may be prudent to install a set of reverse power
relays in the generating facility. These relays monitor the direction of real power
flow which indicates whether the induction machine is operating as a generator or as a
motor. Operating an induction generator as a motor will not cause any damage to the
machine. However, the owner may want to minimize the costs associated with
operating the induction generator as a motor from the interconnected utility.
Another concern pertains to the type of prime mover used to drive the
generator. Wind turbines and Pelton wheels are not susceptible to damage by
motoring operation of the machine. However, in rare occasions where an induction
machine and a steam turbine are coupled, heavy damage could be caused to the
turbine blades as the generator starts motoring.
85


As a final thought, another reason for the use of reverse power relays may be
due to a requirement imposed by the interconnecting utility company upon the
induction generating facility. As highlighted above, the need for reverse power
protection is not critical nor necessary in most induction generator applications,
except as noted.
3.10 Overvoltage and Overexcitation Protection
Installation of surge arrestors and surge capacitors in a generating facility of
any size is important. These devices protect electrical equipment, the generator in
this case, against the effects of overvoltages resulting from lightning, switching
surges, and other disturbances. Without such protections, flashovers and insulation
damage to equipment may occur. Surge arrestors limit overvoltages from appearing
across equipment insulation by conducting the surge currents directly into the ground
while surge capacitors are used with rotating machinery to decrease the steepness of
the surge-voltage wave front. To obtain the maximum benefit from these devices,
both surge arrestors and surge capacitors must be installed in close proximity of the
generator terminals they are intended to protect.
Most induction generators are connected to such a large sum of capacitive
load bank for power factor correction that if the generator loses its load, the resulting
overspeed will cause a serious overexcitation voltage problem across the generator
terminals. To minimize overexcitation voltage problems, design provisions should be
implemented to disconnect the capacitor bank from service immediately upon loss of
generator load.
86


3.11 Typical Overall Protection Schemes
At the beginning, it may not be evident how much protection is adequate for a
particular generating plant. The degree of protection required will require careful
study of the generating plant and its unique design features as well as a review of the
interconnecting utilitys power system. Nevertheless, some general examples of
protection schemes for induction machines of different sizes are presented here to
serve as the starting point for the design engineer.
Based on earlier discussions of induction generator protection, a large-size
single-unit induction generator plant might be arranged as shown in Figure 3.2. The
principle protection for the generator is achieved by using a solid-state multifunction
motor protection relay (ANSI device no. 99). The various relay functions included in
this protection package are listed to the side of the one-line diagram shown in Figure
3.2. The generator system bus in this example is grounded through a fully rated high-
resistance grounding system. As soon as a ground fault is detected by the relay, the
generator breaker will be tripped. Detection of a ground fault on the utility line is
provided at the neutral of the main step-up transformer. The standard over/under
frequency and voltage relays are also shown. In this example the main step-up
transformer is protected by a fused disconnect switch.
Additional examples of induction generators arrangement, metering and
protection requirements from the perspective of the utility companies are shown in
Figures 3.3 through 3.6. These Figures are extracted from Public Service Company
of Colorados publication [16] titled Safety, Interface and Interconnection
Guidelines for Co-Generators, Small Power Producers and Customer Owned
Generators. The utility companies require only that protective equipment that would
protect their interest and the interest of their customers. Therefore, protective
87


equipment specific to the protection of the induction generators (e.g., ANSI device
nos. 12, 38, 39,46,47,49) are not shown in Figures 3.3 through 3.6.
Figure 3.2 Typical protection scheme for a large-size, induction generator.
88


SA

UTILITY
CUSTOMER
r
ft
r
O|l'
1 PHASE


LOAD CENTER
'~n

/""N
.MOLDED .
[CASE BREAKER (OR FUSE)__|
LOCKABLE
UTILITY ACCESSIBLE
DISCONNECT SWITCH
INDUSTRIAL GRADE RELAYS
27 UNDERVOLTAGE TRIP V £ 80%; TIME £ 0.5 SEC.
59 OVERVOLTAGE TRIP V 2 115%; TIME 0.1 SEC.
81-0 OVERFREQUENCY TRIP F 2 63Hz; TIME £ 0.5 SEC.
81 -U UNDERFREQUENCY TRIP F £ 57Hz; TIME £ 0.5 SEC
-H- REQUIRED FOR EXPOSED GENERATORS
1 1 SUCH AS WIND.
SA SURGE ARRESTER
WH WATT HOUR METER
gg CONTACTOR
NOTE: RELAYS DO NOT HAVE TO BE INDIVIDUAL.
FUNCTIONS MAY BE IN CORPORATED IN THE
INTERFACE PROTECTION PACKAGE. OR AS
PART OF AN INVERTER.
Figure 3.3 Typical protection and metering scheme for an induction generator less
than 10 KW.
89


J
l>
GROUNDING
BANK
a
X
QUANTITIES FOR 1PH GENERATORS
QUANTITIES FOR JPH GENERATORS
UNOERVOLTAGE TRIP V .S BOX; TIME .S 05 SEC
REVERSE POWER
PHASE-SEQUENCE OR PHASE-BALANCE VOLTAGE
TIME OVERCURRENT, t/PHASE
RESIDUAL (GROUND) TINE OVERCURRENT
OVERVOLTAGE TRIP V 2 11SX; TIME i 0.1 SEC
UNDERFREQUENCT TRIP F £ 57Hl; TIME i. 0.5 SEC
OVERFREQUENCT TR* F 2 BJHi; TIME i 0.5 SEC
SURGE ARRESTER
WATT HOUR METER
SUGGESTED
POWER FACTOR CORRECTION FOR INDUCTION
GENERATORS. P.F. 2 0.95
REQUIREMENTS DEPEND ON GENERATOR
CROUNOMG ANO STEP UP TRANSFORMER
REOCTREE FOR EXPOSED GENERATORS
SUCH AS WMO GENERATORS
CONTACTOR
1PH GENERATOR MAXIMUM SIZE 20KW
Figure 3.4 Typical protection and metering scheme for an induction generator
lOKWto less than 100 KW.
90


Full Text

PAGE 1

ELEMENTS OF INDUCTION MACHINE APPLICATION IN HYDROELECTRIC POWER GENERATION by Hamid Zanjani B.S., University of Colorado, 1984 A thesis submitted to the University of Colorado at Denver in partial fulfillment of the requirements for the degree of Master of Science Electrical Engineering 1998

PAGE 2

This thesis for the Master of Science degree by Hamid Zanjani has been approved by b Date

PAGE 3

Zanjani, Hamid (M.S., Electrical Engineering) Elements of Induction Machine Application In Hydroelectric Power Generation Thesis directed by Professor Pankaj K. Sen ABSTRACT This study identifies and investigates major aspects of utilizing induction machines in hydroelectric power generation as a viable alternative to the use of conventional synchronous machines. The study begins by first introducing the elementary operating principles and features of induction machines in clear and easy to understand terms. Once a solid technical foundation is built, the author introduces and explores major technical engineering topics that are of concern in a successful assessment and application of induction machines in micro( <1 00 KW) and mini(> 100 KW and <5000 KW) size power plants driven by hydroelectric turbines. Some of the major topics in this study include: voltage and -....-.-, frequency regulation of induction machine, stand-alone versus grid connected machines, comparison of various capacitive var compensation techniques, electronic load governors versus turbine governors, monitoring and protection of induction machines, and utility requirements for induction machines connected to the power grid. This abstract accurately represents the content of the candidate's thesis. I recommend its publication. lll

PAGE 4

ACKNOWLEDGMENTS I am finally getting out of the strain of doing and into the peace of being done. It seemed like a long journey from the time I took my first graduate class to this point where I can finally see the light at the end of the tunnel with my endeavor. I must confess that my journey had its share of passing through construction zones, detours, and stopping at scenic sites along the way. Nevertheless, this journey has been an exciting, educational, and fruitful experience for me. The valuable lessons I learned as a student were that: 1) The purpose of an education is not to learn, but to learn how to learn and 2) learning should be viewed as a life long goal extending beyond college years. I am grateful to all my professors at the University of Colorado at Denver for making this education possible for me. I am especially indebted to Dr. Pankaj Sen and Dr. Bill Roemish for their efforts, inspiration, patience and for the trust that they invested in me throughout this educational program. IV

PAGE 5

CONTENTS Chapter 1. Introduction ................................................................................................................ I 1.1 Operating Principles of Squirrel-Cage Induction Machines .......... ......................... 5 1.2 Design and Performance oflnduction Machine .................................................... 10 1.2.1 Design ............................................. ........................................................... 1 0 1.2.2 Performance ........................................................................................................ 11 1.3 Equivalent Circuit oflnduction Machine .............................................................. 16 1.3 .1 Development of Equivalent Circuit .......................... :. : ...................................... 17 1.3.2 Analysis of the Equivalent Circuit.. .................................................................... 23 2. Application oflnduction Generators in Microand Mini-Hydro Stations .............. 30 2.1 Introduction ............................................................................................................ 30 2.2 The Choice ofTurbine and Governor for a Microand Mini-Hydro Station ........ 32 2.2.1 Turbine .......................................................................... ..................................... 32 2.2.2 Standard Governor Versus Load Governor ........................................................ 38 2.3 Induction Generators in Micro-Hydro Stations as Stand-Alone Units .................. 51 2.3.1 De-Rating oflnduction Machine for Unbalanced Operation .............................. 51 2.3 .2 Optimizing Capacitor Size for a Stand-Alone Induction Generator ................... 61 2.4 Induction Generators in Mini-Hydro Stations ....................................................... 75 v

PAGE 6

3. Protection of Induction Generators ......................................................................... 77 3.1 Circuit Breaker (ANSI device no. 52) ................................................................... 78 3 .2 Instantaneous, Overcurrent and Differential Protection Devices (ANSI device nos. 50 and 87) .................................................................................................................... 79 3.3 Thermal and Overload Protection Devices (ANSI device nos. 49 and 51) ........... 81 3.4 Ground Fault Protection Device (ANSI device no. 64, 50G, 51 G) ....................... 82 3.5 Rotor Overheating Protection (ANSI device nos. 46 and 47) ................................ 83 3.6 Bearing Protective Device (ANSI device no. 38) ................................................. 84 3. 7 Vibration Detection (ANSI device no. 3 9) ....................................... : .................... 84 3.8 Mechanical Overspeed Protection (ANSI device no. 12) ...................................... 84 3.9 Reverse Power Protection (ANSI device no. 32) ................................................... 85 3.10 Overvoltage and Overexcitation Protection ..................................... .' ................... 86 3.11 Typical Overall Protection Scheme ..................................................................... 87 3.12 Utility Interface Protection Requirements ........................................................... 93 3 .12.1 Lockable Disconnect Switch ............................................................................. 95 3.12.2 Detection and Clearing a Fault on the Utility System ............. ........................ 95 3 .12.3 Islanding ............................................................................................................ 98 3 .12.4 Energizing a Dead Circuit. .............................................................................. 1 00 3.12.5 No Manual Synchronization ........................................................................... 100 3 .12.6 Modification Costs to the Utility .................................................................... 1 02 Conclusions ................................................................................................................ 1 03 Vl

PAGE 7

Appendix A. Constants C and D for equations (2.21) and (2.22) defined ................................. .1 05 References ........................................................................................... ....................... 1 06 Vll

PAGE 8

FIGURES Figure Figure 1.1 Relative motion of stator and rotor magnetic fields .................................. 7 Figure 1.2 Induction machine torque-slip curve showing braking, motor, and generator regions .................................................................................................. 12 Figure 1.3-Typical short circuit waveform of a synchronous generator ................... .l5 Figure 1.4 Typical short circuit waveform of an induction generator ...................... .15 Figure 1.5 Stator equivalent circuit of an induction motor with rotor circuit opencircuited ................................................................................................................ 1 7 Figure 1.6-Rotor equivalent circuit at stator frequency ............................................. 20 Figure I. 7 Equivalent circuit of an induction machine (motor or generator) ............ 22 Figure 1.8-Direction of power flow in a motor and a generator operation ................ 22 Figure 1. 9 Typical phasor diagrams for an induction motor and a generator ............ 22 Figure 1.10 Thevenin equivalent circuit of an induction machine ............................ 26 Figure 2.1 -Circuit configuration of the impedance controller with rectifier and chopper. ............................................... ..................... ........................... .... ... ...... 43 Figure 2.2-Rectifier line current for phase A with a=30 without the chopper. ...... .46 Figure 2.3 -Impedance controller line current for phase A with a=30, five-pulse chopping and PW=0.5 ..................................... ........ ..... ...... .... .............. .............. 46 Figure 2.4 -Rectifier line current for phase A with a=90 without the chopper ....... .4 7 Vlll

PAGE 9

Figure 2 .5-Impedance controller line current for phase A with a=90, five-pulse chopping and PW=0.5 .......................................................................................... 47 Figure 2.6-Area of current control offered by the impedance controller for Rdc=1.0 pu at rated voltage ................................................................................... ............ 49 Figure 2.7 Sequence network circuits of an induction machine ................................ 53 Figure 2.8 -Direction of power flow in a self excited induction generator. ................ 55 Figure 2.9-Derating oflnduction machine due to unbalanced voltages .................... 60 Figure 2.10 Derating of induction machine due to unbalanced voltages ...... ......... ... 60 Figure 2.11 Block diagram of a short shunt self-excited induction generator ........... 63 Figure 2.12 Block diagram of a long shunt self-excited induction generator ........... 63 Figure 2.13 -Equivalent circuit of a short shunt self-excited induction generator. ..... 64 Figure 2.14 Variation of Xm versus ElF ................................................................... 65 Figure 2.15 Variation of terminal voltage with shunt capacitor C5h .......................... 68 Figure 2.16 Effect of Cse variation on load voltage regulation ..................... .... ....... 69 Figure 2.17-Variation ofload voltage V L with load ................................................... 70 Figure 2.18 Variation of air-gap voltage E with load ........................................... .... 70 Figure 2.19 Variation of stator current 15 with load ................................................... 71 Figure 2.20Overall characteristics of a short shunt SEIG ....................... : ................ 72 Figure 2.21 Effects of changing shunt capacitor Csh on load voltage ..... ...... ........... 73 Figure 3.1 -Induction generator connections with instantaneous and differential overcurrent protections ................. ...................................................................... 80 lX

PAGE 10

Figure 3.2-Typical protection scheme for a large-size, induction generator ............ 88 Figure 3.3 Typical protection and metering scheme for an induction generator less than 10 KW ................................................................................... ... ......... . .... 89 Figure 3.4-Typical protection and metering scheme for an induction generator10 KW to less than 100 KW ..................................................................................... 90 Figure 3.5 Typical protection and metering scheme for an induction generator 100 KWto 1 MW ....................................................................................................... 91 Figure 3.6 Typical protection and metering scheme for an induction generator 1 MW to less than 10 MW ................................................................. ............................ 92 Figure 3. 7 High voltage ground detection using "comer of delta" voltage relay ..... 98 X

PAGE 11

1. Introduction Rising capital cost of building large power plants, lack of interest from large financial institutions, increased scrutiny by the environmental agencies, changes in federal and legislative laws, in addition to recent deregulation of power utilities have encouraged design and development of smaller power plants. The design objectives of these smaller power plants are simplicity, reliability, and especially cost effectiveness. In meeting these objectives, the lower initial cost, low maintenance, simplicity and ruggedness associated with induction generators (in comparison with synchronous generators) have advanced the widespread application of induction units in smaller power plants. Most induction generators in these applications are driven by either hydraulic or wind turbines in a stand-alone or grid-connected configuration. The most common size of turbine/induction generator set found in commercial use ranges from 50 KW to an upper limit of 1500-2000 KW. Within this range, power plants built with induction generators by far outnumber similar plants built with synchronous generators [ 1]. An induction generator is simpler and cheaper to build since its rotor cage is a symmetrical structure built using an automated fabrication process. Since the rotor winding does not need an excitation system, control equipment requirements for the generator are greatly reduced Furthermore, an induction generator does not need a separate power source for its rotor winding. The induction generator, like a synchronous generator, requires excitation in order to produce voltage and become a source of electrical power. While a synchronous generator derives its excitation from a separate source (e.g., an on-shaft

PAGE 12

rotating excitation system), the induction generator must draw its magnetizing current from the intertied utility system or from capacitors located near the unit. Without excitation from some source, the induction generator cannot sustain a terminal voltage with any load connected to it. In order to familiarize the reader with overall characteristics of induction generators and synchronous generators, the following table makes a condensed comparison of the two different types of generators. Some of these characteristics will be discussed in more detail in the forthcoming sections. 2

PAGE 13

INDUCTION GENERATORS SYNCHRONOUS GENERATORS Rugged and simple construction Sensitive and sophisticated construction Negligible harmonics Harmonics are inherent Excitation var must come from the grid or Needs a separate de source for capacitor banks excitation Does not contribute sustained fault-current Contributes sustained fault-current under short-circuit conditions under short-circuit conditions Capital cost is less because system requires Capital cost is more because of higher less design and construction efforts and design and construction efforts and control devices and machine costs are lower complicated control devices Operating cost is less because these plants Operating cost is high because these can be unattended and maintenance cost is plants need operators and maintenance low rate is high Under load rejection conditions, selfOverspeed and overvoltage conditions excitation condition may result in are under control as a result of overspeed and over voltage governor and automatic voltage regulator actions Control equipment and governor are Control equipment and governor are simple. No A VR t and field breakers are sophisticated A VR and field breakers required are required Starting is comparatively simple Synchronization must be precise Synchronization usually not required No problem in system stability. The system In case of system disturbance, the disturbance will generally not affect machine may lose synchronism from machine connected to the grid the system Table 1 -Comparison of mductwn versus synchronous generators There are two types of induction generators available in the market. One is the wound-rotor induction generator and the other is the squirrel-cage-rotor induction generator The stators of these two types of generators are very similar in design and construction. The major difference between the two types of generators is in the way their rotors are constructed. t AVR is an acronym for Automatic Voltage Regulator. 3

