IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 55, NO.

1, JANUARY/FEBRUARY 2019 273

A Simple Method to Determine Double-Cage Rotor

Equivalent Circuit Parameters of Induction Motors
From No-Load and Locked-Rotor Tests
Shu Yamamoto , Member, IEEE, Hideaki Hirahara, Member, IEEE, Akira Tanaka, and Takahiro Ara

Abstract—This paper presents a simple method to determine

double-cage rotor equivalent circuit parameters of induction mo-
tors (IMs). A new formula to calculate the double-cage rotor equiv-
alent circuit parameters is proposed. The formula is sufficiently
simple to be calculated by a calculator but can determine the
parameters of outer- and inner-cage conductors. The proposed
method takes the magnetic saturation effect of a closed-slot rotor
into account and applies to not only double-cage but also single-
cage IMs. All equivalent circuit parameters can be determined us-
ing only classical no-load and locked-rotor tests. Thus, the proposed
method is suitable to be used as a standardized testing method for
IMs. The validity of the proposed method is demonstrated by exper-
imental results on four 5.5 kW-200 V-4 P-22 A-50 Hz IMs: double-
and single-cage IMs with semi-closed- and closed-slot rotors.
Index Terms—Closed-slot rotor, double-cage rotor equiva-
lent circuit, induction motor (IM), leakage reactance, magnetic
saturation.

I. INTRODUCTION
HE double-cage rotor is effective in construction to im-
T prove the starting performance of induction motors (IMs),
realizing high starting torque and low starting current. It is im-
portant to determine the parameters of the equivalent circuit
model of double-cage IMs, as shown in Fig. 1 [1], [2], to calcu-
late and analyze the performance of double-cage IMs. Methods
to measure or estimate r1 , x1 , xm , and rf e in Fig. 1 are well
established [3], [4]. However, no standardized methods exist to Fig. 1. Equivalent circuit models of double-cage IM. (a) Conventional model
based on physical image of double-cage rotor. (b) Reduced model in which x2
determine the parameters of the double-cage rotor equivalent is assumed to be zero.
circuit.
To address this situation, many researchers have studied meth-
ods to obtain the double-cage rotor parameters. In [5], a method on the rotor leakage reactance. In [6]–[10], iterative parameter
to determine a correct set of parameters in Fig. 1(b) for the identification using tests and manufacturer name plate data is
condition x1 = x23 is presented. However, it is difficult to take proposed. In [11] and [12], an operational impedance obtained
account of the effect of magnetic saturation of the slot bridge from standstill frequency response test is used to identify the
parameters of an IM equivalent circuit model having two ro-
Manuscript received April 6, 2018; revised June 17, 2018; accepted July 8, tor circuits such as in Fig. 1. In [13]–[15], parameter calculation
2018. Date of publication August 6, 2018; date of current version December 12, strategies based on finite-element analysis are presented. In [16],
2018. Paper 2018-EMC-0260.R1, presented at the 2017 IEEE Energy Conver- free acceleration data are utilized to identify the parameters.
sion Congress and Exposition, Cincinnati, OH, USA, Oct. 1–5, and approved
for publication in the IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS by Furthermore, the works in [6] and [15]–[17] point out that the
the Electric Machines Committee of the IEEE Industry Applications Society. magnetic saturation of the narrow bridge of the closed-slot rotor
This work was supported by the Japan Society for the Promotion of Science core should be taken into account. However, these methods re-
KAKENHI under Grant 16K06254. (Corresponding author: Shu Yamamoto.)
The authors are with the Faculty of Human Resources Development, Poly- quire not only complicated computer-based calculation but also
technic University, Tokyo 187-0035, Japan (e-mail:, [email protected]; extra driving test data (or motor dimension data in case of finite-
[email protected]; [email protected]; [email protected]). element analysis). If it is possible to determine all double-cage
Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org. rotor equivalent circuit parameters of IMs only from no-load
Digital Object Identifier 10.1109/TIA.2018.2864105 and locked-rotor tests standardized in [3] and [4] while taking
0093-9994 © 2018 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See http://www.ieee.org/publications standards/publications/rights/index.html for more information.

