ATP Motor Starting PDF
ATP Motor Starting PDF
ATP Motor Starting PDF
in EMTP-ATP
Executive Summary
The work described in this paper has been developed for
the purposes of medium voltage motor starting analysis in
customer projects. EMTP-ATP today offers highly advanced
modeling capabilities for motors and starting equipment.
This paper will present models for several increasingly
complex motor starting methods, such as direct on line
(DOL), star-delta, wound rotor resistance, autotransformer,
soft starter, and variable speed drive (VSD) with
synchronous bypass.
Schneider Electric White Paper 2
Introduction Motor starting is one of the key analyses necessary to validate the integration of
a motor in an electrical installation. Such analyses can often be omitted, particularly
when the motor involved is of low power or remains below 10% of the transformer or
generator.
However, if the motor power is higher, then analyses, typically by simulation, will be
necessary in order to define the most economic and convenient manner to start for the
system and parallel loads. The main motor starting methods are well known and most
of them have been widely discussed and presented in previous papers1, 2, 3.
In this paper the authors review the different starting methods, focusing on the specific
modeling challenges encountered in EMTP-ATP, and compare the performance of each
method in a case study.
Typical Motor starting impacts three main elements: the industrial system, the driven load,
and the motor itself.
Constraints for Voltage drop during starting is the main impact on the industrial system.
Motor Starting It is produced by the current drawn by the motor and is relative to the short-circuit
in Industrial impedance of the system. A higher impedance is characteristic of a power system
with lower short-circuit power. Consequently, it will produce a higher voltage drop.
Systems Voltage drop up to 15% is usually acceptable but lower values are preferable.
The driven load is mainly impacted by the intensity of the start or the applied torque.
Some high inertia loads like fans, or specific applications like water pumps, will suffer
a sudden application of high torque, which may result in a high torsional effect or water
hammer. In such cases, progressive starting is recommended.
The motor is impacted through mechanical stress, starting duration, and heating.
It is recommended that thermal stress is limited to 80-90% of the motor thermal
capacity to avoid premature aging.
1
J. Nevelsteen, H. Aragon, Starting of Large Motors - Methods and Economics, IEEE Transactions On
Industry Applications, Vol. 25, No. 6, November - December 1989
2
J. Larabee, B. Pellegrino, B. Flick, Induction Motor Starting Methods and Issues, in proc. of PCIC
Conference 2005
3
S. Rusnok, P. Sobota, V. Mach, P. Kacor, S. Misak, Possibilities of Program EMTP – ATP to Analyze the
Starting Current of Induction Motor in Frequent Switching, EPE Conference 2015
Case Study for A case study has been defined in order to compare the performance of the various
starting methods.
Comparative
General Data
Analysis The electrical network is composed of a power source with limited short-circuit power,
power transformer, motor cable, and the motor. Details are provided in Table 1.
3.1MW@STD
2
IM
(A) (B)
Table 2 The motor model has been defined as a best fit to match most datasheet performance
Motor Datasheet Data values. The results are shown below in Table 2.
and Resulting EMTP-ATP
Characteristic Datasheet Value EMTP-ATP Model
Model
Rated power (MW) 3.1 3.1
Rated voltage (kV) 6.3 6.3
Rated current (A) 320 310
Rated speed (rpm) 1489 1483
Rated torque (Nm) 19735 19735
Slip (%) 0.73 1.1
Starting current (pu) 5.6 5.6
Starting torque (pu) 0.7 0.615
Maximum torque (pu) 2.5 2.51
Power factor 0.92 0.94
Efficiency 96.7 97
Rs (Ω) 0.06 0.12
Rr (Ω) 0.17 0.21
Xm (Ω) 51.36 62.8
Xs (Ω) 1.453 1.522
Xr (Ω) 0.4622 0.4622
Motor & pump inertia (kgm²) 219.4 219.4
This best fit has been achieved through a simulation trial and error approach. The lack
of saturation modeling will generally increase the maximum torque crest with speed, 4.
The motor is originally designed in star connection. When the starting modes require
the motor to be designed in delta connection, or to be wound rotor, the motor model
has been transformed accordingly, always keeping the same electro-mechanical
behavior.
The motor load is a pump. It has been modeled using the nonlinear current-dependent
resistor TYPE 99. The load curve has a parabolic evolution, however, in EMTP-ATP
it is necessary to define a monotonically increasing curve, which produces a more
conservative curve.
