You are on page 1of 24

Technical

Topics
Volume V Issue 1

Starting Large Motors from an Onan Engine-Generator


Set

Larry Bey
Technical Marketing Specialist
Jim Iverson
Manager, Technical Sales
Synopsis:

A common problem in sizing engine-generators is starting large motors (large relative to the
generator set capacity). Typical problems include holding coils dropping out or chattering of the
motor starter, and stalling of the motor due to insufficient torque for acceleration. The transient
performance of an engine-generator set when starting large motors is a function of the complete
generator set system, including the engine power available, the generator and its excitation
system, and the energy stored in the rotating inertia of the generator set. Proper sizing for a
successful start requires consideration of all these factors, more than just the generator only
data.

Publication No. 900-0286A


1997 by Onan
Starting Large Motors from an Onan Engine-Generator
A common problem in sizing engine-generators is starting large motors (large relative to the
generator set capacity). Large voltage and frequency deviations may occur if the generator set
is not sized properly. Typical problems include holding coils dropping out or chattering of the
motor starter, and stalling of the motor due to insufficient torque for acceleration. Other
connected loads may be more sensitive to voltage and frequency deviations than the motor or
motor starter. The rate of change in generator frequency can also be a problem for some
static uninterruptible power supplies. If the load on the generator set consists of a single large
motor, particularly motors requiring high starting torque, a number of problems may occur if the
generator set does not have sufficient capacity. These problems may include sustained low
voltage operation, extended load acceleration times, tripping of circuit breakers or motor
protective devices, engine-generator protection shutdowns, etc.

The engine generator set needs to be sized for both starting and running the loads, and it is
very common for the load starting kW or kVA requirements to be the dominant factor in sizing.
This is particularly true when a large percentage of the connected load is motors. Unlike a
typical utility source, an engine-generator set is a limited power (kW) source, basically
determined by the available engine horsepower. The transient performance of an engine-
generator set when starting large motors is a function of the complete generator set system,
including the engine power available, the generator and its excitation system, and the energy
stored in the rotating inertia of the generator set. Proper sizing for a successful start requires
consideration of all these factors, more than just the generator only data. Be careful not to use
generator manufacturer’s data for transient voltage dip which does not take into account the
complete system, e.g., the engine performance. Onan’s Gensize ‘96 uses a database of
representative test data on complete engine-generator set models for selection and prediction
of transient performance.

Motor Starting Characteristics


Induction motors have starting characteristics as shown in Figure 1. During starting the motor
draws approximately six times its full load current. This current causes a significant dip in
output voltage from a high impedance source like an engine generator set until the motor
reaches nearly full speed. Initially the starting power factor of the motor is very low which
means that the engine power required is also low, but the power required by the motor peaks
at about 80% of rated speed. The low power factor at motor standstill means that the correct
calculation of generator locked rotor kVA capacity would use vector addition. The locked rotor
kVA of the motor would be added vectorially to the running load kVA vector. However, since
simple arithmetic addition will always give a larger result than vector addition, it will error on the
conservative side. For a successful start the motor must develop greater torque than required
by the load. The difference in torque between the motor and the load determines the rate of
acceleration.

Premium efficiency motors (Design E) have considerably higher starting currents than typical
Design B, C , and D motors. For example, a 15 HP premium efficiency motor has a locked
rotor kVA approximately 50% higher than a standard motor. The higher starting current must
be taken into account to maintain specified transient voltage dip and recovery voltage of at
least 90% of rated voltage during acceleration. If premium efficiency motors are powered by
the generator set, there can be a significant increase in generator set size required to maintain
voltage. The increase in size may be offset somewhat by the use of reduced voltage starting
methods.

1
Figure One. Induction Motor Starting Characteristics - High inrush current causes the generator voltage to dip.
Real power required from the engine peaks at about 80% of motor speed.

Voltage Dip
When starting a motor across the line with a generator set, the motor represents a low
impedance load while at locked rotor or stalled condition, causing a high inrush current,
typically six times the rated running current. The high inrush current causes the generator
voltage to drop. This voltage dip is comprised of two main components, the transient voltage
dip and the recovery voltage dip. A maximum transient voltage dip of 30% is important to
maintain the holding force on coils and a recovery voltage of 90% is important to provide
sufficient torque to accelerate the motor and its load. It is the RMS (root-mean-square) or
effective voltage that provides the holding force and torque during starting. Onan’s Gensize
‘96 sizing software will not select an engine-generator set with less than 90% recovery voltage.
The software does permit the user to adjust the transient voltage dip to the requirements of the
connected load equipment.