Authorized licensed use limited to: MALAVIYA NATIONAL INSTITUTE OF TECHNOLOGY. Downloaded on February 03,2021 at 14:07:03 UTC from IEEE Xplore. Restrictions apply.
 274 IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 55, NO. 1, JANUARY/FEBRUARY 2019

in account the magnetic saturation effect of the closed-slot ro- x23 even when r2 , r3 , x3 , and x23 are determined so that input
tor, they can be used to evaluate the performance of double-cage and rotor impedances, ZS and ZR , in Fig. 1(b) are equal to those
IMs. However, such methods have not been presented so far. in Fig. 1(a). Thus, the authors have distinguished them using the
Thus, the main objectives of this paper are to propose a sim- accent as in Fig. 1(a).
ple method to determine the double-cage rotor equivalent circuit The authors propose a new method to determine x23 , r2 , r3 ,
parameters of IMs only from no-load and locked-rotor tests and and x3 from rotor equivalent circuit impedances ZR at two
demonstrate the validity of the proposed method with experi- different frequencies f1 and f2 . Here, f1 is the rated frequency
mental results. and f2 should be less than f1 .
The authors have previously presented a method to determine A set of simultaneous equations based on the relation-
the parameters of double-cage IMs [18]. In [18], a new formula ships between four unknowns (x23 , r2 , r3 , and x3 ) and the ro-
to calculate the parameters of the reduced double-cage rotor tor impedances (ZR 1 ( = rR 1 + jxR 1 ) at f = f1 and ZR 2 ( =
circuit model, as shown in Fig. 1(b), is proposed. Although the rR 2 + jxR 2 ) at f = f2 ) can be set up as follows:
formula is sufficiently simple so that it can be calculated using  
1
a calculator, it allows one to determine the parameters of the rR 1 = Real of jx23 + (1)
outer- and inner-cage conductors accurately. We also described 1/r2 + 1/ (r3 + jx3 )
 
a method to determine the double-cage rotor equivalent circuit 1
parameters considering the magnetic saturation effects of the xR 1 = Imaginary of jx23 + (2)
1/r2 + 1/ (r3 + jx3 )
closed-slot rotor. We found that, in Fig. 1(b), magnetic saturation  
of the narrow slot bridge of the rotor core greatly affects the 1
rR 2 = Real of jx23 /k + (3)
mutual leakage reactance between the outer- and inner-cage 1/r2 + 1/ (r3 + jx3 /k)
conductors x23 , but has little influence on the leakage reactance  
1
of the inner-cage conductor x3 . On the basis of this feature, xR 2 = Imaginary of jx23 /k + .
1/r2 + 1/ (r3 + jx3 /k)
we presented a method to obtain the parameters of Fig. 1(b) (4)
considering the magnetic saturation effects of the bridge of the
closed-slot rotor. We also described a fast calculation algorithm It is difficult to solve the simultaneous equations directly.
to predict the driving performance. We then considered whether However, it is possible to solve an equation in which x23 , r3 ,
these methods can be applied to not only double- but also single- and x3 are eliminated to obtain r2 , as shown in (5). The obtained
cage IMs in which the rotor equivalent circuit impedance is r2 can then be used to derive r3 and x3 , which are, respectively,
frequency dependent and magnetic saturation of the closed-slot given by (6) and (7). Finally, x23 is obtained by (8). One can see
bridge affects the rotor leakage reactance. However, the validity that (5)–(8) are very simple formulas to determine the double-
for such single-cage IMs was not discussed in [18]. cage rotor circuit parameters in Fig. 1(b)
This paper presents an improved procedure to determine the  
1 k 2 (rR 1 − rR 2 )3
double-cage rotor equivalent circuit parameters. Extended for- r2 = 2 k rR 1 − rR 2 +
2
(5)
mulas that can employ an arbitrary frequency in the locked-rotor k −1 (kxR 2 − xR 1 )2
test are derived. An improved parameter determination method  
r2 (k 2 r2 rR 2 − k 2 − 1 rR 1 rR 2 − r2 rR 1 )
that suppresses the unevenness in the locked-rotor test current r3 = (6)
is shown. This procedure has been arranged so that the parame- (k 2 − 1) (r2 − rR 1 )(r2 − rR 2 )

