Matching Generators to Power

Systems

Allen Windhorn <[email protected]>

24 October 2019
Contents:
1. Ratings and General Considerations
2. Generator Construction
3. Reactances and Fault Currents/Decrement Curves
4. Grid Codes and Effect on Generators
5. Synchronization and Paralleling
6. Reference Frame Theory (Short Version)
7. Excitation System Models and Exciter Response
8. Synchronous Condensers (Compensators)

1
1. Ratings and General Considerations

2
Frequency, Voltage, Current, & Power
• Rated frequency: determined by locale or
application (usually 50 or 60 Hz but 400 Hz
for aircraft ground power, other special)
– Frequency variation: determined by
governor and engine characteristics
– Frequency is determined by RPM and
poles:
𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 × 𝑅𝑅𝑅𝑅𝑅𝑅
• 𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹 = 120
𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹 ×120
• RPM = 𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃
– 120: 60 seconds/minute, 2 pole pairs/cycle
3
Frequency, Voltage, Current, & Power
• Rated voltage (=>flux density) affected by:
– Saturation flux density of steel, geometry
– Frequency
– Number of stator coils (i.e. number of slots)
– Number of turns in each coil
– Pitch of the coils (i.e. number of slots span)
– Stator parallel connections and hookup
– Length and diameter of the stator
lamination stack
𝑑𝑑∅
• 𝑉𝑉 = 𝑁𝑁𝑆𝑆 , 𝑁𝑁𝑆𝑆 is total series turns, ∅ is flux
𝑑𝑑𝑑𝑑
4
Frequency, Voltage, Current, & Power
• Most of the adjustable parameters (e.g. turns,
parallels, slots) have discrete values (we
can’t design to an exact voltage, so there are
tradeoffs)
• Others may be determined by specifications,
physics, or economic considerations (e.g.
pitch, steel characteristics)

5
Frequency, Voltage, Current, & Power
• Voltage constraints:
– Poles depend on engine speed, frequency
– Slots limited by phases, parallels, balance
– Turns usually integer (maybe half turn)
– Pitch limited by harmonics, reactances
– Stack length limited by available frames
– Ability to fit insulation into a reasonable
size coil limits maximum voltage to around
15-20 kV
– Ability to fit in reasonable size busbars
limits maximum current to around 5000A
6
Frequency, Voltage, Current, & Power
• Voltage tolerance: depends on voltage
regulator, generally ±1% or better
• Voltage adjustment range: usually ±5% but
±10% possible, required for some grid codes
– Machine designs may be “broad range”, for
example a generator may be sold at the
same kVA rating for 12.5-13.8 kV, with
±5% adjustment
– Done by making the machine somewhat
oversize and relying on higher quantity to
keep costs down
7
Frequency, Voltage, Current, & Power

From ISO-9528-3 2005 (no NEMA equivalent)

8
Frequency, Voltage, Current, & Power
Q: Can I run a generator rated for xxx volts at
yyy volts?
A: Maybe. If you are going down in voltage,
you will need to derate power by the voltage
ratio. The reactance in % will remain the same
at the new voltage and power rating. Voltage
overshoot will probably increase some.
If you are going up in voltage, you need to
consider if the insulation is adequate, and
unless you increase frequency proportionally,
saturation will increase, leading to higher core
loss and rotor I2R loss. Best to consult vendor.
9
Frequency, Voltage, Current, & Power
Q: Can I run a generator rated for xxx volts at
yyy volts?
A: (cont’d) Some generators are reconnectable.
Changing from parallel to series windings
doubles the voltage (if the insulation is
designed for it) and vice-versa, changing from
wye to delta reduces the voltage to 58%. Also
depends on whether designed “broad range”
You also need to consider what the inspector
will say, if the nameplate doesn’t match the way
you are using it. Some manufacturers may be
able to supply a replacement nameplate.
10
Frequency, Voltage, Current, & Power
Q: Can I run a generator rated for 50 Hz at 60
Hz (or vice-versa)?
A: Maybe again. Changing the frequency also
changes the saturation voltage by the same
proportion, so a 3300V 50 Hz machine can be
used at 4160V 60 Hz, if the insulation is rated
for it, but you also need to ask if the unit can
mechanically run at the higher speed. The
bearings may overheat, or there may be a
resonance vibration close to the 60 Hz speed.
Frequency reduction requires voltage reduction,
maybe further derating due to lower fan speed.
11
Frequency, Voltage, Current, & Power
• Rated current determined by:
– Size of conductor (which is determined by
size of slot and number of turns/coil as well
as amount of insulation required)
– Stator parallel connections and hookup
• Wider slots for larger conductor => narrower
teeth, lower flux capacity, lower voltage
• Deeper slots for larger conductor => thinner
yoke, lower flux capacity, lower voltage, also
increased reactance, smaller coil spacing
• Optimum when slot width ~ tooth width
12
Frequency, Voltage, Current, & Power
• Current determines rated apparent power
VA = Voltage*Current *√3
– Not very dependent on real power (kW)
– Higher power factor => lower rotor heating
– Rated kW depends on engine rating
– Nominal power factor usually 0.8, though
0.85 or 0.9 becoming more common, may
be more appropriate for cogeneration
where islanding is not intended
• Higher power factor => smaller machine, lower
cost (for a given kW) (but other penalties)
13
Power Rating
• What power rating do you need?
– For standby or emergency usage, load is
primary determining factor
– In USA, NEMA 70 (National Electric Code)
and NFPA 110 (Emergency and Standby
Power Systems) dictate sizing
• Generator manufacturer is not expected to be
expert on these standards – application may
vary with jurisdiction
• Many good webinars and white papers are
available online covering these

14
Power Rating
• What power rating do you need?
– For data centers, special standards may
affect requirements, such as Uptime
Institute, federal, state, and local standards
– For cogen or peaking applications, power
requirements will be negotiated with the
power system operator.
– Grossly oversizing a generator set may
lead to problems such as wet-stacking,
high short-circuit current, or inefficiency

15
Power Rating
• Multiple parallel generators preferable
– Run generators at optimum power for best
efficiency, reliability
– Redundancy for failures or scheduled
maintenance

16
Power Rating
• Thermal capacity factors: • Electrical capacity factors:
– Ambient temperature – Available fault current
– Temperature rise/Load – Ride-through
factors: – Load factors:
• Duty cycle • Harmonic voltages
• Power factor • Voltage unbalance
• Unbalanced load • Leading PF load
• Harmonic load
– Motor starting and
• Airflow
voltage dip
– Cooling method – (Above all related to
(discussed later) reactances of machine)
– Efficiency

17
Power Rating
• Thermal capacity factors: • Electrical capacity factors:
– Ambient temperature – Available fault current
– Temperature rise/Load – Ride-through
factors: – Load factors:
• Duty cycle • Harmonic voltages
• Power factor • Voltage unbalance
• Unbalanced load • Leading PF load
• Harmonic load
– Motor starting and
• Airflow
voltage dip
– Cooling method – (Above all related to
(discussed later) reactances of machine)
– Efficiency

18
Ambient Temperature
• Higher-temperature air obviously doesn’t cool
a generator as well as cooler air, so the
maximum air temperature limits power.
• Normal maximum ambient temperature is
40ºC (104ºF). This is common to NEMA,
IEEE, IEC, most every other commercial
standard.
• Marine agency standards generally increase
this to 45-50ºC. If the machine is air-cooled
(but most marine units are water-cooled), a
derating factor of about 88-90% must be
applied to maintain temperature rise.
19
Ambient Temperature
Rated Power vs. Coolant Temperature
110%

105%
Rating % of Nominal

100%

95%

90% Notes:
1. From IEEE Std C50.13-2005
85% 2. Typical values, not necessarily
representing any particular generator
80%
25 30 35 40 45 50 55 60 65
Coolant °C

Rating % Special Lube

20
Ambient Temperature
• Generators that use water or other
substances as a coolant have power ratings
dependent on the temperature of the cooling
substance, rather than the surrounding air
• Heat exchangers increase the temperature:
– Typical air-to-water heat exchanger has an
air output temperature 8-15ºC above the
water inlet temperature
– Typical air-to-air heat exchanger has air
output temperature 25-30ºC above the
external air inlet temp. (derate ~35%)
21
Power Rating
• Thermal capacity factors: • Electrical capacity factors:
– Ambient temperature – Available fault current
– Temperature rise/Load – Ride-through
factors: – Load factors:
• Altitude • Harmonic voltages
• Duty cycle • Voltage unbalance
• Power factor • Leading PF load
• Unbalanced load
– Motor starting and
• Harmonic load
voltage dip
• Airflow
– (Above all related to
– Cooling method reactances of machine)
(discussed later)
– Efficiency

22
Temperature Rise
• NEMA, IEEE, IEC, and other standards
groups have defined “classes” of insulation
systems with different temperature capabilities
• Extensive testing at elevated temperatures,
cyclic stresses (including vibration) and high
humidity to prove that the insulation can
survive a long time at these operating
temperatures (20,000 hours is sometimes
used as a nominal lifetime, but not official).
IEEE Std 117 and 101 define testing methods.
• Why? Running machine warmer means you
can tolerate more loss, get more output.
23
Temperature Rise
Machine Part Method of Temperature Class Class Class
Determination B F H
Indirectly cooled stator Embedded detector (RTD 85 110 130
windings or thermocouple)
Directly cooled stator Air discharge from bar 80 100 N/A
windings Water discharge from bar 50 50 50
Indirectly cooled rotor Resistance 85 110 135
windings
Directly air-cooled rotor Resistance 60-80 75-95 N/A
windings

Cores and mechanical Detector or thermometer Not detrimental to

parts, whether or not in insulation of that part
contact with insulation or any adjacent part

Temperature rise over 40ºC ambient. From IEEE Std C50.13, but IEC
60034 and other standards are quite similar.
24
Temperature Rise
• Most insulation systems are Class F or H now
• Insulation class maximum temperature may
be exceeded during overload conditions, and
to some extent at “hot spots”
• Some standards (e.g. IEEE Std 11) allow
higher temperature rises. NEMA “Standby”
rating allows additional 25ºC rise for
presumed short-term operation.
• Some standards (especially agency approval
standards like ABS, Lloyds, etc.) have more
restrictive temperature rise requirements
25
Temperature Rise
• User may choose a lower temperature rise
than specified for the class of insulation:
– Because the ambient temperature may be
greater than 40ºC (Middle East location,
ship’s engine room, etc.)
– To obtain longer expected life from the
insulation
• Class B rise with Class F or Class H insulation
is a common specification for gas turbine
application

26
Temperature Rise
Arrhenius equation: 2X life for every 10º C the temp is
lowered (based on 20,000 hour life as the thermal index).
This is a rough rule of thumb, not rigorous.