2
Transient voltage dip may also have two components, the instantaneous voltage dip and
possibly a “roll off” voltage dip. The instantaneous dip is a solely a function of the relative
reactances of the generator and motor, and occurs instantaneously upon connecting the motor
to the generator output. Because the instantaneous component of the transient voltage dip is
the product of the motor starting current times the subtransient reactance of the generator (Ims
x X”d), the benefit of a low reactance generator is less instantaneous voltage dip. The
instantaneous voltage dip is predicted on curves published by the generator manufacturer.
These voltage dip curves give an approximation of what might be expected for the
instantaneous dip assuming frequency is constant, that is, with unlimited kilowatts available
from the driver such as a test stand dynamometer fed by the utility. With actual engine-driven
equipment, if the engine slows down due to a heavy starting kilowatt requirement, the transient
voltage dip would include an additional dip as the torque matching characteristic of the voltage
regulator rolls off excitation to help the engine recover speed. This is shown in Figure Two.

Figure Two. Transient Voltage Dip - Transient voltage dip includes a reactance drop and may or may not include a
“voltage roll-off” depending on the engine speed. Following the transient voltage dip the regulator excitation system
“forces the field” to recover voltage to rated.

3
Following the transient voltage dip, the generator excitation system detects the low voltage and
responds by increasing excitation to recover to rated voltage. The excitation level depends on
the field forcing capability of the excitation system. Field current cannot be changed
instantaneously in response to load change. The regulator, exciter field, and main field all have
time constants. The voltage regulator has a relatively fast response. The main field has a
significantly slower response than the exciter field because it is many times larger. Field
forcing is designed into all excitation system components to optimize recovery time. It must be
enough to minimize recovery time, but not so much as to lead to instability (overshoot) or
overcome the engine (which is a limited source of power). See Figure Two. Modern
generators typically have the ability to force field current at about three times normal full load
field current. Generators with permanent magnent generator (PMG) excitation have a
constant excitation power source that is independent of the generator output. Onan
recommends the use of PMG excitation for starting heavy motor loads, because a low voltage
condition on the generator output would not affect the excitation power available for forcing the
field.

At the same time the motor begins to accelerate to rated speed, assuming that the motor
develops enough torque. For induction motors, motor torque is directly proportional to the
square of the applied voltage. The rate at which the motor accelerates to rated speed is a
function of the difference between the torque the motor develops and the torque requirements
of the load. In order to avoid problems with excessive acceleration time or possibly stalling the
motor, it is important for the generator to recover to rated voltage as quickly as possible. The
manner in which the generator voltage recovers is a function of several factors including the
relative sizes of the generator, the motor, the kilowatt capacity of the engine, and the
generator excitation forcing capability. If the excitation system responds too fast and too stiff
the generator can actually overload the engine when starting large motors. Depending on the
severity of the load, several cycles to a few seconds later, the generator recovers to rated
voltage.

The maximum locked rotor kVA capability of a generator, while still maintaining 90% of rated
voltage is derived from its sustained overload characteristic, shown in Figure Three. The
characteristic is graphed based on short circuit testing done at near zero power factor. Since
the motor’s locked rotor kVA is expressed based on full voltage starting it is necessary to
correct the generator kVA at 90% voltage to kVA at 100% voltage. To do this graphically a line
is drawn from zero through the overload characteristic curve at 90% voltage. The line is
extended to the intersection at 100% voltage. From that intersection a vertical line is dropped
down to read the maximum motor starting kVA that corresponds to the motor’s locked rotor
kVA. A typical kVA overload capability is 2.5 to 3.0 times the generator’s rating.

4
100
90
80
70
% Rated Volts

60
50
40
30
20
10
0
1.0 2.0 3.0
kVA at
kVA or Amperes PU 100% Voltage
Figure Three. Generator Overload Characteristic - Maximum motor starting kVA at 90% recovery voltage corrected
to motor nameplate voltage (100%).