Figure 3 Load torque, Nm
Load Torque 0.45
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05 Load in EMTP
0 Load curve
0 20 40 60 80 100 120 140 160
Although in practice each starting mode and motor will be designed for a specific load,
in this paper, for comparison purposes, the same load has been used for all starting
modes and motors.
4
G. J. Rogers, D. Shirmohammadi, Induction Machine Modelling for Electromagnetic Transient Program,
IEEE Transactions on Energy Conversion, Vol. EC-2, No. 4, December 1987
Modeling and The focus on the results presented in this paper will be on motor torque evolution and
transients, network voltage, and network current fed to the starter. For autotransformer
Analysis of starting, the current on the motor side will also be presented.
Motor Starting Direct On Line Starting
Methods Direct on line (DOL) starting is the simplest and most economical way to start an induction
motor, but it causes a considerable starting current, almost 5 to 7 times the rated current.
In practice, a DOL start is achieved using a contactor or circuit breaker. In EMTP-ATP it is
modeled by a 3-phase switch. The corresponding model and results are shown in Figure 4:
3.1 MW@STD
Figure 4 500 MVA BCT
2
(A) DOL Starting Model ∆ Y IM
(B) Motor Torque
(C) Network Voltage
and Current.
11 KV
(A)
The results obtained show expected torque oscillations and a significant voltage drop (15%).
Star-Delta Starting
This is a very economical starting method that reduces the starting current.
The equipment normally consists of three contactors and the motor must be designed
in delta connection. The starting current is about one third of the direct on line start,
as shown in Figure 5. (B). However, the process of switching from star to delta
connection generates a significant oscillation torque.
Figure 5 500 MVA BTC
∆ Y IM
Model
R(i)
Although the oscillation might be exaggerated in the simulation, it shows that this
starting method produces high mechanical stress on the motor and may require
more frequent motor maintenance. In practice, the star-delta method is reserved for
infrequent starting and for smaller and inexpensive motors.
M
M
and Current 0 S 0 S 0 S 0 S 0 S
11 KV
IM
Y
UI
RN
R(i)
(A)
5
Fahai L. 电机与拖动基础第三版 (Motor and drag), 2005
Autotransformer Starting
An autotransformer start, also called RVAT (Reduced Voltage Autotransformer),
is another starting method that reduces the starting current, as the voltage across the
motor is reduced during starting. The torque is reduced as the square of the applied
voltage. When the motor has almost reached its rated speed, the motor is progressively
connected in direct on line.
Three main steps are considered in this case. In the first step, the motor is connected
to the starting voltage with autotransformer. The autotransformer neutral is then opened
and the motor is connected to the network through the autotransformer impedance.
In the last step, the motor is connected in direct on line.
(A) Autotransformer
Starting Model 500 MVA BTC
3.1MW@STD
Due to the limited current, the motor takes longer to accelerate to full speed.
In the transition from step 1 to 3, it is predominantly torque oscillations that are
observed, which last much longer than those of current and voltage.
The current ramp and limitation phase can be clearly observed in Figure 8.
The voltage is evolving in the opposite manner. This is due to the constant impedance
of the source. In the case of a generator, the voltage profile will change and the steady
voltage drop will be limited through the voltage regulation. As expected, the motor
torque has a smooth increase during the ramp time.
6
T-T-H. Pham, D. Penkov, S. Heighington, using EMTP/ATP for transient and stability analysis of an energy
efficient LNG ship power system, EEUG Meeting 2015
Variable With a variable speed drive, the speed of the asynchronous motor is controlled and,
compared with other methods, it has the smallest starting current but a considerable
Speed Drive starting torque. However, this starting equipment is the most expensive and complex of
all the methods reviewed in this paper. When the drive is used for starting only, there is a
Modeling parallel bypass device that connects the motor in direct on line once the start is finished.
The architecture of the VSD model, as shown in Figure 9, includes the rectifier, inverter,
and controller:
Figure 9
DOL Bypass
Principle of Starting Using
a VSD and a Bypass V grid
Contactor V motor Start Controller Block
I motor
AC DC Reactance Motor
Grid
INPUT DC AC OUTPUT
Rectifier Inverter
VSD
Once the motor reaches a reference speed, the control is switched to scalar control
and the voltage on the motor side is smoothly aligned with the network voltage.
The bypass switch is then closed and the VSD is progressively disconnected from
the network.
A cascaded H-bridge inverter was considered for the purposes of this paper because
of the advantages it offers, such as low switching stress, high quality load wave,
simple algorithm, easy packaging, and high equivalent switching frequency. The use
of cascaded inverters is becoming increasingly popular in variable frequency speed
regulation systems7.