PAGE 14

A wound rotor in an induction generator is similar in shape to the rotor in a synchronous generator. The leads of the wound rotor are brought out through slip rings mounted on the rotor shaft for adding external impedance or changing the number of poles and hence the speed of induction generator. These features enhance the performance and flexibility of a wound rotor induction generator in certain applications where speed torque, power and efficiency controls are important. The disadvantages of a wound rotor induction generator are in its higher initial capital cost (which compares to that of a synchronous generator) and the extra maintenance required for the slip rings. So, in meeting the power plant's feasibility and cost effectiveness objectives mentioned earlier, a wound rotor induction generator would not be a good candidate. A squirrel-cage rotor in an induction generator consists of parallel bars that are located on the periphery of the rotor shaft and form a cylinder that resembles a squirrel cage. There is a ring that goes over the top and another ring that goes over the bottom of these parallel bars which keep the bars in place and also serve to provide a path for the current from the bars to flow through. The construction of this type of rotor is relatively simple and cheap. Furthermore, this type of rotor is rugged and has low maintenance compared to a wound rotor. The lower cost of this type of induction generator makes it attractive and economically viable in small power plant applications when compared to a synchronous or a wound rotor induction generator. For this reason, the scope of this thesis is focused on the squirrel cage induction generator and very little time is dedicated to exploring the wound rotor induction generator. 4

PAGE 15

1.1 Operating Principles of Squirrel-Cage Induction Machines An induction machine can either be a motor or a generator. For clarity and ease of understanding, the discussion here will first concentrate on an induction motor and then the concepts developed will be expanded to cover an induction generator. At first, it seems puzzling how a squirrel-cage induction motor is capable of producing rotation and torque. In comparison with other types of motor that we are familiar with (such as de and synchronous motors) it seems as though there is something missing from the rotor circuit before it can start to tum--either a permanent magnet that can interact with the stator's magnetic circuit, or an electrical winding connection in the rotor that can serve as a magnetic circuit. A magnetic circuit produces a magnetomotive-force (mmf or f) based on the number of turns in the circuit and the current flowing through it (f= NI). According to magnetic field theory, the presence of two such mmfs acting upon one another and separated by an angle between them produces rotational torque. This torque is zero when the rotor mmf and the stator mmf are in phase with each other and maximum when these mmfs are in quadrature. But, since there is not a permanent magnet nor an electrical winding circuit that we are aware of in the rotor, then how is torque produced in a squirrel-cage induction motor? The answer lies in "Induction". An electromotive-force voltage (emf or speed voltage+) is induced in the rotor bars when voltage is applied to the + + There are two ways that voltage can be induced. One is through transformation which is what takes place in the windings of a transformer and another method is through emf or speed voltage which takes place in rotating machinery. Based on Faraday's law a voltage e is induced if a conductor oflength I moves with linear velocity v in a non-time varying magnetic field of flux density B by "cutting-of-flux" equation: e=Blv according to the right-hand rule. The same hold true ifthe conductor is held stationary and the magnetic field moves past the conductor. The 5

PAGE 16

stator circuit. Many books and technical papers [2, 3] describe this action to be similar to what takes place in a short-circuited secondary winding of a transformer when voltage is impressed on its primary winding; except, in an induction motor the magnetic core is not continuous and has a small air-gap that separates the rotor from the stator. Furthermore, because the rotor is moving with respect to the magnetic field frequency of the stator, the voltage induced in the rotor is of a different frequency than that in the stator. Based on these phenomena, the induction machine may also be compared to a transformer and a frequency changer combined. The concept of slip (S) is important in the operation and analysis of an induction motor. Slip is defined as the difference between the synchronous speed (N5 ) and rotor speed CNr) expressed as a fraction of the synchronous speed. Slip (S) = N s -N r Ns Slip multiplied by stator's synchronous frequency (F) gives the frequency induced in the rotor's electrical or magnetic circuit, and (1-S) multiplied by synchronous speed (N5 ) gives the rotor's rotational speed CNr) So as an example, if it is given that in a four-pole motor the synchronous frequency (F) is 60 Hertz, slip (S) is 0.03, and corresponding synchronous speed (N5 ) is 1800 rpm, then the electrical or magnetic frequency of the rotor circuit is 0.03x60 = 1.8 Hertz. Furthermore, the mechanical speed of the rotor is: (1-0.03)x1800 = 0.97x1800 = 1746 rpm 1746 rpm= 1746 rpm = 29.1 revolutions or cycles per second mechanical 60sec/ min relative m_9tion of the conductor and the magnetic field causes the flux lines to be cut and hence the voltage in the conductor is induced. 6 (1.1)

PAGE 17

In a four-pole motor, one complete mechanical cycle corresponds to two complete electrical cycles. Therefore: 29.1 mech. cycles per second= 29.lx2 = 58.2 elec. cycles per second or Hertz As can be noted from the above example, the sum of the rotor's magnetic field frequency and the rotor's speed (when expressed in terms of electrical frequency) always add up to equal the synchronous frequency ofthe stator (58.2+1.8=60). With a little bit of imagination (refer to Figure 1.1 ), it can be seen that the rotor's magnetic field frequency is riding over the rotor's mechanical speed, both traveling in the same direction as the stator's magnetic field frequency. Magnetic fields of the rotor and the stator have the same frequency and are stationary in space with respect to each other at any given motor speed. (This is analogous to stating that if a train is traveling at 58.2 mph and a person standing on top of the train is walking toward the front of the train at 1.8 mph, then a car which is traveling alongside the train at 60 mph would appear stationary in space to the person on top of the train.) With two magnetic fields stationary with respect to each other, the torque produced will be constant and no torque pulsation will exist in an induction motor. This constant torque is what turns the rotor at a steady speed. -------Rolor speed expressed in Herlz plus -----Rolor magnetic field frequency in Herlz Equals --------Slalor magnetic field frequency in Herlz Figure 1.1 Relative motion of stator and rotor magnetic fields. 7

PAGE 18

For a moment, let's reiterate how an induction motor functions in a slightly different way for further clarity. An induction motor develops useful torque since its rotor speed is slightly less than the synchronous speed of the magnetic field present in the stator winding. That relative motion means that the rotor bars are "cutting" the stator's magnetic flux lines The emf voltage induced in the rotor bars will cause bar currents to flow since rotor bars are short-circuited together at their ends. In tum, this current flowing in the rotor produces its own magnetic field. Finally the interaction of magnetic field with stator's magnetic field produces torque. Contrary to common belief, the design and operation of an induction generator is no more complicated than that of an induction motor. In most cases, an induction generator is nothing more than a standard squirrel-cage induction motor driven above its synchronous speed by a prime mover connected to its rotor shaft (note: machines specifically designed to serve as a generator are slightly different as will be explained in the next section). In an induction generator, the prime mover drives the rotor slightly faster than the synchronous frequency in the stator winding. The relative motion is still present, but now the direction is reversed. The rotor is moving faster, not slower than the stator magnetic field. The direction of the resultant magnetic field torque is nearly reversed The rotor field leads the stator field instead of trailing it. The machine, instead of producing mechanical output torque from electrical input power, now produces electrical output power from mechanical input torque supplied by the prime mover. For the induction generator to supply Kilowatts, the rotor must be driven slightly faster than synchronous speed. This differs from a synchronous generator, which when paralleled with the utility, maintains a constant speed fixed by the utility The faster the induction generator is driven above its synchronous speed rating, the 8

PAGE 19

more power it will produce. Typical induction generators operate at speeds in the range of 1 to 2 percent above synchronous speed. Induction motors, by contrast, operate fully loaded at 1 to 2 percent below synchronous speed. Another consideration in an induction generator operation is the excitation source. The excitation current which is used to magnetize the core must still come from an external source since there is no other source of field excitation available. To meet this excitation current requirement, lagging var must be supplied to the generator. In a case where the induction generator is connected to the utility grid, the lagging var is supplied by the utility's overexcited synchronous machines or capacitor banks. Since induction generator's magnetizing current or lagging var come from the utility grid, the grid controls both voltage and frequency of the generator. There is nothing in the simple induction generator's design, control, and operation that can control either voltage or frequency at all. In an induction generator, whose frequency and voltage are fixed by the utility, generator output power is the only thing that can be controlled by the speed of the prime mover. In a case where the induction generator is operating in a stand-alone or "isolated" mode, such lagging var must be furnished by capacitors (properly-sized for the optimal operating condition) connected across generator terminals. Since capacitors furnish a fixed amount of var, as the generator Kilowatt output demand changes so does the generator terminal voltage. In order to minimize such terminal voltage variations, different schemes have been developed and implemented to change the amount of capacitors connected to the generator terminals by means of mechanical and/or electronic switching devices. 9

PAGE 20

As already mentioned, induction generators can deliver only Kilowatts by consuming kilovars from the system. For this reason, induction generators are only rated in terms of Kilowatt output and not in terms of Kilovars. 1.2 Design and Performance of Induction Machine 1.2.1 Design There are no major differences in electrical or mechanical design of an induction motor versus an induction generator. In many cases, a standard induction motor can very satisfactorily serve as an induction generator. Nevertheless, an induction machine built specifically to serve as an induction generator may have a few features specifically tailored for improving its operation and efficiency. Since the generator does not have to accelerate as a motor, it can be designed with an especially low rotor cage resistance which results in higher efficiency at the sacrifice of normal starting torque. Furthermore, deep rotor bars or double-cage rotors which provide good starting torque and high full load efficiency in an induction motor do not provide an advantage in an induction generator and therefore are not needed. Another area where the induction generator may differ from an induction motor is in its voltage rating. In an induction generator, the internal air-gap voltage is always higher than the terminal or bus voltage whereas in an induction motor the reverse is true. Therefore, induction generators must be specified to have a higher voltage rating than the bus voltage. For example for a 480Volt system, the rating of the generator voltage should be specified as 500 Volts and that for a motor should be specified as 460 Volts. 10

PAGE 21

One last design detail that may be incorporated into an induction generator is its higher overspeed rating. In many cases an induction generator may experience severe overspeed conditions resulting from loss of electrical load or separation from the utility system. Such overspeed conditions may reach twice the rated speed of the generator. The presence of capacitor banks at the generator terminals does make further contributions to this overspeed condition. Some generator manufacturers build a stiffer rotor and support structure to handle the mechanical forces involved under overspeed conditions. Others provide centrifugal switches on the rotor shaft to trip the generator at a preset overspeed condition. 1.2.2 Performance A typical torque-speed or torque-slip characteristic for an induction machine which is connected to the utility line with constant frequency and constant voltage is shown in Figure 1.2. The normal operating region of the machine as a motor or as a generator is between motor breakdown and generator pushover limits at approximately .1 slip. The torque-speed characteristics of an induction motor and an induction generator are almost duplicate of one another; except, one is an upside down mirror image of the other one. 11

PAGE 22

Current_/ Torque %250 0 Braking ----Region ---.... ll "' c: "' <.;> %250 2.0 1.8 1.6 14 1.2 Figure 1.2 Induction machine torque-slip curve showing braking, motor, and generator regions. In an induction motor, as the mechanical load connected to the motor shaft increases, speed drops (slip increases) and torque goes up, until the point of "breakdown torque" on the curve is reached. Further load increase past this point will quickly reduce speed down to a stall while line current rises to its locked rotor value. In an induction generator, as the electrical load connected to the generator terminals suddenly drops, speed increases (slip become more negative) and net accelerating torque increases, until the point of "pushover torque" on the curve is reached. If speed increases past this point, the generator resisting torque suddenly drops away rapidly, and the shaft speed increases until prime mover can no longer keep up, or until the generator self destructs from over-speeding. Once an induction generator passes the pushover speed, it can reach twice its rated speed in less than a few seconds which is far beyond what most induction machines are designed to endure. 12

PAGE 23

What takes place in an induction generator may be best described by an analogy. Consider that someone is gradually pushing a large sphere uphill toward a cliff. At any point before reaching the cliff, if the person lets go of the sphere, it will roll backward to lower ground and stability. However, once the sphere is pushed over the cliff, it continues to free fall until it hits the ground and self destructs. Short-circuit performance of an induction generator is significantly different from that of a synchronous generator. Under fault conditions, the current drawn by the induction generator to provide its excitation is removed or severely reduced. For a three-phase fault, current cannot be transferred past the fault point to the induction generator, and therefore the unit looses its excitation. If capacitors are provided for the unit, a three-phase short would quickly cause these capacitors to be discharged, again resulting in removal of excitation from the generator. Once the induction generator has lost its excitation source, the output current will quickly decay. The speed of the current decay under fault conditions is a function of the machine's ability to store magnetic energy. After all of the stored energy is used, the output will go to zero. The maximum and/or initial current magnitude for a short-circuit on the terminals of an induction generator is determined by the impedance of the device. This current magnitude will be of the same order as that seen when initially energizing the unit (five to six times full-load current). All this differs from the performance of a synchronous generator due in most part to the separate source of excitation. Figures 1.3 and 1.4 provide graphical representations of the performances of an induction and a synchronous generator each connected to a three-phase fault. The phase with the highest magnitude of current is shown and de offset is neglected. From these Figures it can be observed that the decay of short-circuit current for an induction generator is so rapid, compared to that 13

PAGE 24

of a synchronous unit, that it is not possible to use this current as an indicator of system fault conditions. 14

PAGE 25

' I I I I I ,\ I I I I I .. ,. t -: Figure 1.3 Typical short-circuit waveform of a synchronous generator. r r r r r ' r r t Figure 1.4 Typical short-circuit waveform of an induction generator. 15

PAGE 26

1.3 Equivalent Circuit of Induction Machine The steady state equivalent circuit of an induction machine will be developed in this section. The extent of this development will be limited to the point where the reader will feel comfortable understanding the equivalent circuit and be able to make use of it in machine performance calculations. Some useful equations may be presented without actually being derived. The reader will be spared from the rudimentary development of the equivalent circuit which starts off with Faraday's law and rotating magnetic field theory. Discussions on these topics can readily be found in most classic machinery textbooks. For simplicity, only machines with a symmetrical-three-phase winding excited by a symmetrical-three-phase voltage are considered in this section. On a per-phase basis it may be easier to consider a Y connected machine so that winding current is the same as the line current and the phase voltage is the same as the line-to-neutral voltage. The analysis of induction machines is based on the following standard simplifying assumptions: 1. All three phases in the rotor and stator are assumed to be balanced. 2. It is assumed that the coefficient of mutual inductance between any stator winding and any rotor winding is a cosinusoidal function of the electrical angle between the axes of the two windings. 3. It is further assumed that the rotor is smooth and that the self-inductances of all the windings are independent of the rotor position. 4. The effects ofhysteresis and eddy currents are neglected [4] The values of resistances and reactances on the stator and rotor sides ofthe induction machine can be determined from a combination of field tests and calculations. Once these values are assigned to the corresponding impedances of the 16

PAGE 27

equivalent circuit, the stator and rotor currents can then be calculated for any desired speed or load condition. The copper losses associated with the stator and rotor circuits can readily be calculated from the equivalent circuit. When data is available on other types oflosses (i.e., core loss, stray load loss, friction and windage loss), then all other performance characteristics such as speed-torque, speed-current, efficiency, power factor, and so on, can be calculated. It is important to note that stator leakage reactance Xs, rotor leakage reactance Xr, magnetizing reactance Xm, and rotor resistance RP are not constants and will vary somewhat with saturation caused by a change in load current, frequency, and rotor speed. 1.3.1 Development of Equivalent Circuit First assume that the rotor winding of a three-phase induction motor is open circuited and the bus voltage is applied to the stator winding terminals. Then the only magnetomotive-forces acting are those produced by the alternating current flowing in the stator winding. Figure 1.5 shows the equivalent circuit that represents this condition. Rs Xs I 5 L4> 1 VsLo-Xn E I j j Figure 1.5 Stator equivalent circuit of an induction motor with rotor circuit open circuited. 17

PAGE 28

Where: V s = stator terminal voltage per phase E1 = E = counter emf generated by resultant air-gap flux Is = stator current Rs. = stator effective resistance Xs = stator leakage reactance Xm = magnetizing reactance From the above circuit, stator current Is can be calculated from the stator terminal voltage and impedance as follows I = Vs amps 5 Rs + j(Xs + Xm) (1.2) For the condition of an open-circuited rotor winding, it is evident that stator current (I5 ) equals the magnetizing current (Im) Magnetizing current Cim) sustains air-gap flux wave ( which rotates at an angular velocity of m =2 7t F as seen by the stator, and induces the air-gap emf voltage E1 in the winding. From the viewpoint of equivalent circuit, the magnitude of the induced voltage E1 is equal to the difference of stator terminal voltage and the voltage drop across the stator impedance (Rg +jX5). Assume that the rotor winding of the machine is now closed and that bus voltage is applied to the terminals of the stator winding. The rotating flux wave ( in the stator, produces an induced voltage E2 in the rotor winding which in turn (since rotor winding is now closed) causes a current to flow in the rotor. The current flowing in the rotor produces its own magnetomotive force which when it interacts with the stator magnetomotive-force causes the rotor to turn. The actual voltage 18