Real Power and Frequency Dip


“If I can start it, I can run it.” That’s the way a senior engineer once summed up sizing an
engine-generator set for a motor load. He was making the point that real engine power is what
starts and accelerates motors, and that this starting kW (SkW) requirement often determines
what size the engine-generator needs to be. The starting kW requirement of the motor can be
as high as 3 times its running kW requirement. It is particularly important when starting large
motors that the engine has sufficient power (SkW) to meet this demand, or the motor may not
start successfully. Any engine preload subtracts from the available engine power. If the engine
cannot develop sufficient power, the motor being started will rotate at a low speed determined
by the balance between the power developed by the engine and the power required by the
motor. The stored energy in the rotating inertia of the engine generator set can overcome the
maximum power required by motors that are small compared to the engine, but on larger
motors the engine may slow significantly during the momentary overload or a generator set
protective device may operate. Onan generator sets include a “torque matching’ voltage
regulator function which senses the slowing down of an overloaded engine and rolls off
excitation power to reduce output voltage enough to allow the engine to recover. Since power
is proportional to the square of the voltage, a small reduction in voltage will assist the engine in
picking up the kilowatt load.

Motor Moment of Inertia


The moment of inertia for a rotating mass is its resistance to acceleration. To start the motor
and its load rotating this inertia must be overcome by an accelerating torque which translates
to engine power (SkW). The load connected to the motor shaft has its moment of inertia and
in practical situations for specific equipment this may or may not be available information.
Fortunately, for the purpose of sizing the engine-generator set, or more specifically to
determine the engine power needed to start and accelerate a rotating motor load, the motor
load moment of inertia need only be broadly categorized as low or high inertia. High inertia
loads are characterized by high breakaway torque requiring prolonged acceleration times
and/or pulsating or unbalanced loads. Low inertia loads are characterized by low starting

5
torque at standstill, increasing as motor speed increases resulting in high acceleration to rated
speed. High inertia loads are harder to start than low inertia loads.

Examples of high inertia loads include:

• Single and multi-cylinder pumps


• Single and multi-cylinder compressors
• Crushers
• Hydraulic Elevators without unloading valves

Examples of low inertia loads include:

• Fans, Centrifugal and Blower


• Rotary Compressors
• Pumps, Rotary and Centrifugal
• Motor-Generator Elevators

Note: Pumps starting into high head pressure, large diameter fans or fans starting into high
restriction should be classified as high inertia loads.

If the motors are driving low inertia loads with low starting torque requirements, Gensize will
multiply SkW by a factor of 0.6 because these loads accelerate so quickly that the engine’s
inertia provides sufficient energy to maintain near rated speed until the motor achieves rated
speed and running kilowatts.

Power Factor Correction


Loads that include power factor correction or filters for power quality improvement should not
be applied to a generator set operating at light load levels. The capacitive elements of these
loads can cause the voltage of a generator set to rise uncontrollably at light load levels. Power
factor correction capacitors for large motors should either be switched with the motor or be
connected to the utility side of the transfer switch, not the load side.

Regenerative Loads
Loads such as elevators, cranes and hoists often rely on the capability of the source to absorb
power during certain sequences of operation, typically for braking purposes. Since the utility is
essentially an infinite power source serving diverse loads, this is not a problem when operating
from utility power. A generator set, however, is a limited power source and has limited
capability to absorb power, especially if no other loads are connected. The regenerative power
capability of a generator set is information that is available from the manufacturer. A typical
regenerative capability is about 15% of the generator set rating.

If the regenerative power of the load exceeds the capacity of the generator set, the generator
set may overspeed and shutdown on overspeed protection. The overspeed limit of a generator
set rated at 1800 RPM is usually about 2100 +/- 50 RPM. Applications that are most likely to
be a problem are where an elevator is the major load on the generator set. Generally, the
regeneration problem can be solved by adding loads which can absorb the regenerative
power. For example, transfer the lighting load to the generator first before transferring the
elevator.

