Paper title:

“A Setting-Free Approach to Detecting Loss of Excitation in

Synchronous Generators”

Prepared by :
Ghaith Mohammad Alqudah

Presented to :
Dr. Ayman Alquran
 Outline:

 Introduction

 Excitation system
 Loss of excitation
 Loss of excitation protection schame
 Proposed method
 Simulation
 Introduction
 Generators are machines that produce electric energy in the form of voltage and
current. They produce electricity based on the principle of electromagnetic induction.

 There are two fundamental types of generators known as AC (alternating current) and
DC (direct current) generators.

 AC machines can be further classified as Induction machines and Synchronous machines.

And hence, AC generators as  synchronous generators  (commonly referred as
alternators) and  Induction generators  (or asynchronous generators).
 Protection of synchronous generators

involves the consideration of faults and harmful abnormal operating conditions, more so
than the protection of other power system elements. The range of abnormal conditions and
faults that can occur is extensive, and many could result in serious damage to the
alternator and associated plant. In a properly protected generator, automatic protection
against harmful abnormal conditions is required.

 Unlike most other areas of a power system where faults can be cleared by opening all
electrical sources to the faulted element, a synchronous generator is a system with
mechanical (prime mover, inertia), DC electrical (field), and AC electrical (power
system) sources of energy. An understanding of the interaction of these various energy
sources is necessary to properly protect a synchronous generator.
 Alternators need to be protected not only from short circuits but also from abnormal
operating conditions such as overexcitation, overvoltage, underexcitation, loss of field,
unbalanced currents, loss of prime mover and abnormal frequency conditions. When
subjected to these abnormal conditions, damage or complete failure can occur within
seconds thus requiring automatic detection and isolation.

 Generator operation can be affected by both faults within the machine and disturbances
occurring in the power system to which it is connected. A generator protection system,
therefore, has a dual objective: to protect the machine and protect the power system.

 Choice of protection portfolio

A wide choice of protection and measuring functions is available, and the protection portfolio will
usually depend on the size and the function of the generator being protected. Basing generator
protection on machine size is difficult because the desired protection may be determined more by
the importance of the generator to the power system than by the size of the generator. This
figures give examples of protection for different size generators.
Small generator protection Medium generator protection
 Generator protection schemes

Protection schemes are designed to guard against two classes of condition:

• Fault conditions, either within the generator or external to the generator. Faults can
cause serious damage to the generator and generally result in shutdown of the
installation.

• Abnormal conditions, which are caused by conditions external to the generator, but
which can seriously affect the operation of the generator and also result in damage.
 Abnormal operating conditions protection
1. Abnormal frequency protection (over or under frequency).
2. Over excitation and overvoltage protection.
3. Field loss (excitation loss) and under- excitation protection.
4. Unbalanced stator loading protection.
5. Reverse power or motoring (loss of prime mover ).
6. Over speed protection.

 Protection against fault conditions

1. Stator phase to phase fault protection.

2. Stator earth fault protection.

3. Field fault protection.

4. Generator breaker failure.

5. Stator overheating (overload) protection.

 Excitation System
 Definition: The system which is used for providing the necessary field current to the rotor
winding of the synchronous machine, such type of system is called an excitation system.

 In other words, excitation system is defined as the system which is used for the production of
the flux by passing current in the field winding. The main requirement of an excitation system
is reliability under all conditions of service, a simplicity of control, ease of maintenance,
stability and fast transient response.

 The amount of excitation required depends on the load current, load power factor and speed
of the machine. The more excitation is needed in the system when the load current is large,
the speed is less, and the power factor of the system becomes lagging.

 Types of Excitation System

1. DC Excitation System
2. Static Excitation System
3. AC Excitation System
 Rotor Excitation System.
 Brushless Excitation System.
 DC Excitation

AC Excitation
 Loss of Excitation

 loss of excitation occurs at these condition:

1. Field Open Circuit (field current is zero).

2. Field Short Circuit (field current is too high).

3. AVR Control Failure.

4. Accidental Tripping of Field Breaker.

5. Loss of Supply to the Main Exciter.

6. Field Circuit Breaker Latch Failure.

7. Slip Ring Flash Over.

 The effect of generator loss of excitation on the system:

1. Low-excitation and magnetic-loss generators absorb reactive power from the system,
causing the voltage of the power system to decrease.

2. When a generator is demagnetized, due to the voltage drop, other generators in the
power system will increase their reactive power output.
Mechanical
input

Pe
generator Grid
jQe

Speed= Ns

Field current = If Voltage

  =

Mechanical
input
Pe
generator Grid
jQe

Speed > Ns

Field current = zero Voltage =

 
Active and reactive power after LOE Voltage drop and rotor acceleration
 Loss Of Excitation – Protection Schemes
 The simplest method by which loss of excitation can be detected is to monitor field
current of the generator.