100000
Hours

10000

1000
80 100 120 140 160 180 200 220
Total Winding Temperature - Degrees C

27
Temperature Rise
Svante August Arrhenius
(19 February 1859 – 2 October
1927) was a Swedish scientist.
Originally a physicist, but often
referred to as a chemist,
Arrhenius was one of the
founders of the science
of physical chemistry. He
received the Nobel Prize for
Chemistry in 1903, becoming
the first Swedish Nobel laureate.
In 1905, he became director of
the Nobel Institute, where he
remained until his death.
Discoverer of global warming.
--Wikipedia

28
Temperature Rise
• Conditions that impact temperature rise:
– Altitude
– Overloads and duty cycle
– Current unbalance
– Harmonics
– Low power factor
– Airflow restrictions
– Dust and other debris in air passages

29
Temperature Rise
• Airflow required for cooling is determined by
the machine losses and allowable air
temperature increase
– For 18ºC air temperature rise, 100 cu. ft.
per minute (about 170 m3/hour or 2800
liters/min) is required for every kW of loss
• For ducted air, the fan must be able to
produce enough pressure to maintain
required flow – special fan may be needed

30
Altitude Effect on Temperature Rise
Power Rating vs. Altitude
6,000 20000

5,000
16000

4,000 Notes:
1. Does not consider decreasing 12000
Altitude (m)

Altitude (ft)
temperature with altitude
3,000 2. Lower coolant mass can be
compensated by increasing 8000
2,000 airflow volume (larger fan) for
high-altitude applications
3. For 4 kV+, insulation may have 4000
1,000
to be increased

0 0
0% 20% 40% 60% 80% 100% 120%
% of Load Rating

IEC34 Kato

31
Altitude
• IEC 60034-1 Table 10 suggests assumption
that the ambient temperature decreases with
altitude, at a rate that may be sufficient to
make generator derating with altitude
unnecessary below 4000 m (use at own risk)

32
Overloads and Duty Cycle
• The materials that make up a generator have
a certain amount of heat capacity or “thermal
inertia”. If the machine is overloaded, the
temperature will take a while to reach the
limits of the insulation. And insulation can
operate safely above its rated temperature for
short periods, provided the average “aging
rate” is low enough to give adequate lifetime.
• Calculations may done by treating the
machine as a “thermal R-C network” with
losses as input, aging rate per Arrhenius eqn
• Need accurate estimate of duty cycle profile
33
Overloads and Duty Cycle
Generator Heating with Duty Cycle
2.0
1.8
1.6
Power (PU)

1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
0 50 100 150 200 250 300 350 400
225.0
200.0
175.0
Temp (°C)

150.0
125.0
100.0
75.0
50.0
25.0
0.0
0 50 100 150 200 250 300 350 400

4
Aging Rate (PU)

3
Average Aging Rate = 0.78
2

0
0 50 100 150 200 250 300 350 400

Time (min)
34
Overloads and Duty Cycle
• What do the different classes of rating mean?
– ISO 8523-1 defines four load categories:
• Continuous power (COP) – runs at full rating
24/7, e.g. cogeneration. Most data sheets will
be for this rating unless otherwise stated.
• Prime power (PRP) – runs at full rating some of
the time but average power is limited to 70% of
rating in a 24-hour period. 10% overload for 1
hour out of 24 is allowed. Backup power.
• Limited time power (LTP) – continuous but
limited to 500 hours/year, e.g. peaking
• Emergency standby (ESP) – same as PRP but
200 hours/year (may allow add’l 25ºC rise?)
35
Overloads and Duty Cycle
• What do the different classes of rating mean?
– IEC 60034-1 duty types S1-S10. These
are intended mostly for motors, but if
specified for a generator:
• Duty Type S1 – eq. to continuous power
• Duty Type S2 – eq. to limited time power, but
specific durations must be given in specification
• Duty Type S9 – similar to prime power, but
cyclic operation must be given explicitly
• Others should be negotiated with the vendor on
a case-by-case basis. Thermal analysis will
probably be needed.
36
Current Unbalance
• Unbalanced load causes extra heating in rotor
due to negative-sequence currents in stator
– The stator magnetic field due to negative-
sequence current rotates opposite of that
caused by normal load current
– Induce current in rotor bars and winding
– Increase I2R loss in rotor
• But as load becomes more unbalanced (with
no more than rated current in any phase), the
total load decreases, compensating to some
extent for the increase in rotor loss
37
Current Unbalance
• With fully-connected copper cage, 58% of
rated kVA single-phase may be available under
these conditions (full rated current)
• Special windings like dogleg or double-delta
help to balance stator currents and may allow
up to 66% of three-phase rating (but these do
nothing for rotor heating)
• Some machines are wound specifically for
single-phase operation and have only one
phase winding (ditto)

38
Current Unbalance
Allowable Load vs Negative Sequence
100%

95%

90%
Overloaded
85%
Load % of Rated

80%

75%

70%
Acceptable Load
65%

60%

55%

50%
0% 10% 20% 30% 40% 50% 60% 70%
Negative Sequence % of Total Load Current

39
Current Unbalance
Unbalanced Load Rating
110%

100%
Derate Factor

90%

80%

70%

60%
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Unbalance

L-N Load L-L Est

40
Harmonic Loads
• High harmonic load currents produce rotating
magnetic fields that are not synchronous with
the rotor, so they induce current in the rotor
– Mostly in the damper cage, if there is one
– Causing additional rotor heating
– This is similar to the extra heating caused
by unbalanced loading, and is treated in a
similar manner to calculate heating
– Effective harmonic “negative sequence
current” I2E used for calculation (weighted
sum of harmonics – see paper by Rice,
cited later, for details) 41
Harmonic Loads
• Some sources propose extreme oversizing
(like 250%) as a precaution against high load
harmonics – this is usually not necessary and
may lead to other problems
• If no unbalance, 10% I2E should be thermally
tolerable without increasing the size, if the
load is well-balanced (10% I2 is the limit for
load unbalance, which has the same effect)
• Up to 30% (normal six-pulse rectifier)
harmonics requires no more than 10% derate
(see following graph)

42
Harmonic Loads
• Properly sizing a generator for harmonic load
requires knowledge of the harmonic
frequencies and the current at each frequency
(spectrum), at the worst-case combination of
all harmonic loads
• Rather than oversizing the generator to control
harmonics, other methods may be more
effective or economical:
– Using lower-harmonic equipment (e.g. 12-
or 18-pulse rectification)
– Applying phase-shift transformers to cancel
lower harmonics, or using harmonic filters
43
Harmonic Loads (cont.)

For machine with robust

cage fully connected.

Assumes “typical” six-pulse harmonic spectrum

and that harmonics have the same effect as
negative-sequence current. Modified to account for
no derating with up to 10% I2E.

44
Power Factor (review)
• In a power system carrying sinusoidal AC
voltages and currents, the voltage on each
wire may not be “in phase” with the current
• The instantaneous power carried by that wire
is the product of the voltage and current at
that instant. If the peak of the voltage and
current don’t occur at the same time, the
amount of power carried by the wire is
reduced. The fraction to which it is reduced
is called the power factor (PF).
• The PF is equal to the cosine of the electrical
phase angle between voltage and current
45
Power Factor

Red – voltage, blue – current. Current lags voltage. 46

 Power Factor
• Active (real) power (W) -.
Eventually produces a
tangible result like heat or Reactive
light (= I2R). Power
(VAR)
• Reactive power (VAR) -
Surges back and forth
between the source and load
(zero average value). This Active power (W)
power produces alternating
magnetic fields in devices
( = I2X).
• Apparent power (VA) - Is the
vector sum of active and
reactive, the total needed
(= I2Z).

47
Power Factor
• In synchronous generators a stator (armature)
winding produces voltage in response to a magnetic
field produced when a rotor winding is “excited” by
field current.
• Increasing field current increases the magnetic field
resulting in increased voltage.
• Adding load to the generator changes the amount of
field current required to maintain the voltage setpoint.
• Lagging power factor load increases the required field
current; Leading power factor load decreases the
required field current.
• Leading power factor loads will be discussed in more
depth later

48
Power Factor
• NEMA, IEEE, IEC use 0.8 (overexcited, i.e.
lagging load) as nominal power factor
• Lower lagging power factor requires
oversizing due to two factors:
– For a given kW rating, the kVA will
increase at lower power factor, requiring
more stator current (larger machine)
– For a given kVA rating, the rotor current
increases as power factor decreases, due
to stator reaction flux. A larger machine is
needed to prevent rotor overheating.
49
Power Factor
Machine Ratings with Lagging Power Factor

18000 Decrease due to armature

16000 reaction (field heating)

14000
Rating kW/kVA/kVAR

12000
Decrease due to power kW Rating
10000 factor directly
kVAR Rtg
8000 kVA Rtg

6000

4000

2000

0
0.00 0.20 0.40 0.60 0.80 1.00
Power Factor

50
Reactive Capability Curve
3000

(Lagging) Limited by
Inductive Load
(Power Factor)
0.4
0.6
0.7 rotor heating
2000
0.8

1000
0.9
(Either) Limited by
stator heating
Leading -- Reactive Load kVAR -- Lagging

0
0 500 1000 1500 2000 2500 3000
Limited by control
stability
Capacitive Load

-1000
0.9

(Leading) Limited by
0.8
dynamic stability
-2000
0.6 0.7
0.4

-3000 Different view, same data

Power (kW)

51
Power Factor Nomograph

Email
[email protected]
or [email protected] for
electronic copy.

52
Leading Power Factor
• Operation with load at a more leading power
factor (less inductive, or more capacitive)
reduces the required field current, so reduces
the rotor temperature rise. If the rotor
temperature is the limiting factor, the machine
could be operated at a higher kVA rating with
this load. But engine limits kW anyway.
• But normally, we say the kVA rating is limited
by stator current, so it is constant at lagging
power factors greater than 0.8.
• Leading power factor load is usually stability
limited, and will be discussed more later.
53
Airflow Restrictions
• Some cooling systems and enclosure types
(for example, filters, or the multi-bend
ductwork for WP-II enclosure) can restrict the
flow of cooling air, so cause derating of the
generator. We will cover this when we talk
about enclosures and cooling systems.
• Inadvertent restrictions, like long narrow
ducts, air intakes or outlets too close to
obstructions, lack of room air provisions, or
debris collected in air passages or on
generator surfaces, can cause generator
overheating.
54
Power Rating
• Thermal capacity factors: • Electrical capacity factors:
– Ambient temperature – Available fault current
– Temperature rise/Load – Ride-through
factors: – Load factors:
• Altitude • Harmonic voltages
• Duty cycle • Voltage unbalance
• Power factor • Leading PF load
• Unbalanced load
– Motor starting and
• Harmonic load
voltage dip
• Airflow
– (Above all related to
– Cooling method reactances of machine)
(discussed later)
– Efficiency

55
Cooling Method
• Cooling methods that restrict airflow, or result
in higher-temperature coolant, require
derating of the generator
• We will discuss derating factors for different
cooling methods further on in the program,
when we talk about enclosures

56
Power Rating
• Thermal capacity factors: • Electrical capacity factors:
– Ambient temperature – Available fault current
– Temperature rise/Load – Ride-through
factors: – Load factors:
• Altitude • Harmonic voltages
• Duty cycle • Voltage unbalance
• Power factor • Leading PF load
• Unbalanced load
– Motor starting and
• Harmonic load
voltage dip
• Airflow
– (Above all related to
– Cooling method reactances of machine)
(discussed later)
– Efficiency

57
Electrical Capacity Factors
• Often it is necessary to oversize a generator
for reasons not connected with temperature
• Some of these require controlling generator
reactance or impedance:
– To increase or reduce available fault
current (e.g. for breaker coordination)
– To reduce harmonic voltages or voltage
unbalance
– To meet a grid code ride-through spec
• Meeting grid codes may also require
increased inertia (hence a larger rotor)
58
Electrical Capacity Factors -- Reactance
• Reactance is controlled by the same factors
that control voltage and power ratings
• Reactance is normally given in per-unit or
percent of a base impedance so that it is
independent of voltage and power rating
• Lower reactance requires fewer turns, fewer
slots, or lower pitch, which all reduce voltage
– So additional magnetic material has to be
added to bring the voltage back to rated
– Makes for a larger machine, higher cost

59
Electrical Capacity Factors -- Reactance
• Higher reactance requires more turns, more
slots, or higher pitch (1st 2 of which reduce
the area available for copper), or else added
lamination stack
– So a larger lamination may be needed to
allow the extra room for the windings
– Or the added stack also increases size
• Lower reactance improves voltage dip, but
increases fault current, and vice-versa
• Lower reactance reduces voltage unbalance
and harmonic voltages
60
Unbalanced Voltage
• The impedance of a generator to unbalanced
load current is primarily negative-sequence
reactance X2 (with a small R2 component)
– Controlled mostly by damper cage design
• Voltage unbalance is calculated by means of
symmetrical components
– Use known current amplitudes and phases
to calculate negative-sequence current I2
– Multiply I2 by X2 to get negative-seq voltage
– Add to positive-sequence voltage to obtain
unbalanced voltage vector set
61
Unbalanced Voltage
• Voltage drop in generator is mostly inductive
• Lagging load power factor (more inductive)
means voltage drop is more “in-phase” with
respect to line voltage.
• So lagging power factor load increases the
negative-sequence voltage across the
generator reactance, so increases unbalance

62
Unbalanced Voltage
• Rough rule of thumb: 20% current unbalance
(between minimum and maximum currents)
gives about 10% negative-sequence current
Negative-Sequence vs. Unbalance Definition: Worst-case
60% X% unbalance is
50% 100% current in two
phases and X% less
Negative-Seq %

40%
in the third.
30%

20% Assumes equal power

10%
factor (120º current
angles) for L-N load.
0%
0% 20% 40% 60% 80% 100%
Unbalance % Assumes no neutral
L-N Load L-L Load current for L-L load
63
Harmonic Voltage
• High harmonic load currents also interact with
the generator impedance to produce harmonic
voltages. These can cause excess heating of
motors and transformers, as well as failures of
sensitive equipment. IEEE Std 519
recommends maximum harmonic levels for
power systems.
– The effective impedance for harmonics is
the negative-sequence reactance, X2,
multiplied by the harmonic order
– X2 should be close to the subtransient
reactance, if the machine has a fully
connected damper cage 64
Harmonic Voltage
• Harmonic voltage of a particular harmonic is
equal to harmonic current at that frequency
times reactance at that frequency
• Root of the sum of the squared individual
harmonic voltages gives the effective RMS
harmonic voltage – divide by the fundamental
to get total harmonic distortion (THD)