PAGE 29

induced in the rotor winding E2 is equal to E1 at blocked-rotor conditions. In general, E2 is equal to E1 multiplied by rotor slip (S) as defined in equation (1.3). E2 = SE1 Volts/phase The resulting current flowing in the rotor circuit therefore is I= E2 r Rr + jXr(@slip) amps/phase Rr + jXr(@slip) Where Rr = resistance of rotor winding Xr(@ slip)= leakage reactance of rotor winding at slip frequency (1.3) (1.4) As mentioned before, rotor slip (S) is defined as the difference in speed between synchronous speed and rotor speed, expressed in percent or in per unit of synchronous speed. The actual rotor frequency (Fr) is equal to (1.5) Where:Fr =rotor frequency F5 = stator frequency which is the same as synchronous frequency (F) Accordingly, ifXr is the rotor winding leakage reactance at 60 Hertz, then the rotor winding leakage reactance at any other rotor frequency (or slip) is given by Xr(@ slip)= SXr or X = Xr(@slip) r S ohms/phase (1.6) Equation ( 1. 7) for rotor current may then be written as I = SEI = SEl amps/phase r Rr +jSXr )2 + (SXr )2 (1.7) Dividing the numerator and denominator of the above equation by slip (S) yields 19

PAGE 30

I= Et r R r x s+J r (1.8) The rotor impedance as seen from the stator side at synchronous frequency would then be Z = El = Rr +J'X ohms/phase r I S r r (1.9) In terms of stator emf voltage (E1), and stator load current (same as rotor current Ir), the impedance as seen across stator air-gap is (R/S)+jXr. The corresponding rotor equivalent circuit as seen from the stator at synchronous frequency is shown in Figure 1.6. Resistive quantity (R/S) represents the combined effects of rotor copper loss and the real power consumed by the load connected to the shaft. An equivalent way of showing the same rotor circuit would be to divide resistance (R/S) into two parts. Resistance Rr would then account for the copper loss in the rotor and variable resistance Rr[(l-S)/S] would account for the power consumed by load connected to the shaft which varies with slip. l Figure 1.6-Rotor equivalent circuit at stator frequency. 20

PAGE 31

Combining equivalent circuit of the rotor with equivalent circuit of the stator results in the overall equivalent circuit for the induction motor as shown in Figure 1.7. All rotor quantities are referred to the stator side and the designation E1 is replaced by plain E from here on to refer to the air-gap voltage. The current flowing into the rotor equivalent circuit Clr) is the same as the load current flowing out of the stator equivalent circuit. Except that, when this current is viewed from the stator side, it appears to have the same synchronous frequency present in the stator winding. The voltage across the impedance in the rotor equivalent circuit is the same as the voltage across the magnetizing reactance (Xm) in the stator equivalent circuit. It should be noted again that when rotor currents and voltages are reflected into the stator, their frequency is also changed to stator frequency. All rotor electrical phenomena when observed from the stator side have the stator frequency because what the stator winding sees is just magnetomotive-force and fl11x waves traveling at synchronous speed. In Figure 1.7, shunt resistance Rc (across magnetizing reactance Xm) which represents the core loss effect is omitted for simplicity. In most calculations this small loss is deducted from the machine at the same time when friction, windage, and stray load losses are subtracted from the machine. This equivalent circuit is the same for a motor and a generator. For a motor, slip is positive and for a generator, it is negative. Thereby, a negative resistance in the equivalent circuit indicates a source of generation. In Figures 1.7 and 1.8 directions of currents and power flow for a motor and a generator are indicated and in Figure 1.9 their corresponding phasor diagrams are drawn. 21

PAGE 32

Rs X s Rr Xr ISL> 1 x,., Er r j ' pd zeq b Note: Current flow direction of Is and Ir is shown for a motor. In case of a generator, the direction of Is and Ir is reversed. slip S lakes on a negative value, and variable resistor lakes on a negative value. Figure 1. 7 Equivalent circuit of an induction machine (motor or generator). Q--MOTOR Q--P Mech. --GENERATOR Figure 1 8 Direction of power flow in a motor and a generator operation. I,., Generator Note: Is and I r are shown wilh currents flowing oul of lhe generator I,., Figure 1.9-Typical phasor diagrams for an induction motor and a generator. 22

PAGE 33

I 1.3.2 Analysis of the Equivalent Circuit In equations ( 1.1) through ( 1. 9) important relationships related to currents, voltages, and impedances in the rotor and stator circuits of an induction machine were provided. From the equivalent circuit presented at Figure 1. 7, additional important equations regarding the steady-state performance characteristics of an induction machine can be derived. These equations can be used to calculate speed, losses as load/torque changes, starting torque, maximum torque, etc., in an induction machine. The following equations are presented for an induction motor, consistent with the conventions used in the equivalent circuit of Figure 1. 7. The same equations can be used for an induction generator by using a negative value of S for slip. In turn, a negative value of S would result in a negative resistor value in the equivalent circuit; a negative resistor value results in negative power which is an indication of generator operation instead of motor operation. In an induction motor, copper losses, core losses, friction and windage losses, stray load losses, etc., are subtracted from the input electrical power to arrive at the output mechanical shaft power. Whereas, in an induction generator, all above losses are added to the output electrical power in order to arrive at the input mechanical shaft power. In a three-phase induction motor, total power (P g) transferred from stator circuit to rotor circuit across the air-gap is P = 3(1 )2 Rr Watts g r S ( 1.1 0) A portion of this power transferred across the air-gap is consumed as I2R loss or copper loss in the rotor circuit 2 ., Rotor I R loss= 3(1rfRr Watts (1.11) 23

PAGE 34

The difference of the two equations is called developed or internal mechanical power P d delivered to the rotor shaft P d = P g-rotor I2R loss= 3(Ir )2 Rr -3(Ir )2 Rr Watts s 2 1-S = 3(Ir) Rr-8 =(1-S)Pg Watts (1.12) Output power Pout is the net power available to do work Pout= P d-(friction, windage, core, and stray load losses) (1.13) In an induction motor, a fraction of the air-gap power (SP g) is dissipated as copper loss in the rotor circuit and the rest of it (1-S)P g is converted to mechanical power delivered to the motor shaft. It can be seen that when a motor operates at high slip values there is more copper loss in the rotor circuit that reduces the efficiency of the motor. Similarly, motor torque (T) in Newton-meters at various points can be calculated by dividing power (P) in Watts by angular velocity ( ro) in radians per second. p T = Newton-meters (1.14) 0) T = Pg = -1-3(1 )2 Rr Newton-meters (1.15) g 0) 0) r S s s Tct = Pct = 1 (1-S)P = 1-3(1 )2 Rr = T Newton-meters (1.16) ro 0) (1 S) g 0) r S g r s s T = P d -(friction, windage, core, and stray load losses) ( 1 .1?) out 0) r Where: T g = torque developed at the air-gap where ro is at synchronous speed 24

PAGE 35

T d = torque input to the shaft where ro is at (1-S) times synchronous speed Tout= output torque of the shaft where ro is at (1-S) times synchronous speed hr I I 47tf CO 5 = sync onous angu ar ve octty = --poles COr= rotor angular velocity = (1-S) 41tf poles Upcoming equations (1.20) through (1.23) will be better understood ifthe subject of equivalent Thevenin circuit is first introduced. The general application of Thevenin theorem permits replacing a network of linear circuit elements and constant phasor voltage sources with only one single equivalent voltage source in series with one single equivalent impedance as seen from arbitrary terminal points "a" and "b" of the original network circuit. The equivalent voltage source is the voltage appearing across terminals "a" and "b" of the original circuit when these terminals are open circuited. The equivalent impedance is what appears across the same terminals looking into the original circuit with all the voltage sources short-circuited lfThevenin theorem is applied to the circuit ofFigure 1.7 with terminals "a" and "b" as marked and the network circuit of interest being everything to the left of terminals "a" and "b" (which includes the stator circuit and magnetizing branch reactance), then Thevenin equivalent voltage source and equivalent impedance can be calculated by the following two equations A A jXm v = v la 5 Rr + j(Xm + Xs) Where: vI a= Thevenin equivalent voltage as a phasor quantity V5 = supply or bus voltage as a phasor quantity 25 (1.18) ( 1.19)

PAGE 36

Ze = Thevenin equivalent impedance The equivalent circuit of an induction motor as simplified by the use of Thevenin theorem is presented in Figure 1.1 0. Notice that now there is only one current, namely rotor current Or), flowing in the entire circuit. This circuit configuration makes it easier to calculate maximum torque, slip at maximum torque, and so on as will be shown in the following equations b Figure 1.1 0 Thevenin equivalent circuit of an induction machine. The following useful equations presented here will not be derived step by step; however, their use will be explained. Most machinery textbooks do a good enough job of providing a detailed account of their derivations so that there is no need in re-inventing the wheel here. What's important is knowing that these equations exist and being able to use them effectively in this thesis or on any another assignment. Internal torque of an induction motor in terms of Thevenin equivalent circuit IS 26

PAGE 37

(1.20) This equation is equivalent to torque equation (1.15) presented earlier with Vfa I [ (Re +Rr/S)2 + (Xe +Xr ) 2 ] replacing Clr ) 2 From equation (1.20), the torque slip curve of Figure 1.2 can be drawn. Internal torque is maximum when the power delivered to the resistive component of the rotor circuit (R/S) is a maximum. By use of the principle of impedance matching from circuit theory, power delivered to (R/S) will be maximum when its magnitude equals the magnitude of impedance in the rest of the Thevenin equivalent circuit of Figure 1.1 0. Rr = Smax T Where: Smax T refers to the value of slip at the point of maximum torque From the above equation the value of slip at maximum torque is (1.21) (1. 22) The value of maximum torque T max can now be calculated by substituting equation ( 1.22) into equation (1.20) for torque as follows 1 1.5Vfa Tmax ffis (1.23) If torque, current, and slip are known for a particular operating point (i.e., starting maximum, full load), the following equations may be used to calculate the same quantities at a different operating point. 27

PAGE 38

Tstart = [Ir start ] 2 S 'L I f.l. f.l. r f.l. Tstart 2 X Smax T Tmax = T Tn 2 x Smax T x Su 2 2 max Smax T + Su (1.24) (1.25) (1.26) (1.27) Note: Slip values in equation (1.27) must be entered as a percentage number (i.e.: a slip value of 0.035 or 3.5% must be entered as 3.5) in order for this equation to work. T 2 Tmax = ( S ) +(Smax T) Smax T S (1.28) Note: Above equation is only valid for small values of slip (S). Rr Rr + Rr external S2 (1.29) Note: Above equation is only valid for small values of slip (S). This equation is only useful for wound rotor induction machine where slip (hence speed) can be changed from S1 to S2 by adding external resistors Rr external to the rotor circuit. 1 + ( r Sr.L S x max T T Ir start = I r f.l. (1.30) 28

PAGE 39

Note: Slip values in equation (1.30) must be entered as a percentage number (i.e.: a slip value of0.035 or 3.5% must be entered as 3.5) in order for this equation to work. Now that the basics of an induction machine have been covered, it is time to explore some of the important engineering topics required for a successful assessment and application of induction machines in microand mini-sized power plants driven by hydroelectric turbines. 29

PAGE 40

2. Application of Induction Generators in Microand Mini Hydro Stations 2.1 Introduction The discussion in the preceding chapter outlined the design and operational characteristics of an induction generator. Its relative simplicity, reliability, and low cost compared to a synchronous generator are major factors responsible for the induction generator's popularity among small power producers. Despite these advantages, application of an induction generator in a micro and mini-hydroelectric project is not without its own share of obstacles that require careful considerations. Some of the more generic problems encountered in developing microand mini-hydro stations include: economic feasibility issues, high cost of fulfilling numerous requirements of regulatory agencies, difficulty in obtaining permits, the need for hydrological information to determine power potential, water rights, lack of standardized equipment that could reduce costs, and difficulties in transmitting the generated power to the utility or consumers in a relatively small quantity. There are also specific problems such as frequency and voltage regulation associated with using an induction generator in a stand-alone configuration that must be addressed. The majority of the above problems are related to the construction and licensing parts of the project which are unique for each site. In the limited scope, time and resources available for this thesis, these problems cannot all be discussed. Nevertheless, a selected few of the issues directly associated with an induction 30

PAGE 41

generator in a hydro setting will be addressed and solutions and/or alternative options will be proposed. A major portion of the total cost of the project is the electrical and mechanical equipment cost for the powerhouse. In an effort to make such small microand mini hydro projects economically feasible, it is sometimes necessary to consider low cost alternatives for power production equipment. In this process, it is necessary to look for energy conversion equipment which is easily available at a low cost, is reliable and can be easily operated and maintained. In stating low cost, the benchmark is the cost of conventional equipment. Although the conventional equipment may not be easily available nor most reliable, still, it offers high efficiency at a premium price. With some of the alternatives proposed here, high efficiency may be compromised in order to gain lower equipment cost and higher reliability. This is especially important in a micro-hydro applications where the main concern is economic feasibility and reliability instead of high efficiency. In order to make the proper selection, an economic feasibility study must be performed to take into account cost of equipment, payback period, life of the plant, tariff rates, etc., for each set of options available. In searching for easily available equipment, the criterion is to look for equipment from proven and well-established lines of products commonly available in the market with well-defined operating characteristics. In utilizing such commonly available products, the intention is to use them within their capability limits but not necessarily in the same manner or convention as that for which they were intended. For example, an induction machine which is built to run as a motor may serve as a generator and similarly a pump may be substituted for a turbine. By doing so, the nominal operating mode of operation has been changed; however, the capability limits are not exceeded. A proper selection of such equipment could have a large impact on the viability of a small project. 31

PAGE 42

The conventional drive mechanism for a hydroelectric miit is a turbine governor set which is responsible for converting hydraulic energy to mechanical energy and at the same time regulating shaft speed (i.e., real power output of the generator). As will be covered in the following section, alternative low cost equipment that can perform the same function as a turbine is a suitable centrifugal pump operating in reverse mode. Furthermore, instead of having a slow acting and expensive mechanical governor to control shaft speed (by controlling the mechanical input power), it is possible to control speed by controlling the electrical load on the generator. Electrical devices suited for this application come with various designs and features. The most common names given to such devices are an electronic load governor or an impedance controller which have the function of compensating for the fluctuations in the consumer load so that the total load seen by the generator remains constant and hence speed and frequency are kept the same. 2.2 The Choice of Turbine and Governor for a Microand Mini-Hydro Station 2.2.1 Turbine A hydraulic turbine consists primarily of a runner connected to a shaft which converts hydraulic power into mechanical power from the energy stored in water. The function of a governor is to control the turbine's operation, and mechanical speed and the real power produced by the generator. The hydraulic turbine is the most important element in a hydroelectric power plant and its proper selection is crucial. There are two general groups of hydraulic turbines [5]. One is the impulse type where water enters the turbine with high kinetic energy (in the form of velocity) and a relatively low value of potential energy (in the form of pressure); and, the other one is the reaction type where the water enters the turbine with high potential energy 32

PAGE 43

and a lesser amount of kinetic energy The reaction-type turbine can be further subdivided into Francis type and propeller-type turbines. Finally, the propeller type turbine has the following sub-categories: 1. Fixed blade propeller turbine. 2. Adjustable-blade propeller (Kaplan) turbine. 3. Axial-flow propeller turbine (tube, pit or bulb). 4. Diagonal-flow turbine. Simplistically speaking, in the aerated housing of an impulse (Pelton) turbine, one or more water jets hit the bowl-shaped turbine runner buckets with high kinetic energy (high velocity) and lose most of their kinetic energy (low velocity). The forces on the buckets are a result of the impulse or momentum change of the water as its absolute velocity is reduced to near zero in the buckets. The impulse turbine thus utilizes the kinetic energy of the fluid entering the turbine to generate power. Impulse turbines are used for high heads ranging from 1 000 to 5000 feet. In a reaction-type turbine, water enters the intake side of turbine housing with high potential energy (high pressure) and relatively low kinetic energy (low velocity). The water then leaves the outlet side of the turbine, which is much larger in diameter than the intake side, at a lower pressure The pressure difference across the top and bottom of the turbine runner exerts a force that causes the runner to rotate and deliver mechanical energy to the shaft. In a Francis turbine, water enters the spiral case (curled around the turbine shaft) from the intake or penstock. Water then passes through the stationary stay rings (which guide the water flow) and through the adjustable wicket gates (which control the water flow to the runner and thus power output of the turbine). Finally, water passes through the runner before it enters the draft tube and into the tailrace. A Francis turbine runner looks somewhat like a squirrel cage rotor with deep and 33