ters can be determined by a common process applicable to both k r2 4 (rR 1 − rR 2 )(rR 2 − k 2 rR 1 + (k 2 − 1)r2 )
double- and single-cage IMs. The proposed method is imple- x3 = 2
k −1 (r2 − rR 1 )2 (r2 − rR 2 )2
mented on four 5.5 kW-200 V-4 P-22 A-50 Hz IMs: double- and
single-cage IMs with semi-closed- and closed-slot rotors. The (7)
validity of the proposed method is demonstrated by comparing r2 2 x3
experimental and calculation results of the driving performance x23 = xR 1 − . (8)
(i.e., torque–slip and current–slip curves). (r2 + r3 )2 + x3 2
In addition, standstill testing methods using adjustable single- Since there are four parameters (x23 , r2 , r3 , x3 ) in the ro-
phase ac voltage sources and application to deep bar rotor IMs tor circuit model shown in Fig. 1(b), these parameters can be
are discussed. identified uniquely when two rotor circuit impedances in dif-
ferent frequencies are obtained at least. The rotor circuit model
shown in Fig. 1(b) physically corresponds to the rotor construc-
II. FORMULA TO CALCULATE DOUBLE-CAGE ROTOR
tion and can express the behavior of the frequency effects on
EQUIVALENT CIRCUIT PARAMETERS the dual cage. For instance, to improve the starting torque, the
In Fig. 1(a), the same rotor impedance ZR can be expressed outer circuit resistance r2 should be larger than the inner cir-
by an infinite number of combinations of the rotor parameters cuit resistance r3 since the starting current is concentrated on
x23 , r2 , x2 , r3 , and x3 [19]. Among these combinations, the au- the outer conductor. The mutual leakage reactance between the
thors focus on a method to find the parameters of Fig. 1(b), which outer and inner conductors x23 exists. As a result, it is found that
is a combination when x2 = 0 in Fig. 1(a). Here, r2 , r3 , x3 , and the frequency effects on the dual cage for entire slip range (from
x23 in Fig. 1(b) do not completely correspond to r2 , r3 , x3 , and start-up to no-load operations) are not directly measured in the

Authorized licensed use limited to: MALAVIYA NATIONAL INSTITUTE OF TECHNOLOGY. Downloaded on February 03,2021 at 14:07:03 UTC from IEEE Xplore. Restrictions apply.
 YAMAMOTO et al.: SIMPLE METHOD TO DETERMINE DOUBLE-CAGE ROTOR EQUIVALENT CIRCUIT PARAMETERS OF INDUCTION MOTORS 275

Fig. 2. Modeling of the slot. (a) Definition of slot dimension. (b) Leakage flux
paths of double-cage rotor.

Fig. 3. Equivalent circuit of IM considering magnetic saturation in the bridge

of closed-slot rotor.
proposed method, but can be indirectly considered through the
circuit model in calculating the driving performance for the
entire slip range. as follows:
Theoretically, the rated slip frequency is the best for f2 . In H
the real world, however, accurate impedance measurement be- Bs = (10)
α + βH
comes difficult when the slip frequency is very like the rated
slip frequency. Taking this tradeoff into account, the authors set where H is the magnetic field intensity and α and β are constants
that f2 = f1 /4 (namely, k = 4). used for approximating BS − H characteristics. This is the so-
called Frohlich equation. After substituting γIR for H in (10),
the authors substitute (10) for Bs in (9) and multiply (9) by ω. γ
III. DETERMINATION OF DOUBLE-CAGE ROTOR EQUIVALENT is a constant used for approximating the relation between IR and
CIRCUIT PARAMETERS CONSIDERING MAGNETIC SATURATION H and ω the angular frequency of the power source. Assuming
IN THE ROTOR-SLOT BRIDGE that x23 can be expressed as ωls , (9) can be rewritten as follows
[21]:
A. Derivation of an Approximate Expression of Mutual
Leakage Reactance x23 B
x23 (IR ) = A + (11)
IR + C
In this section, the authors derive a proper approximation
equation capable of expressing the change in x23 with respect where A, B, and C are the constants determined from the locked-
to IR . rotor tests. The procedure is explained in the next section. Note
A narrow bridge of iron core exists in the closed-slot rotor, that α, β, γ, a, b, c, and S are not required in determining not
as shown in Fig. 2. The authors noted that the mutual leakage only A, B, and C but also the equivalent rotor circuit parame-
flux φ23 does not pass through any air gap clearances. Thus, ters.
x23 is thought to vary with rotor current IR since magnetic
saturation occurs in the bridge, while there is little influence B. Determination of the Double-Cage Rotor Equivalent
of magnetic saturation on x3 . This is because the leakage flux Circuit Parameters
of the inner-cage conductor φ3 does not pass the bridge. The Fig. 3 shows the proposed equivalent circuit in which (11)
leakage flux of the outer-cage conductor φ2 is very small since is applied to x23 to take magnetic saturation of the bridge of
it has a short magnetic path and passes through high magnetic the closed-slot rotor into account. Fig. 4 shows the procedure to
reluctance parts such as the bridge and air gap clearance. This determine the parameters in Fig. 3.
suggests that the effect of magnetic saturation of the bridge on In Fig. 4, stator and exciting impedances (Z1 and Zm ) are first
x23 should be mainly taken into account in Fig. 1. determined. Note that the equivalent circuit remains accurate
The leakage inductance of a semi-closed-slot conductor tak- even when one chooses an arbitrary value for x1 . This is because
ing account of saturation in the slot bridge can be written as the circuit model in Fig. 1(a) can be equivalently transformed
follows [20]: by converting the rotor currents [22]. Second, input impedances