65
Harmonic Loads (cont.)

Notes:
• Six- and twelve-pulse load
values are the fraction of
the load that is harmonic-
generating, in per cent
• Align X2 and % of load
values with a straightedge,
and read voltage THD
from middle scale, or align
% of load and THD to get
the maximum value for X2
• Estimate only, since THD
depends on system factors
other than the generator
impedance

66
Operation at Leading Power Factor

67
Excitation of Synchronous Machines
• Remember leading power factor load decreases field
current. At some value of leading power factor load,
the field current required to maintain voltage
decreases to zero. At this point, there is no ability to
control voltage and the generator voltage will increase
until saturation stops it.
• If the generator is paralleled with a grid, the effect is
different. Since voltage is controlled by the grid, it
doesn’t increase. But without field current, the
generator can’t accept the torque delivered by the
engine, so will become unstable and slip poles. “Pole
slipping” causes severe voltage disturbances and
pulsating torque, and results in rapid heating of the
generator.
68
Excitation of Synchronous Machines
• Leading power factor is sometimes referred to as
“reverse VARs” or “negative VARs”
• This is not the same thing as “negative power factor”
or “reverse power”, which is real power flow into the
set, and is controlled by the engine, not the generator.
• Power factor is positive for either leading or lagging
condition. As long as power is being delivered, the
PF is greater than zero. A different quantity, usually
called reactance factor (RF), becomes negative when
the power factor is leading.
𝑅𝑅𝑅𝑅 = 1 − 𝑃𝑃𝑃𝑃 2 (Pythagoras)

69
Reactive Capability Curves
• The interaction between field current and output voltage
and current is complex, especially when saliency and
saturation of the magnetic material are considered, so
usually simplified (but conservative) rules are used to
determine generator sizing.
• The tool most commonly used to determine excitation
limits is the reactive capability curve, which is derived
from these simplified rules. We saw this earlier – let’s
take a closer look.
• Kato uses the method from IEEE Std 67-1971 for
lagging power factor load, and Rubenstein and
Temoshok, AIEE Trans. Dec.1954, V73-III-B, p. 1434
for leading power factor load. Just so you know.

70
Reactive Capability Curves
• The generator parameters that control this are
saturated short-circuit ratio (SCR or, in Europe,
Kcc) and saturated synchronous impedance Xds.
These both represent the same quantity and are
reciprocals of each other.
• High SCR and low Xds make the machine more
stable. To a first approximation, if Xds is less than
1, or SCR greater than 1, the machine will be
stable with any leading power factor down to zero.
• To achieve this usually requires a considerably
oversized machine.

71
Reactive Capability Curves
This shows the reactive
capability curve from
Rubenstein and Temoshuk.

Xe is system reactance for

grid paralleled machine.
Smaller Xe makes the
curve flatter – limit with Xe
equal to zero is horizontal
line (zero torque angle δ).
Kato uses curve A for Xe =
Kato curve
0.4, which is very
conservative for salient-
pole machines.

72
Reactive Capability Curves
We will build the full reactive
capability curve in pieces.

This curve assumes that the

machine is thermally limited
to 1.0 per unit current. In
reality, there is usually some
margin, so the curve is for a
nominal arbitrary current
limit rather than the actual
machine capability. This
saves bother and allows the
curves to be drawn
automatically.
Distance from (0,0) is load
current in PU.

73
Reactive Capability Curves
Distance from center point
represents approximate value
Max rotor of excitation (saturated).
current Assumes that the machine is
perfectly thermally balanced
at 0.8 PF, and becomes rotor-
limited below 0.8 PF. This is
not generally true, so this
curve represents what we will
guarantee the machine to do,
not what it is actually capable
of doing.

Per IEEE Std 67-1971

Assumes cylindrical rotor,
which is not true for our sets,
but is conservative
74
Reactive Capability Curves
Blue curve is the theoretical
torque limit for a round-rotor
machine (δ = 0).
Orange curve is for a more
reasonable δ limit of 15º.
Purple curve is for the limit
from the Rubenstein and
Temoshuk paper, also for
round-rotor machine.
These are conservative for
salient-pole, because they
ignore saliency torque, which
improves stability.
Limits for paralleled machines,
same for islanded machines,
but for different reasons.

75
Reactive Capability Curves

These two equations are the

same

76
Reactive Capability Curves

Point of zero excitation

77
Reactive Capability Curves

What if you
add saliency?

78
Reactive Capability Curves
• With lagging power factor, the limits on generator
output are caused by heating. At leading power
factor, on the other hand, the limits are concerned
with the stability of the machine, meaning its ability to
supply power at a steady rate at constant speed and
voltage.
• For large or high-speed machines there may be a
limit on leading power factor due to heating of the end
regions of the rotor. We have never observed this in
the kind of machines we build, but it should be kept in
mind as we build larger machines.

79
How to Specify Generators for Leading PF

Preferably, the generator OEM and the user (or site

engineer) should work together to select a generator
design that will meet the requirements of the site. In
addition to the normal rating data, the manufacturer will
benefit from the following information:
• Whether the site will run in island or grid-connected
mode, or both.
• In island mode, what sort of capacitive load will be
connected, and how it will be coordinated with other
loading.
• In grid-connected mode, the maximum leading VARs
or minimum leading power factor required

80
How to Specify Generators for Leading PF

• For example, large UPS sets will generally have a

fixed capacitive reactance on the input due to the
input filters, but these UPS sets will have variable
real power load depending on how their output is
loaded. Other loads on the circuit (e.g. chillers)
may provide lagging load to counterbalance the
leading load, but if they are started after the leading
load, the system must be stable without them.

81
How to Specify Generators for Leading PF

• In grid-connected mode, the user has control over

the VAR loading, but if leading power factor
operation is required (possibly for local voltage
control at light load), the generator manufacturer
needs to know the possible range of real and
reactive loading.
• Recent grid codes require operation at least at 0.95
PF leading, possibly 0.9 PF leading.
• Also required is operation at 90-95% reduced
voltage (increased current), increasing the per-unit
reactance and decreasing stability limits.

82
How to Specify Generators for Leading PF

To meet the site requirements, we may:

• Propose an oversize generator in order to keep the
synchronous reactance low.
• Use a special design that is more saturated than
normal for the same reason.
• Include special testing to insure that the generator
will meet requirements.
• Propose additional protective relaying and controls
to detect or prevent unstable conditions.

83
How to Specify Generators for Leading PF

• At a minimum, a reactive capability curve as well as

saturation curves and V-curves should be requested
for the proposed generator before ordering, and
carefully examined by the site engineer to make
certain the generator will always operate within the
safe region.
• Because the leading power factor capability
depends strongly on factors that vary by
manufacturer, frame size, pole count, and model, it
is not possible to use a “rule of thumb” for sizing the
generator.

84
How to Specify Generators for Leading PF

** End of Detour **

85
Motor Starting and Voltage Dip
• For generators rated ~50 MW or less with
electronic voltage regulators, the transient
reactance X’d controls the voltage dip or rise
due to short-term load transients
• On load application, generator reactance
forms a voltage divider circuit with the load
impedance, reducing the voltage until the
regulator can respond and increase the
internal excitation voltage to compensate
• Similarly on load rejection, when the load is
removed, voltage rises to value of (increased)
internal excitation voltage
86
Motor Starting and Voltage Dip

(Complex)

Voltage Divider Effect

87
Motor Starting and Voltage Dip
• For low power factor load (esp. motor
starting, typically 0.2-0.4 PF), ZL and X’d are
both mostly inductive, and voltage dip can be
calculated using the magnitudes of the
impedances
• For higher power factor (> 0.4), complex or
vector calculation must be used
• To simplify, an approximate power factor “dip
factor” can be used (see following slide)

88
Motor Starting and Voltage Dip
Voltage Dip Factor vs. PF for 0.3 PU X'd
1.0

0.9

0.8

0.7 Based on E.L. Harder & R.C. Cheek paper, “Regulation of

A-C Generators With Suddenly Applied Loads,” AIEE
Dip Ratio

0.6

0.5

0.4

0.3

0.2
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Power Factor

89
Motor Starting and Voltage Dip
• Preload of the generator affects the voltage
dip, and more so if it is motor load or other
constant-power load
• Resistive load like heaters or incandescent
lighting will draw reduced current during a
voltage dip (still, the dip increases – why?)
• Motors or electronic load maintain constant
power draw, so if the voltage decreases, the
current will increase, worsening the dip
• An approximate preload “dip factor” is used to
estimate the increase in the voltage dip
90
Motor Starting and Voltage Dip
Preload Multiplier for Voltage Dip
1.25

Based on Harder & Cheek paper, (AIEE Trans. June 1944)

1.20

1.15
Multiplier

1.10

1.05

See paper
1.00
0% 20% 40% 60% 80% 100%
Preload (% of Rated)

Resistive Motor/Electronic

91
Motor Starting: Examples

• NEMA MG1 32.18.5.3 Motor starting equation

𝑋𝑋 ′ 𝑑𝑑 Where D is PU dip, X’d is in per unit
D= 𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵 of the KVA base, BKVA is base kVA,
𝑋𝑋 ′ 𝑑𝑑 + SKVA is motor starting kVA
𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆
Example: 2000 kW @ 0.8 PF with X’D = 18.2% starts a 750
HP code F motor – What is the approximate voltage dip?
– Base kVA = 2000 kW / 0.8 PF = 2500
– Starting kVA = 750 HP × 5.6 kVA / HP = 4200
0.182
– %Dip = 100% × 2500 = 23.4%
0.182 +
4200

• Note: To reduce the dip, reduced voltage starters can be used.

Engine transient response will increase dip, but the engine
response is usually much slower than the generator. This
assumes a very low starting power factor.
92
Motor Starting and Voltage Dip
Generator Sizing for Required Voltage Dip
500%

Note:
400% Better results can be obtained if the
generator is specifically designed
for low voltage transient. This is for
300% illustrative purposes only.

200%

100%

0%
0% 5% 10% 15% 20% 25%
Voltage Dip (%)

93
Electrical Capacity Factors -- Efficiency
• Increasing the size of a generator allows
reduction of the number of coil turns or slots,
so the winding resistance can be reduced
– This reduces winding losses, improving
efficiency
– It also increases core and other fixed
losses, so a balance must be achieved
between the losses to obtain best
efficiency.
• Highest efficiency is usually achieved at or
slightly below full load for normal generator
94
Electrical Capacity Factors -- Efficiency
Expected Efficiency (Typical Generator)

100%

99%

98%

97%

96%
% Eff'y.

95%

94%

93%

92%

91%

90%
0% 25% 50% 75% 100% 125%

Load (% of FL)

95
Electrical Capacity Factors -- Efficiency
• The first step to increasing the efficiency of a
generator is usually to reduce the core loss
– Change stator steel to lower-loss material
(more expensive)
– Change rotor steel to electrical steel, or
core-plate, to reduce pole-face loss
• Second, increase the machine size to allow
larger conductors/fewer turns for lower
resistive (I2R) loss in stator and (indirectly) in
rotor windings

96
Electrical Capacity Factors -- Efficiency
• Since losses are now lower, a smaller fan
may be used to reduce friction and windage
• The machine will also run cooler, reducing
the resistance and hence resistive loss
• Double-end-vent construction also allows use
of smaller and more efficient fans (but not as
suitable for reciprocating engine application)
• All other things being equal, slower and larger
generators generally have higher efficiency

97
Electrical Capacity Factors -- Efficiency
• Applying advanced techniques like variable
electric blowers, rotor shrouds, optimization
techniques, and special aerodynamic design,
can reduce loss still further
98% eff’y!