PAGE 44

skewed bars that are tapered in diameter towards the top. Francis turbines are normally used for medium heads ranging from 100 to 1500 feet. The runner of a propeller-type turbine has blades similar to a ship's propeller. The number of blades varies from 3 to 10, and they could be either fixed or adjustable (variable pitch) as their names imply. In most propeller-type turbines the axis of each individual blade is at a right angle to the turbine shaft; except, in a diagonal-flow propeller turbine where the axis of each blade makes a 45 angle with the turbine shaft. Fixed-blade and adjustable-blade vertical propeller turbines are normally used for medium heads ranging up to 150 and 200 feet respectively. The operating range of a propeller turbine partly overlaps the operating range of a Francis turbine. In an axial-flow propeller turbine, the turbine shaft is either slightly inclined (as in a tube turbine) or horizontal (as in a bulb or a pit turbine). A horizontal or nearly horizontal turbine shaft allows for a straight-through or nearly straight-through water passageway from intake to draft-tube discharge. Turbines like these are often used in tidal and other low head hydro plants and can be designed to operate as a pump or as a turbine. The shaft speed of a smaller tube turbine is very slow. In order to reduce the size and cost of a generator for a tube turbine, a gearbox for increasing speed is mounted between the turbine shaft and the generator shaft. Since the shaft of a tube turbine is slightly inclined, it permits the generator be located outside the water passageway. Whereas in a pit or a bulb turbine the generator is housed inside a submerged and watertight enclosure in the water passageway. In general, axial flow turbines are used for low heads ranging from the lowest head practical to 75 feet. The main advantages of a horizontal bulb turbine in comparison with a verticalshaft-propeller turbine (with the generator mounted overhead) are as follows: 34

PAGE 45

1. Because ofthe straight-through water passageway, a bulb turbine is more efficient than a vertical-shaft-propeller turbine of the same size and ratings. In a bulb turbine water is not subjected to directional changes in intake, semispiral casing, and draft tube. 2. The length, width, and especially height of the powerhouse structure can be kept to a minimum. 3. The excavation and construction costs are lower. Economics and efficiency dictate that the speed of a turbine should be selected as high as practical. At higher speeds, the overall size of the turbine and its cost are reduced. Furthermore, since a hydraulic turbine is usually directly coupled to a generator, higher speed means higher generator efficiency and lower generator cost (due to smaller number of poles). The number of generating units in a power station should be kept to a minimum. A couple of larger units are more efficient, require less auxiliary equipment and maintenance compared to several smaller units. Although other factors such as flexibility of operation, higher-efficiency during low-load demands, and minimum loss of capacity during shutdown for repair and maintenance may argue the need for multiple smaller units. The limiting factor in the size of a turbine generator unit is based on the largest size turbine runner that can be shipped via ground, railroad, or sea transportation. Still, this is seldom a problem for the turbine runner of the miniand micro-hydro units discussed in this thesis. 2.2.1.1 Reverse Mode Centrifugal Pumps as Turbines A hydraulic machine converts hydraulic energy to mechanical energy or vice versa. A machine may be designed to perform one of these tasks with maximum efficiency, this duty being defined as the nominal operating condition. If the same 35

PAGE 46

machine is operated with a reversed direction of energy flow, it may be said to be in a reverse mode operation. It is certain that in reverse mode operation the performance of the machine will change significantly from that of its nominal operating mode. For example, hydro turbines are used as pumps in pumped storage projects and reverse mode centrifugal pumps are used as turbines in small hydro projects. In the frrst case, the hydraulic machine is designed to operate most efficiently in the turbine mode, and in the second case in the pump mode. Despite these departures from optimum performance and efficiency, reverse mode operation may still be economically viable since the capital cost involved in providing a separate pump motor set for the pumped storage facility will be much greater than the savings involved through the use of an efficient machine. Likewise in a small hydro project, where the price of a conventional turbine is about five times that of a reversible centrifugal pump of essentially similar rating, the additional cost for a regular turbine will be too large to offset the advantage gained from the energy savings involved. For these economic reasons, a reverse mode centrifugal pump may be used in certain micro-hydro schemes as an alternative to the conventional turbine. However, a study of the complete characteristics of the pump is necessary to determine its suitability to function as a turbine. A radial-flow centrifugal pump is generally used to convert mechanical energy into fluid energy. In reverse mode operation, the pump works as a turbine where the fluid enters the discharge port at a high pressure, drives the impeller (runner), and leaves the suction port at a low pressure. The fluid energy is thereby converted into mechanical energy which can be utilized to drive a generator. Pump manufacturers generally provide characteristics only for pump mode operation, which relate head, speed, efficiency, power input and flow rate. However, these characteristics are different in the reverse mode operation. Still, the general 36

PAGE 47

characteristics of a pump operating in reverse mode can be obtained from well documented papers like Buse's [6]. A few of the generalized salient features of operating a pump in reverse mode (turbine mode) are summarized below. 1. The turbine's peak efficiency point occurs at a lower flow rate than where the pump's peak efficiency point occurs. 2. The turbine's Best Efficiency Point (BEP) takes place at a higher head and flow rate than the pump's BEP. 3. The turbine's power output is higher than the pump's input power at BEP of each. 4. The turbine's maximum efficiency seems to occur over a wider range of head and flow rate than the pump's maximum efficiency. 5. At a certain low head, water may be passing through the turbine, but no power will be produced. 6. If the turbine's performance curve is known at one speed, its performance curve at other speeds may be calculated by the use of affinity relationships. However, the usual degree of accuracy should not be expected. 7. The runaway speed of the turbine may be calculated from the performance curves and the use of affinity relationships. 8. The turbine's mechanical operation is smooth and quiet like the pump's operation. To select a pump for a particular reverse mode duty, it is necessary to specify the head, flow rate, speed and power output for the turbine duty. Usually the manufacturer's data sheet provides data on the pump's characteristics only. Still, there are couple of options available in determining the pump's reverse mode operating characteristics. One option is the use of conversion factors such as those 37

PAGE 48

proposed by Buse. These conversion factors relate the turbine BEP performance with pump BEP performance. A second option is to test a potentially suitable pump in a test laboratory or at a factory. The test results obtained will be much more accurate than from any mathematical conversion factors available. Availability of hydraulic pumps (new or used) far exceeds those of hydraulic turbines and the cost savings involved could very well make a micro-hydro station economically feasible. 2.2.2 Standard Governor Versus Load Governor In the past, many studies have been done to regulate the voltage and frequency of a self-excited induction generator under varying consumer loads. In most such studies, the frequency regulation is achieved by regulating the speed of the prime mover by using a mechanical governor. The control of supply voltage is usually performed by controlling a variable reactive power source. However, changes in prime mover speed do not result in a linear change in frequency under varying loads due to changes in slip speed of the machine. Furthermore, difficulties in building a smoothly variable reactive power source at a low cost have restricted the performance of voltage regulators. Moreover, the transient response ofthe generator system to a load change is limited by the large time constant of the mechanical governor. Based on several studies [7, 8], regulation ofboth voltage and frequency can be achieved by controlling the current drawn from the generator while operating with an unregulated turbine and a constant excitation capacitance. Regulation of the load current is accomplished with the aid of an impedance controller connected at the generator terminals which is capable of drawing a controllable lagging current. Unlike in large hydro power stations, the conservation of water in the reservoir is not usually a constraint in microand mini-hydro stations. Therefore, full power is generated with a fully opened gate valve, regardless of the demand on 38

PAGE 49

consumer load. The unnecessary power is dissipated as heat in the impedance controller. If required, the conservation of water in the reservoir is possible by closing the gate valve slowly during long periods wich low power demands. The dissipated heat may be used for other purposes such as water heating and steam making. Thus, the mechanical speed governor responsible for slow transient response and a significant percentage of total cost of a small power station are eliminated. A regulator is designed to control the impedance controller such that the load current of the generator is always at a value which results in rated voltage and rated frequency under all operating conditions. When the current drawn by the consumer load drops below the calculated value, the difference in currents is drawn by the impedance controller connected in parallel to the consumer load. An impedance controller which provides fast response to control commands can be implemented with fast switching power electronic devices. As a consequence, the proposed load current regulator is capable of improving the transient performance of the generator system significantly. The quality of an ac power supply is primarily measured by its ability to supply real and reactive power demand of the consumer load at the rated voltage and frequency under all operating conditions. The main cause of difficulty in maintaining the voltage and the frequency at their rated values is the ever-changing nature of consumer demand. In order to maintain constant voltage and the frequency, the supply and the demand of real and reactive power should be exactly balanced. 2.2.2.1 Control of Terminal Impedance A different method to maintain the balance of real and reactive power of the system with varying consumer loads would be to keep both the power supply and the demand of power from the consumers constant. The supply of real power is kept 39

PAGE 50

constant by using a constant power source such as a hydro turbine operated with a fixed gate valve position at a constant head. The reactive power supply is maintained constant by using a constant bank of excitation capacitors. The demand of real and reactive power from the generator can be kept constant despite the variations of consumer load by controlling an additional load at the generator terminals, connected in parallel with the consumer load. This auxiliary load is controlled to consume the difference between the supply and the demand of both real and reactive powers. In another words, the impedance seen by the generator at its terminals is maintained constant. Therefore, the auxiliary load at the generator terminals can be identified as the impedance controller of the generator system. There are several advantages in using this control strategy. First, the regulation of both voltage and frequency ofthe system is performed by a single controller instead of two controllers required in the standard method. This results in reduced complexity of control. Secondly, a mechanical governor regulating the speed of the turbine is not required for this method of generator control. Thus the cost of the generating station drops significantly. Thirdly, the slow response of the mechanical governor system is replaced by the fast response of an electronically switched impedance controller. As a consequence, the terminal impedance controller has the potential to provide excellent transient performance which cannot be achieved by a conventional mechanical governor. 2.2.2.1.1 Proposed Regulator Due to the availability of power semiconductor devices capable of switching at high frequencies, an impedance controller with desired properties can be realized in practically many different configurations. Also, the control of such an impedance controller is now easier than ever before, mainly due to the development of powerful 40

PAGE 51

micro-controllers at lower costs. Despite the availability of powerful micro controllers, the control strategies described in this thesis are fairly simple and do not fully exploit the capabilities of such powerful devices. Considering a small hydro power station operating with an unregulated turbine and a constant excitation capacitance bank, the types of disturbances experienced can be divided into two main categories-fast and slow disturbances. The dominant cause of fast disturbances is sudden changes in consumer load an its impedance. The quality of the power supplied mainly depends on the ability of the regulator to act against such fast disturbances. In addition, slow disturbances are also to be expected due to many causes such as fluctuations of the water level of the reservoir and drifts in system parameters with aging and changes in ambient conditions. A regulator with a feed-forward control loop which would monitor the rate of change in load current would provide an effective control strategy for regulation against fast disturbances. Similarly, addition of a proportional-integral feedback control loop acting on the steady-state errors in voltage and frequency is necessary to regulate the induction generator through the impedance controller against slow disturbances. 2.2.2.1.2 Impedance Controller For the successful implementation of the proposed regulator, the impedance controller must fulfill a number of requirements. First, the controller must be able to control the flow of real and reactive currents independently in order to be able to compensate for arbitrary changes in real and reactive power demands from the consumer load. This can only be achieved when the impedance controller has at least two control variables. Second, it must provide a fast response to a given control 41

PAGE 52

command so it can have high performance during load changes. With feed-forward regulation, the delay in acting against changes in load current depends only on the delay in calculating the desired control commands and the delay in responding to these commands. The calculation of the proper values of the control commands is mainly determined by the speed of the microprocessor. Since present day microprocessors can perform this task within a few hundreds of a microsecond, the regulator performance is primarily limited by the response time of the impedance controller. Third, a continuous control of real and reactive currents drawn from the impedance controller minimizes any jumps in voltage and frequency due to actions of the regulator. Therefore, an impedance controller with high control resolution is important. Other important features of an impedance controller include: a simple and rugged circuit with high reliability, minimum number of circuit components to keep the cost down, and the ability to minimize harmonic distortion from the fast switching of power electronic devices. There are several different circuits that could satisfy the basic requirements of an impedance controller. In most cases, solid-state switching devices such as phase controlled thyristors are used in some form of a static var circuit configuration to control and vary the effective amount of resistive, capacitive, and inductive loads on the generator. After a careful examination of all the available options, the circuit configuration that best suits our current application and meets the above requirements is the one which uses three fixed capacitors, a controlled bridge rectifier, and a single quadrant chopper (operates in the first quadrant only) and a resistor on the de side of the rectifier. This circuit configuration is shown in Figure 2.1. 42

PAGE 53

Consumer Load GENERATOR Reclifier Ia Figure 2.1 -Circuit configuration of the impedance controller with rectifier and chopper. The resistively loaded bridge rectifier draws both real and reactive currents determined by the rectifier delay angle. The function of the chopper is to change the magnitude of the current drawn by varying the conduction period as is done in pulse width modulation. This configuration requires only one resistor, seven power semiconductor switching devices, and three fixed excitation capacitors. The thyristors of the bridge rectifier are line commutated and the single-quadrant chopper can be implemented with a switching device such as GTO or BJT which have gate tum-off capabilities. The reduction of harmonic distortion on the ac supply voltage is accomplished by switching the chopper at high frequencies. The delay angle of the rectifier and the pulse width of the chopper can be continuously and independently controlled to achieve high resolution. Furthermore, with high frequency switching of the chopper, this configuration provides the fastest response of all alternatives under considera1ion. 43

PAGE 54

In implementing the selected impedance controller, the first element to be evaluated is the controller resistance The size of this resistance determines the current ratings of all power semiconductor devices in the circuit. The selection of Rdc is made such that the impedance controller provides the desired range of control for the consumer load. This impedance controller draws both real and imaginary current components through the bridge rectifier and chopper. Both direction and magnitude of the current drawn could be varied as changes are made to the delay angle of the rectifiers a and pulse width of the chopper PW. As a result, a and PW are used as the two control commands to control of power factor and magnitude ofthe current drawn from the circuit. One of the common problems with power electronic devices is the harmonic distortion injected into the supply voltage waveform. Similar to a bridge rectifier, the impedance controller under consideration introduces 5, 7, 11, 13, ... ,etc., harmonics into the ac voltage supply. The excitation capacitors at the input terminals of the rectifier act as a low pass filter to reduce some the distortion of the ac voltage waveform caused by the impedance controller. Further reduction of the harmonics can be achieved by increasing the switching frequency of the chopper analogous to the pulse width modulation of an inverter. The high switching frequency of the chopper reduces the magnitudes of lower order harmonics such as 5, 7, 11 at the expense of increased higher order harmonics. Since the ac system acts like a low pass filter, the net result is a reduction in distortion in the voltage waveform. Although as the switching frequency ofthe chopper is increased, the overall voltage distortion gets reduced; this benefit diminishes as the switching frequency is increased beyond a certain limit. 44

PAGE 55

Furthermore, practical problems such as the maximum switching frequency of power semiconductor devices such as GTO's limit the number ofswitchings ofthe chopper that can take place during a 60 period that each of the six rectifiers is conducting. The results of a study performed by Hoops [9] show that five switchings of the chopper per each conduction period of the rectifier (every 60) to be the optimum switching frequency. For easier understanding, Figures 2.2 and 2.4 show the current waveforms of the bridge rectifier alone without the presence of the chopper circuit for delay angles of 30 and 90 respectively. Figures 2.3 and 2.5 show the complete operation of the bridge rectifier and the chopper combined with the same delay angles, five pulse chopping per each conduction period, and a chopper pulse width of 0.5 (50% on, 50% off). With five-pulse chopping there are 30 choppings per each cycle ofthe ac supply. This amounts to chopper frequency of 1800 Hertz for a fundamental frequency of 60 Hertz on the ac supply. 45

PAGE 56

z.o Phase angle (degrees) Figure 2.2 Rectifier line current for phase A with a.=30 without the chopper. 2.0 1.5 1.0 ;;-0.5 ..::: -0. 5 -1.0 -1.5 -2.0-60 90 120 Phase angle (degrees) Figure 2.3 -Impedance controller line current for phase A with a=30, five-pulse chopping and PW=0.5. 46 360

PAGE 57

2.0 2 0 -60 -30 330 360 Phase angle (degrees} Figure 2.4 Rectifier line current for phase A with a=90 without the chopper. 2 0 -60 -30 210 240 270 300 330 360 Phase angle (degrees) Figure 2.5-Impedance controller line current for phase A with a=90, five-pulse chopping and PW=0.5. 47

PAGE 58

In a case where there is a strict requirement for reducing the harmonic distortion to a minimum, a harmonic filter can designed by using some or all of the excitation capacitors and extra inductors to form an LC filter. Typically, the harmonic distortion of voltage without any filtering is only about five percent for five-pulse chopping The contribution of an LC filter is fairly limited. In reported studies, it was able to reduce the total harmonic distortion from about five percent to about 1.4 percent with added cost and complexity to the project. Therefore, the use of such LC filters is not recommended for this application. In the interest of brevity, most of the mathematical analysis, computer simulation studies, and intermediate experimental results associated with performance and characteristics of the impedance controller are excluded from this thesis. Instead, a summary is presented on the effects of varying delay angle (a) and pulse width (PW) on real and imaginary currents drawn from the impedance controller and how well the closed loop control regulator performs in conjunction with the impedance controller. The primary effect of the rectifier delay angle (a) is to change the power factor of load current and the primary effect of the chopper pulse width (PW) is to change the amplitude of the current. The pulse width of the chopper has almost no effect on the power factor of the current drawn. As a result, power factor of the impedance controller is almost completely determined by the rectifier delay angle (a) and the amplitude of current is almost directly proportional to the per unit pulse width (PW). In short, after the power factor has been determined by a, PW determines the amplitude of the current. Consequently, a change in the amplitude of the load current can be corrected much faster than a change in the phase angle of the load since chopper frequency is five times faster than the rectifier frequency. 48