A cascaded inverter is composed of several power cells. Each power cell has its own
DC power, produced by a multi-pulse rectifier. Figure 11 illustrates the topology of a
single-phase cascaded inverter. It is made up of 4 cells and the output waves are 2n+1,
or 9 levels. These multiple levels of voltage, hence power cells, help to reduce the output
harmonics and increase the equivalent switching frequency8. This higher switching
frequency means that the oscillation torque is also reduced. Respectively, the higher
number of power cells means that the input rectifier is also of a higher level. The harmonic
currents on the network side are also reduced through an appropriate coupling of the
secondary windings on the input transformer 9. There are 4 secondary windings in the
model presented below. The phasing is obtained using zigzag transformers.
It is worth mentioning that the input transformer model varies from an actual transformer
in that it does not consider mutual impacts from secondary windings, given that it has an
individual primary winding for each.
Figure 11
(A) Cascaded H-Bridge
Multilevel Inverter Model
(B) Topology
(A) (B)
7
ZHANG Jingjun, H 桥级联型多电平逆变器的研究 (Cascaded H-bridge multilevel inverters research), 2011
8
江友华 (JIANG Youhua), 曹以龙 (CAO Yilong), 龚幼民 (GONG Youmin),
基于载波相位移角度的级联型多电平变频器输出性能的研究 (Research on Output Performance of
Cascaded Multi - level Inverter Based on Carrier Phase Shift Angle), 中国电机工程学报 (Journal of China
Electromechanical Engineering), 76-81, 2007
9
WANG Zhaoan王兆安. (n.d.), 电力电子技术 (第四版) (Power Electronics).
Simulation Results
For motor starting, the torque is set to 60% of the rated value. Once the motor is close
to rated speed, the synchronization sequence is initiated. Figure 12 illustrates the
model obtained in ATPDraw and the starting results:
Figure 11 Switch
The results show very short and limited transients in the torque, at the change of the
control type and at the closing of the bypass contactor. During parallel operation in
direct on line and VSD mode, the consumed current is higher.
The motor consumption remains unchanged, but the variable speed drive and bypass
switch form a current loop and there is a consumption of reactive power.
When the VSD is stopped, the loop is opened and the consumed current from the
network and that of the motor become equal.
From the results shown above, it can be concluded that longer acceleration times are
obtained with progressive starting methods, such as star-delta, autotransformer and
soft starter, where the motor torque is reduced. Generally, longer starting times mean
also higher motor heating, except in the star-delta method where reduction in current
is higher than the increase in starting time. Motor torque is higher with rotor resistance
starting and variable speed drive. Voltage drop is exhibited most predominantly in
direct on line starting, with all other methods being very efficient in reducing the voltage
drop below 10%.
Conclusion Several motor starting methods are modeled and compared in this paper.
These models are intended for use in real projects for evaluating appropriateness of a
given starting method, or to be used as a basis for reference for developing simplified
comparison tools for non EMTP-ATP users. During the development of the models,
some EMTP-ATP limitations have been identified, including the difficulty of fitting the
motor model to datasheet data, and autotransformer operation with an open neutral.
However, in the majority of cases, EMTP-ATP and ATPDraw have allowed an easy
model implementation, and the results obtained match the expected behavior
for the various motor starting methods.
Zhimin Qin graduated from the Institut National Polytechnique de Grenoble in 2017. He is currently
a validation engineer, based in Shanghai.
Jerome Guillet graduated with a Master’s degree in mechanical and industrial engineering from the
Arts et Métiers ParisTech in 2004. He started his career at the electrolytic capacitors manufacturer
SICSAFCO as a technical engineer in 2005. He joined Schneider Electric Power Quality as an R&D
engineer in 2008. Since 2015, he has worked in the Motor Management Competency Center. He
currently assists during the early stages of large motor projects, and contributes to popularize motor
knowledge with dedicated calculation tools and training.
Wang Weisheng was born in Gansu, China in 1983. He graduated with a Bachelor degree from Tongji
University in 2005. He is now employed by Schneider Electric as a solution architect implementing
drive application solutions. His specialist field of interest is synchronous motor control.
Appendices
Denominations of variables
Annex used in vector control:
P Laplacian operator
Lr Rotor inductance
Lm Mutual inductance
Lr/Rr (rotor resistance)
a Number of pole pairs
Coefficient of friction
J Inertia
Rotor flux in D axis
Contact us
For feedback and comments about the content of this white paper:
motor.management@schneider-electric.com
998-2095-11-08-18AR0_EN