6
Running Surge Voltage Dip
Once all the loads on the generator set have been started, some of those loads may cycle ON
and OFF when called on by an automatic control, such as an air conditioner or an elevator. If
there are motors that cycle, a running surge voltage dip should be calculated. The worst case
condition would be when all cyclical loads start simultaneously with the rest of the connected
load running. Taking cyclical loads into account could significantly increase the size of the
required generator set depending on the required transient performance and could invalidate
the process of placing loads in a step starting sequence. The worst case calculation for
running surge voltage dip starts with the total preload in running kW and kVA and adds the
total starting kW and kVA of all of the cyclical loads. Onan’s Gensize ‘96 asks the user to
specify a maximum running surge voltage dip and to select which loads cycle ON and OFF.
With this information the software will calculate the running surge voltage dip and select an
appropriate generator set.

Reduced Voltage Starting Methods


Although voltage dip often causes various problems, a controlled reduction in voltage at the
motor terminals can be beneficial when it is used to reduce the starting kVA of a motor in
applications where the reduced motor torque is acceptable. It is important to determine that an
acceptable level of motor torque is achieved. If not, loads will be slow to accelerate or may fail
to reach full speed and motor damage can result. Reducing motor starting kVA can reduce the
size of the generator set required, lessen the voltage dip and provide a softer start for the
motor loads. All of the following reduced voltage starting methods are included in the Gensize
‘96 program.

A Comparison of Motor Starting Methods:


Table 1 compares the effects of full voltage, autotransformer and resistor starting on a 50
horsepower, Code G motor. As can be seen, autotransformer starting requires less motor
starting capacity from the generator set.

TABLE 1. REDUCED VOLTAGE MOTOR STARTING COMPARISON

TYPE OF STARTER
AUTOTRANSFORMER RESISTOR FULL VOLTAGE
% of applied voltage 65 42 100
(tap)
% of full voltage 42 42 100

Starting kVA 124 124 295


Starting kW 47 106 110
Run kVA 46 46 46
Run kW 42 42 42

Full Voltage Motor Starting


Full voltage, across–the–line starting is typical unless it is necessary to reduce motor starting
kVA because of the limited capacity of the generator set or to limit voltage dip during motor

7
starting. There is no limit to the HP, size, voltage or type of motor. This method is most
common because of its simplicity, reliability and initial cost. Note on the kVA and torque curves
that starting kVA remains fairly constant until the motor almost reaches full speed. Also note
that kW peaks at about 300 percent of rated near 80 percent of synchronous speed.
MOTOR STARTING DIAGRAM TYPICAL TORQUE AND KVA CURVES FOR
SQUIRREL CAGE INDUCTION MOTORS
600

kVA & TORQUE (% F. L.)


500
3 KVA
400
LINE

2
MOTOR
1 300

200
TORQUE
100
START: CLOSE 1-2-3
RUN: NO CHANGE 20 40 60 80 100
SPEED (% SYNCHRONOUS)
Autotransformer Motor Starting, Open Transition
The autotransformer is in the circuit only during starting to reduce voltage to the motor. The
opening of the circuit during transition can cause severe transients. Open transition switching
of reduced voltage starters should be avoided in generator set applications, especially when
the motors are not brought up to full speed at the time of transition. The reason for this is that
the motor slows down and is then out of synchronization during the switching transition. The
result is similar to paralleling generator sets out of phase. The kVA drawn immediately after
switching can exceed starting kVA. Also note that the starting power factor is lower when an
autotransformer is used.
MOTOR STARTING DIAGRAM TYPICAL TORQUE AND KVA CURVES FOR
SQUIRREL CAGE INDUCTION MOTORS
8 600
kVA & TORQUE (% F. L.)

6 500

400
LINE

5
MOTOR KVA
4 300

3 200

100 TORQUE
1 2

START: CLOSE 2-3-5-6-7 20 40 60 80 100


RUN: OPEN 2-3-5-6-7; CLOSE 1-4-8 SPEED (% SYNCHRONOUS)

Autotransformer Motor Starting, Closed Transition


The circuit is not interrupted during starting. During transfer, part of the autotransformer
winding remains in the circuit as a series reactor with the motor windings. Closed transition is
preferred over open transition because of less electrical disturbance. The switching, however,
is more expensive and complex due to three elements. It is the most commonly used reduced
voltage starting method for large motors with low load torque requirements, such as M–G sets,
sewage lift pumps and chillers. The principle advantage is more torque per current than with
other reduced voltage starting methods. Operation can be automatic and/or remote. Also note
that the starting power factor is lower when an autotransformer is used.