98
2. Generator Construction

99
Generator Construction

100
Generator Construction
• Types of enclosure
– Open drip-proof (ODP)
– Totally enclosed air-to-air (TEAAC or CACA)
– Totally enclosed water cooled (TEWAC or
CACW)
– Water Protected II (WP-II)
• IC cooling codes
• IP protection codes
• Double- vs single-end ventilation

101
Generator Construction
• Open drip-proof:
– Most common type of construction
– Protects against water from above

Double-end
ventilation
(draws air from
Air out both ends and
exhausts in
middle)
Air in Air in

102
Generator Construction
• Open drip-proof:

Single-end
ventilation
(draws air from
one end and
exhausts at
other)

Air in
Air out

103
Generator Construction
• Totally-Enclosed Air-to-Air (or TEFC)
– Larger machines have top heat exchanger
– Smaller may just have fins
– Water or dust ingress is restricted

May be single-
or double-end
vent internally

104
Generator Construction

105
Generator Construction
• Totally-Enclosed Air-to-Air (or TEFC)
– Cooling air temperature is increased
because of the heat exchanger – internal
air may be up to 25ºC warmer than outside
– Rating is reduced on account of this (about
25% reduction for Class B rise)

106
Generator Construction
• Totally-Enclosed Water Cooled
– Usual for marine applications
– Heat exchanger on top

107
Generator Construction

108
Generator Construction
• Totally-Enclosed Water Cooled
– Internal cooling air will be about 8ºC
warmer than the incoming water (18ºC for
double-tube, required for seawater cooling)
– Typical water temperature is 32ºC, so air is
at around 40ºC, or 50ºC for double-tube
– Some derating is required from the 40ºC
rating for double-tube heat exchanger

109
Generator Construction
• WP-II (Weather-Protected)
– Uses large ducts for low-velocity air
– Uses right-angle bends to deter water from
entering and divert any that does enter
– About 5% derate required due to restricted
airflow

110
Weather Protected II (WPII)

Inlet Air
Three 90 degree
direction changes
and <600 fpm air
speed. Optional air
filters

111
Other Ambient Conditions
• Humidity:
– “Coastal” overcoating for moist locations
– Use space heaters to keep temperature
above condensation point
– API-546 construction (double VPI plus
overcoat, windings immersed in water for
dielectric test)
– Totally-enclosed (fan or water cooled)
• Outdoor installations:
– WP-II IEC IP-44 construction

112
Other Ambient Conditions
• Dust and debris:
– Inlet air filters keep dust out
– Filter back pressure detectors to indicate
that filters need to be changed
– Increase size by about 5%

113
Air Filtered Unit
Filtered Air Inlet

Optional IP44 Sealed Outlet Box Construction

114
IEC 60529-02 IP Codes (Sim to IEC 60034-5)
– Protection of persons against contact with hazards

– Protection of machine against ingress of solid

objects

– Protection of machine against harmful effects due

to ingress of water

115
IEC 60529-02 IP Codes
1st digit –
2) Protected against 12 mm objects
4) Protected against 1 mm objects
5) Dust protected
6) Dust tight
2nd digit –
1) Protected against dripping water
2) Protected against 15 deg water spray
3) Protected against 60 deg water spray
4) Protected against splashing water
5) Protected against water jets
6) Protected against powerful water jets (fire hose)
Most Common IP Codes
– IP21 (ODP) IP22 IP23 IP44
IP54 IP55 IP56
116
IEC 60034-6 IC Classification
• Sequence of numerals and letters to identify the following:
– cooling type, TEWAC, TEAC
– cooling medium, air, water, hydrogen

117
IEC 60034-6 IC Classification
Table 1 - Circuit
arrangement defines
machine construction
specifics, ODP, totally
enclosed.

Cooling medium routing is

also defined for cooling
circuits.

118
Table 1 - Numeral for Circuit Arrangement
0) Free Circulation
1) Inlet pipe or inlet duct circulated
2) Outlet pipe or outlet duct circulated
3) Inlet and outlet pipe or duct circulated
4) Frame surface cooled
5) Integral heat exchanger (using surrounding medium)
6) Machine-mounted heat exchanger (using
surrounding medium)
7) Integral heat exchanger (using remote medium)
8) Machine-mounted heat exchanger (using remote
medium)
9) Separate heat exchanger (using surrounding or
remote medium)
119
IEC 60034-6 Classification
Table 2: Coolant defines coolant type; water, air, oil
TABLE 2
A Air
F Freon
H Hydrogen
N Nitrogen
C CO2
W Water
U Oil
S Any other
Y Reserved

120
IEC 60034-6 Classification
Table 3: Method of
Movement defines
coolant circulation type,
convection, fan drive
(internal or external),
pumps.

121
Table 3 - Numeral for Method of Movement
0) Free Convection
1) Self-circulation
2) Reserved for future use
3) Reserved for future use
4) Reserved for future use
5) Integral independent component
6) Machine-mounted independent component
7) Separate and independent component or coolant
system pressure
8) Relative displacement
9) All other components

122
IEC 60034-6 Classification

Typical Kato code:

TEWAC, IP44, IP54
Equivalent IEC 60034-6 code:
IC8 A1 W7 = Totally enclosed
air-to-water cooled.

8 = Machine-mounted heat exchanger with remote

medium
A = Primary circuit cooling medium, air
1 = Self circulation of medium
W = Secondary cooling medium, water
7 = Separate circulation not mounted to machine.
123
IEC 60034-6 Classification

Typical Kato code:

TEAAC, CACA, IP44, IP54
Equivalent IEC 60034-6 code:
IC6 A1 A1 = Totally enclosed
air-to-air cooled with ambient
air

6 = Machine-mounted heat exchanger using

surrounding medium
A = Primary circuit cooling medium, air
1 = Self circulation of medium
A = Secondary cooling medium, air
1 = Self circulation of medium

124
IEC 60034-6 Classification
• IC not used by Kato to describe cooling
• But we can interpret if found in a spec

125
NEMA MG 1, Section 1, Part 1
• NEMA MG 1, Section 1, Part 1
– Defines Classification according to environmental protection
and methods of cooling
• Consolidates IP and IC codes (from Parts 5 and 6)
• Classifications
– Open Machines
• Drip proof
• Splash-proof
• Semi-guarded
• Guarded
• Drip proof guarded
• Open independantly ventilated
• Open pipe-ventilated
• Weather Protected Type I / Type II

126
NEMA MG 1, Section 1, Part 1
• NEMA MG 1, Section 1, Part 1 Classifications (Continued)
– Totally Enclosed Machines
• Totally enclosed nonventilated
• Totally enclosed fan-cooled
• Totally enclosed fan-cooled guarded
• Totally enclosed pipe-ventilated
• Totally enclosed water-cooled
• Water-proof
• Totally enclosed air-to-water cooled
• Totally enclosed air-to-air cooled
• Totally enclosed air-over machine
• Explosion-proof machine
• Dust-Ignition-proof machine

127
NEMA MG 1, Section 1, Part 1
• NEMA MG 1, Section 1, Part 1 Classifications (Continued)
– Machines with encapsulated or sealed windings
• Moisture resistant windings
• Sealed windings
– Most common classifications
• Dripproof guarded machine (ODP)
• Weather-protected machine: Type I & Type II (WPII)
• Totally enclosed fan-cooled guarded (TEFC)
• Totally enclosed air-to-water-cooled (TEWAC, CACW)
• Totally enclosed air-to-air-cooled (TEAAC, CACA)
• Machine with sealed windings

128
Hazardous Location
• Two systems in common use:
– NEC Article 500 “Class/Division” system
(mostly North America)
• Kato machines normally Class I, Div 2, T3
temperature
– IEC 60079 “Zone” system (Europe and rest of
world)
• Kato machines normally Zone 2, Ex nA IIB
T3 temperature
• Similar (not identical) requirements for both
systems

129
Hazardous Location
• Require special generator design, construction,
and testing – examples:
– No open conductors or sparking components
– Maximum temperature limits for machine surfaces
and space heaters
– Voltage limiters on CTs
– Depending on site requirements, may require
totally-enclosed construction with purging facility
– Manufacturing drawing approval by agency
– Witnessed type testing by agency including heat
runs and possibly enclosure testing
• Needs close coordination with site engineer and
engine manufacturer
130
3. Reactances and Fault Currents

131
Measurement of reactances
• D-axis synchronous reactance Xd is
calculated from the saturation curve
• What is this “D-axis” business?
– “D” is for “direct”, “Q” for “quadrature”
– Physically, the D-axis is aligned with the
center of the salient pole (or the center of
the field winding for cylindrical rotors)
– D-axis is involved with reactive current –
for short-circuits, current is mostly reactive,
so D-axis most important for short-circuit
– More later in reference frame theory
132
Measurement of reactances
• D-axis synchronous reactance Xd = field
current @ rated current SC / rated voltage NL
• The traditional method of measurement for
transient reactances is to apply a sudden
short-circuit to a generator operating at no
load and full voltage (some methods require
+5% voltage).
• An oscillograph trace (digital these days) is
taken of the current and interpreted to give
the D-axis transient (medium-term) and
subtransient (short-term) reactances and time
constants.
133
Reactances
• Reactances are used to describe the behavior of a
generator during certain operating conditions
• Transient reactance X’d is used in motor starting
calculations
- lower X’d results in better motor starting (i.e. lower
voltage dip)
• Subtransient reactance X”d is used in short-circuit
current and arc flash calculations
- lower X”d results in higher short circuit currents.
• There are other reactances for different situations.
- e.g. direct and quadrature axis, zero and negative
sequence, saturated and unsaturated
• The reactances and time constants that are calculated
or measured by test determine the generator model
parameters
134
Reactances (Cont.)

• Test method to obtain

X’d, X”d:
– Machine operating
at rated voltage, no-
load, manual

Sample data
excitation
– Short circuit is
applied across all
three phases
simultaneously
– Current envelope in
at least one phase is
captured over time
as it decays
135
Measurement of reactances
• The (saturated) synchronous D-axis
reactance determines the steady-state short-
circuit current (it is the reciprocal of the short-
circuit ratio). It is taken from short- and open-
circuit saturation curves.
• If required, separate tests measure the
negative- and zero-sequence reactances,
used for unbalanced fault calculations.
These are defined in IEEE Std 115.
• Tests for Q-axis quantities are difficult, and
not as important, so are usually calculated

136
Time Constants
• Time constants characterize the rate of decay of the
generator fault currents
• The two most common are the subtransient time
constant T”d and the transient time constant T’d.
– These are applied as two exponential decay
terms that combine to form the decrement curve
– Sub-transient time constant T”d determines length
of time sub-transient (damper) current flows.
– Transient time constant T’d determines length of
time transient current flows.
• Armature time constant TA defines the decay of the
DC offset term, and is calculated separately, from
the curve with AC component removed

137
Time Constants

138
Time Constants (cont.)
• Another time constant example: Open circuit
transient T’d0 (related but not equal to T’d).
– This represents the decay in armature output
voltage when main field excitation is removed
– Test requires slip rings to manually control the
main field voltage

139
Time Constants (cont.)

140
Reactances, Time Constants: Why are they Important?

• Very useful to predict performance in large system

simulations.
• Now, for many users, the reactances and time
constants are plugged into simulators used for
system modeling and coordination studies.
• Generator manufacturers use this information to
produce three curves commonly used for system
coordination:
– Thermal damage curve
– Short-circuit decrement curve
– Reactive capability curve

141
Decrement Curves
• Decrement curves show the fault currents of
a synchronous machine under different fault
and excitation conditions
• They are used for sizing and coordination of
interrupting devices such as circuit breakers
and fuses
• Following is a bit more detailed look at how
the reactances and time constants combine
to produce the estimated fault current curve

142
Fault Currents of a Synchronous Machine

143
Fault Currents of a Synchronous Machine
• IEEE Std 242 provides a simplified method to calculate
generator current delivered into a “hard” three-phase
fault applied suddenly and with constant excitation.
• Symmetrical 3-phase fault is the sum of two sinusoidal
decaying exponential functions plus a constant AC term.
• Modifications of this method provide for regulator action
that increases excitation following application of the
fault. These are not part of any known standard.
Basically, we just increase the steady-state excitation
relative to the initial value, using an exponential function
with short-circuit time constant, whose limit is
determined by the exciter ceiling voltage. In the method
shown below, this function is represented by eT, which is
constant for fixed excitation.
144
Fault Currents of a Synchronous Machine
• Equation for symmetrical AC bolted
fault current (three terms):
𝑡𝑡 𝑡𝑡 Subtransient
− " − ′
term (~0-10
𝑖𝑖𝑎𝑎𝑎𝑎 = 𝑖𝑖𝑑𝑑" − 𝑖𝑖𝑑𝑑′ 𝑒𝑒 𝑇𝑇𝑑𝑑 + 𝑖𝑖𝑑𝑑′ − 𝑖𝑖𝑑𝑑 𝑒𝑒 𝑇𝑇𝑑𝑑 + 𝑖𝑖𝑑𝑑 𝑒𝑒𝑡𝑡
cycles)
where:
𝐼𝐼
ei terms
𝑒𝑒𝑡𝑡 𝐹𝐹 represent
𝐼𝐼𝐹𝐹𝐹𝐹 𝑒𝑒𝑖𝑖
𝑖𝑖𝑑𝑑 = = (constant) excitation to get
𝑋𝑋𝑑𝑑 𝑋𝑋𝑑𝑑
𝑒𝑒𝑒𝑖𝑖
initial output.
𝑖𝑖𝑖𝑑𝑑 = (transient)
𝑋𝑋𝑋𝑑𝑑 𝐼𝐼𝐹𝐹
is the ratio of
𝑒𝑒"𝑖𝑖 𝐼𝐼𝐹𝐹𝐹𝐹
𝑖𝑖𝑖𝑑𝑑 = (subtransient) loaded field
𝑋𝑋𝑋𝑑𝑑
excitation to no-
load value et.

