You are on page 1of 6

Analysis of efficiency characteristics of squirrel-

cage induction motor for pump applications



L. Szychta, R. Figura
Technical University of Radom, Poland, Faculty of Transport and Electrical Engineering


Abstract. Criteria of improving energy efficiency of water system
pumping plants are formulated. Impact of pumps on a control
system is defined. Efficiency characteristics of a squirrel-cage
induction motor at varying rotational speeds and load torques
are presented. The modified Air Gap Torque (AGT) method is
introduced to determine efficiency of a squirrel-cage induction
motor. Results of testing which continues the authors work on
application of AGT method to estimating induction motors
efficiency are discussed. The definition of estimated losses is
presented. The proposed method is verified in a number of
selected low-power motors.
Keywords: AGT method, motor efficiency, motor efficiency
testing, efficiency estimating.
I. INTRODUCTION
The share of pumps in electricity consumption is
estimated at (25-30)% of national electricity consumed to
power machinery. A variety of methods are applied to
reduce the electricity consumption. Manufacturers use
solutions to improve efficiency of pumps and electric
motors. A method of pump operation control with regulated
rotational speed limits the energy consumption typical of
throttling or discharge methods. Some automatic solutions
feature not only the classic regulators but also expert
systems such as neural networks, fuzzy logic or genetic
algorithms.
The article assumes regulation of the pumps rotational
speed with the classic PI regulator. Additional savings on
electricity can be implemented by designing a method of
control that ensures pump operation in the vicinity of the
point of optimum pump efficiency. The article intends to
present a method of pump operation control under varied
conditions at best possible efficiencies, thus resulting in
reduced energy consumption of pumping.
II. ENERGY CONSUMPTION OF PUMPING UNIT
Application of modern power electronic elements in
drives of pumping sets allows for development of efforts to
rationalise energy consumption. Quantity of consumed
electricity depends on parameters of a pumping set that
comprises a pump and its electric drive. Drive of a pumping
set consists of a motor and a feeding device, for instance,
a frequency converter or any other electrical device that
powers the motor directly from the three-phase mains
system.
Pumping sets in serial or parallel connections between
suction and discharge branches make up a pumping unit.
A pumping unit including one pumping set is a special case.
Pumps in a unit are assumed to have identical rated
efficiencies Q
N
, and heads H
N
.
Components of a pumping unit and its control systems
ought to limit consumption of electricity, which means
minimising momentary value of electric power P
el
used by a
unit, which is defined as follows:

_
=
=
m
i
zi
i i
el
H gQ
P
1
q



(1)
where:
f s z
q qq q =
(2)

q pump efficiency,
q
f
efficiency of frequency converter,
q
s
efficiency of induction motor,
q
z
efficiency of pumping set,
density of medium,
m number of working pump sets.
III. PUMP EFFICIENCY CURVE
In practice, it is important to control relative pump
rotational speed n
-
in order to stabilise head of water h
-
,
where:
N
n
n
n =
-
relative rotational speed,
N
H
H
h =
-
relative head,
n rotational speed,
n
N
rated rotational speed,
H head [m],
H
N
rated head.
It also influences pump efficiency q. Thus, knowledge
of characteristics of pump efficiency q as a function of
rotational speed n
-
and capacity q
-
at h
-
=const, becomes a
significant element of analysis of pumping energy
consumption. Characteristics of pump efficiency q=f(n
-
),
and q=f(q
-
) at h
-
=const are to a certain degree dependent
on the pump type, e.g. one-stream, two-stream, and multi-
grade.
where:
N
Q
Q
q =
-
relative capacity of pumping set,
Q capacity of pumping set,
Q
N
default capacity of pumping set.
Most applications include the condition of maintaining
constant pressure in outlet of pump stations. In case of
constant head of inflow
-
zN
h constant head of water h
-

obtains for the pump [22]. Constant head of water h
-
is
arrived by changing capacity q
-
of one or several pumps in a
unit. Capacity q
-
is most often changed by shifting the
pump rotational speed n
-
.
978-1-4673-0141-1/12/$26.00 2012 IEEE
71
a)

q
0,9
0,8
0,7
0,6
0,5
0,4
0,2 0,4 0,6 0,8 1,0 1,2 1,4 1,6
h*=0,6
h*=0,8
h*=1
h*=1,2
q*