ZS are measured by locked-rotor tests. Third, rotor impedances
μ0 b − a cBs ZR are calculated from Z1 , Zm , and ZS . An additional process
ls = +a−c + (9)
S 3 IR is then carried out to smooth the change in ZR with respect
to f for each IR . This is done to suppress unevenness in the
where μ0 is the magnetic permeability of vacuum, S is the slot locked-rotor test. Finally, the set (r2 , r3 , x3 , x23 ) is calculated
width, a and b are the distances of the conductor top and bottom for each IR from (5)–(8). However, x23 is considered to vary
from the slot top, c is the depth of the semiclosed overhang of with IR when the tested IM has a closed-slot rotor. In this case,
the slot bridges, and Bs is the flux density of the slot bridge. A, B, and C are determined by the following method.
The authors use (9) to derive an approximate expression By substituting three sets of (x23 , IR ) in (11), we obtain
for x23 as follows. An approximate equation for Bs is written a set of simultaneous equations with three unknowns (A, B,

Authorized licensed use limited to: MALAVIYA NATIONAL INSTITUTE OF TECHNOLOGY. Downloaded on February 03,2021 at 14:07:03 UTC from IEEE Xplore. Restrictions apply.
 276 IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 55, NO. 1, JANUARY/FEBRUARY 2019

Fig. 4. Flowchart to determine the equivalent circuit parameters of Fig. 3 when the locked-rotor tests are conducted at three currents and four frequencies.

and C), from which we obtain (12)–(14) shown at the bot- C. Calculation of Current and Torque With Respect to Slip
tom of this page. Here, X1 , X2 , and X3 are the values of x23
Fig. 5 shows the procedure to calculate the driving perfor-
when the rotor currents IR are equal to IR 1 , IR 2 , and IR 3 , mance with respect to slip using the equivalent circuit shown
respectively. in Fig. 3. In Fig. 5, an initial value of x23 can be obtained by

IR 1 X1 (X2 − X3 ) + IR 2 X2 (X3 − X1 ) + IR 3 X3 (X1 − X2 )

A= (12)
IR 1 (X2 − X3 ) + IR 2 (X3 − X1 ) + IR 3 (X1 − X2 )
(IR 1 − IR 2 )(IR 2 − IR 3 )(IR 3 − IR 1 )(X1 − X2 )(X2 − X3 )(X3 − X1 )
B= (13)
−(IR 1 (X3 − X2 ) + IR 2 (X3 − X1 ) + IR 3 (X2 − X1 ))2
IR 1 IR 2 (X1 − X2 ) + IR 2 IR 3 (X2 − X3 ) + IR 3 IR 1 (X3 − X1 )
C= (14)
IR 1 (X2 − X3 ) + IR 2 (X3 − X1 ) + IR 3 (X1 − X2 )

Authorized licensed use limited to: MALAVIYA NATIONAL INSTITUTE OF TECHNOLOGY. Downloaded on February 03,2021 at 14:07:03 UTC from IEEE Xplore. Restrictions apply.
 YAMAMOTO et al.: SIMPLE METHOD TO DETERMINE DOUBLE-CAGE ROTOR EQUIVALENT CIRCUIT PARAMETERS OF INDUCTION MOTORS 277

Fig. 6. Photograph of the test motor. (a) Stator. (b) Semi-closed-slot double-
cage rotor. (c) Closed-slot double-cage rotor. (d) Semi-closed-slot single-cage
rotor. (e) Closed-slot single-cage rotor.