145
Fault Currents of a Synchronous Machine
• Equation for AC symmetrical bolted fault
current:
𝑡𝑡 𝑡𝑡 Transient term
− " − ′
(~10-100 cycles)
𝑖𝑖𝑎𝑎𝑎𝑎 = 𝑖𝑖𝑑𝑑" − 𝑖𝑖𝑑𝑑′ 𝑒𝑒 𝑇𝑇𝑑𝑑 + 𝑖𝑖𝑑𝑑′ − 𝑖𝑖𝑑𝑑 𝑒𝑒 𝑇𝑇𝑑𝑑 + 𝑖𝑖𝑑𝑑 𝑒𝑒𝑡𝑡
where:
𝐼𝐼𝐹𝐹 e'i terms
𝑒𝑒𝑡𝑡 𝐼𝐼 represent initial
𝐹𝐹𝐹𝐹
𝑖𝑖𝑑𝑑 = excitation.
𝑋𝑋𝑑𝑑
𝑒𝑒𝑒𝑖𝑖 𝐼𝐼𝐹𝐹
is the ratio of
𝑖𝑖𝑖𝑑𝑑 = 𝐼𝐼𝐹𝐹𝐹𝐹
𝑋𝑋𝑋𝑑𝑑 loaded field
𝑒𝑒"𝑖𝑖 excitation to no-
𝑖𝑖𝑖𝑑𝑑 =
𝑋𝑋𝑋𝑑𝑑 load value et.

146
Fault Currents of a Synchronous Machine
• Equation for AC symmetrical bolted fault
current:
𝑡𝑡 𝑡𝑡
− " − ′
Constant term
𝑖𝑖𝑎𝑎𝑎𝑎 = 𝑖𝑖𝑑𝑑" − 𝑖𝑖𝑑𝑑′ 𝑒𝑒 𝑇𝑇𝑑𝑑 + 𝑖𝑖𝑑𝑑′ − 𝑖𝑖𝑑𝑑 𝑒𝑒 𝑇𝑇𝑑𝑑 + 𝑖𝑖𝑑𝑑 𝑒𝑒𝑡𝑡
where:
𝐼𝐼
𝑒𝑒𝑡𝑡 𝐼𝐼 𝐹𝐹 e'i terms
𝐹𝐹𝐹𝐹
𝑖𝑖𝑑𝑑 = represent initial
𝑋𝑋𝑑𝑑 excitation.
𝑒𝑒𝑒𝑖𝑖
𝑖𝑖𝑖𝑑𝑑 = 𝐼𝐼𝐹𝐹
𝑋𝑋𝑋𝑑𝑑 𝐼𝐼𝐹𝐹𝐹𝐹
is the ratio of
𝑒𝑒"𝑖𝑖 loaded field
𝑖𝑖𝑖𝑑𝑑 =
𝑋𝑋𝑋𝑑𝑑 excitation to no-
load value et.

147
Fault Currents of a Synchronous Machine
• The preceding terms are AC, meaning they have a
sinusoidal variation at the generator output frequency
as shown below.
Subtransient time constant

Constant-current region
Transient time constant

148
Fault Currents of a Synchronous Machine
• Asymmetrical fault current adds the DC offset term,
which varies depending on the instant of fault
initiation and the phase being considered
• For decrement curves, we usually assume the worst
case for the DC offset term, which is:
𝑡𝑡
−
3𝑖𝑖𝑑𝑑" 𝑒𝑒 𝑇𝑇𝐴𝐴
𝑖𝑖𝑑𝑑𝑐𝑐 =
• Note that this is a DC term and has no sinusoidal
variation. It produces a decaying offset of the AC
waveform that increases the peak fault current
• This offset occurs in differing magnitude and direction
in the three phases. Since it sums to zero in the 3
phases, it is often ignored for breaker sizing
purposes (depending on the particular standard)
149
Fault Currents of a Synchronous Machine
• Adding the DC term gives the complete short-circuit
current waveform:
16.000

14.000

12.000
Subtransient time constant

10.000
DC Offset – Armature time constant
8.000

6.000 Transient time constant

Constant-current region
4.000

2.000

0.000

-2.000

-4.000
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

DC Term Sum

150
Typical Published Decrement Curve

151
Fault Currents of a Synchronous Machine

** End of Detour **

152
4. Grid Codes Effects on Generators

153
Grid Codes Effects on Generators
• REPORT ON COORDINATION OF GRID
CODES AND GENERATOR STANDARDS:
Consequences of Diverse Grid Code Requirements
on Synchronous Machine Design and Standards [GCTF]
– IEEE Electrical Machinery Committee Task Force
on Grid Code Impacts on Generator Standards
• Chair: Robert Thornton-Jones
• Members and Contributors: Evert Agneholm, William Bloethe,
Edson Bortoni, Kevin Chan, Kay Chen, Bob Cummings,
Robert F. Gray, Randall Groves, Les Hajagos, Joe Hurley, Relu
Ilie, Chavdar Ivanov, Ana Joswig, Jason Kapelina, Ruediger
Kutzner, Jim Lau, Kevin Mayor, Bill Moore, Lon Montgomery,
Nils Nilsson, Ryan Quint, Steve Richards, Michel Rioual, Luis
Rouco, Mike Sedlak, Uwe Seeger, Nico Smit, Fabian Streiff,
Robert Thornton-Jones, John Yagielski, Marc Zeidman,
Carsten Zuehlke
154
Summary

• What is a Grid Code, why new grid codes and

consequences ?
• Impact of the Grid Code on sizing:
– Static
– Dynamic
– LS guide lines and basic recommendations
• Fault Ride Through risk and behavior

155
What is a grid code ?

• A grid code is a technical specification which defines the

requirements for a generating set to be connected to a
public supply network. How it meets, safely, securely and
economically a correctly functioning electrical system,
• The grid code is specified by an authority responsible for
the system integrity and network operation. This is either
a rule, a law or a standard, depending on each country.
• All generators are concerned whatever the driving system
and whatever the power rating.

156
Why is there a new grid code ?
• The increasing intermittent power generation (due to increasing
wind farm installations or photovoltaic plants) and more micro-power
plants leads to more instability of the grids,
• Previously, these power suppliers were supposed to protect
themselves first in case of fault conditions (short circuit…)
• The grid supervisors demand from power suppliers to remain
connected to the grid in case of trouble on the distribution line bus
bars and also to contribute to the network stability when there are :
– Instabilities (regarding voltage & frequency)
– Voltage sags
– (Micro) short circuits, etc
• This means that, at least, the Protection strategy & the Genset
control sequence must change

157
 New grid code consequences

• Under the “Grid Code” label, we include all

possible requirements related to the constraint of
connecting a genset to a public supply network.
• Rules have changed over the past years and will
continue to change
• New regulations appear in each country.
• Impact on the genset is related to both running
conditions :
– static : enlarged running condition in steady state
– dynamic : Fault Ride Through (FRT),

158
New grid code consequences

• IEEE 1547-2018 - IEEE Standard for

Interconnection and Interoperability of
Distributed Energy Resources with Associated
Electric Power Systems Interfaces
– And related “dot” standards
• 1547.1 Test Procedures
• 1547.2 Application Guide
• 1547.3 Guide for Monitoring, Information Exchange, &
Control
• 1547.4 Guide for Design, Operation, and Integration
• 1547.6 Recommended Practice for Interconnecting DR
• 1547.7 Guide for Conducting Distribution Impact Studies
• 1547.9 Interconnection of Energy Storage
159
Grid Code Impact on Generator Design

GCTF
160
Grid Code Impact on Generator Design

Continuous operation region

Short-term operation

• IEC 60034/IEEE Std C50.12/13 voltage and

frequency operating range GCTF
161
Grid Code Impact on Generator Design
• Grid codes call for wider operating ranges
– Increased size leads to higher cost
IEEE 1547 Category II Operating Range Comparison
1.15
IEEE 1547
1.1

1.05
Voltage (PU)

1
IEEE C50.13

0.95

0.9

0.85
56 57 58 59 60 61 62 63
Frequency (Hz) GCTF
162
Grid Code Impact…: Power requirement
New Grid Code imposes new operating limits…
• Regarding reactive power production and absorption,
• Lagging & leading power factor requirement respectively…
Direct impact on alternator sizing… meaning :
• Xd max value criteria
Capability Curve
or
• Kcc min value criteria

More on this later…

Reminder: Xd = 1/Kcc

Maximum voltage

Minimum voltage
163
Grid Code Impact on Generator Design

• IEEE Std 1547-2018 voltage vs. time operating GCTF

range for Category II systems – more complex
164
Grid Code Impact on Generator Design
• Limits apply at grid connection point, not at
generator terminals
– Line and transformer drops may expand
required voltage operating range
• Grid codes are not harmonized (not even
within Europe) – each must be dealt with on a
case-by-case basis
• Even IEEE 1547, targeting North American
grid, is not uniform
– Three categories, negotiable between end
user and system operator GCTF
165
Grid Code Impact on Generator Design
• Power factor:
– May be required to supply 0.9 lagging
through 0.9 leading power factor – 0.85
lagging on generator side of transformer
– Less strict VAR requirement than normally
called for in island mode (0.8 PF lagging)
– May apply over voltage range (unrealistic)
• Short-circuit ratio:
– Requirement for (e.g.) 0.5 minimum may
lead to oversized machine, increasing cost
GCTF
166
Grid Code Impact on Generator Design
• Ride-through requirements:
– Voltage vs. time profile for remaining
connected

NERC PRC-024

GCTF
167
Grid Code Impact on Generator Design
• What happens?
– During a fault, not much power is being
delivered by the generator (low voltage)
– The prime mover can’t react, continues to
supply power at the same rate
– Extra energy goes to speed up the
generator at a rate determined by inertia
– Generator voltage gets out of sync with the
grid. If not too far out, it will slow down
once the grid comes back. If not, it may
continue to accelerate and slip poles.
168
Grid Code Impact on Generator Design
• Must remain connected and in synchronism
during and after fault
– May require higher inertia (to reduce speed
variation), lower reactance (= higher short-
circuit ratio SCR, to help pull back into
sync on recovery), or both
– Causes high stresses on windings and
shaft, may require stronger construction
– High-response excitation will improve
ability to resynchronize
GCTF
169
Grid Code Impact on Generator Design
• Generators providing grid support may
encounter new service factor requirements:
– More frequent stopping and starting
– More load variability while running (both
MW and MVAR)
• Make sure generator vendor is aware of all
the requirements – simply calling out the grid
code is not sufficient
• Helpful to involve generator manufacturer in
negotiation with power company
GCTF
170
5. Synchronization and Paralleling

171
Synchronization and Paralleling
• Why operate generators in parallel?
– Add capacity:
• Limitations on engine/generator size
• Modularity for expansion, flexibility of operations
• Multiple locations
– Redundancy
• N+1 allows one set to fail or be maintained while others
deliver required load power
• Improves reliability from 98% to 99.96% (typical system)
– Efficiency
• During light load conditions, sets can be shut down so
remaining ones operate closer to full load
• Prevent “wet-stacking” and engine damage

172
Synchronization and Paralleling
Synchronization
• Involves preparation of two power buses for being
connected together. One is usually just called the
“bus” and the other, “oncoming”, is the machine or
set of machines being added to the bus
• Transient currents and torques occur that depend on
the instantaneous voltage difference between the two
buses at the moment of connection, and the
impedance of the two buses
• Voltage and phase angle contribute to this voltage
difference
• Speed (i.e. frequency) difference is also important

173
Synchronization and Paralleling
• Difference is due to
combination of angle
and voltage.
• 240V difference is
about half of rated
voltage, so will produce
about half of normal
short-circuit current and
torque.
• 10-15º and 5% voltage
difference are typical
generator requirements
for synchronization – if
system will exceed this,
extra bracing needed

174
Synchronization and Paralleling
Parallel Operation
• When a generator is operating in parallel with a bus
(or another generator), it must be controlled in order
to balance both reactive (VAR) and real (power) load.
• Real power is controlled entirely by the engine or
prime mover via the governor. The generator itself
has no control over power.
• Reactive load is controlled by the generator
excitation alone. The prime mover can’t control the
VARs (much).