8
MOTOR STARTING DIAGRAM TYPICAL TORQUE AND KVA CURVES FOR
SQUIRREL CAGE INDUCTION MOTORS
5 kV600
A
&500
4 T
7
LI O400
N 3 R
MOTOR KVA
E Q300
2 6 U
E 200
(%100 TORQUE
1 F.
START: CLOSE 6-7-2-3-4
TRANSFER: OPEN 6-7 20 40 60 80 100
RUN: CLOSE 1-5 SPEED (% SYNCHRONOUS)

Reactor Motor Starting, Closed Transition


Reactor starting has the advantage of simplicity and closed transition, but results in lower
starting torque per kVA than with autotransformer starting. Relative torque, however, improves
as the motor accelerates. Reactor starting is generally not used except for large, high–voltage
or high–current motors. The reactors must be sized by HP and voltage and may have limited
availability. Typically, reactor starting costs more than autotransformer starting for smaller
motors, but is simpler and less expensive for larger motors. Starting power factor is
exceptionally low. Reactor starting allows a smooth start with almost no observable
disturbance on transition and is well suited for applications such as centrifugal pumps or fans.
MOTOR STARTING DIAGRAM TYPICAL TORQUE AND KVA CURVES FOR
SQUIRREL CAGE INDUCTION MOTORS
600
kVA & TORQUE (% F. L.)

6
3 500

5 400
LINE

2 300
MOTOR KVA
4 200
1 100
TORQUE
START: CLOSE 1-2-3 20 40 60 80 100
RUN: CLOSE 4-5-6
SPEED (% SYNCHRONOUS)

Resistor Motor Starting, Closed Transition


Resistor starting is occasionally used for smaller motors where several steps of starting are
required and no opening of motor circuits between steps is allowed. Also available as a
stepless transition starter which provides a smoother start. Resistor starting is usually the least
expensive with smaller motors. Loads accelerate faster because the voltage increases with a
decrease in current. Resistor starting has a high starting power factor, so it may actually
increase the size of the generator set.

9
MOTOR STARTING DIAGRAM TYPICAL TORQUE AND KVA CURVES FOR
SQUIRREL CAGE INDUCTION MOTORS
6 9
600

kVA & TORQUE (% F. L.)


3
5 8 500

400
LINE

2
MOTOR KVA
4 7 300
1 200

100
START: CLOSE 1-2-3
SECOND STEP: CLOSE 4-5-6 TORQUE
THIRD STEP: CLOSE 7-8-9 20 40 60 80 100
SPEED (% SYNCHRONOUS)
Star–Delta Motor Starting, Open Transition
Star–Delta starting requires no autotransformer, reactor or resistor. The motor starts star–
connected and runs delta–connected. This starting method is becoming more popular where
low starting torque’s are acceptable. It has the following disadvantages:
1. Open transition. Closed transition is available at extra cost.
2. Low torque.
3. No advantage when the motor is powered by a generator set unless the motor reaches
synchronous speed before switching. In applications where the motor does not reach
synchronous speed, the generator set must be sized to meet the surge.
MOTOR STARTING DIAGRAM TYPICAL TORQUE AND KVA CURVES FOR
SQUIRREL CAGE INDUCTION
MOTORS
3 MOTOR 600
kV
2 4 5 6 A
7 500
&
1 T 400
LINE
O
8 R 300
Q
U 200
9 E KVA
(% 100
F. TORQUE
START: CLOSE 1-2-3-4-5-6 L.) 20 40 60 80 100
RUN: OPEN 4-5-6; CLOSE 7-8-9
SPEED (% SYNCHRONOUS)

Part Winding Motor Starting, Closed Transition


Part winding starting is less expensive because it requires no autotransformer, reactor or
resistor and uses simple switching. Available in two or more starting steps depending on size,
speed and voltage of motor. Automatically provides closed transition. First, one winding is
connected to the line; after a time interval, the second winding is paralleled with the first.
Starting torque is low and is fixed by the motor manufacturer. The purpose of part winding is
not to reduce starting current but to provide starting current in smaller increments. There is no
advantage to this method if the motor is powered by a generator set unless the motor can
reach synchronous speed before transition to the line.