The overview of rotor-slot shapes of the tested single-cage

IMs is as follows. The maximum outer slot widths for the semi-
closed- and closed slots are both 4.6 mm. The distances of the
slot bottom from the slot top are both 14.1 mm. The thickness
of the closed-slot bridge is 0.3 mm. The slot opening of the
semi-closed-slot is 1.4 mm.
Fig. 7 shows the on-load test circuit used for measuring torque
and current for each slip. Here, the test motor and torque detector
are connected to an 11-kW dc generator (DCG) loaded with a
speed-controlled thyristor Leonard system. Using this circuit,
the steady-state torque and current can be measured for a given
slip s (0 < s < 1). The no-load and locked-rotor tests are carried
Fig. 5. Flowchart to calculate the torque and current with respect to slip. out using a three-phase 12-kVA linear power amplifier that can
generate a three-phase pure sinusoidal-wave voltage in several
frequencies.
assuming a certain value of IR . Next, a renewed rotor current
IR C is calculated. This process is repeated until IR C converges
B. Double-Cage IM With Semi-Closed-Slot Rotor
to IR . Then, the torque and current for a targeted slip can be
obtained. After installing the semi-closed-slot double-cage rotor to the
stator, the authors carried out a no-load test at the rated frequency
IV. EXPERIMENTAL VERIFICATION AND DISCUSSION and locked-rotor tests at four frequencies (50, 37.5, 25, and
12.5 Hz) and four currents (5.5, 11, 16.5, and 22 A). Note that
A. Test Motor and Experimental Setup
although only two frequencies and three currents are required
The tested machine is a 5.5 kW-200 V-4 P-22 A-50 Hz to determine the rotor circuit parameters, the locked-rotor tests
delta-connected IM. Fig. 6(a) shows a photograph of the sta- are conducted at four frequencies and four currents. This is due
tor. Fig. 6(b)–(e) shows photographs of the rotors. These rotors to confirming how rR and xR vary with frequency f and how
can be installed to the common stator shown in Fig. 6(a). x23 changes with rotor current IR .
The overview of the rotor-slot shapes of the tested double- Fig. 8(a) shows calculation results of rotor impedances ZR .
cage IMs is as follows. The maximum outer slot widths for the From the figure, it is confirmed that ZR changes with frequency
semi-closed- and closed slots are 3.5 and 4.2 mm, respectively. f , but does not depend on rotor current IR .
The distances of the slot bottom from the slot top are 23.3 and Next, the authors determined four sets of (r2 , r3 , x3 , x23 )
20.5 mm. The thickness of the closed-slot bridge is 0.3 mm. The with respect to IR using the proposed method shown in Fig. 4.
slot opening of the semi-closed-slot is 0.8 mm. In the present case, x1 was assigned a design value, calculated

Authorized licensed use limited to: MALAVIYA NATIONAL INSTITUTE OF TECHNOLOGY. Downloaded on February 03,2021 at 14:07:03 UTC from IEEE Xplore. Restrictions apply.
 278 IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 55, NO. 1, JANUARY/FEBRUARY 2019

Fig. 7. On-load test circuit for measuring torque and current for each slip.

Fig. 8. Test results on double-cage IM with semi-closed-slot rotor. (a) Change in rotor impedance Z R (= rR + jxR ) with rotor frequency f . (b) Change in
rotor circuit parameters x23 , r2 , r3 , and x3 with rotor current IR . (c) Comparison between measured and calculated performances (torque–slip and current–slip
curves) when the stator voltage is 0.5 p.u.