b)
Fig. 1. Pump efficiency at constant heads of water h
-


Sample pump efficiency characteristics at constant
heads of water h
-
, of flow q
-
, and constant rotational speed
n
*
are shown in Figure 1. For purposes of the analysis,
pump efficiency q was assumed to vary within the range
0.48-0.8. Pump efficiency q varies slightly at different
heads of water h
-
and at constant flow q
-
less than the rated
flow of one (Fig.1.a). In case of capacity q
-
below the rated
value, efficiency variations are below 5%. At flow q
-
> 1
values of efficiency q vary considerably. Differences
among the pump efficiencies at constant capacity q
-
over
the rated flow reach as much as 15% within the range of
head h
-
e(0,6; 1,2).
Pump efficiency q is heavily dependent upon variations
of rotational speed n
-
(Fig.1.b). A reduction of rotational
speed n
-
by 0.1, at constant head of water h
-
, leads to a
reduction of pump efficiency by 0.3. A steep decline in
pump efficiency is counter-effective in terms of LCC [6].
In order to fully analyse the nature of changes of pump
efficiency q, the efficiency characteristic curve was divided
into two slopes: lower and upper. Both parts of the
efficiency characteristic curve have their domains, defined
by boundary values of rotational speed n
-
and the
corresponding boundary pump efficiencies q. The lower
slope is defined in the following way [20]:
( ) c bn a
N
+ =
- -
ln q
(3)
The upper slope is approximating by function:
( ) ( )
1 1
2
1
c n b n a
o
+ + =
- - -
q
(4)
The equations (3) and (4) allow to control pump speed to
achieve the highest pump efficiency q.
Speed changes also affect the efficiency of the squirrel-
cage induction motor. This means that the efficiency of the
pump set depended on pump and motor efficiency (2).
Therefore, in the following section will present modified
non-intrusive method for determining on-line the efficiency
of the motor.

IV. INDUCTION MOTOR EFFICIENCY CURVES
Determining pump efficiency as a function of rotational
speed is not sufficient to define consumption of electricity
by a pumping set. In line with the dependence (1),
efficiency variations of squirrel-cage induction motor must
be defined as well. Most pump drives employ squirrel-cage
induction motors whose rated efficiency q
s,
at power of
above 2kW, is most often greater than 80%. High efficiency
is maintained across a wide variation range of the relative
torque T
*
loading the motor's shaft, at rated frequency and
supply voltage (Fig.2).
Rated frequency voltage motor supply limits the choice
of methods of controlling a pump station's productivity. The
commonly applied, variable rotational speed pump drives
require that squirrel-cage induction motors be supplied with
voltage of variable: frequency and rms value. These
parameters influence rotational speed and electromagnetic
torque gained by the motor [12, 17].

Fig.2. Efficiency of induction motor q
s
of 7.5kW INDUKTA as a function
of relative load torque T
*
[23], motor is supplied by sinusoidal voltage
Harmonics of current and voltage are also greater, which
induces additional losses in the motor, reducing its
efficiency q
s
[1, 9, 12, 13, 18, 19, 22, 23]. [14] presents a
method of calculating losses in the induction motor powered
from a frequency converter, applying the induction machine
model, where each harmonic of the supply voltage
corresponds to an independent equivalent circuit.
Figure 2 shows efficiency of induction motor q
s
as a
function of relative load torque T
*
, where:
N
T
T
T =
*

T motor torque,
T
N
rated motor torque.
The impact of all losses on efficiency variations of the
motor q
s
supplied with sinusoidal voltage is presented.
Efficiency q
s
of both the supply systems has similar values
in the frequency range fe(2550)Hz. Usually the same
range frequency of voltage supply is applied for water pump
set drives. The possibility of computing the efficiency of a
motor in a working pump set by measurements of electrical
variables is considered in the article.
A number of methods are known of determining
efficiency q
s
of an induction motor. They can be divided as
follows [3]:
methods employing indirect slip measurements,
methods based on measurements of stators phase
current,
72
methods based on equivalent diagrams,
methods of determining partial losses,
methods of determining air gap torque,
methods of motor shaft torque measurement.
Efficiency
s
is defined in the following dependency:
1
2
P
P
s
= q
(5)
where: P
1
mean active power consumed by a motor, P
2