10
MOTOR STARTING DIAGRAM TYPICAL TORQUE AND KVA CURVES FOR
SQUIRREL CAGE INDUCTION MOTORS
3
600

kVA & TORQUE (% F. L.)


LINE

2
1 500

400
MOTOR
300 KVA
6
200
5
4 100
TORQUE
START: CLOSE 1-2-3 20 40 60 80 100
RUN: CLOSE 4-5-6
SPEED (% SYNCHRONOUS)

Wound Rotor Motor Starting


A wound rotor motor can have the same starting torque as a squirrel cage motor but with less
current. It differs from squirrel cage motors only in the rotor. A squirrel cage motor has short
circuit bars, whereas a wound rotor motor has windings, usually three–phase. Starting current,
torque and speed characteristics can be changed by connecting the proper amount of external
resistance into the rotor. Usually, wound rotor motors are adjusted so that the starting kVA is
about 1.5 times running kVA. This is the easiest type of motor for a generator set to start.

MOTOR STARTING DIAGRAM TYPICAL TORQUE AND KVA CURVES FOR


WOUND ROTOR MOTORS
3
600
kVA & TORQUE (% F. L.)
LINE

2
1 500

400
MOTOR
ROTOR 300
KVA
4 6 8 200

100
5 7 9 START: CLOSE 1-2-3 TORQUE
STEP #1: CLOSE 4-5
STEP #2: cLOSE 6-7 20 40 60 80 100
RESISTORS SPEED (% SYNCHRONOUS)
RUN: CLOSE 4-5-6

Solid-State Starters
Solid-state starters can adjust the starting torque, acceleration ramp time, and current limit for
a controlled acceleration of the mechanical load while starting motors. For the purpose of
sizing the generator set, the current limit adjustment reduces the inrush current and may be
used to reduce the starting kW and kVA requirement on the generator. The range of available
current limit settings are typically from 150 to 600% of full load current. A 600% current limit
setting would be sized just the same as across-the-line starting. A 300% current limit setting
would reduce SkW and SkVA by 50%. Use of the current limit setting also reduces the motor
torque available to the load.

The sequence of operation of a solid-state starter is shown in Figure 4. From standstill the
starter applies an initial voltage to provide sufficient torque to begin to rotate the motor.
Typically this voltage (torque) is adjustable from 0 to 100%, with a typical setting of 60% or at
the minimum effective starting torque. The acceleration ramp time is adjusted to control the
time it takes the starter to reach current limit. A typical range of ramp time adjustment is from 0
to 60 seconds. The setting is based on the desired time to accelerate the mechanical load.
The current limit setting caps the inrush current until the motor comes up to full speed. From a
generator sizing perspective, an extended acceleration ramp time and a low current limit

11
setting (if also appropriate for the motor and the mechanical load) would have the minimum
impact on voltage and frequency excursions.

CURRENT LIMIT

P
RAM
CURRENT

N
ATIO
ER
CEL
AC

INITIAL TORQUE RUN CURRENT

TIME
Figure 4. Solid-state Starter Sequence of Operation - The initial torque, acceleration time, and current limit may be
adjusted to provide a smooth start for the load.

Solid-state starters use SCRs and are a non-linear load that will cause voltage distortion. To
compensate for this distortion the generator capacity should be increased. The recommended
generator capacity required is two times the running kilowatt load, except where an automatic
bypass is used. If the solid-state starter is equipped with an automatic bypass, the SCRs are
only in the circuit during starting. Once the motor is running the bypass contactor closes and
shunts the SCRs. In this case, the voltage distortion may be ignored during starting and there
would be no additional generator capacity required. Gensize 96, Version 2.0, includes a
selection for automatic bypass with solid-state starters.

Variable Frequency Drives

All variable frequency drives are current limiting and reduce both starting kW and kVA. The
current drawn by these drives is non-linear harmonic currents, which cause a distorted voltage
drop across the reactance of the generator. Since VFDs are non-linear an additional
generator capacity sizing factor needs to be used to keep voltage distortion to a reasonable
level, 15% total harmonic distortion or less. For six-pulse VFDs a typical generator sizing
factor would be two times the running kW of the drive. The increase in size of the generator
for the non-linear current will more than offset any reduction in starting kW and kVA. If the
VFD is the pulse width modulated type, or includes an input filter to limit current distortion to
less than 10%, then the additional sizing factor may be reduced to 1.4 times the running kW of
the drive. Gensize 96, Version 2.0, includes a selection for PWM drives with input filters.