from the winding specifications and slot dimensions. Fig. 8(b) Fig. 9(a) shows the calculation results of the rotor impedance
shows the results, demonstrating that the rotor circuit parameters ZR . It can be seen that the resistance component rR does not
do not vary with IR . Thus, magnetic saturation of the slot bridge change with IR , but xR changes with both f and IR . This result
appears to have little effect. suggests that it is necessary to consider not only the frequency-
The authors then calculated the torque–slip and current–slip dependent variation of ZR but also the change in rotor leakage
curves, following the procedure shown in Fig. 5. To verify the reactance with respect to rotor current IR .
calculation results, the authors measured the torque–slip and Fig. 9(b) shows the changes in x23 , r2 , r3 , and x3 with respect
current–slip curves by an on-load test, when the stator voltage to IR . In calculating A, B, and C using (12)–(14), the authors
is 0.5 p.u. The main reason why a full voltage is not employed is employed IR 1 , IR 2 , and IR 3 , when the stator currents in the
temperature rise. It was difficult to operate the motor in steady locked-rotor test are 5.5, 11, and 22 A, respectively. In Fig. 9(b),
state under the stable temperature at higher slip frequency. The it is confirmed that only x23 varies with IR , and that (11),
other reason is due to properly verifying the accuracy of the rotor whose parameters A, B, and C are calculated from (12)–(14),
equivalent circuit parameters under the situation that variation expresses the variation of x23 accurately.
of x1 , caused by stator leakage flux saturation occurred when a Fig. 9(c) shows the starting performances when the voltage
full voltage is applied at higher slip frequency, is little. is 0.5 p.u. In the low slip range, one can see that the proposed
Fig. 8(c) shows the measured and calculated results, which method is greatly superior to the classical equivalent circuit
agree well, thus demonstrating the validity of the proposed method.
method. It is also confirmed that the calculated results obtained Meanwhile, errors between the measured and calculated
by the classical equivalent circuit (dashed curves) do not accu- torques are observed in the high slip range. Since the calcu-
rately reflect performance except in the vicinity of slip s = 1. lated torque in Fig. 8(c) corresponded to the measured one, it
This is because the frequency of the locked-rotor test for de- can be seen that leakage flux path saturation occurring on the
termining the parameters rR and xR is chosen to be the rated semi-closed stator slots of the test motor was little under 50%
frequency 50 Hz. voltage. However, the currents in high slip range in Figs. 8(c)
and 9(c) were both above two times of the maximum locked-
rotor test current. Thus, it is thought that this error in torque
C. Double-Cage IM With Closed-Slot Rotor
in Fig. 9(c) can be reduced by employing the result of a large-
The rotor was next replaced with the closed-slot double-cage current locked-rotor test. This is because the accuracy of the
rotor and the same test and calculation were carried out as the extrapolation of x23 improves in the higher current range. How-
double-cage IM with a semi-closed-slot rotor. ever, this requires verification in a follow-up study.

Authorized licensed use limited to: MALAVIYA NATIONAL INSTITUTE OF TECHNOLOGY. Downloaded on February 03,2021 at 14:07:03 UTC from IEEE Xplore. Restrictions apply.
 YAMAMOTO et al.: SIMPLE METHOD TO DETERMINE DOUBLE-CAGE ROTOR EQUIVALENT CIRCUIT PARAMETERS OF INDUCTION MOTORS 279

Fig. 9. Test results on double-cage IM with closed-slot rotor. (a) Change in rotor impedance Z R (= rR + jxR ) with rotor frequency f . (b) Change in rotor
circuit parameters x23 , r2 , r3 , and x3 with rotor current IR . (c) Comparison between measured and calculated performances (torque–slip and current–slip curves)
when the stator voltage is 0.5 p.u.

Fig. 10. Test results on single-cage IM with semi-closed-slot rotor. (a) Change in rotor impedance Z R (= rR + jxR ) with rotor frequency f . (b) Change in
rotor circuit parameters x23 , r2 , r3 , and x3 with rotor current IR . (c) Comparison between measured and calculated performances (torque–slip and current–slip
curves) when the stator voltage is 0.5 p.u.