mean active power discharged by a motor.
In the special case of symmetrical and sinusoid variable
voltages supplied to a three-phase symmetrical motor, P
1

can be assumed to equal [7,8]:
W W V V U U
i u i u i u P + + =
1
(6)
where: u
U
, u
V
, u
W
momentary values of voltages supplied
to the motor; i
U
, i
V
, i
W
momentary values of conduction
currents.
P
2
in laboratory conditions was determined by direct
measurements of mean torque T and mean rotational speed
n of the rotor according to the formula [7,17]:
60
2
2
Tn
P
t
= (7)
V. METHOD OF ESTIMATING P
2

The method of determining q
s
by means of nonintrusive
estimation of P
2
has been adopted in this article, which
involves calculation of the electromagnetic torque in the
motors air gap. This solution does not require installation
of specialist measurement instrumentation in industrial
drive systems.
The division of induction motors power losses (Fig. 3)
results in a dependency between P
2
and power P

of the
rotating field [17]:
m dodr Fer Cur
P P P P P P A A A A =
2
(8)



Fig. 3. Division of squirrel-cage induction motors power losses; P
Cur

losses across rotors winding; P
Fer
losses in rotors core; P
dodr

additional loading losses of rotor; P
m
motors mechanical losses, P
Cus

losses across stators winding, P
Fes
losses in stators core

P

of the rotating field is determined on the basis of


mean electromagnetic torque T
ag
in the air gap and
rotational speed n
s
of the magnetic field:
60
2
s ag

n T
P = (9)
where: mean T
ag
is defined according to a momentary value
of electromagnetic torque:
}
=
T
ag ag
dt t
T
T
1

(10)
Considering the formula
Cur
sP P = A for power losses
across rotors winding and the equation (7), P
2
can be
defined as:
m dodr Fer
ag
P P P
n T
P A A A =
60
2
2
(11)
(11) provides the basis for a method of estimating P
2

across the motor shaft at the time of its operation with the
control system of electric drive (11) provides the basis for a
method of estimating P
2
across the motor shaft at the time
of its operation by means of the control system of electric
drive. T
ag
is a result of momentary value of t
ag
defined as a
module of the product of momentary values of stator
streams vector
s
times momentary values of stator currents
vector i
s
[11, 15, 16]:
s s
i = p t
ag
(12)
where: p number of pole pairs in the motor;
s
vector of
values of stator streams; i
s
vector of values of stator phase
currents.
The stream
s
(12) occurs in the motors air gap and
determines power P

of the rotating field (Fig.3). Its value


is based on the equivalent diagram of the induction motor
[7] which ignores losses P
Fes
in the rotor core. Their
knowledge is necessary for determining P

according to
Figure 5. Therefore, the concept of an equivalent vector
st

of momentary values of stator stream
st
is introduced:
( )dt
}
=
s s s st
R i u (13)
where: R
s
matrix of phase resistances of stator winding; u
s

vector of momentary values of phase voltages supplied to
the motor.
This equivalent vector
st
has produced an equivalent
power P
t
of the rotating field (Fig. 3), which conforms
with the analysis in [11, 15, 16]. Dependences (12) and (13)
for three-wire, sinusoidal, and symmetrical power supply
system of a three-phase motor produce the following
expression of momentary electromagnetic torque t
agt
:
( ) ( ) | |
} }
= dt i R u i dt i R u i p t
V s V U U s U V agt
3
(14)
With regard to inter-phase voltages, (14) can be
formulated as:
(
(
(
(

|
.
|

\
|
+
+
+
+ |
.
|

\
|

=
}
}
dt i R
u u
i
dt i R
u u
i
p t
V s
WU UV
U
U s
WU UV
V
agt
3
2
3
3

(15)
t
agt
(15) helps to compute the equivalent power P
t
of the
rotating field (Fig. 3) which is, according to the earlier
assumption, defined as:
Fes t
P P P A + =
+
(16)
Available literature [11, 15, 16] employs P
t
to compute
the discharged power P
2
. Losses P
Fes
in the stators core
and P
Fer
in the rotors core are ignored. Additional loading
losses P
dodr
are assumed to comply with IEEE 112 [24].
The assumed P
dodr
have tabulated values and are
73
percentages of P
2
discharged by a motor (Tab. 1).
Mechanical losses P
m
are assumed to equal 1.2% P
2
.