Sizing examples using Gensize ‘96:

Sizing the engine-generator set may be done with manual calculations using a worksheet
available in Onan Application manual T-030 or with GenSize ‘96 software. The basic process
is the same. It is always best to use actual data if known, otherwise one advantage of GenSize

12
‘96 is that much of the required information on typical load characteristics is available as
default information.

One commonly used approach is to assume that all connected loads will be started in a single
step. This assumption will result in the most conservative (largest) generator set selection.
Unless something has been done, such as multiple transfer switches with staggered time
delays or a step load controller, then a single step load should be used for sizing purposes.
However, once all of the loads have been brought up on line with the generator set, the load
equipment may be frequently stopped and started by automatic controls. In such cases, the
generator set will have to be sized to start the largest motor last, with all other connected loads
on line.

Following are five sizing examples. The first three examples have been selected to show the
effect on recommended generator set size of transient voltage dip specifications and
autotransformer reduced voltage starting. The last two examples have been selected to show
the effect of multiple steps compared to a single step when starting several large motors.

Example 1. 200 HP motor across-the-line with 10% transient voltage dip. Refer to Figures Five
and Six. In this example the specification for no more than 10% transient voltage dip when
starting a large (200 HP) motor across the line results in a selection of a 1100 kW generator
set with an oversized 80 degree C rated rise generator. The running load on this generator
set is only 14% of rated standby kilowatts which is well below the 30% recommended
minimum. This would not be a good selection. If used this diesel engine would probably
experience problems with wet stacking, carboning, oil diluted with fuel, etc.

Figure Five. 200 HP Motor Load Characteristics - The motor is started across the line and is low inertia.

13
Figure Six. Example #1 Generator Set Selection - With 10% transient voltage dip a 1100 kW generator would be
required.

14
Example 2: Figure Seven. A 200 HP motor across-the-line with 30% transient voltage dip. In
this example starting the same motor as in the first example, but allowing the transient voltage
dip to be 30% maximum (29% estimated actual), the generator set selection goes down to a
200 kW set with a 125 degree C rated rise generator running at 81% of rated load. This would
be a reasonable selection.

Figure Seven. Example #2 Generator Set Selection - With 30% transient voltage dip allowed, a 200 kW generator
set is selected.

15
Example 3: Figures Eight and Nine. 200 HP motor with autotransformer start and 30%
transient voltage dip. In this example use of the autotransformer reduced voltage starting at
the 65% tap further reduces the generator set selection to a 175 kW with a 125 degree C rated
rise generator. The estimated actual transient voltage dip is down to 18% and the running
load is up to 93% of rated. If no future load increases are expected, this would also be a
reasonable selection.

Figure Eight. 200 HP Motor Load Characteristics - The motor is started with autotransformer starter at 65% tap and
is low inertia.

16
Figure Nine. Example #3 Generator Set Selection - With autotransformer starter, a 175 kW generator set is
selected.

17
Example 4: Figure Ten. Three 200 HP high inertia motors started across-the-line in a single
step with 30% transient voltage dip allowed. Refer to Figure Ten. In this example a selection
is made for a 800 kW standby rated generator set running at 61% of rated load.

Figure Ten. Three 200 HP motors started in a single step - The starting kilowatts (SkW) required results in
selection of an 800 kW standby generator set.

18
Example 5: Figure Eleven and Twelve. Three 200 HP motors in three steps. A step start
sequence is accomplished using three automatic transfer switches with staggered time delays,
or other step sequence controller. Use of three steps reduces the generator set selection to a
500 kW model.

Figure Eleven. Three 200 Hp motors started in three steps - Using three step starting sequence reduces the size of
the generator set from 800 kW to 500kW standby.

19
The running load of 488 kW, which conservatively assumes that the motors are loaded
to nameplate rating, equals about 97% of the generator standby rating. The generator
set is rated to carry this load in applications within the standby rating definition. Since
the margin, if any, between motor nameplate power and the actual load on the motor is
seldom actually known, a more definite margin may be added by stepping up a size to
the 600 kW model. Gensize ‘96 allows the user to select models as shown in Figure
Twelve. With a “nameplate” running kW of 488, the 600 kW generator set would run at
81% of rating or less.