175
Synchronization and Paralleling
Parallel Operation – Droop Mode
• Droop mode is the simplest method of parallel
operation.
• For control of VARs, the voltage regulator is set so
that as the voltage decreases, it applies more
excitation (attempting to increase the voltage), but
allows the voltage to decrease slightly in proportion
to the VARs being drawn by the load. Requires a
current transformer to sense the reactive current.
• The voltage serves as the signal that controls the
generator excitation, and varies with reactive load.

176
Synchronization and Paralleling
Parallel Operation – Droop Mode
• Similarly, for the engine, if the speed decreases, the
governor applies more fuel (attempting to increase
the speed), but allows the speed to decrease slightly
in proportion to the power being drawn by the load.
• The speed serves as the signal that controls the
engine, and it varies with power load.
• Droop mode is best for systems with widespread or
diverse generation, or paralleled with the grid.

177
Synchronization and Paralleling
Parallel Operation – Cross-Current Mode
• Cross-current mode is more complex and requires a
separate channel of communication between
generators.
• For control of VARs, the voltage regulators are cross-
connected to that as the reactive current becomes
unbalanced, more excitation is applied to the
machine producing fewer VARs and less to the
machine producing more, rebalancing the currents.
The voltage remains constant.
• The signal that controls the generator excitation is
carried on a separate pair of wires.

178
Synchronization and Paralleling
Parallel Operation – Isochronous Mode
• For the engine, this mode is known as isochronous or
constant-speed. If the power becomes unbalanced,
the governor applies more signal to the lower-power
machine, and decreases the power to the higher
power machine. Speed does not vary.
• The signal that controls the engine is carried on a
separate channel between the governors.
• As far as I know, nothing prevents these two types of
control from being mixed (e.g. droop voltage and
isochronous speed).
• Isochronous/cross current modes are best for
systems of like machines in close proximity.
179
Synchronization and Paralleling
• Paralleling of generators with different third harmonic
may require special attention as explained
elsewhere.
• Accidentally synchronizing generators with phase
angle or voltage difference beyond the recommended
range may cause internal damage to the generators
due to high currents, or to the shaft or engine due to
high torques.
• A method of synchronizing several generators exists
where the generators are connected to the (dead)
bus and excitation is applied before the engines are
started. As the engines come up to speed, the
generators are drawn naturally into synchronism.

180
Grounding Methods (Short Diversion)
• Grounding affects circulating currents
• IEEE Std 142, “Recommended Practice for
Grounding of Industrial and Commercial Power
Systems” (“Green Book”) shows several
methods for grounding power systems
– Ungrounded (no intentional ground)
– High- or low-resistance grounding
– Reactance grounding
– Ground-fault neutralizer (resonant ground)
– Solid grounding
• Each of these has advantages and detriments
181
Grounding Methods
• Ungrounded (no intentional ground)
– Allows operation with a single ground fault
– Simple and inexpensive
– No 3rd harmonic circulating current
– Used on Navy ships for LV power system
– May lead to overvoltages with arcing faults
– Uncontrolled neutral voltage

182
Grounding Methods
• Low-resistance grounded
– Limits ground fault current to no more than three-
phase fault current (protects generator from
physical stresses due to high current)
– Ground fault cleared by coordinated breaker
– Prevents transient overvoltages
– Multiple sources multiply available ground fault
current unless switched grounds used
– Reduces third-harmonic current – but grounding
resistor must be sized for steady-state 3rd harmonic
current if present (standard requires 15%
continuous)

183
Grounding Methods
• High-resistance grounded
– Has same benefits as low-resistance
– Also allows operation for limited time with
one ground fault present
– Limits current in generator internal fault,
reducing damage
– Reduced continuous 3rd harmonic current
(reduces power dissipation)
– Multiple grounds usually are no problem
– Requires separate ground-fault detection
means (resistor voltage or current will do)
184
Grounding Methods
• Reactance grounded
– Similar to resistance grounded, lower loss
• Ground-fault neutralizer
– Mainly on systems above 15 kV, excluding
most generators
• Solidly grounded
– May allow very high fault current to flow in
the generator windings in case of a single-
phase-to-ground fault – possibly causing
generator damage (NEMA MG-1 says
windings braced for 3-phase fault only)
185
Grounding Methods
• Hybrid grounded (recent introduction)
– External or switched low-resistance ground
combined with high-resistance generator
neutral ground
– Combines benefits of both systems
– Requires careful design

186
Grounding Methods -- IEC
• IEC 60364 Grounding Arrangements
– First letter: T for “earthed”, I for “isolated”
– Second letter: T for directly earthed, N for connected
to the earthed neutral at origin
– TN-C: Combined ground and neutral, like old-style
dryer hookup
– TN-S: Separate ground and neutral wires connected
together at source, similar to normal NA practice
– TT: Like TN-S but not requiring bonding neutral to
common ground, using GFCI for protection
– IT: Isolated from ground with ground fault detector

187
Harmonics and Paralleling
Fundamental & Harmonic Voltages
vs. Pitch (Knoltons Handbook)

188
Harmonics and Paralleling
Effects of Generator Pitch
• Pitch factors for reduction of various
harmonics:
Harmonic Number
Pitch Fund 3 5 7 9 11 13
2/3 0.866 0.000 -0.866 0.866 0.000 -0.866 0.866
11/15 0.914 -0.309 -0.500 0.978 -0.809 0.105 0.669
3/4 0.924 -0.383 -0.383 0.924 -0.924 0.383 0.383
7/9 0.940 -0.500 -0.174 0.766 -1.000 0.766 -0.174
4/5 0.951 -0.588 0.000 0.588 -0.951 0.951 -0.588
5/6 0.966 -0.707 0.259 0.259 -0.707 0.966 -0.966
6/7 0.975 -0.782 0.434 0.000 -0.434 0.782 -0.975
7/8 0.981 -0.831 0.556 -0.195 -0.195 0.556 -0.831
Full 1.000 1.000 1.000 1.000 1.000 1.000 1.000

(Harmonic in flux wave is multiplied by reduction factor.)

(Also reduced by distribution, skew)
189
Circulating Harmonic Currents
• Circulating current: When two
generators (or generator and utility) are
connected in parallel and third
harmonic current flows between them + Ic +

depending on zero-sequence V3A V3B

_ _
impedance and difference of 3rd
harmonic voltages. Per phase:
XoA XoB
V3A - V3B
IC =
3 • (X0A + X0B)
Where:
- V3 is L-N third harmonic voltage.
- X0 is zero-sequence reactance of each
source (including external grounding
reactance).
190
Circulating Harmonic Currents

• Higher harmonics (5th and up)

are not usually significant:
– 5th, 7th, etc. don’t flow in + +
Ic
neutral lead.
V3A V3B
– Higher impedance at _ _
higher frequencies means
harmonic current is lower.
XoA XoB
• Third, 9th, etc. harmonic flows
in neutral.
• May cause overheating of
neutral lead or grounding
resistor, false tripping of
differential protection relay
and other problems. 191
Circulating Harmonic Currents
• What to do about this?
– It may not be a problem
– Neutral current splits three ways, so 100A in
neutral is only 33A in phase leads
– Adds as RMS, so 10% extra phase current
only adds 2% to heating
– Neutral grounding device will reduce current
• Must be rated for continuous 3rd harmonic ~15%
– Otherwise, third harmonic voltages, not pitch,
should be matched to reduce circulating
current – consult manufacturer
192
Selection of Correct Pitch (cont.)

• Why not just use 2/3 pitch?

– Low zero-sequence impedance will cause generator to attempt

to absorb any third-harmonic current on bus from other sources.
Must balance third harmonic voltage to prevent circulating
current.

• Why do other manufacturers have a different recommendation?

– If building mainly low-voltage machines, these are usually solidly

grounded and will more likely require 2/3 pitch for utility
paralleling.

– May have made design decision to allow high third harmonic in

flux wave, and reduce by making all units 2/3 pitch.

193
Selection of Correct Pitch (cont.)
2/3 Pitch
• Kato generators are designed for optimum pitch
unless otherwise specified
• Pole pitch is the angle between adjacent poles, i.e. 4
pole = 90º mechanical
• Winding pitch is the coil span divided by the pole
pitch, i.e. 60º / 90º = 2/3
• Optimum pitch balances generator performance while
effectively utilizing generator active materials
• 2/3 pitch used when paralleling different generators
and in conjunction with nonlinear L-N loads. May
require derating.
194
Selection of Correct Pitch (cont.)
• Paralleling directly with utility bus, solid ground, 480V up:

– Normally no line-neutral loads, no load third

harmonic generated locally.

– No third harmonic from utility due to transformer

connection.

– Use 2/3 pitch to minimize circulating current with

utility.

195
Selection of Correct Pitch (cont.)
Selection of Correct Pitch (cont.)
• Paralleling with utility, high impedance grounding:

– Third harmonic circulating current cannot flow

through neutral.

– Use optimum pitch to minimize cost.

– In most cases systems with high-impedance

ground will not require 2/3 pitch. Consulting
engineer for project should determine allowable
neutral voltage. Adjust pitch to reduce 5th and
7th harmonics.

196
Selection of Correct Pitch (cont.)

• Paralleling with other generators, same type:

– Since all generators matched, no circulating

current.

– Use optimum pitch to minimize cost.

– Note: if also paralleling with external grid

through grounded primary transformer, 2/3
pitch may be required.

197
Selection of Correct Pitch (cont.)
• Paralleling with other generators, different type, solidly
grounded:
– Must know third harmonic voltage on bus.
– Select pitch and slots to match third harmonic
voltage, minimize circulating current.
– May not be same pitch as other generators
on bus. Depends on pole shape, saturation,
other factors.
– Must involve manufacturer and end user in
design.
– Possibly use switchable grounds, or
grounding transformer

198
Selection of Correct Pitch (cont.)
• Not paralleled, but high harmonic line-neutral load:

– Normally 120/208 volt or possibly 277/480V

systems.

– Use 2/3 pitch to minimize zero-sequence reactance,

to reduce third harmonic L-N voltage.

• Not paralleled, but only line to line loading:

– Many industrial 480V systems and almost all

medium- and high-voltage systems.

– Use optimum pitch to minimize cost.

199
6. Reference Frame Theory

200
 Intro to Reference Frame Theory
Short Version

• Three phase balanced set of AC currents

creates approximately a moving wave of
sinusoidal flux which travels around the stator
at a rate determined by the frequency.
• In a synchronous machine, the rotor rotates
at the same (average) speed as this flux
wave. If it lags behind, it is pulled by the flux
and acts as a motor. If it is pushed out ahead
of the flux, it delivers power as a generator.

201
 Intro to Reference Frame Theory
Short Version (from Gen School)
Imagine you are on a new theme park
ride (“The Synchronous Generator!
Amazing!”)
• You take your seat on the rotor of the
generator and put on your magic 3D glasses
which allow you to see the magnetic lines of
force, and the currents.
• The machine starts and comes up to 1800
RPM…*
*Only 1500 in Europe, which is why EuroDisney is boring.

202
 Intro to Reference Frame Theory
Short Version

When excitation is applied to the generator

with no load you see lines of flux going
straight up from the rotor to the stator.
• They are most concentrated at the center of
the rotor.
• There are few between the poles.
• On the adjacent poles they go from the
stator to the rotor.
• There is no current in the stator.