Table 1. Additional loading losses as per IEEE 112
Range of motor power
Percentage share of additional losses
P
dodr
in the discharged power P
2

1 90 kW 1.8%
91 375 kW 1.5%
376 1850 kW 1.2%
1851 kW and up 0.9%

These authors have verified the model of motor losses
adopted in [11,15,16] by comparing the actual motor
efficiency for P
2
calculated according to (7) with its
estimate, ignoring the iron losses. The results indicate great
estimation errors, therefore, the simplifications accepted in
[11,15,16] cannot be applied to practical calculation of
motor efficiency at the time of operation.
These negative results of motor efficiency estimation
[4,5] led the authors to formulate new dependencies to
compute P
2
discharged by a motor. As part of the
assumptions for this testing method, the theoretical and
laboratory values of the equivalent power P
t
of the rotating
field were demonstrated to match.

Fig. 4. A simplified division of power losses of squirrel-cage induction
motor
It was therefore assumed that stator copper losses P
Cus

are constantly identified. Due to the variety of calculation
methods, the remaining losses were assumed to belong to a
single group, defined as estimated losses P
est
(Fig. 4).
On the basis of such assumptions, Sankey diagram is
simplified to a form shown in Figure 6. (11), which defines
P
2
, is expressed as:
est
agt
P
n T
P A =
60
2
2
(17)
where: P
est
estimated losses defined by [2, 4]:
m dodr Fer Fes
s
est
P P P P
n
n
P A A A A = A
(18)
P
est
are assumed here to be a function of rotational
speed n (P
est

= f(n)) and expressed as:
( )
o
* *
n P
est
= A
(19)
where: P
est
*
relative estimated losses; n
*
relative
rotational speed of the rotor; o powers exponent, where
o e R,
with:
estN
est
est
P
P
P
A
A
= A
*

(20)
N
n
n
n =
*
(21)
where: P
est
estimated losses; P
estN
estimated losses
under rated operating conditions of the motor; n rotational
speed of the motor; n
N
rated rotational speed of the motor.
According to (19), P
est
are zero for n equal zero. Where
the rated rotational speed n=n
N
of the rotor, estimated losses
correspond to the rated P
est
(n
N
)=P
estN
which is expressed
as:
N N agtN estN
P n T P
2
60
2
= A
t
(22)
where: P
2N
rated motor power; T
agtN
mean rated torque
t
agtN
in the motors air gap, defined by:
(
(
(
(
(

|
.
|

\
|
+
+
+
|
.
|

\
|

=
}
}
dt i R
u u
i
dt i R
u u
i
p t
VN s
WUN UVN
UN
UN s
WUN UVN
VN
agtN
3
2
3
3

(23)
where: i
UN
, i
VN
momentary currents across the stators
winding for rms value of current I=I
N
; u
UVN
, u
WUN

momentary conduction voltages for rms value of voltage
U=U
N
.
The estimated rated losses P
estN
are the second point of
the characteristic P
est
=f(n). There is no rule for
determining the exponent o (19). It is assumed to be
verified by empirical results. At this stage, its value has
been assumed as o=1 [5].
VI. LABORATORY TESTING
The testing employed a laboratory stand whose flow
diagram is illustrated in Figure 7. IM2 motors (Tab. 2),
were supplied with sinusoidal variable voltage from an SG
of rated power S
GN
=4.0 kVA. IM2 motors were loaded with
a DCG. The drive assembly DML and a programmed
rotational speed meter helped to produce a desired braking
torque across the motor shafts.
The testing process employed a script IT control system
for automatic industrial equipment [10] developed by the
Department of Electric Drive and Industrial Electronics,
Faculty of Transport and Electrical Engineering, Technical
University of Radom. It provides for simultaneous reading
of measured values.
Table 2. Ratings of tested motors
No. Manufacturer
P
N

[kW]
U
N

[V]
I
N

[A]
n
N

[rpm]
cos
[-]

[-]
1 INDUKTA 2.2 400 4.8 1425 0.80 0.82
2 INDUKTA 2.2 400 5.0 2870 0.77 0.82
3 TAMEL 1.5 380 3.7 1420 0.80 0.77
4 TAMEL 1.5 380 3.5 1410 0.80 0.81

The testing covered the frequency range of 25 - 50Hz as
operation of pumping sets or compressors in wider
regulation ranges is not practicable. Motor efficiency was
determined by direct measurements of shaft torque in line
with PN-EN 60034-2-1 [24]. These measured efficiencies
of the test motors were treated as actual values.