PAGE 59

Figure 2.6 shows the control range of the impedance controller as delay angle (a.) and pulse width (PW) are varied for terminal voltage (V) and impedance controller resistance (Rdc) set to one per unit each. 0.8 0.7 0: 0.5 c.. E 8 0.4 i::' "" "' .CD 0 3 "" .. 0.2 0.1 0 Figure 2.6 Area of current control offered by the impedance controller for pu at rated voltage. Furthermore, experimental results using a thyristor bridge rectifier and a GTO chopper showed close correspondences among theoretical, computer simulation, and laboratory experimental results. The feed-forward regulator displayed a superb performance against load changes. When the generator system is regulated only by the feed-forward regulator, the steady-state errors ofvoltage and frequency are 0.04 pu and 0.01 pu respectively. However, when the same load change is applied to the generator system regulated with both feedback and feed-forward regulators, the errors of voltage and frequency 49

PAGE 60

are slowly brought back to zero. Thus the small errors resulting from the use of approximately calculated control commands can be eliminated by the action of the feedback regulator. The results of computer simulation studies demonstrate the superior performance of the chosen regulator The action of the feed-forward regulator results in excellent transient performance for changes in consumer load. Except for the small steady-state errors resulting from the use of approximately calculated control commands, the transients in voltage and frequency are over within two cycles The remaining steady-state errors are corrected by the slow-acting feedback regulator. The feedback regulator guarantees errorless operation in steady-state even when the operating parameters change from their normal values. According to the proposed strategy, the generator is operated with an unregulated prime mover eliminating expensive mechanical speed governors. The regulation of voltage and frequency is achieved by drawing the correct amount of load current resulting in rated voltage and rated frequency under any changes in operating conditions. An impedance controller connected at the terminals of a generator draws the difference between the generator current and the consumer load current. The fast disturbances resulting from changes in consumer load impedances are quickly compensated by the feed-forward regulator. Compensation for slow-acting disturbances and steady-state errors resulting from feed-forward regulation is made by a feedback regulator. Thus the regulator draws the correct amount of current even when operating parameters slowly drift from their normal values. Therefore, the described regulator and impedance controller have the potential to improve the performance of the induction generator in great proportions. 50

PAGE 61

2.3 Induction Generators in Micro-Hydro Stations as Stand-Alone Units 2.3.1 De-Rating of Induction Machine for Unbalanced Operation 2.3.1.1 Introduction A major difficulty of power production in micro range (less than 100 KW) is that most of these units are located in remote areas, operate as stand-alone units, and serve only the local loads. Types of loads found in the majority ofthese remote locations are either single-phase or unbalanced. Yet, the generators used for these applications are balanced three-phase generators that require balanced three-phase loads in order to produce their rated outputs. One way of handling this problem is simply by de-rating the induction generator. But from an economic point of view, this is not attractive. Another solution would be to study each load profile and redistribute the electrical loads. The goal is to redistribute the electrical load in such a way that for any given time period, each phase of the generator has an equal share of the total load. Most likely this won't be a one-hundred-percent solution and a small amount of de-rating of the induction generator still has to be done based on the worst case of unbalance condition. Another method suggested in reference [ 1 0] proposes the use of symmetrical components to calculate reactance values for an unbalanced three-phase excitation bank such that when connected to the unbalanced load bank, the combined impedance of the two banks as seen by the generator appears balanced. This concept is also further developed to make a single-phase load appear as a balanced three-phase load to the induction generator. Quite often, de-rating of an induction generator seems simpler and more straightforward compared to the other alternatives available. So it would be prudent 51

PAGE 62

to discuss and develop a method for handling de-rating of an induction generator in this section. The following analytical methods are applicable for de-rating an induction motor as well as a generator. It is widely known that induction machines are sensitive to unbalanced operation. The National Electrical Handbook states that if the supply voltage has an unbalance greater than five percent, the induction motor either must be taken out of service or properly de-rated first. A relatively small unbalance in the voltage of an induction motor causes a rather large unbalance in line currents which result in excessive machine temperature rise. The formulas developed in the following section can be used to de-rate an induction machine subject to unbalanced three-phase voltages. 2.3.1.2 Balanced Induction Machine Subjected to Unbalanced Voltages The effects of unbalanced voltages on a balanced three-phase induction machine can be studied by the use of symmetrical components. In a three-phase, three wire system (no neutral wire) the unbalanced voltages can be resolved into a set of positive voltages and another set of negative voltages. Since the induction machine is assumed to be symmetrical, the positive sequence voltages give rise to the positive sequence currents in the positive sequence equivalent circuit of the machine, and similarly (yet independently) the negative sequence voltages give rise to the negative sequence currents in the negative sequence equivalent circuit of the machine. The following Figure shows the positive and negative sequence equivalent circuits of an induction machine. 52

PAGE 63

Rs Xs x,. Rs Xs x,. 1 1 v;j XM v;j X M Positive sequence equivalent circuit Negative sequence equivalent circuit Figure 2.7-Sequence network circuits of an induction machine. In the positive sequence circuit the direction of power flow is in for a motor and out for a generator. In the negative sequence circuit the direction of power flow is always inward since the value of variable resistive R/(2-S) is always positive under either motoring or generating operating conditions. z+ and z-reflect the equivalent positive and negative sequence impedances of the circuits. Next, their relative magnitudes at different operating points are estimated for later use in the calculations. At standstill (S = 1 ): z+ = z-or z+ I z-= 1 At synchronous speed (S = 0): z+ >> z-At no load speed (S:: 0 005): z+ I z-:: 12 At full load speed (S :: 0.025): z+ I z-:: 7 As will be demonstrated by the following example, the large ratios of z+ I z-at no load and at full load indicate that an induction machine is highly sensitive to voltage unbalances. A small percentage of voltage unbalance will cause a 53

PAGE 64

much larger percentage of negative sequence current to flow For example, ifthe unbalance factor (ratio of v-I y+) in a typical motor operating at full load is I 0%, then : I-v1 z-v-z+ = = -x-= O.lx7 = 0.7 per unit I+ v+ 1 z+ v+ z-(2.1) So for a 10% unbalance factor, the negative sequence current (I-) is about 70% of positive sequence current (I+). The positive and negative sequence currents produce positive and negative sequence torques respectively. The positive sequence torque is productive and does the work in a motor or the generator while the negative sequence torque is always counterproductive. It acts to brake or slow down the machine causes heating problems, and consumes energy from the electrical system. 2.3.1.3 Balanced Self-Excited Induction Generator Subjected to Unbalanced Loads In the case of a self-excited induction generator the capacitors connected across generator terminals furnish excitation reactive power requirements of the generator The capacitors must also meet the var requirements of the load. When an induction generator is operating in the self-excited mode, the generator frequency adjusts itself to a value where the following two power equilibrium conditions are met. 1. The terminal voltage assumes a value where the net reactive power flow in the circuit is zero. 2. The slip assumes a value where the net real power flow in the circuit is zero. 54

PAGE 65

UnbalanceAf-__1...,---H---f7.----f Load GENERATOR Figure 2.8 Direction of power flow in a self-excited induction generator. In this stand-alone (self-excited) mode of operation, the only energy source in the circuit is the variable negative resistance shown in the equivalent circuit above. Positive sequence power is fed from the generator to the load. Some of this power is absorbed by the unbalanced load and the remainder is converted to negative sequence power and fed back to the induction generator. Some of negative sequence power fed back to the generator is absorbed as copper loss, core loss, and the rest is converted into parasitic negative sequence torque. Figure 2.8 shows how real and reactive power are accounted for in a balanced self-excited induction generator feeding an unbalanced load 2.3.1.4 Calculation for De-rating of an Induction Machine Calculation for de-rating of a balanced three-phase induction machine subjected to unbalanced three-phase voltages is done by studying the positiveand negative-sequence equivalent circuits separately. The total copper loss due to unbalanced voltages equals the sum of copper losses produced in the positiveand negative-sequence equivalent circuits independently. 55

PAGE 66

In calculating an approximate expression for the copper loss, the magnitude of rotor current Or) is assumed to be equal to the magnitude of the stator current (!5). This is a fair assumption since the magnetizing current Om) is on the order of 0.4 per unit and it is nearly in quadrature with the rotor current ( Is 2 = Im 2 + Ir 2 ) Therefore, the magnitudes ofl5 and Ir are within 90% of each other. Then, the approximate expression for the total copper loss can be written as follows. (2.2) For an unbalance factor ofO.l and a typical ratio of sequence impedances of z+ I z-=7 at full load, it follows that I -v 1 zv -z+ s = s = _s_x-= O.lx7 = 0.7 per unit Is+ Vs + I z+ Vs + zPcu = 0.49 Ptu (2.3) (2.4) (2.5) (2.6) This last expression shows that total copper loss will be increased substantially even for a small value of unbalance factor. The negative sequence copper loss may reach even higher values if the voltage unbalance in one phase is significantly higher than the average unbalance of the three phases. Under the above unbalance conditions, the motor cannot operate continuously at its rated output level and it will sustain damage from overheating. Therefore, it is necessary to establish a method for de-rating the induction machine. Before a method can be developed, a guideline or a criterion for de-rating must be established. One limiting criterion can be that the current in the most heavily 56

PAGE 67

loaded phase should not exceed the rated value. This criteria is somewhat limiting because it does not make any allowance for temperature equalization between phases to take place. A better criterion which is more acceptable for de-rating of smaller induction machines states that the total positiveand negative-sequence copper losses should not exceed those losses experienced under balanced rated conditions. Under this criterion, even if one phase experiences a higher temperature rise than the other two phases, temperature equalization between phases should help to bring down the temperature of the hottest phase. Experimental test results reaffirm this assumption and therefore this criterion will be adopted as the basis for the calculation that follows. +_ J 2 2 Is IuIs-Where: 15+ is the maximum allowable positive-sequence stator current I5 -= vI z-is the negative-sequence stator current In= V/Z+ is rated stator current under balanced condition and V is the voltage under rated balanced condition I -v-z+ _s_= -xIu v z-v-vNote: for small unbalance factors = -v y+ [ v-z+J Is-=If! -x.. v+ z-By substituting equation (2.1 0) into equation (2.8) an expression for maximum allowable output under the unbalanced voltage condition results 57 (2.7) (2.8) (2.9) (2.1 0)

PAGE 68

+ v-z+ [ ] 2 Is = If.!. 1--x y+ z-[ ] 2 v-z+ Maximum Allowable Output= 1--x-x rated output v+ z-(2.11) (2.12) This equation for de-rating of an induction machine is plotted in Figure 2.9. + -Some textbooks as well as Gleason and Elmore [11] refer to the rat1o of Z /Z as being essentially equivalent to the easily obtainable ratio of the starting current to the full load current in an induction machine. A better approximation results if the above equation is written in terms of rotor current Cir) Let X be the no load magnetizing current expressed in per unit. Then, from approximation ( I5 2 = Im 2 + Ir 2 ), per unit value of rotor current It under rated balanced condition can be written as Ir(p.u/ x2 Under unbalanced conditions the same equation in per unit transforms to [ It ]2 -X 2 per unit If .I. (2.13) (2.14) At Ir.1.=1.0 equation (2.8) can be written as It= J1-I?and using equation (2.14) the per unit rotor current can be written as = per unit (2.15) lf.I. [ ]2 X 2 lf.I. 58

PAGE 69

Ii Ir.t. = 1-x2 -[ Is ]2 per unit Ir.t. Ii = [ ] 2 v-z+ 1-x2 ---x y+ z-per unit Ir.t. I+ Maximum allowable output = _r_ = Ir.t. Where UF is the unbalance factor. [ UFx r 1-per unit 1-x2 (2.16) (2.17) (2.18) This equation for de-rating of an induction machine which also takes into account the per unit magnetizing current (X) is plotted in Figure 2.1 0. 59

PAGE 70

0.9 0.8 0.7 a :s g_ 0.6 0.2 0.1 ... ..... ... .. \ ',, \ \ ... \ \ \ ' \ \ I I Unbalance Factor I I I I I I I ZF.L.=S ZF.L.=7 ZF.L.=9 Figure 2.9 De-rating oflnduction machine due to unbalanced voltages. __ ww __ 0 0.02 0.04 0 .06 0.08 0.1 0.12 0.14 0.16 0.18 0.2 Unbalance Factor ZF.L.=S ZF.L.=7 Zp.L.=9 Figure 2.10 De-rating of induction machine due to unbalanced voltages. 60

PAGE 71

One area that had been not been discussed until now is the effect of unbalanced voltages on increasing the core loss. Fortunately, the core loss of an induction machine under balanced operation is relatively small. Furthermore, the unbalanced voltages will have very little effect in increasing this core loss in an induction machine. The results of William's study [12] show that the increase in temperature rise due to this increase in core loss will be negligible. Therefore, in this approximate calculation for de-rating of an induction machine due to unbalanced voltages, the core loss will be considered unchanged and it will not have a temperature rise effect on the induction machine. Worth mentioning, unbalanced voltages also affect a few other performance characteristics of an induction machine. For instance, unbalance voltages cause locked-rotor and breakdown torques to decrease. This reduction in torque must be taken into account before selecting an induction machine for a particular application. Also, unbalanced voltages cause a lowering in full-load speed of an induction motor The motor, as well as a generator, experience higher losses which translate into a higher slip at full load. Finally, the locked-rotor current will be unbalanced to the same degree that the voltages are unbalanced at starting (at standstill, z+ I z-= 1 ); however, the current will be unbalanced at about 6 to 10 times the unbalanced voltages at normal operating speed (at full load speed, z+ I z-:: 7). 2.3.2 Optimizing Capacitor Size for a Stand-Alone Induction Generator The use of capacitors as an excitation var source for stand-alone (self-excited) induction generators has already been mentioned in previous sections of this thesis. It should be clear that when a sufficiently large three-phase capacitor bank is connected to the terminals of an externally driven induction generator, a small voltage tends to be in the windings. This voltage is generated due to the residual magnetism 61

PAGE 72

of the rotor or the initial charge on the capacitor bank The resulting currents flowing in the capacitors provide a positive feedback to increase the induced emf voltage in the generator windings. This process is called "self-excitation." The voltage continues to increase until it reaches its steady state value. A steady state voltage is reached when the magnetizing reactance Xm saturates to a level high enough, determined by the excitation capacitance and the rotor speed. One ofthe major application setbacks of a self-excited induction generator (SEIG) is its poor voltage regulation even under controlled speed. In order to maintain a constant terminal voltage while the load is being increased, the amount of capacitive var must be increased proportionally as well. There are a number of voltage regulating schemes that reportedly work satisfactorily. Some of these schemes include static var compensators, switched capacitors, saturable core reactors, or any combination of the above. Most of these schemes :use closed control loop, relays, contactors, and semiconductor switching operations that are complex, expensive, require maintenance, and produce harmonics. In effect, these types of voltage regulating schemes negate some of the advantages of choosing an induction generator for use in a miniand micro-power generating unit in the first place. Besides the above options and the impedance controller option already discussed, there are two simple and inexpensive capacitive var compensating schemes that exhibit almost a flat voltage regulation profile as the load changes. As discussed by Shridhar in reference [13], the two schemes are referred to as short shunt and long shunt SEIG schemes. These schemes eliminate the need for an additional voltage regulator since the schemes have inherent self-regulating features. Figures 2.11 and 2.12 show the block diagrams for these schemes. 62

PAGE 73

0 prime mover SE!G Shunt Capacitors Series Capacitors l------'----+-1 ft----i 1-------'---j(t----l load Figure 2.11 -Block diagram of a short shunt self-excited induction generator. prime mover SE!G Shunt Capacitors Series Capacitors load Figure 2.12-Block diagram of a long shunt self-excited induction generator. Shridhar [13] has demonstrated through his calculations and lab results that it is possible to have an almost flat voltage regulation across the load with a particular combination of capacitors connected in short shunt and long shunt configurations. The results of his experiments conclude that the performance of a short shunt configuration is superior to a long shunt configuration. The analytical procedures for selecting the optimum capacitor sizes and predicting performance for the short shunt scheme is explained in the following paragraphs. The objective of the following procedure is to find an ideal combination of series and shunt capacitor sizes that would maintain the load voltage within acceptable limits from no load to full load power output. The actual study itself involves modeling, computer simulation, analysis, and estimation of desired parameters. Equations (2 19) through (2.29) represent the highlights of the actual 63

PAGE 74

process involved. It will become evident that once the procedure is established and understood the design of a self-regulating self-excited induction generator is fairly straightforward. The following procedure takes into account variation of magnetizing reactance Xm with saturation as the basis of this calculation. The steady state equivalent circuit of Figure 1. 7 is modified and redrawn in Figure 2.13 to reflect the variation of reactances with frequency. This step is necessary, because in a stand-alone induction generator, speed (and hence frequency) is not regulated by the utility. As the operating speed changes so will the values of the circuit reactances--especially the value ofXm. The rotor resistance quantity R!S is rewritten in terms of per unit frequency and per unit speed as Rr[FI(F-V)] where the terms inside the bracket are equivalent to liS. Finally the extem:alload resistance, series, and shunt capacitances are added to the original equivalent circuit. -iXse /F Is Rs jX5 F iXr F RL -iXsh/F iXmF E IL ) Figure 2.13 Equivalent circuit of a short shunt self-excited induction generator. Where: X5h = Per phase capacitive reactance corresponding to shunt capacitance, C5h X5e =Per phase capacitive reactance corresponding to series capacitance, C5e RL = Per phase load resistance 64