203
 Intro to Reference Frame Theory
Short Version
• If a lagging zero-power-factor load* is applied.
– You see stator current ahead of you, going
to the right, and behind you going to the
left, but none near the center of the pole.
– The location of the current is steady from
your point of view, even though the
surface of the stator is passing by at high
speed.
– The lines of flux are reduced, because the
stator current is in the opposite direction
from the rotor current.
* Lagging power factor means that the peak of the current occurs after
the peak of the voltage. Leading power factor means the peak of the
current occurs before the peak of the voltage.
204
Intro to Reference Frame Theory
Short Version

205
 Intro to Reference Frame Theory
Short Version

• If a leading zero power factor load is applied:

– You see stator current ahead of you, going
to the left, and behind you going to the
right, but none near the center of the pole.
– The location of the current is still steady
from your point of view.
– The stator current flows the opposite
direction, and the flux is strengthened.
• In both cases, flux is high where the current is
low, and vice versa, so no torque developed.
206
Intro to Reference Frame Theory
Short Version

207
 Intro to Reference Frame Theory
Short Version

• If a real (unity power factor) load is applied:

– You see current from left to right over the
center of the pole head -- where there is
high flux -- and low current between poles.
Current x flux generates force (power).
– The current in the stator pulls the lines of
flux backward, crowding them into the
trailing edge of the pole head.
– Unity power factor load does not have a big
effect on the total flux.
208
Intro to Reference Frame Theory
Short Version

209
 Intro to Reference Frame Theory
Short Version
• Summary of generator loading:
– Lagging power factor (inductive) current
demagnetizes the rotor, requiring higher field
current to maintain flux and voltage. It does not
produce torque or power.
– Leading power factor (capacitive) current aids the
magnetization of the rotor. So less field current is
required. It does not produce any torque or power.
– Unity power factor (resistive) current neither
magnetizes not demagnetizes the rotor, so field
current is not affected (much). The retarding
torque produced uses mechanical power and
produces electrical power.
210
Intro to Reference Frame Theory
Rotational Transformations in 2-Space
Let x⃗ ∈ R2. We will need to find equations
relating x⃗ =(x,y) to its image w⃗ =(w1,w2) under the a
rotational transformation T. Let ϕ be the angle between
the positive x-axis and x⃗ , and let θ be the angle
between x⃗ and w⃗ . The length of both vectors is R.

Skip ahead ->

211
Intro to Reference Frame Theory
Note that we can calculate the components of our
vector x⃗ =(x,y) with the polar equations
x = R cosϕ and y = R sinϕ (where R = ||X||) - both of
which were derived by basic trigonometry. We can also
calculate the components of w⃗ =(w1,w2) from these
equations:
w1=Rcos(θ+ϕ) and w2=Rsin(θ+ϕ).

Using the following trigonometric identities:

(1) cos(α+β)=cosαcosβ−sinαsinβ
(2) sin(α+β)=sinαcosβ+cosαsinβ
we can write w1 and w2 as follows:
w1=Rcosθcosϕ−Rsinθsinϕ
w2=Rsinθcosϕ+Rcosθsinϕ
212
Intro to Reference Frame Theory
Lastly we substitute x=Rcosϕ and y=Rsinϕ to get:
w1 = x cosθ−y sinθ
w2 = x sinθ+y cosθ

It thus follows that if w=Ax, then our standard matrix A =

cos θ − sin θ
, and transformation in matrix form is:
sin θ cos θ

𝑤𝑤1 cos θ − sin θ 𝑥𝑥1

W = Ax or =
𝑤𝑤2 sin θ cos θ 𝑥𝑥2

The zero-axis terms don’t rotate, so don’t have to be

transformed.

213
Intro to Reference Frame Theory
Use the Clarke transform to convert to direct/
quadrature/zero axes:
The Clarke transform (named after Edith
Clarke) converts vectors in the ABC reference
frame to the XYZ (often αβz) reference frame.
The primary value of the Clarke transform is
isolating that part of the ABC-referenced vector
which is common to all three components of the
vector; it isolates the common-mode
component (i.e., the Z component). The power-
invariant, right-handed, uniformly-scaled Clarke
transformation matrix is [from Wikipedia]:
214
Intro to Reference Frame Theory

uxyz = Kcuabc so uabc

(three-phase) becomes
uxyz (orthogonal ).

The non-power-invariant form is more common in older

texts, it just removes the square root sign, and preserves
voltages or currents instead of power. Note the quantities
are amplitudes, not RMS values, in fundamental system
(dimensionless in per-unit system).

This is a similar process with vectors to what we do with

phasors in symmetrical components. Or is it the other way
round? Anyway…
215
Intro to Reference Frame Theory

Then we change all stator quantities to rotor

reference frame (stop the rotation).
This is the Park transformation. It multiplies the
instantaneous vector quantities by reference vectors
rotating (normally) at the same speed as the rotating
MMF from the field winding:

So uxyz (rotating) becomes udqz (stationary).

216
Intro to Reference Frame Theory
• When θ = ωt, the reference frame is rotating
at ω radians/sec. If this is the rotor speed,
the new reference frame is stationary with
respect to the rotor, and all steady-state
stator quantities become DC.
• This is similar to the way traditional DC
motors actually work: the field is stationary,
and the commutator effectively performs a
transform on the rotating armature currents
so that they become stationary as well.
• There are a lot of good articles online for
further reading (including Park’s original
1926 article).
217
Intro to Reference Frame Theory
• By making orthogonal D and Q axes
stationary with respect to the rotor, we get rid
of all of the time-varying inductances and
align the analysis with the physical structure
• Note that we can rotate the reference frame
at any speed we want, and declare any axes
we want – but for synchronous machines
only the rotor reference frame and D-Q axes
allow us to save labor.
• With high-speed computing power available,
it is possible to analyze a synchronous
machine in the stator reference frame, with
all three phases, but it’s still a pain.
218
Reference Frame Theory – the Details
• Winding distribution in ideal machine
– Current in a conductor creates magneto-
motive force (MMF, or magnetic field H),
which is the driving force for magnetic flux
– Units of MMF are ampere-turns, i.e.
multiple turns with current in each behave
as the sum of the individual currents.
– MMF at a location equals the net current
enclosing that location
– Winding distribution defines a winding
function giving MMF over a path (gap)
219
Intro to Reference Frame Theory
• Here we will skip over a lot of important
material including:
– Turns and winding functions, winding factor
(for harmonics and fundamental), MMF
– Winding distributions and pitches
– Inductances of cylindrical machine
– Effective gap, flux distribution, and how to
deal with shape of salient pole
– Leakage inductance calculations
• These are properly covered in a semester
class
220
Generator Equivalent Circuits
• The Clarke and Park transforms can be
applied to any kind of electrical circuit –
including passive networks – but provide the
most benefit with synchronous machines.
• Equivalent circuits for the transformed
synchronous machine have been developed
• These are models of the machine – the
elements of the model don’t exist inside the
machine as discrete entities. There are
many circuits that can be used to model a
particular machine. Some may provide a
better fit than others. All have their uses.
221
Generator Equivalent Circuits
Traditional model for a salient pole machine:
We can write
this circuit as a
set of
differential
equations, then
solve them to
find the
response of the
circuit.
Riaz (UofM)
uses Spice to
simulate the
circuit.
222
Generator Equivalent Circuits
• Equations of this model:
From Krause, Analysis of
Electric Machinery
Ψ values are “flux linkages
per second” (~voltages)
The “r” subscript means rotor
reference frame, the prime
(“’”) means transformed to
stator base. In per-unit, as
(solid only)
we normally use, this is
always the case, so I leave
them off.
“p” is differential operator
(solve for those terms).
223
Generator Equivalent Circuits
• Equations of this model (psi equivalences):
Normally we solve the
differential equations
for the “psi” quantities,
then calculate the
currents. This means
you have to invert the
inductance matrix used
to calculate ψ

(Disregard “kq2” quantities for laminated rotor)

λ is flux linkage, which is inductance times current, Li. Ψ is
flux linkage per second, ωLi. Since ωL is X, and Xi is
voltage, these are effectively voltages.
224
Generator Equivalent Circuits
• Let’s look at one of the equations:

The last term represents the transient behavior of the d-

axis flux linkage – a change in either the inductance or
the current will result in transient voltage (product rule).
The previous term is called a “speed” voltage, because
it is proportional to speed. It represents the generator
output due to the rotation of the rotor. The first term is
just the voltage drop in the resistance.
Notice that Q-axis flux results in (negative) D-axis
voltage, and D-axis flux results in Q-axis voltage. The
flux is changing fastest when its value is zero.
225
Generator Equivalent Circuits
• Showing added rotor circuit for solid pole

226
Generator Equivalent Circuits
Steady-state (synchronous) terms:

227
Generator Equivalent Circuits
Transient terms add field circuit (and one rotor
circuit for solid pole)

(Close to Xls + Xlfd)

Field source is low
impedance, so “shorted” as
far as transient goes

228
Generator Equivalent Circuits
Subtransient is result of all windings:

229
Generator Equivalent Circuits
Adding “Canay” reactance:
Better prediction
of rotor field
currents.
Represents the
flux that links
cage, but not
winding, and
vice-versa.
Can’t be
measured
accurately using
short-circuit test.

The model is not

the machine!

230
Solving the model equations:
• First rearrange them so the derivative term is on the
left hand side.

231
Solving the model equations:
• Then change to matrix form:

You can use currents directly as state variables. The equations get
more complicated, and the solution is slower, but you don’t have to do
the current conversion. 232
Solving the model equations:
• A similar treatment is used to obtain the transforms
between flux linkages and currents in matrix form
• These equations may be made into a function that
takes the V (input) and Ψ (state) vectors as inputs,
and produces the derivatives of Ψ as outputs
• If voltage (V) vector is not known, it must be
determined by iterative methods (if load is known, it
may be incorporated into the model and V set to 0)
• The initial values are determined from steady-state
conditions, by setting the derivative terms to zero
and solving for state variables
• A numerical differential equation solver can be used
to solve these equations (Matlab/Octave/Numpy)
233
Solving the model equations: Saturation
• Saturation of the magnetic circuit has a very
significant effect on the behavior of the machine
• Basic models only consider D-axis saturation, by
changing the value of the magnetizing reactance as
a function of the excitation
• More elaborate models consider effect of D-axis
saturation on the Q-axis reactance
• Different fitting functions are used for saturation
curve
– The best one is E = AV 1 + α𝑒𝑒 β𝑉𝑉 where A is
i
the slope of the air-gap line, α and β are fitting
constants, and V is the terminal voltage

234
Solving the model equations: Saturation
• You will see requirements for NERC code names:
“The GENROU, GENSAL, GENTPF, and GENTPJ models represent round rotor
and salient pole synchronous machines. The predominant difference between the
GENROU/GENSAL and GENTPF/GENTPJ models is how they account for
saturation.
• The GENSAL model uses simplifying approximations that significantly
compromise treatment of magnetic saturation. The GENSAL model ignores
saturation on the q-axis completely. In both the GENROU and GENSAL models,
saturation is a single additive terms. The GENROE and GENSAE models use the
same treatments of saturation as GENROU and GENSAL; the only difference is
that they fit saturation with an exponential rather than quadratic curve.
• The GENTPJ and GENTPF models use approximations in their treatments of
saturation, but are more accurate than GENSAL and GENROU. In these models,
saturation is multiplicative on all inductance terms. GENROU and GENTPF do
not fully recognize the effect of stator current on saturation.
• The GENTPJ1 model recognizes the effect of stator current on saturation by
including an additional parameter, Kis, which appears in the saturation function as
shown in [3] and [5].”

235
Solving the model equations
• Refer to IEEE Std 1110 and the NERC documents
for more details on these models
• Different power system analysis packages have
different requirements for entry of generator
parameters. We aren’t intimately familiar with any
of them, so can’t always advise as to the correct
value to put in a particular field.
• Fortunately, many of the more obscure ones don’t
make a lot of difference to the response

236
Simulink simulation of model:
• Matlab/Simulink allows you to build a model by
graphically connecting function blocks. Other
software packages (Labview, Simula, Scilab) have
similar capability).
• The software does the job of assembling the
differential equations and solving them “behind the
scenes” for you
• Simulink has an extension “Simscape” and further,
“Simscape Electrical” that contain pre-assembled
model blocks for synchronous and other electrical
machine types. These also take care of the
Park/Clarke transforms, loads, etc..

237
7. Excitation System Models, IEEE Std
421.x, and Exciter Response

238
Exciter AC8C Model
• The models of excitation systems defined in IEEE Std 421.5
(most recent 2016) are appropriate to different kinds of
exciters and regulators. The AC8C (latest) model is becoming
popular for PID regulators used with synchronous brushless
exciters such as are normally used on Kato generators. It
represents the exciter adequately for system modeling
purposes.
• The exciter model is common to all the AC(N) models, so the
following is applicable to any of these.
• The definition of exciter constants in the Standard is left to the
manufacturer of the exciter. Unfortunately, no guidance is
given for how to calculate those constants. Correspondence
with the Working Group for this standard has yielded no
further clues, so we are apparently on our own as regards the
interpretation of the standard.