Fig. 5. Flow diagram of the measurement station

74

a)


b)


c)


d)


Fig. 6. Motor efficiency as a function of electromagnetic torque =f(T
*
): a)
motor 1, b) motor 2, c) motor 3, d) motor 4.

a)


b)


c)


d)


Fig. 7. Motor efficiency as a function of electromagnetic torque =f(T
*
) on
introduction of factor |: a) motor 1, b) motor 2, c) motor 3, d) motor 4.
75

Estimated efficiency q
sest
of an induction motor was
based on equivalent power of the rotating field P
t
(16)
according to the modified AGT method on consideration of
estimated losses P
est
(19). Calculation results of efficiency
q
sest
and actual efficiency q
s
are shown in Figure 8. q
s

determined by direct measurements of shaft torque of a
tested motor is plotted with black dots. Estimated efficiency
q
sest
is shown with a broken line. The results vary
substantially across motor models. In respect of motor 1
(Fig. 6a), the estimation error of efficiency is satisfactory
and below 4% while the divergences are unacceptable for
the remaining motors (Fig. 6b,c,d). Accordingly, the
definition of estimated losses P
est
in (19) must be verified
by means of a correction factor |:
( )
o
|
* *
n P
est
= A
(24)
| varies for the test motors. It includes variations of
power losses over the lifecycle of a motor. Application of
an individually selected | for each motor reduced the error
of estimated efficiency (Fig. 7).
Efficiency estimation results for motor 1 do not require
application of a correction factor | (Fig. 6a and 7a). In the
circumstances, | becomes 1. The error estimation is
comparable to the systematic error determined according to
accuracy classes of the measurement instruments.
Efficiency estimation results for motors 2, 3, and 4 differ
from measured values and the estimation error rises as the
motor shaft torque reduces (Fig. 6b, c, and d).