PAGE 75

F = Per unit (PU) frequency V = Per unit speed Is, Ir, IL =Per phase stator, rotor, and load currents respectively The ratio of air-gap voltage over frequency (E/ F) is an indication of flux density in the generator air-gap. Figure 2.14 shows a typical curve demonstrating variation of Xm versus E/ F. The magnetizing characteristic of the machine which are shown in terms ofXm here is critical in analyzing the SEIG. E/F (P U ) 1.6 ..:-..:....__..:....__ __________ -, Kt Saturated region 1.2 0.8 0.4 0 0 5 1.0 1.5 2 0 2.5 xm (P.U.) Figure 2.14-Variation ofXm versus ElF. The saturated region of the above curve can be expressed mathematically by the following equation (2.19) Where K 1 and K2 are constants that correspond to the starting point and slope ofthe above curve respectively. For the purpose of solving the circuit unknowns in Figure 2.13, let's assume the values of Xse and Xsh are given. Furthermore, speed of the prime mover and the 65

PAGE 76

load RL connected to the generator are known. Then the only unknowns remaining in the circuit of Figure 2.13 would be the values of Xm and per unit base frequency (or stator frequency) F. These two unknowns can be evaluated by writing a current loop equation in the stator circuit as follows l5Z5 = 0 Where: Zs = zl + Zsh ZL I( Zsh + ZL) + Z2Zm I(Z2 + Zm) Zsh = -jXsh IF A F z2 = Rr--+ JXrF F-V ZL = RL-jX5e IF (2.20) When the generator is operating under load, stator current Is is greater than zero. Therefore, from equation (2.20), the value of Zs must equal zero. Zs is a complex quantity which could be broken down into its real and imaginary parts. By setting each of the real and imaginary parts of Zs to zero, now there are two equations and two unknowns which could be solved for Xm and F. The following two equations show the general form and order of the equations obtained. A 3 2 Real Z5= (DIXm+D2)F +(D3Xm+D4)F +(D5Xm+D6)F+ (D7Xm+Ds) = 0 Imaginary Z5 = (C 1 Xm +C2) F 4 +(C3Xm +C4) F3 +(CsXm +C6) F2 + (C7Xm+Cg)F+(C9Xm+C10) = 0 (2.21) (2.22) The C and the D constants in the above equations are defined in appendix-A. The above equations can be solved using a number of numerical methods including the Newton Raphson or the Secant method. There are numerous papers published on the above topic. In some of the other papers, the order of the equations derived may be different, but the general approach is very similar to what is presented here. This 66

PAGE 77

is mostly because of the assumptions and simplifications made in the papers. For example, in this study for simplicity it is assumed that the value of rotor reactance Xr and stator reactance X5 are equal and that the load is purely resistive. Once values of Xm and Fare found, then equation (2.19) can be used to calculate E. Subsequently, once E is known, the following equations can be used to calculate voltages and currents in the generator circuit. l5 = El [ZI + ZLZsh I (ZL + Z5b)] (2.23) Ir = E I [RrF I (FV) + jFXr] (2.24) fL = IsZsh I(ZL +Zsh) (2.25) Vs = ILZL (2.26) VL = ILRL (2.27) 2 F (2.28) P =-3Ir R --m rF-V pout = 3 I L 2 R L (2.29) As mentioned earlier, if the values of shunt and series capacitors, prime mover speed, and load are given, then above equations can be used to solve the remaining circuit parameters (i.e., Xm, F, voltages, currents, and power). The next step would be to use the above equations and perform mathematical and/or experimental trial studies to determine what size capacitors can keep the voltage regulation within an acceptable limit. From the equivalent circuit of Figure 2.13, it is clear that under no load condition, load current IL flowing through the series reactance Xse and load resistance RL are zero. Therefore, Xse (or series capacitor C5e ) has no effect on generator terminal voltage under no load condition. Therefore, under no load condition, shunt 67

PAGE 78

reactance Xsh (or shunt capacitor C5h) is solely re.sponsible for maintaining generator terminal voltage. The effect of varying shunt capacitor Csh on terminal voltage can be studied to determine the range of values for a shunt capacitor that will maintain the terminal voltage to within percent of rated terminal voltage. A typical curve which resulted from the study of a laboratory 3.7 KW, three-phase induction generator is presented in Figure 2.15. 1.6r-----------------------, 1.4 1.2 1.061.0 0.94 0 8 =i c.: 0 6 "' 5' 0.4 -cl ] 0.2 Shunl capacitance C5h (micro Farads) Figure 2.15-Variation ofterminal voltage with shunt capacitor Csh The above graph shows that as the value of Csh is increased, so will the no load terminal voltage V5 of the generator. As a first trial, a value ofCsh equal to 16.7 1-1F is chosen. This value corresponds to a no load voltage of 1.0 per unit. With the value of Csh fixed, another mathematical and/or experimental trial study is performed to determine regulation of load voltage V L (from no load to full load power) as the value of series capacitor Cse is changed. Figure 2.16 shows the results of such study. The two vertical dashed lines in Figure 2.16 represent the range of values for series capacitor Cse that would maintain the generator terminal voltage V s to within 68

PAGE 79

percent of rated value. This Figure also shows that load voltage regulation can be minimized to about 2 percent when the value of the series capacitor is set at 40 j.I.F. 12 10 "' 8 :3 "' :; CD 6 = "' <.> ct4 2 0 0 Csh =16. 7 micro Farad computed lab results 120 Figure 2.16 Effect of C5e variation on load voltage regulation. The first set of suitable values for C5h and C5e have now been computed. Once the mathematical routine is set up to calculate one set of values (such as those calculated above), then it is relatively simple to play a game of"what if?" with the computer and change the value of series capacitor Cse In the next step, the effects of C5e at 20, 40 and 60 !J.F are studied to determine impact on load voltage V L> terminal voltage V 5 and stator current I5 at different power output levels. The objective of this study is make sure generator voltage and current ratings are not being exceeded while the capacitors maintain load voltage regulation to within an acceptable limit. The following three Figures graphically show the results of such study. 69

PAGE 80

1.2 VL (P.U.} 0.9 Cse Computational Experimental 20 40 60 0 2 0 + 0 4 0 6 0 8 1.0 Output power in per unit 1.2 Figure 2.17 Variation of load voltage V L with load. 1.4 1.4 1.2 Cse Computational Experimental 0.6 20 0 40 + 60 0 6 0 0 2 0 4 0 6 0 6 Output power in per unit 1.0 14 Figure 2.18 Variation of air-gap voltage E with load. 70

PAGE 81

!.4 l5(P. U ) 1.2 0 C5e Computational Experimental 20 l 0 40 2 + 60 3 0 0 2 0.4 0 6 0.8 1.0 1.2 1.4 Output power in per unit Figure 2.19 Variation of stator current l5 with load. From examining the above Figures, the following observations could be made. When a lower value of C5e corresponding to 20 !J.F is used, load voltage regulation is high (Figures 2.16 and 2.17), generator terminal voltage rises above its rating (Figure 2.8), and rating of stator current is somewhat exceeded (Figure 2.19) at rated output power. When a higher value ofCse corresponding to 60 !J.F is used, load regulation is slightly poor (Figures 2.16 and 2.17), generator terminal voltage is normal (Figure 2 18), and the stator current drops below one per unit at rated power output. So with a higher value of C5e the generator output capability can be increased at the expense of higher voltage regulation. There is also an extra cost involved for the purchase of additional capacitors and the generator may experience higher overspeed conditions in case it loses its load. A further adjustment of the series capacitor size to C5e = 45 !J.F demonstrates the best overall performance of the system. At this setting, the voltage regulation is as low as two percent and the generator can be loaded to I 07 percent without 71

PAGE 82

exceeding the voltage and current ratings of the generator beyond manufacturer's recommendations. The low value of voltage regulation makes the SEIG attractive when compared to the synchronous generator. The synchronous generator, in addition to being more expensive, also has complex control systems as covered earlier. Figure 2.20 combines all the important characteristics of the circuit in one graph for this new setting. 0.6 C se =45 micro Farads C sh= 16. 7 micro Farads 0.4 0.2 0 14 0.4 0.6 Q B Output power (P. U ) 0 0 2 Figure 2.20 Overall characteristics of a short shunt SEIG. After going through this entire process, there may be one other aspect of the design which may be worth investigating. It should be of no surprise that the shunt capacitor C5h also has some effects on load voltage regulation. What would happen if instead of setting C5h to 16.7 J.l.F, it was changed to 15.3 and 18 J.l.F. These values correspond to 0.94 and 1.06 per unit ofno load terminal voltage in Figure 2.15. From the results of such a study, it is concluded that C5h set at 16.7 (corresponding to unity no load voltage) still produces the best performance on load voltage regulation. The results are shown in Figure 2.21. As seen from this Figure, for smaller values of 72

PAGE 83

Csh an undesirable voltage dip appears across the load at partial generator output levels. VL per unit 1.2r--=-'--------------------, 0.2 0.4 0.6 0 6 1.0 1.2 1.4 Output power in per unit Figure 2.21 -Effects of changing shunt capacitor C5h on load voltage. The last topic to be covered in this section is investigating the performance of a long shunt SEIG (see Figure 2.12) versus a short shunt SEIG. Derivation of equations and circuit analysis of a long shunt SEIG is fairly lengthy yet similar to what has already been covered for a short shunt SEIG here. Therefore, the reader will be spared from a detailed coverage oflong shunt SEIG. Nevertheless, mathematical and experimental results from the study are discussed ahead. Analyses and experiments similar to those were conducted for a short shunt SEIG were also repeated for the long shunt configuration. The results show that like a short shunt SEIG, the long shunt SEIG also has self-regulating features. However, the best voltage regulation that a long shunt SEIG can produce is about six percent compared to two percent that was obtained from a short shunt SEIG. The size of the series capacitor required for a long shunt configuration is about 100 --more than double size and cost of the capacitor used for the short shunt arrangement. 73

PAGE 84

Furthermore, in the long shunt configuration, the series capacitor must have a higher current rating to be able to handle not just the load current, but the entire generator line current. So with minimal time spent on long shunt SEIG, it can be concluded that performance characteristics and economics make the short shunt SEIG more attractive than the long shunt SEIG. In a follow up study presented in reference [14] the transient performance of short shunt SEIG is compared to the long shunt and the simple shunt (no series capacitor) configurations. Here, the conclusion drawn reinforces the performance superiority of short shunt SEI G configuration over that of the other two configurations during steady state as well as transient state. Some of the relevant and noteworthy findings ofthis study are summarized below. After a short-circuit, a simple shunt SEIG cannot build up voltage and self excite itself if the load remains connected to it. The residual magnetism of the machine reduces to such a low level that an external jolt is required before re excitation and voltage build up is possible. A simple shunt SEIG has poor voltage regulation, low overload capability, and problems with voltage build up after a short circuit on its terminals. On the contrary, the short shunt SEI G not only can re-excite itself after a short-circuit, it can do it while the load remains connected to the generator. The series capacitor reinforces the voltage build up process and prevents the generator from becoming totally de-magnetized. The short shunt SEIG has good steady state as well as transient performance. It has a high overload capability and can withstand switching transients of loads up to 1.6 per unit without losing self-excitation. 74

PAGE 85

2.4 Induction Generators in Mini-Hydro Stations In previous sections the use of induction generators in micro-hydro stations was emphasized. Hydro power generation of capacity less than 100 KW is categorized as micro-hydro generation. The economics as well as the technical problems encountered in micro-hydro schemes are different from those of the mini hydro's. The economic aspect of a micro-hydro scheme may suggest the use of a reverse mode centrifugal pump along with an impedance controller instead of a conventional turbine governor set. However, in mini-hydro schemes the problems are of a different nature. An induction generator may be cheaper, more robust, and relatively maintenance free compared to a synchronous generator in the medium ratings (1 00 KW to 5000 KW), making it a better choice than a synchronous generator; but the generation must be controllable--much more so than in a micro-hydro scheme. Satisfactory parallel operation must be ensured by giving attention to the stability conditions. To make the scheme economically viable the generators must work in parallel with the grid system where there are other synchronous generators available to supply the excitation power. These problems make the mini-hydro scheme different from the micro-hydro schemes and suggest a different approach. A coordination study of an induction generator in an interconnected grid system is necessary to determine system's long term dynamic stability. The short term transient stability aspect is kept out of the picture assuming the system is fairly strong. Two different methods are widely used for performing such studies--either computer simulation of the complete system or a real time simulation. A typical study involves investigating many different systems. Each of these systems in turn has its own sub-system. For example a thermal power generating system has a boiler, turbine, governor, turbogenerator, voltage regulator, etc. 75

PAGE 86

If in lieu of real time simulation study a computer simulation method is selected, then each system is represented by its mathematical model. However, in this approach simplifications and approximations are made in order to keep the size of the simulation realizable on a computer. Furthermore, one has to make several trial runs, study several computational results, and compare numerous graphs before being able to get a feel for the characteristics of the complete generating system. In medium range, use of induction generators in mini-hydro schemes is contingent on economic aspects of meeting their excitation requirements. The cost of providing the excitation capacitors offsets their economic viability as compared to synchronous generators with larger ratings However, in situations where induction generators are used in a power system network in conjunction with other generators which can provide the required excitation power, use of induction generators could be made economically viable even in medium output range. Most induction generators in mini-hydro application today operate in parallel with the interconnected network. DeMello has reported [15] that induction generators are technically feasible in interconnected power systems and where the network is strong enough from the transient stability point of view. The var control could be implemented at the source of excitation where it might be at some distance away from the induction generator. 76

PAGE 87

3. Protection of Induction Generators Protection of induction machines has been extensively covered in numerous IEEE publications and in protective relay manufacturers' literature [ 1]. These protection schemes may vary from bare essentials (those required by the National Electrical Codes) to very elaborate and expensive schemes (often implemented or required by the larger utility companies). The choice of protection for each induction generator should be based on sound engineering judgments and the specific circumstances that surround each induction generator installation. One determining factor is the size of the induction generator and its replacement cost. The amount ofprotection provided for a 1000 KW generator should not be the same as that provided for a 50 KW generator. A larger and more expensive generator can justify a more elaborate and expensive protective scheme than that can be afforded for a small generator. Other important factors that should be evaluated in considering protection include: the location of the induction generator and ease of access to replacement parts and qualified technicians, the presence (or lack of) an operator on duty who can respond to a problem quickly, the impact of prolonged unavailability of a generator in lost revenue and the dependency of a remote community on its operation as its main source of electric power. In the forthcoming paragraphs different protective devices that make up a typical induction generator protective scheme are discussed. Each protective device's intended applications, benefits and limitations are underscored. Afterward, these individual devices are consolidated together to form a typical induction generator protective scheme. 77

PAGE 88

Please bear in mind that the protective devices discussed in this section will be augmented by additional operational and/or protective requirements irn:posed by the utility on induction generators interconnected to the power grid. The additional requirements and their basis will be covered in the following section under "Utility Interface Protection of Induction Generators." 3.1 Circuit Breaker (ANSI device no. 52) Similar to all other motors and generators, there is a need for a switching mechanism to connect and disconnect an induction generator from the rest of the electrical circuit under normal and fault conditions. This can be accomplished in one of two ways: either by a set of motor starter contactors (ANSI device no. 42) in series with fuses or by a power circuit breaker. The combination of motor starter contactors and fuses costs less than the power circuit breaker. However, this option requires more maintenance and once a fuse is blown due to a fault it must be replaced before generator operation can be resumed. Furthermore, the contactors have been known to open unexpectedly (because of their control relay dropping out) when their control voltage reaches to about 70 percent of its normal value during voltage transients. Power circuit breakers cost more than the above option; however, they provide the advantage of being able to interrupt the fault current They offer greater flexibility and control functions versus a set of motor starter and fuses and they should be utilized on medium or larger generator applications. 78

PAGE 89

3.2 Instantaneous, Overcurrent and Differential Protection Devices (ANSI device nos. 50 and 87) It is essential to provide instant protection for an induction generator against excessive current, or excessive rate of rise of current normally present in a fault within the generator or its protected circuit. Fuses, if they are used, provide this type of protection very well inherently. However, if power circuit breakers are used, then they must be supplemented with instantaneous overcurrent (ANSI device no. 50) or differential overcurrent relays (ANSI device no. 87). Instantaneous overcurrent relays must be set high enough to prevent the generator breaker from tripping because of the inrush currents during startup. Instantaneous overcurrent relays are not as sensitive as the differential relays and they provide only the minimum amount of protection against short circuits. These relays are adequate for smaller induction generators where the generator neutral leads are not externally available and a differential relay scheme can not be implemented. Differential type relays provide a more accurate and positive way of monitoring fault currents within the protected zone. Their application is strongly recommended for medium and larger size induction generators. The percentage differential setting of these relays should be set to detect a fault from the phase leads to neutral with a coverage of at least 80 percent of the entire length of stator winding. Additional cost is the main drawback in applying differential relays versus instantaneous relays. Additional costs appear in the form of the more expensive relays, the cost of additional wiring required for differential circuit wiring, the cost for doubling the number of current transformers, and the additional manufacturing costs in providing induction generator neutral leads for the differential wiring circuit. 79