239
Exciter AC(N)(X) Model

Exciter model from

the AC series of
excitation system
models, IEEE Std
421.5

Vr is the regulator
output voltage, Ifd is
the main field current,
Efd is the exciter
output voltage to the
main field (all in PU)

240
Exciter AC(N)(X) Model

SE is the exciter
saturation function.
We use an
exponential function
if it is needed, but
most of our exciters
have such low
saturation under
normal conditions,
that it can be ignored
(set to zero).

241
Exciter AC(N)(X) Model

VEMAX is the exciter

voltage limit. This is
already defined by
the saturation
function (if used), and
anyway, the regulator
usually saturates
before the exciter, so
this is usually set to a
high value like 99.

242
Exciter AC(N)(X) Model

TE is the exciter time

constant. We use the
open-circuit value, to
be conservative. It is
typically 0.1-0.5
seconds depending
on the size of the
exciter.

243
Exciter AC(N)(X) Model

KE is the exciter
(feedback) gain
constant.

244
Exciter AC(N)(X) Model

KD is the exciter
droop constant. This
is a nonlinear term,
but the model uses a
linear approximation.

245
Exciter AC(N)(X) Model

FEX is the rectifier

commutation
function. It is ~
constant for a given
frequency and value
of main field
resistance. We don’t
use this function at
the moment (we set
KC equal to 0.0), but
we include the effect
in the KD constant.

246
Exciter AC(N)(X) Model Parameters
Three regions
of operation,
depending on
X/R ratio
Most Kato
exciters
operate in or
near MODE 1,
where the
effect can be
(mostly)
subsumed into
KD

247
Exciter AC(N)(X) Model Parameters
• Real rectifier function equations:

248
Exciter AC(N)(X) Model

For convenience I
1−𝐼𝐼
use Fex ~ 1−𝑢𝑢𝑢𝑢𝑁𝑁 ,
𝑁𝑁
where u ~ 0.563
(fitted curve)

249
Exciter AC(N)(X) Model

KC is the rectifier loading

factor

250
Exciter AC(N)(X) Model Parameters
• The standard defines 1.0 PU input (VR) as the input voltage
required to produce 1.0 PU output on the unloaded air gap
line of the exciter saturation curve.
– This means that under normal conditions, 1.0 PU input
does not produce 1.0 PU output from the exciter!
– A per-unit conversion is needed to make this come out
right.
– This definition is equivalent to requiring the constant KE to
be equal to 1.0, assuming the value of SE tends to zero at
low values of excitation (which it should).
• SE is the saturation function for the exciter (different programs
use different saturation functions, and IEEE 421.5 doesn’t
prescribe a particular function)
• Best SE formula we have found is 𝑆𝑆𝐸𝐸 = 𝛼𝛼𝑒𝑒 𝛽𝛽𝐸𝐸𝐹𝐹𝐹𝐹 , which fits
tested curves very well

251
Exciter AC(N)(X) Model Parameters

SE represents the PU increase

in exciter input voltage to
produce a particular no-load
output, above the value on the
air gap line.
(from IEEE Std 421.5-2016)
Shown as function of EFD, it
must be inverted numerically
If the exciter is not very
saturated at the regulator
ceiling voltage, SE can be
disregarded (set to zero)

252
Exciter AC(N)(X) Model
• Kd in the AC models appears as a linear term,
whereas in the actual exciter it produces a curve
depending on the effective field impedance and the
effective displacement power factor of the rectifier
(which are variable in the transient condition).
• We assume if we match the exciter performance
when it is unloaded and fully-loaded the errors in-
between won’t matter much.
• Using KC for part of the feedback will put some
“bend” in the curve.

253
Exciter AC(N)(X) Model

254
Exciter AC(N)(X) Model
• We believe that normally the exciter constants should be
calculated hot as this is the usual operating condition.

255
Exciter AC(N)(X) Model
• With ceiling voltage input, call output on air gap line = Eocu,
and on unloaded saturation curve = Eoc. If they differ by more
than (say) 5%, then the exciter is saturated at normal
operating conditions and Kd and Ke calculations may have to
consider saturation (this is a judgement call depending on how
critical the requirement for accuracy of the model is).
• Calculations: From the exciter saturation curves, determine
the exciter field current values for:
– Ceiling voltage unloaded (in case the regulator is not
known, use the standard values).
– From the unloaded ceiling voltage, determine Se(max)
unloaded.
– From 75% of the unloaded ceiling, determine Se(0.75
max)

256
Exciter AC(N)(X) Model
• Calculations:
– KD may be calculated from steady-state saturation curves:
– Since at steady-state the integrator gain is infinite, and its
input will be zero,
– 𝑉𝑉𝑅𝑅 − 𝑆𝑆𝐸𝐸 𝑉𝑉𝐸𝐸 − 𝐾𝐾𝐸𝐸 𝑉𝑉𝐸𝐸 − 𝐾𝐾𝐷𝐷 𝐼𝐼𝐹𝐹𝐹𝐹 = 0 from the model diagram
– 𝐼𝐼𝐹𝐹𝐹𝐹 = 𝐸𝐸𝐹𝐹𝐹𝐹 at steady state, and FEX is constant, so
𝐸𝐸
𝐼𝐼𝐹𝐹𝐹𝐹 = 𝐸𝐸𝐹𝐹𝐹𝐹 = 𝑉𝑉𝐸𝐸 𝐹𝐹𝐸𝐸𝐸𝐸 and 𝑉𝑉𝐸𝐸 = 𝐹𝐹𝐹𝐹
𝐹𝐹𝐸𝐸𝐸𝐸
𝑉𝑉𝑅𝑅
– 𝑉𝑉𝑅𝑅 − 𝑆𝑆𝐸𝐸 + 𝐾𝐾𝐸𝐸 + 𝐾𝐾𝐷𝐷 𝐹𝐹𝐸𝐸𝐸𝐸 𝑉𝑉𝐸𝐸 = 0 so = 𝑆𝑆𝐸𝐸 + 𝐾𝐾𝐸𝐸 + 𝐾𝐾𝐷𝐷 𝐹𝐹𝐸𝐸𝐸𝐸
𝑉𝑉𝐸𝐸
𝑉𝑉𝑅𝑅 𝑆𝑆𝐸𝐸 +𝐾𝐾𝐸𝐸 +𝐾𝐾𝐷𝐷 𝐹𝐹𝐸𝐸𝐸𝐸 1+𝐾𝐾𝐷𝐷 𝐹𝐹𝐸𝐸𝐸𝐸
– = = if SE = 0 and KE = 1
𝐸𝐸𝐹𝐹𝐹𝐹 𝐹𝐹𝐸𝐸𝐸𝐸 𝐹𝐹𝐸𝐸𝐸𝐸
𝑉𝑉𝑅𝑅 1
– 𝐾𝐾𝐷𝐷 = −
𝐸𝐸𝐹𝐹𝐹𝐹 𝐹𝐹𝐸𝐸𝐸𝐸
𝑉𝑉𝑅𝑅
– can be determined from points on the unsaturated part
𝐸𝐸𝐹𝐹𝐹𝐹
of the unloaded vs. loaded saturation curves
257
Exciter AC(N)(X) Model
• Calculations:
– KC and hence FEX is a function of the X/R ratio of the
exciter (KC ~ XL/RLOAD), and 𝐹𝐹𝐸𝐸𝐸𝐸 𝑆𝑆𝑆𝑆 ~𝑒𝑒 −0.504𝐾𝐾𝐶𝐶 for
reasonable values of KC (<=1)
– If result gives a negative value for KD, you need to
decrease KC
– TE will be the exciter value of T’do during transients only
when the exciter output current is close to zero, but will be
somewhat lower otherwise. For conservative estimation of
recovery time etc., we recommend using T’do for transient
response calculations.

258
 Exciter Response Ratio
• Exciter response ratio is a measure of how much output is
available from the exciter after half a second under forcing
conditions, compared to its normal output.
• Based on 1959 AIEE article by V.C. Strode. Now embodied
in IEEE Std 421.2-2014.
• Meant as a metric for comparison of excitation systems.
• 0.5 second is a long time for a modern excitation system,
except maybe for power-plant size machines.
• Can’t really be calculated accurately without a full magnetic
model of generator and exciter, including rotating rectifier.
• Measuring response ratio requires a fairly elaborate setup
• Some useful approximations can be made.

259
Exciter Response Ratio
1. We assume that the exciter is not highly saturated. This is a
good approximation for most Kato exciters, and is
conservative if the exciter is saturated.
2. We assume that the generator field is purely resistive. This
is definitely not the case, but that is the way the exciter is
tested on its own, and the actual field inductance will result
in higher response ratio, so it is conservative.
3. Absent any better value, we use the exciter T’do as the
exciter time constant. The loaded time constant will be
shorter so the response will be faster. This is also
conservative.

260
 Exciter Response Ratio
• Procedure: operate the exciter (usually with resistive load) at
its nominal full-load point, then suddenly increase the exciter
field voltage to the regulator ceiling value.
• Take a recording of the exciter output voltage under these
conditions, for at least half a second.
• Draw a straight line from the initial point such that the area
under the line, between time = 0 and time = 0.5 second, is the
same as the area under the exciter response curve (in the
past this was generally an “eyeball” estimation). The value of
voltage at the end point of the line, divided by the full-load
starting voltage, less one, times 2, is the response ratio
(second-1).

261
Exciter Response Ratio

Typical Response Ratio Test

262
 Exciter Response Ratio
• Nowadays we would do a numerical integration based on
data acquisition output to determine the test result.
• Normally we do not perform this test (it requires slip rings and
is expensive). We can get a good approximation of the result
by calculation based on exciter data sheet. A spreadsheet
can be used to perform this calculation.
• Alternatively, a nomograph is available to calculate response
ratio.

263
Exciter Response Ratio

Typical Response Ratio Test

264
8. Synchronous Condensers (or
Compensators)

265
Synchronous Condensers (Compensators)
• What is a synchronous condenser?
– Nothing more or less than a synchronous motor or
generator (NB: they are the same thing), when
operated without a prime mover in order to deliver
(or absorb) reactive “power” (volt-amperes)
– May be optimized for low losses and zero power
factor operation
– Surplus generators are often pressed into service
as synchronous condensers
– Probably need some changes in protection and
controls
– Will need some kind of starting system

266
Synchronous Condensers (Compensators)
• Purpose of synchronous condensers
– Distributed generation using inverters, which until recently
have had difficulty supplying VARs (volt-amperes reactive)
at the level required
– Providing line “stiffness” for momentary overloads and fault
support
– Providing inertia, which is lacking in most distributed
resources
– Solid-state devices such as System VAR Controllers (SVCs)
and Static Compensators (StatComs) can provide or sink
VARs but do not provide inertia or fault current, and are
often more expensive
– Recent changes in “grid codes” mandate ability to supply
and absorb VARs to stabilize the grid

267
Synchronous Condensers (Compensators)
• Starting methods (see also M-G sets)
– Asynchronous starting, using cage or solid rotor to
start as induction motor
• Across-the-line or reduced-voltage start
• Energize field at around 95% speed to pull machine into
synchronism
– Variable-frequency drive
• Low starting current and fast starting time
• Most expensive option, but drive sized for partial rating
can be used
– Pony-motor start
• Low starting current, may run on auxiliary source
• Drive sized for pony motor rating so can be economical

268
Synchronous Condensers (Compensators)
• Design and construction
– Because the synchronous condenser is simply a
generator, often a surplus machine repurposed
from a power plant, the design process is very
similar to a generator design. For a purpose-built
condenser, there are opportunities for
optimization. In particular, losses should be
minimized, and ability to accept VARS may be
useful.

269
)
Comparison of Synchronous Condenser with FACTS
Characteristic Synchronous Condenser Fixed Reactance Static VAR Compensator STATCOM
Cost Moderate Inexpensive Moderate Expensive
Speed Moderate Slow Fast Fast
Reliability High High Moderate Low-Moderate
Durability High Moderate Moderate Low
Longevity Very high High Moderate Low
Size Large Small Moderate Moderate
Inertia Moderate (lg w/ flywheel) None None None
Lagging Cap'y Low As required High High
Controllability High Low High High
Efficiency Moderate High High High

270
 Questions?
[email protected] or [email protected]

271