Introduction of | reduced the estimation error
(Fig. 7b, c, and d) to the level of the systematic error which
is equal 1% for rated power P
N
and 7% for 0.3P
N
. The
correction factor is in the range 0.6 0.9 with regard to
motors 2, 3, and 4.
VII. CONCLUSION
The test results demonstrated that the modified
AGT method together with estimated motor losses can be
applied to determining efficiency of a squirrel-cage
induction motor at the time of its operation. Simultaneous
estimation of pump efficiency helps to select such a range
of rotational speeds as to maximise energy efficiency of a
pumping set. The results are of particular importance to
water supply systems where water heads and distributions
are variable. Research must continue to improve accuracy
of estimating , which defines estimated losses.
VIII. REFERENCES
[1] D. Elton: An Application At a Model Based Measurement
Technique to the Inspection of Induction Machines, IMEKO
1988,
[2] R. Figura, E. Szychta, L. Szychta, In-Service Efficiency
Estimation with the use Modified Air Gap Torque Method for
Squirel Cage Induction Motor, Electronics and Electrical
Engineering,No 8, 2011,
[3] R. Figura, L. Szychta: Extreme controlling for pump set of
irrigation systems. Monographs Faculty of Transport, 2008,
[4] R. Figura, L. Szychta, E. Szychta: Wyznaczanie wspczynnika
sprawnoci silnika klatkowego pracujcego ze zmienna
prdkoci obrotow (Determining Efficiency of a Squirrel-
Cage Motor Operating At Variable Rotational Speed). Zeszyty
Problemowe Maszyny Elektryczne, 2011,
[5] R. Figura, L. Szychta: Estymacja sprawnoci silnika
indukcyjnego klatkowego pracujcego w zespole pompowym
(Estimating Efficiency of a Squirrel-Cage Induction Motor
Operating As Part of a Pumping Set). Zeszyty Problemowe
Maszyny Elektryczne, 2011,
[6] L. Frenning, et alia: Pump Life Cycle Costs a Guide to LCC
Analysis for Pumping Systems, 2001.
[7] Z. Grunwald: Napd elektryczny (Electric Drives). WNT,
(1987),
[8] J. Hsu, J.; Kueck, J. D.; Olszewski, M.; Casada, D. A.; Otaduy,
P. J.; Tolbert, L. M.: Comparison of induction motor field
efficiency evaluation methods, IEEE Transactions on industry
applications, 34 1998,
[9] E. Howard: Energy efficient electric motors and their
application, USA 1983,
[10] R. Kwiecie, L. Szychta, R. Figura: Skryptowy informatyczny
system sterowania urzdzeniami automatyki przemysowej
(Script IT System for Control of Industrial Automatic
Equipment), Przegld Elektrotechniczny Electrical Review,
(2010),
[11] J.D. Kueck, M. Olszewski, D.A. Casada, J. Hsu, P.J. Otaduy,
L.M. Tolbert: Assessment of methods for estimating motor
efficiency and load under field conditions, Oak Ridge National
Laboratory ORNL, (1996),
[12] P.C. Krause: Analysis of Electric Machinery, McGraw-Hill
International Editions, Singapore 1987,
[13] R. Kaczmarek, M. Amar, F. Protat: Iron Loss Under PWM
Voltage Supply on Epstein Frame and In Induction Motor Core,
Tramsaction on Magnetics 1/1996,
[14] J. Kokotkiewicz: Projektowanie silnikw indukcyjnych
klatkowych o zmniejszonych stratach mocy spowodowancy
zasilaniem poprzez przemiennik czstotliwoci (Designing
Squirrel-Cage Induction Motors with Reduced Power Losses
When Supplied From Frequency Converters), BOBRE Komel
Katowice. 1995,
[15] B. Lu, T.G. Habetler, R.G. Harley: A Survey of Efficiency
Estimation Methods of In-Service Induction Motors with
Considerations of Condition Monitoring Requirement,. Electric
Machines and Drives, 2005,
[16] B. Lu, T.G. Habetler, R.G. Harley: A Nonintrusive and In-
Service Motor Efficiency Estimation Method using Air-Gap
Torque with Considerations of Condition Monitoring, Industry
Applications Conference, 2006,
[17] A.M. Plamitzer: Maszyny Elektryczne (Electric Machines),
WNT, Warszawa 1986,
[18] J.R. Przygodzki: Minimisation of Energy Loses In Magnetic
Cores of Electric Motors and Transformers, Energy Saving in
Electrical Engineering, Warsaw 2001,
[19] Pump Save 2,0.: Energy Saving Calculator for Pump Drivers,
ABB 1999,
[20] L. Szychta: Zasady doboru system sterowania pompowni
wodocigowych (The rulse of water pumping stadion control),
Radom, 2006,
[21] L. Szychta, System sterowania pompowni wedug algorytmu
pompy wiodcej (Pumping plant control system following the
lead pump algorithm), IV Krajowa Konferencja Postpy w
Elektrotechnice Stosowanej, Kocielisko 2003,
[22] K. Zakrzewski: Overloss coefficient for dynamo sweet during
axid magnetization with nonsinusoidal flux, Archives of
Electrical Engineering Warsaw, Poland 3/1997,
[23] K. Zakrzewski: Porwnanie charakterystyk ekesploatacyjnych
silnika indukcyjnego przy zasilaniu napiciem sinosoidalnym i
napiciem falownikowym MSI (Comparison of Operation
Characteristics of Induction Motor Supplied With Sinusoid
Voltage and MSI Converter Voltage), Politechnika dzka,
d. 1993,
[24] IEEE Standard test procedure for polyphase induction motors
and generators. IEEE 112, 2004, Polish Standard PN-EN
60034-2-1, Rotational electrical machines Part 2-1:
Standardised methods of determining losses and efficiencies by
testing (excluding traction vehicle machinery), PKN, (2010).


Authors: Prof. Leszek Szychta, Eng., Technical University of Radom,
Poland, Faculty of Transport and Electrical Engineering; ul.
Malczewskiego 29, 26-600 Radom, E-mail: l.szychta@pr.radom.pl;
Radosaw Figura, M.Eng., Technical University of Radom, Poland,
Faculty of Transport and Electrical Engineering; ul. Malczewskiego 29,
26-600 Radom, E-mail: r.figura@pr.radom.pl.

76
Powered by TCPDF (www.tcpdf.org)