PAGE 90

Most induction machine manufacturers charge extra for bringing out and terminating the neutral leads. Instead, the neutral leads of most induction machines are tied together internally. This is because the fault contribution period of induction generators is relatively short and there is no need for grounding the neutrals of these generators. Grounding of the system, if required, can take place at the neutral of the step up transformer associated with the generator. A more cost effective alternative is also available where the instantaneous relays are applied in a self-balancing differential method as shown in Figure 3 .1. In this method the neutral lead as well as the phase lead passes through the same current transformer. In the absence of an internal winding fault the currents in the two leads cancel each other out and the instantaneous relay detects a net zero current passing through the current transformer. The zone of protection for this scheme is less than the percentage relay. It can also be implemented for lower cost and it is sufficiently sensitive enough for use with medium size machines. Ll l2 Generator Windings Three lead machine with lnstantanous overcurrent protection. No dillerential protection 50 External connected neutral r-----------, ' ' iR i Generator! 87 Windings R ' ' ' '-------------..l Six lead machine with Di!!erenlial protection. External connected neutral Six lead machine with sell balancing di!!erenlial protecion. Figure 3.1 Induction generator connections with instantaneous and differential overcurrent protections. 80

PAGE 91

3.3 Thermal and Overload Protection Devices (ANSI device nos. 49 and 51) Thermal and overload protections ensure that the generator will not operate in excess of its thermal ratings for a prolonged period of time, thus preventing the generator from insulation damage and loss of life. The thermal protection can be accomplished in a number of methods. In one method, a series combination of thermally actuated switches and generator starter contactors is provided in the same enclosure. The load current passes through the thermal switches as well as the generator starter contactors. The time versus current characteristics of the bimetalic thermal switches is designed and set to prevent the generator winding from getting too hot before the bimetalic switches open and take the generator out of service. This method does not actually monitor the temperature of the generator winding; instead, it estimates the winding temperature by monitoring the load current magnitude and its duration. In other situations where power circuit breakers are present instead of generator starter contactors, an inverse time overcurrent set of relays can be used to trip the breakers. The inverse time overcurrent relays (ANSI device no. 51) and instantaneous time overcurrent relays (ANSI device no. 50) can be purchased as one unit from most relay manufacturers. Settings of both of these relays have to be coordinated against starting transient conditions of the generator and other downstream relays. Time overcurrent relays do not monitor the actual winding temperature of the generator either. Instead, they make an estimation of winding temperature by monitoring the load current. Time overcurrent relays are most effective when there is a generator overload condition or a persistent short-circuit with limited fault current magnitude present. These relays would not work effectively in case of a solid short-circuit because the decay rate of the generator fault current magnitude is too fast for the time overcurrent relay to detect it. 81

PAGE 92

Still, the most effective means of monitoring and protecting the generator winding against high temperature is by use of resistive temperature detectors (R TDs) or thermocouples. These devices are embedded in the generator stator winding slots and provide direct and accurate readings of winding temperatures from several points to the meters, relays, and/or microprocessor units with which they are interfaced. The relays and the microprocessor units can be set to trip the generator either based on the hottest temperature reading or based on an average reading from the RTDs or thermocouples. It is a fairly standard practice for most generator manufacturers to provide eight RTDs (two per phase plus two more for shaft bearings) for medium and larger size machines. 3.4 Ground Fault Protection Device (ANSI device no. 64, 50G, 51 G) Type and location of the ground fault protection devices depend on generating station configuration and the type of equipment used. As was mentioned earlier, fault current magnitude of an induction generator decays very rapidly and therefore the generator is not considered to support or make a significant contribution to a ground fault. Consequently, neutral leads of most induction generators are not grounded. Nevertheless, remaining electrical equipment should be grounded in order to stabilize and prevent overvoltages on the non-faulted phases of the system during a fault. A grounded Yon the low voltage side of the step-up transformer or a dedicated grounding transformer could serve this function very well. The grounding method (i.e. high impedance, low impedance, or solid) requires some studying and evaluation of the generating plant, the interconnecting utility system's requirements, and the protective relaying scheme employed. In most small and single unit generating stations a high-resistance grounding method is used to limit the fault current to less than five amperes and minimize consequent damage to 82

PAGE 93

generator and related equipment. High resistance grounding takes place by first grounding the system through a dedicated grounding transformer that is connected with grounded Y on the primary side and broken delta on the secondary side. Then a suitable grounding resistor is connected across the broken delta on the secondary side of the transformer. A ground protective relay is usually connected in parallel with the grounding resistor across the broken delta terminals. In case of a ground fault, a current passes through the grounding transformer's primary side which will induce a voltage across its broken delta secondary side terminals. This voltage causes a current to flow through the grounding resistor that will be sensed by the ground relay as a ground fault. 3.5 Rotor Overheating Protection (ANSI device nos. 46 and 47) In a squirrel-cage induction generator the rotor is more robust and less susceptible to damage from overheating than the stator winding. Nevertheless, in earlier sections it was shown that only a few percentage of voltage or load unbalance can cause a large magnitude of negative sequence current to flow in the rotor. This negative sequence current produces a negative sequence (or braking) torque in the rotor which is responsible for rotor overheating. Rotor temperature cannot be readily measured. Instead, where economically feasible, protective relays are installed that would measure the amount of voltage unbalance, current unbalance, and/or negative sequence current flowing in the system. Unbalanced voltage conditions could especially present a problem for induction generators connected to a utility grid. Despite the utility's best efforts to keep the phase voltages balanced, there will always be some amount of voltage unbalance due to daily load fluctuations. As discussed earlier, The National Electrical 83

PAGE 94

Codes does not recommend operating an induction machine with more than five percent of voltage unbalance on the system. 3.6 Bearing Protective Device (ANSI device no. 38) Overheating of the generator bearings can be caused either due to lack of lubricant or excessive mechanical wear. Bearing temperature should be monitored closely--especially in unattended generating stations. Then, if the temperature reaches above a certain setpoint, the generator must be tripped. Protection against high bearing temperature can be performed by either indicating thermometers that are equipped with contacts to trip the generator, or RTDs that are interfaced with auxiliary trip relays. With either method, the important thing is to place the sensors right at the bearing surface and not at some remote location like the lube oil reservoir. 3. 7 Vibration Detection (ANSI device no. 39) High vibration caused by a mechanical imbalance can escalate and cause serious damage to the generator in a short time. The best locations to monitor for vibration are on the generator bearings. Here, vibration detectors can monitor any excessive sign of shaking and quickly shut down the generator. Installation of a vibration monitor is highly recommended--especially in unattended generating stations. 3.8 Mechanical Overspeed Protection (ANSI device no. 12) Induction generators frequently experience overspeed conditions as a consequence of load rejection or disconnection from the utility (especially when a large capacitive bank is connected across their terminals). The monitoring of overspeed condition is crucial in preventing generator damage. 84

PAGE 95

The most common method in overspeed protection is to use a centrifugal speed switch on generator shaft. At a certain speed, the contact from the switch changes state under the centrifugal force of rotor rotation. This change in contact state is transferred from the rotating shaft to the stationary stator through a magnetically coupled slip ring installed on the generator shaft. The contact in turn initiates the shutdown of the generator through a set of control relays. The generator shutdown procedure may include application of brakes to the rotor an/or rapid reduction of mechanical power delivered from the prime mover In some cases a tachometer with a contact having an adjustable setpoint for speed may be used to monitor generator speed. Alternatively, such a contact can also serve to shut down the generator. The tachometer senses rotor speed through a rotating encoder mounted on the generator shaft and a stationary-electro-optic-speed sensor mounted on the stator in close proximity to the encoder. 3.9 Reverse Power Protection (ANSI device no. 32) Under special conditions it may be prudent to install a set of reverse power relays in the generating facility. These relays monitor the direction of real power flow which indicates whether the induction machine is operating as a generator or as a motor. Operating an induction generator as a motor will not cause any damage to the machine. However, the owner may want to minimize the costs associated with operating the induction generator as a motor from the interconnected utility. Another concern pertains to the type of prime mover used to drive the generator. Wind turbines and Pelton wheels are not susceptible to damage by motoring operation ofthe machine. However, in rare occasions where an induction machine and a steam turbine are coupled, heavy damage could be caused to the turbine blades as the generator starts motoring. 85

PAGE 96

As a final thought, another reason for the use of reverse power relays may be due to a requirement imposed by the interconnecting utility company upon the induction generating facility. As highlighted above, the need for reverse power protection is not critical nor necessary in most induction generator applications, except as noted. 3.10 Overvoltage and Overexcitation Protection Installation of surge arrestors and surge capacitors in a generating facility of any size is important. These devices protect electrical equipment, the generator in this case, against the effects of overvoltages resulting from lightning, switching surges, and other disturbances. Without such protections, flashovers and insulation damage to equipment may occur. Surge arrestors limit overvoltages from appearing across equipment insulation by conducting the surge currents directly into the ground while surge capacitors are used with rotating machinery to decrease the steepness of the surge-voltage wave front. To obtain the maximum benefit from these devices, both surge arrestors and surge capacitors must be installed in close proximity of the generator terminals they are intended to protect. Most induction generators are connected to such a large sum of capacitive load bank for power factor correction that if the generator loses its load, the resulting overspeed will cause a serious overexcitation voltage problem across the generator terminals. To minimize overexcitation voltage problems, design provisions should be implemented to disconnect the capacitor bank from service immediately upon loss of generator load. 86

PAGE 97

3.11 Typical Overall Protection Schemes At the beginning, it may not be evident how much protection is adequate for a particular generating plant. The degree of protection required will require careful study of the generating plant and its unique design features as well as a review of the interconnecting utility's power system. Nevertheless, some general examples of protection schemes for induction machines of different sizes are presented here to serve as the starting point for the design engineer. Based on earlier discussions of induction generator protection, a large-size single-unit induction generator plant might be arranged as shown in Figure 3.2. The principle protection for the generator is achieved by using a solid-state multifunction motor protection relay (ANSI device no. 99). The various relay functions included in this protection package are listed to the side of the one-line diagram shown in Figure 3 .2. The generator system bus in this example is grounded through a fully rated high resistance grounding system. As soon as a ground fault is detected by the relay, the generator breaker will be tripped. Detection of a ground fault on the utility line is provided at the neutral of the main step-up transformer. The standard over/under frequency and voltage relays are also shown. In this example the main step-up transformer is protected by a fused disconnect switch. Additional examples of induction generators arrangement, metering and protection requirements from the perspective of the utility companies are shown in Figures 3.3 through 3.6. These Figures are extracted from Public Service Company of Colorado's publication [ 16] titled "Safety, Interface and Interconnection Guidelines for Co-Generators, Small Power Producers and Customer Owned Generators." The utility companies require only that protective equipment that would protect their interest and the interest of their customers. Therefore, protective 87

PAGE 98

equipment specific to the protection of the induction generators (e.g., ANSI device nos. 12, 38, 39, 46, 47, 49) are not shown in Figures 3.3 through 3.6. OG }-------151 Surge Capacitors [> Utility Distribution Circuit Fused Switch Step-up Transformer t $ grounding system P.T. 99 Functions included in this package 49 Stator overload 36 Bearing over temper ature 50 lnstantanous over current 46 Current balance 51 Time over current G fnduction Generator "'0--@ C T Figure 3.2-Typical protection scheme for a large-size, induction generator. 88

PAGE 99

UTILITY CUSTOMER ;:-:---,-.=...-.=... I f +-r"'\ r"'\-+ -----+--r"'\ MOLDED LOCKABLE UTILITY ACCESSIBLE DISCONNECT SWITCH PHASE 27 UNDERVOLTAGE TRIP V 801.; 0.5 SEC. 59 OVERVOLTAGE TRIP V 2 1151.; TIME 0.1 SEC. 81-0 OVERFREOUENCY TRIP F 2 63Hz; TIME 0.5 SEC. 81-U UNDERFREOUENCY TRIP F 57Hz; TIME 0.5 SEC. tt REQUIRED FOR EXPOSED GENERATORS SUCH AS WIND. SA SURGE ARRESTER WH WATT HOUR METER 83 CONTACTOR NOTE: RELAYS DO NOT HAVE TO BE INDIVIDUAL. FUNCTIONS MAY BE IN CORPORATED IN THE INTERFACE PROTECTION PACKAGE. OR INDUSTRIAL GRADE RELAYS PART OF AN INVERTER. Figure 3.3 Typical protection and metering scheme for an induction generator less than 10 KW. 89

PAGE 100

.I UTILITY ----CUSTOMER _____ __,. -) MOLDED CASE BREAKER GENERATOR (ALT) I I STATION ty1 SERVICE unUTY ACCESSIBLE DISCONNECT SWITCH ) MOLDED CASE I;IREAKER (OR FUSE) ( ) [ ) 27 32 47 !11 !IIG !It 111-U 11-G SA WH EB ** 1t IB __ j ) MOLDED .. LOADS QUAHTTTIES F"OR 1PH GENERATORS OUANTIT1ES F"OR 3PH GENERATORS UHDERV'Ot.TAGE TRIP Y ..S: BOX; TIM[ ..S: 0.5 SEC RMRSE POMR PHASE -SEOIJEHCE OR PtiASE -EWMCE VOlTAGE TniE CJiiERCURR[NT, 1 RESIDUAL (GROUND) TIUE CMRC\IRROO OYERYOLTAGE TRIP V .2 115'; TillE ..S: 0.1 SEC UHOERfliEQU[NCY 1R1P F ..S: 57Hz; TIUE ..S: 0.5 SEC OYERFREOUENC'I' TRP r .2 SlHr. TillE ..s: o.s SEC SURGE ARRESml WATT HOUR IIETER SUGGES1m POWER FACTOR CORRC'IION FOR INOUCllOH GENERATORS. P .F .2 REQUIRDIHTS DD'OIO ON CENEI!ATOR CROUNOCHG NfJ STEP UP 'IIWISrORWER REQUIRED FOR EXPOSED GENERATCI!S SUCH AS MHO GNERATORS CONTACTOR 1 PH GOIERATOR IIAXIVUW SIZE 201CW Figure 3.4-Typical protection and metering scheme for an induction generator 10 KW to less than 100 KW. 90

PAGE 101

--LOCKABLE GENERATOR STATION SERVICE UTIUlY ACCESSIBLE DISCONNECT SWITCH ***0 Xn r1 r-1 ________ _J UTILllY GRADE RELAYS I OPTION "S' I I J _j UTILITY ----CUSTOMER 15 SPEED MATCHING 27 UNDV TRIP V TIU ().5 SEC 1/PHAS 32 REVERSE POWER 45 NEG SEQ OR PH CM:RCURRENT 47 I'K'SE-SEOUENCE OR PH.ISE-EWNiCE VOLTAGE 50/51Y VOLTAGE CQNlROUED Till CM:RCURRENT WI1H INSTNflANEOUS. I /I'K'SE SIC RE.'SI)U.tt. (GROUND) TillE CM:RCURRENT 52 CIRCUIT BRAKER L LOAD T TRANSFORMER G GENERATOR 59 OVERVOLTAGE TRIP V ,1 TIUE 0.1 SEC 81/U UNOERFREO. TRIP r 57Hr: TillE 0.5 SEC 81 /0 OVERI'REQ. TRIP r .1 63Hr; TillE 0.5 SEC SA SURC ARRESTER WH WATT HOUR liETER esiJCC(ST[Il POWER FACTOR CORRECTION FOR INilUCTION GENERATORS. PF .1 0.95 * REOUIREIIENTS DEPENO ON GENERATOR GROUND 6: STEP UP XFIIR Figure 3.?Typical protection and metering scheme for an induction generator 100 KW to 1 MW. 91

PAGE 102

15 SI'EED IIATOflHG (110. CEH.KSEE Si:CT. 5.11) 27 UIClEIMli.TI TRI' Y .S 1101; TilE .S 0.5 SEC. 1/IWSE 32 RE'ttRSE POllOI 41 NEG sm. OR PHASE lllolNICE CMIICUiliiNf 47 PHASE-SEQUENC[ OR FHASE-IIIolNICE VOLTI S0/51 liST. NfJ 1111[ CMRCUIRENf. 1/IWSE S0/51G liST. N> liiiE R[SIOUAI. GIIOIJID a.tRCURIIEliT S0/51N liST. NfJ TilE NEUlRAI. GRDUCI CMRCURRENT SO/SlY Vll.T.IGE camKlJ.) Til CMRCU!IDf 111H IISTANTAHfOUS. 1/IWS[ 52 CIRCIIT IIIAI
PAGE 103

3.12 Utility Interface Protection Requirements Prior to the design of electrical interconnecti0n and protection systems between the small or co-generating facility and the utility, the designer must review and understand the specific requirements of the host utility in detail. It important to establish an open and effective communication link between the parties involved in order to minimize setbacks and future design modifications. There are several requirements that seem to be unified among various utilities worth mentioning here: The co-generating facility is required to furnish details on protection system design, construction, and testing to the host utility for review and approval according to a scheduled time frame. The co-generating facility is required to install a lockable disconnect switch at the intertie accessible to the utility's maintenance personnel. The co-generating facility must be able to detect and clear a fault on the utility system. The co-generating facility is not allowed to remain connected to the utility system in a stand-alone (islanded) situation when power from utility's own generators is interrupted. The co-generating facility is not allowed to energize a deenergized utility line. The co-generating facility is not allowed to manually synchronize with the utility line. Any costs involved in making additions or modifications to the utility's equipment in meeting the established requirements will be covered by the owner of co-generating facility. 93