Figure Twelve. Example #5 Generator Set Selection - Stepped up a size to the 600 kW model.

20
Gensize 96 Glossary of Terms

GkW - Generator capacity provided to compensate for non-linear distortion, usually equal to
RkW times a factor for the type of non-linear load.
Max SkVA - Maximum Surge Kilovolt-amperes, the highest kilovolt-ampere load in any load
step.
Max SkVA Req. - Maximum Surge Kilovolt-amperes Required, the highest total of any surge
kVA load added to the previously running kilovolt-ampere load.
Max SkW - Maximum Surge Kilowatts, the largest block of kilowatt load in any load step.
Max SkW Req. - Maximum Surge Kilowatts Required, the highest total of any surge kW load
added to the previously running kilowatt load.
RAmps - The total running amperes for a step.
Reduced Max SkVA Req. - The Max SkVA Req. times a multiplier to compensate for reduced
voltage. Must be less than or equal to the generator set Max kVA.
Reduced Max SkW Req. - The Max SkW Req. times a multiplier to compensate for reduced
voltage. Must be less than or equal to the generator set Site rated Max SkW.
RkVA - Running Kilovolt-amperes, the running load.
RkW - Running Kilowatts, the running load.
RPF - Running power factor is the load power factor when the load is running at rated steady
state conditions.
RSkVA - The running surge kilovolt-amperes, the total surge kVA of all cyclic loads, added to
any welder loads, and medical imaging loads.
RSkVA Req. - The running surge kilovolt-amperes required, the highest total of any running
surge kW load added to the previously running kilowatt load.
RSkW - The running surge kilowatts, the total surge kW of all cyclic loads, added to any
welder loads, and medical imaging loads.
RSkW Req. - The running surge kilowatts required, the highest total of any running surge kW
load added to the previously running kilowatt load.
SPF - Starting power factor is the power factor of the load at the time it is initially energized or
started. This is a particularly important parameter for motor loads which exhibit a low power
factor during locked rotor conditions, causing a high inrush kVA (SkVA) during starting.

Summary
Sizing an engine-generator set for starting large motors requires an understanding of motor
characteristics, starting methods, and the transient performance of the engine-generator
during motor starting. Because the engine-generator set is a limited power source both in kW
and kVA, voltage and frequency excursions will occur when starting motors. A properly sized
engine generator set will keep the voltage and frequency excursions within reasonable limits
without oversizing the machine. Use Gensize ‘96 to make the sizing problem easier and the
resulting recommendation more accurate.

21
About the Authors
Jim Iverson
Manager, Technical Sales has been with Onan since 1976. He has a Masters in Engineering
Science, Bachelors in Electrical Engineering. Jim spent 7 years at General Electric and 2
years at Electric Machinery before joining Onan. While at Onan, he managed Transfer Switch
Design, Systems Engineering and Switchgear and Controls. Current responsibilities are to
provide technical direction to the Commercial Marketing and Sales Department, participate in
domestic codes and standards development, sales and service training, technical input for and
review of all published literature, published papers on relevant industry topics.

Lawrence A. Bey
Senior Technical Marketing Specialist and a 1978 University of Minnesota Graduate. Larry
has been an Onan employee from 1978 to the present, spending 7 years in Engineering, and
11 years in Marketing and Sales. He represents Onan on Technical Committees of the
National Fire Protection Association for NFPA 110 Emergency and Standby Power Systems
and NFPA 99 Essential Electrical Systems for Health Care Facilities. Larry is active with
National electrical Manufacturers Association (NEMA) on Motor and Generator Subcommittee
and the Automatic Transfer Switch Equipment Subcommittee. He is a member of the
International Association of Electrical Inspectors.

22
See your distributor for more information.

Onan Corporation
1400 73rd Avenue N.E.
Minneapolis, MN 55432
612-574-5000
Fax: 612-574-8087
Onan and PowerCommand are registered trademarks of Onan Corporation Cummins is a registered trademark of Cummins Engine Company

23

You might also like