 If the field current falls below a threshold, a loss of field signal can be raised.

 A complicating factor in this protection is the slip frequency current induced in the
event of loss of excitation and running as an induction generator.

 The quantity which changes most when a generator loses field excitation is the
impedance measured at the stator terminal.

 On loss of excitation, the terminal voltage begins to decrease and the current begins to
increase, resulting in decrease of impedance.

 The loss of excitation can be clearly detected by a Mho relay located at the generator
terminals.
 Loss Of Excitation – Protection Schemes
 In 1949, a single phase “offset Mho Relay” was introduced for the high speed
detection of loss of excitation in synchronous generators.

 There are five LOE protection schemes used today:

1) R-X scheme with single and double relay scheme ( based on generator terminal
impedance measurement).

2) R-X with directional element scheme.

3) G-B scheme ( based on generator terminal admittance measurement).

4) P-Q scheme ( based on generator active and reactive power output).

5) U-I scheme ( based on the measurement of phase angle difference between

phase voltage and current).

 However, R-X schemes is widely used in power systems.

 Impedance Measurement (Single
Element)
 The diameter of the circle set equal to the synchronous reactance (Xd) and
offset will be set equal to one – half of the transient reactance (X’d/2).This
circle is operation zone for LOE relay.

 As viewed from the machine terminals the relay will operate for any
impedance phasor that terminates inside the circular characteristic.

 In normal operation condition, the generator generates active and reactive

power to the system which means both R and X are positive and the terminal
impedance is located in the first quadrant in R-X plane.
 When the excitation is lost, the generator starts to draws reactive power from the system and X
becomes negative from the LOE relay point of view.

 As a result, the terminal impedance loci in R-X plane moves to the forth quadrant and the endpoint
of terminal impedance ranges between the sub transient reactance and synchronous direst axis
reactance.

 When the measured impedance falls into the operating region, the relay function will be picked up
and after a certain time delay to enhance the security for power swing. A trip signal will be sent to
the generator breaker.

Normal condition

Trip
region
 Impedance Measurement (Double
Element)
 This protection scheme applies two offset Mho impedance circles by using the generator
terminal side voltages and stator currents as input signals.

 The offset-mho relay in the impedance plane has two circles with a diameter of direst
axis transient reactance X’d and a negative offset of X’d/2 for the outer circle.

 And the diameter of ‘1.0’ p.u and a negative offset of X’d/2 for the inner circle.

 Zones 1 and 2 are for detecting LOE with full load and light load. The typical time
delays for zone –1 and zone – 2 are about 0.1s and 0.5-0.6s.
 PROPOSED METHOD

 To demonstrate an LOE condition, assume a synchronous

generator connected through a transformer to the main grid Fig. 1.

 the equivalent circuit of system after the LOE occurrence

can be modeled per phase as shown in Fig. 2.

 When the LOE occurs, the magnitude of generator internal

voltage (EG) will decrease. This causes variations in voltages
and currents at the generator terminal resulting in changes in
impedance measured by the relay.
 Assuming
   , the impedance can be expressed as

Where

=++

 By splitting the real and imaginary parts of (5), the resistance and reactance measured
by relay 40 are

  the real part, R(t), of impedance measured by relay 40 will be utilized to develop a

method for LOE detection. When the LOE occurs, k starts increasing while δ can be
safely assumed as a constant since the generator still keeps its synchronism.

 Now, taking derivative of real part in terms of time will then result in

= . .

+ .
 Since k increases, is positive. Therefore, as long as k is less than 1, the first term of
above equation will be positive This condition is not threatening the generator itself
and the power system to which it is connected since it has not yet turned to a “sink” of
reactive power.

  During this period it is not possible to judge the sign of the second term since the

variation of XTot.(t) is non-monotonic and can be either positive or negative.

 after a few seconds when XG(t) settles to its final value, the second term will become
zero. At this time, k has increased far more than 1, that is (1−k 2 ) becomes negative
while still remains positive.

 after a period of time, the first term will become and remain negative and so will . As a
result, the relay will observe the rate of change of R(t) and if becomes and remains
negative for a while, LOE can be detected.

 During a power swing, similar condition can occur which means that becomes and
remains negative for a while. To distinguish the LOE from the power swing, it is then
necessary to determine for how long will remain negative during power swing at the
worst case.
  

 Under the power swing conditions, on the contrary, δ will change and k can be safely
assumed constant. Moreover, XG is very close to Xd' and can be considered as a
constant and so XTot. Therefore, during the power swing will be:

= . = .