PAGE 104

Among all the requirements listed above, the single most important one is furnishing protection system design details to the utility for review and approval. The co-generating facility's design staff may benefit from this review and approval process in two ways. First, it allows veteran utility engineers perform a peer review of co-generating facility's design and offer suggestions in improving it. Second, it entitles the co-generating facility's design staff to review, discuss and familiarize themselves with the utility system's design features--especially in the following areas: Utility's primary and backup protection systems practices Automatic reclosing practices Utility's customer load profile as it may affect the co-generating plant Utility's surge protection practices on the transmission line Details on transmission line types and ratings in the vicinity of co generating plant and their maintenance and outage rates Utility companies in the United States are required by law to purchase any excess power produced by co-generation plants and small independent power producers. Utilities must purchase this power at a rate so called "avoided" cost. This means that the rates paid to small power producers must reflect the energy and cost savings utilities realize by being able to avoid capacity additions of their own. This requirement was put in effect by the Public Utility Regulatory Practice Act (PURP A) in 1978 to help ease the energy-shortage crisis at the time. Since then, most of the utility companies have prepared standard design documents outlining the requirements of integrating small power generators into the utility's power grid. There are also a number of informative IEEE papers published on this subject [17] that discuss the basis for such utility requirements and provide practical solutions in meeting them. 94

PAGE 105

The majority of the utility requirements are related to the protection system. Some of the utility's protection system requirements may appear to be complicated, expensive, and over designed to the small power producer. Nevertheless, most of utility's elaborate requirements have technical justifications that are based on many years of designing, maintaining and operating power production facilities. The utilities need to maintain a safe and reliable power source for their customers. Therefore, for a smooth and reliable integration of a small generating facility into the utility, the equipment and design criteria used in the small generating facility must meet the same quality and standards set forth by the utility. The intention is to form a single seamless power system under utility's ordinance. In the following sections some of the utility requirements will be reviewed in more detail followed by recommendations in how best to fulfill these requirements. 3.12.1 Lockable Disconnect Switch This requirement by the utility company is warranted and its basis stems from a safety standpoint. Before any maintenance work can begin on a segment of a transmission or distribution line, the maintenance crew must ensure that the power line is deenergized and the visible disconnecting devices at the ends of that power line are padlocked in the open position. The crew proceeds by grounding one end of the power line before maintenance work can begin. The co-generating facility can install such a disconnect switch between the high voltage terminals of the step-up transformer and the utility power line. 3.12.2 Detection and Clearing a Fault on the Utility System There may be instances when the co-generating facility gets involved in the fault detection and clearing processes for a fault located on the utility system side. 95

PAGE 106

Whether the protective relays in the co-generating facility act as primary or backup relays will depend on location of the fault. A fault may be located within a zone where it is bounded by at least one branch circuit terminal equipped with protective relays and breakers owned by the co generating facility. In this case, the co-generating facility provides primary protective function, paralleling those provided by the utility, in clearing the fault. However, if a fault is within a zone bounded on all its terminals by utility-owned fault-detecting relays and breakers, then any fault clearing that may be required by the co-generating facility will be to back up the utility equipment. 3.12.2.1 Primary Protective Function In situations where the co-generating facility relays are required to provide primary protective functions in conjunction with utility's protective relays, the protective relays furnished by the co-generating facility must be totally compatible with those relays used by the utility for each protective zone. For example, if the relays are protecting a zone which is at distribution or subtransmission voltage level, then overcurrent relays or directional overcurrent relays will be the common choice. However, if the relays are protecting a zone which is at transmission voltage level, then distance measuring relays with pilot scheme will be frequently used. Again, remember that the co-generating facility must follow through with whatever precedent the utility has sets forth with protective relay types and schemes on its system. 3.12.2.2 Remote Backup Protection The protective functions the co-generating facility provides in backup protection of a fault on the utility system are not as complicated as those provided for 96

PAGE 107

primary protection. In most cases, the co-generating facility is required to provide what is referred to as remote backup protection. The relays at the co-generating facility have the responsibility of detecting a fault on the utility system and taking action (i.e., tripping the tie breaker or shutting down the generator) only if the utility's own protective relays fail to clear the fault within a predetermined time period. In comparison, the function of these remote backup protection relays is very similar to that of the relays used in the co-generating facility in detecting a fault within its own plant boundary except in this case the relays are set to detect a fault on the utility system's side. As a minimum, remote backup protection from the co-generating facility should provide enough coverage to detect a fault in the next immediate protective zone totally owned by the utility. The utility provides backup protection for the remaining zones of its own system. Two types of remote backup protection are commonly required by the utility. They are phase backup protection and ground backup protection. For phase backup protection, a time overcurrent relay (ANSI device no. 51) with either voltage restraint or voltage controlled feature is commonly used. The appropriate relay choice for ground backup protection depends on the generator step up transformer's high voltage winding connection. If this winding is connected in grounded wye arrangement, then a ground fault on the utility system can be readily detected by monitoring the current flowing in the neutral of this winding. Installing an overcurrent relay (designated as 51 G) in the neutral of this winding will take care of ground backup protection (refer to Figure 3.2). However, if the high voltage winding of transformer is connected in delta arrangement, then it will be necessary to install additional potential transformers (PTs) on the high voltage side of the generator step up transformer. These potential transformers will be connected in a 97

PAGE 108

broken delta arrangement. Ground fault condition can be detected by connecting an overvoltage relay with time delay across the terminals of this open delta. The voltage relay measures the zero sequence voltage across the open delta terminals in this arrangement (this connection measures the vector sum of the three phase-to-neutral voltage phasors). Under normal system conditions this voltage is zero, while a ground fault will impose a measurable voltage across the relay. Co-generator +
PAGE 109

generators spin off to form their own little electrical network independent of the rest of the system. Voltage, frequency and power output of the islanded (or isolated) generators adjust themselves to match the remainder of electrical load still connected to them. Islanding is normally initiated by a transient disturbance on the electrical network. In case of the co-generating facility, islanding designates the condition when power from the utility generators is interrupted and the only generators carrying the system load are those belonging to the co-generating facility. Because ofboth legal and technical reasons, islanding operation is not desirable to the utility. Likewise, the co-generating facility should also_avoid operation in islanded mode because most frequently the electrical load reaming on the system is much greater than the output capacity of co-generating facility. To detect an islanding condition, one can assume that the load remaining on the system is either greatly more than or less than the capability of the islanded generators. Under these conditions, the generator's voltage and frequency would either drop or rise respectively. The islanding condition can then be detected by installing both over/under frequency and over/under voltage relays in the co generating facility. However, in less frequent cases, it may be possible that the load remaining on the system is either exactly the same or slightly off from the generator's output rating. Under these conditions, the generator's voltage and frequency would change none at all or slightly respectively. Consequently, the over/under frequency and over/under voltage relays become ineffective under these circumstances. One other option that remains open for the co-generating facility would be to monitor the auxiliary contacts ofthe circuit breakers on the utility system. Opening of certain circuit breakers could indicate an islanding condition. Then from the utility a signal could be sent back to 99

PAGE 110

the co-generating facility initiating a shutdown. The only drawback with this strategy is the complexity of monitoring different combinations of breakers' contacts for islanding condition. 3.12.4 Energizing a Dead Circuit A utility's main concern in not allowing a co-generating facility to energize a dead circuit is safety. A weather or accident related condition may cause a power line to fall down to the ground.where it would become within the reach of the general public. Energizing such a power line without the utility's knowledge would pose great danger to people. It is the responsibility of the utility company to ensure that the power line is safe and repaired before energizing it. Furthermore, if the co-generating facility takes initiative to reenergize a dead circuit, then the utility may encounter problems in synchronizing its own power lines to a live circuit already energized by the co-generating facility. Because of these reasons and more, the co-generating facility must be designed so its circuit breaker cannot close when the utility line is deenergized. One simple method of accomplishing this is to check for a live utility line as a prerequisite of synchronization and breaker closure. 3.12.5 No Manual Synchronization Synchronization is the process of matching the voltage, frequency, and phase sequence (i.e., abc versus acb) of one piece of electrical equipment (for example: a generator, a bus, or a transmission line) to another piece of electrical equipment before closing a tie-breaker to electrically connect them together. If the voltage and frequency of the electrical equipment are not properly matched, upon closing the 100

PAGE 111

circuit breaker, a power transient or a surge occurs in the power system referred to as a bump. During power transients, system voltage and frequency fluctuate up and down This fluctuation may present itself in the form of lights flickering to the residential consumers or as an out of order operation of relays or control circuits in the industrial plants Either way, improper synchronization must be avoided to minimize power transients. In addition, improper synchronization causes transient torque and fatigue to the turbine generator shaft which ultimately results in loss of useful life of the machine. Smaller induction generators normally don't require synchronization. However, a larger induction generator may be required to be brought up to speed using the prime mover first. Once the induction generator is brought up to speed, auxiliary capacitors may be used to build up voltage across its terminals before synchronizing it with the rest of the power system. Synchronization of a larger induction generator in this manner significantly reduces the current transients otherwise present when an induction generator with rated speed and no terminal voltage is connected to a live utility line. Customarily, the utility forbids a co-generating facility from manually synchronizing a generator to the utility power system. With this requirement, a careless manual synchronization from an inexperienced operator can be avoided. Preferably when required, every co-generating facility should be equipped with an auto-synchronization relay. This relay will automatically raise or lower turbine/generator's voltage and frequency (speed) until they closely match the system's voltage and frequency before the relay sends a signal to close the interconnecting breaker. If an auto-synchronization relay cannot be economically justified, then as a minimum a protective relay must be installed to supervise manual synchronization by an operator. The protective relay will provide a control 101

PAGE 112

permissive for the operator to be able to manually close the breaker only if voltages and frequencies are closely matched. 3.12.6 Modification Costs to the Utility In PURPA legislation, just as it orders the electric utility company to buy power from the co-generating facility, it also delegates the financial responsibility to the co-generating facility for all costs involved in modifications of utility owned electrical systems to accommodate the co-generating facility. The reason for this requirement is not so obvious at first. The rate that the utility companies charge to their electrical customers is based on a fixed rate of return on the utility's capital investment. Since the modifications that a utility company makes for accommodating the co-generating company do not benefit the utility's electrical customers (and only benefit the co-generating facility's owner), the utility company can not include these additional costs as part of its capital investment for rate increases in the future. Therefore, the utility company is unable to get reimbursed for this expenditure and must pass on the expenses to the co-generating facility. As demonstrated in this chapter most of the requirements imposed on the co generating facility by the utility have good technical justifications. It is not easy and straightforward to determine the type and amount of protection required for a co generating facility. The decision is made based on thorough understanding of the unit's operation, good engineering judgments, and experience. It is also very important for the co-generating facility to establish open and effective communication with the utility right from the beginning of the project. The co-generating facility is responsible for determining how the addition of an induction generator might impact the utility system and so obtain prior agreement on the interface requirements from the utility. 102

PAGE 113

CONCLUSIONS The demand for electrical energy continues to increase year after year. Environmental, regulatory, and financial issues have made it more difficult to build large power plants. Most suitable sites for building large power plants with hydro, wind and geothermo energy resources have already been developed and it is unlikely that new nuclearor fossil-fueled power plants will be built in this country in the near future. Induction generators are particularly suitable for micro( <1 00 KW) and even mini-hydroelectric (> 100 KW and <5000 KW) power applications as a supplement to existing power generation. Ideal plant locations are those where sufficient reactive power is available from the power grid to meet the excitation requirements of the induction generators. Additionally, the power grid must be strong enough from the transient point of view where induction generators are interconnected to the utility system. The renewable nature of water as an energy resource and the non-polluting feature of hydroelectric power plants provide sufficient justifications to venture upon and utilize this form of power generation wherever possible. Additionally, relative simplicity, reliability, and low cost of induction generators compared to synchronous generators do provide further incentives to consider building smaller hydroelectric power plants; when otherwise, these smaller powerplants would have been considered too cost prohibitive if they were built using conventional synchronous generators. Finally, any kind of technical problems associated with induction generators and hydroelectric plants are worth investigating and resolving since over the long run the conventional forms of fuel become more expensive and not as readily available. 103

PAGE 114

With the advent of electronic commerce in the Internet, locating and purchasing new or used reversible pumps/turbines, induction motors/generators, breakers, and other industrial electrical equipment from surplus dealers and government auctions are made simple. There is considerable cost savings involved in this approach compared to paying retail. Outstanding quantity and selection of reversible pumps and induction machines compared to turbines and synchronous machines are apparent. It seems practical that if a small group of entrepreneurial engineers decide to design and build a small and economic hydroelectric power station, the Internet should provide them with ample choices and cost savings. Development of such power plants in other less developed and less regulated countries, where plenty of untapped hydro resources still exist, has even greater impact than in this country. Since such simple, affordable, and low maintenance form of power generation would make a significant impact in the local economy, productivity and social activity--especially in remote areas where rural electrification is still limited. The various topics on induction generators and hydroelectric power generation covered in this thesis have helped me increase my understanding of and interest in the subject matter through academic reading and exploration. I hope that by sharing what I have learned throughout this process in this thesis, I can inform and raise interest in others who are looking for an economic and viable alternative form of power generation. 104

PAGE 115

Appendix-A Constants C and D for equations (2.21) and (2.22) defined. Assumptions: X5 = Xr =XI c1 = -2 x1 R1 c 2 = -Xt2 RL Cs = Xsh RL-CXse-Xsh) (Rs+Rr) c6 =(XI Xsh+Rs Rr) RL-(Xse-Xsh) (Rs+Rr) xl C7 = -V Xsh RL + V CXse + X5h)R5 c8 =xI c 7 c9 = o.o DI = -2 (Xse-Xsh) Xl+(Rs+Rr) RL D2 = (Rs+Rr) xl RL-(Xse-Xsh) XI 2 D3 = 2 VCXse-Xsh) XI-VRL Rs D4 = V(X5e-X5h) XI-VXI RL Rs D6 =(XI XseRr Rd Xsh+(Xse-Xsh) Rr Rs D7 = -VXse Xsh 105

PAGE 116

REFERENCES 1 James D. Bailey, "Factors Influencing the Protection of Small-to-Medium size Induction Generators," IEEE Trans. on Industry Applications, Vol. 24, No.5, Sept./Oct. 1988, pp. 955-964. 2 Richard L. Nailen, "Watts from Waste Heat--Induction Generators for the Process Industries," IEEE Trans. on Industry Applications, Vol. IA-19, No.3, May/June 1983, pp. 470-475. 3 A. E. Fitzgerald, Charles Kingsley Jr., and Stephen D. Umans, Electric Machinery, Fourth Edition, New York: McGraw-Hill, 1983. 4 H. C. Stanley, "An Analysis ofthe Induction Machine," AlEE Trans., Vol. 57, 1938, pp. 751-757. 5 D. G. Fink, and H. W. Beaty, Standard Handbookfor Electrical Engineers, Twelfth Edition, New York: McGraw-Hill, 1987. 6 F. Buse, ''Using Centrifugal Pumps as Hydraulic Turbines," Chemical Engineering, Vol. 88, Jan. 26,1981, pp. 113-117. 7 R. Bonret, and G. Hoops, "Stand Alone Induction Generator with Terminal Impedance Controller and No Turbine Controls," IEEE Trans. on Energy Conversion, Vol. 5, No.1, March 1990, pp. 28-31. 8 J. L. Woodward, and J. T. Boys, "Electronic Load Governor for Small Hydro Plants," Water Power and Dam Construction, Vol. 32, July 1980, pp. 37-39. 9 G. D. Hoops, "Terminal Impedance Control of a Capacitor Excited Induction Generator," Ph.D. Thesis, University ofToronto, 1988. 10 J. L. Bhattacharya, and J. L. Woodward, "Excitation Balancing of a Self Excited Induction Generator for Maximum Output." lEE Proceeding, Vol. 135, Pt. C, No.2, March 1988, pp. 88-97. 106

PAGE 117

11 L. L. Gleason, and W. A. Elmore, "Protection of ThreePhase Motors Against Single Phase Operation," AlEE Trans. Part III, Vol. 77, Dec. 1958, pp. 11121119. 12 Williams J.E., "Operation of Three-Phase Induction Motors on Unbalanced Voltages," AlEE Tran. Pt. III-A, PAS, Vol. 73, 1954, PP. 125-133. 13 L. Shridhar, et al., "Selection of Capacitors for the Self Regulated Short Shunt Self Excited Induction Generator," IEEE Trans. on Energy Conversion, Vol. 10, No. 1, March 1995, pp. 10-17. 14 L. Shridhar, et al., "Transient Performance of the Self Regulated Short Shunt Self Excited Induction Generator," IEEE Trans. on Energy Conversion, Vol. 10, No 2, June 1995, pp. 261-267. 15 F. P. De Mello, et al., "Application oflnduction Generators in Power Systems," IEEE Trans. on Power Apparatus and Systems, Vol. PAS-101, No.9, Sept. 1982,pp.3385-3393. 16 Safety, Interface and Interconnection Guidelines for Co generators, Small Power Producers and Customer-Owned Generators, Public Service Company of Colorado, April, 1997. 17 Louie J. Powell, "An Industrial View of Utility Cogeneration Protection Requirements," IEEE Trans. on Industry Applications, Vol. 24, No. 1, Jan./Feb. 1988, pp. 7 5-81. 107