 Therefore, during the slowest power

swing will not remain negative for
more than 1.67 seconds. To distinguish
an LOE from the power swings.
 Logic of the Proposed Method

  This figure shows the proposed method in a logical pattern. At
first, the relay calculates . Once is non-zero, it will check
the sign of .
 If becomes and remains negative for more than 1.7 seconds,
an LOE can be detected and a trip signal is sent to the
generator circuit breaker (GCB).
 If remains negative for less than 1.7 seconds and changes to
positive, a power swing or an out-of-step condition can be
identified.
 It can be inferred that the proposed method acts as a setting-
free method since there is no need to set any threshold for
the rate of change of R and only its sign is required. Moreover,
the generator and system parameters are not required to be
known.
 SIMULATION EXPERIMENTS
 the performance of the proposed method will be evaluated. To demonstrate the
superiority of the proposed method over the conventional ones, its performance will be
compared to one of the most famous conventional method.

 To do this a power plant with all lines connected to it were simulated. Figure shows the
single line diagram of the power plant. In the simulated system, there are two types of
generators including gas and steam turbine generators. G1 to G4 are steam turbine
generators and G5 to G12 are gas turbine generators. Moreover, the steam units are
connected to 400-kV voltage level while the gas units are connected to 230-kV voltage
level. The proposed method will be evaluated for the LOE in both types of generators.

 We will do Evaluating the Performance for Proposed Method into two scenarios the first
one is completely LOE and the other partially LOE.

 The performance of the proposed method will be investigated for two load conditions, off-
peak and peak loads because there is a measurable difference between LOEs occur-ring in
off-peak and peak loads.
 Evaluating for off-peak load

  we assume steam turbine generator G1 loses its excitation at t=0.2 seconds in the simulated

system. Fig. 1(a) shows measured by relay 40 of the generator after an LOE. As shown,
following some oscillations, becomes and remains negative. After the last positive cycle,
the relay starts counting and if remains negative for more than 1.7 seconds, a trip signal
will be sent to the genera-tor circuit breaker. Fig. 1(b) shows a trip signal beginning at
t=2.69 seconds.
 Fig. 2(a) shows the impedance locus with respect to the conventional LOE
characteristic. In this method, it is a common practice to present all values including
impedance locus and setting parameters in per-unit (p.u) with the base impedance of
the generator. The impedance locus enters into the characteristic at t=3.56 seconds
and after a time delay which is usually set to 0.5 it will issue the trip signal at t=4.06
seconds as shown in Fig. 2(b).
 Evaluating for peak load

  The similar scenarios are studied for peak load. At t=0.2 seconds, generator G1 loses its

excitation. Fig. 1(a) shows the variations of showing that after a few seconds it be-comes and
remains negative. Fig. 2(b) shows the trip signal issued by the proposed method at t=3.29
seconds.
 Fig. 2(a) shows the impedance locus with respect to the conventional LOE characteristic
for this scenario. The impedance locus enters into the characteristic at t=5.11 seconds
and after a time delay of 0.5 seconds, it will issue the trip signal at t=5.61 seconds as
shown in Fig. 2(b).
 Evaluating the Reliability of Proposed
Method
 It is necessary to scrutinize how the proposed method respond to other disturbances
threatening the reliability of LOE relays. One of the most possible disturbances is the power
swing which has caused mal-operation to the conventional method in some cases. In this
section, the performance of the proposed method will be investigated under power swing
conditions. Numerous power swing scenarios have been studied, but only two of them are
presented here.

 In the first scenario of power swing, a three-phase fault occurs on one of the circuits of line
L7 at t=0.2 seconds. The fault will be cleared by the action of the line relays and opening
the corresponding circuit breakers at t=0.6 seconds. It is assumed that the relays have
detected the fault after 0.4 seconds from occurrence. In the second scenario, a three-
phase fault occurs on line L9 at t=0.2 seconds and will be cleared by the action of the line
relays and opening the corresponding circuit breakers at t=0.3 seconds.

  the second power swing scenario Similar to the first scenario, will not remain negative for

more than 1.7 seconds during the second scenario from viewpoint of both relays. These
results clearly indicate the reliability of the proposed method during other disturbances.
 Figure shows the variations of measured by relay 40 of generator G1 during the first
power swing scenario. As can be seen, during the power swing, there is no time interval
more than 1.7 seconds in which remains negative. Therefore, no trip signal will be issued
by the proposed technique.
  
 Figure shows the variations of measure by relay 40 of generator G5 for the first power
swing scenario. Similar to G1, there is no time interval more than 1.7 seconds in which
remains negative. Therefore, proposed relay 40 of G5 will also not act to this scenario.
 CONCLUSION

 The main advantage of the proposed method is that it is setting-free which means that
there is no need to know about generator and system parameters.

 It acts faster than the conventional methods and thus can isolate the generator
suffering from an LOE with less time delay.

 The proposed technique is reliable and does not respond to the power swings and out-
of-step conditions, which are threatening the operation of the conventional methods.