The authors counted the number of iterations n required for found that these parameters can express the change in ZR with
convergence for each slip in calculating the performance. It was respect to slip frequency.
found that n was less than five regardless of the initial value of Fig. 11 shows the results on the single-cage IMs with the
rotor current IR for the entire slip range. This demonstrates that closed-slot rotor. In Fig. 11(a), it can be seen that xR varies with
a fast calculation can be achieved. IR as in Fig. 9(a), but the change in xR with respect to f is
smaller than Fig. 9(a). This accords with the physical image of
a closed-slot single-cage rotor. In Fig. 11(b), it is confirmed that
D. Single-Cage IMs With Semi-Closed- and only x23 varies with IR as in Fig. 9(b). Fig. 11(c) also suggests
Closed-Slot Rotors that the proposed method can calculate more accurate torque–
Similarly, the authors performed the test and calculation on slip and current–slip curves than the classical equivalent circuit
single-cage IMs with semi-closed- and closed-slot rotors. for the entire slip range.
Fig. 10 shows the results on the single-cage IMs with the The error in torque at higher slip frequency in Fig. 11(c),
semi-closed-slot rotor. Frequency-dependent variation of ZR of calculated by the proposed method, is much smaller than that
the single-cage IM is observed in Fig. 10(a). But this variation is in Fig. 9(c). In Fig. 11(c), it is thought that the correct extrapo-
smaller than that of the double-cage IM in Fig. 8(a). In Fig. 10(b), lation of x23 at large rotor current is unexpectedly done though
it is found that constant rotor parameters are obtained as in the locked-rotor test current is small. To surely improve the pre-
Fig. 8(b). Fig. 10(c) demonstrates that the proposed method can cision, it is thought that a large locked-rotor test current should
calculate more accurate torque–slip and current–slip curves than be employed.
the classical equivalent circuit for the entire slip range. Equivalent circuit parameters of the four tested IMs are sum-
In case of single-cage IMs, r2 , r3 , x3 , and x23 do not phys- marized in Tables I and II. In these tables, there is no x23
ically correspond to the rotor-slot shape. However, it can be for the closed-slot machines since x23 varies with IR and is

Authorized licensed use limited to: MALAVIYA NATIONAL INSTITUTE OF TECHNOLOGY. Downloaded on February 03,2021 at 14:07:03 UTC from IEEE Xplore. Restrictions apply.
 280 IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 55, NO. 1, JANUARY/FEBRUARY 2019

Fig. 11. Test results on single-cage IM with closed-slot rotor. (a) Change in rotor impedance Z R (= rR + jxR ) with rotor frequency f . (b) Change in rotor
circuit parameters x23 , r2 , r3 , and x3 with rotor current IR . (c) Comparison between measured and calculated performances (torque–slip and current–slip curves)
when the stator voltage is 0.5 p.u.

TABLE I
EQUIVALENT CIRCUIT PARAMETERS OF THE DOUBLE-CAGE IMS OBTAINED BY NO-LOAD AND LOCKED-ROTOR TESTS

TABLE II
EQUIVALENT CIRCUIT PARAMETERS OF THE SINGLE-CAGE IMS OBTAINED BY NO-LOAD AND LOCKED-ROTOR TESTS

calculated using A, B, and C. On the contrary, there are no

A, B, and C for the semi-closed-slot machines since x23 is
constant. In Table I, r2 is larger than r3 . This result physically
agrees with double-cage construction. In Table II, r2 is also
larger than r3 . This is because rR increases and xR decreases in
both double- and single-cages IMs in spite of difference between
the rotor shapes. In case of the semi-closed-slot IM, the com-
bined rotor impedance in the proposed circuit model at 50 Hz
corresponds to rR + jxR in the classical circuit model. In case
of the closed-slot IM, however, it does not become like this
because x23 varies.
Fig. 12. Change in rotor circuit parameters x23 , r2 , r3 , and x3 with rotor
current IR when the smoothing process of Fig. 5 is not carried out. (a) Double-
E. Effect of the Process for Smoothing the Rotor Impedance cage IM with closed-slot rotor. (b) Single-cage IM with closed-slot rotor.
In Fig. 5, rR and xR were recalculated after performing lin-
ear approximation of rR − f and xR − f characteristics. The uniform than those of Fig. 12(a) and (b), respectively. In par-
authors verified the effectiveness of this smoothing process. ticular, the parameters in Fig. 12(b) vary widely, in contrast to
Fig. 12 shows calculation results of the rotor circuit parame- those in Fig. 11(b). Thus, it is found that the smoothing process
ters x23 , r2 , r3 , and x3 when this process is not carried out on reduces the influence of the unevenness of rR and xR , which are
the double- and single-cage IMs with closed-slot rotors. One highly sensitive to rotor circuit parameters, and thus contributes
can see that the parameters of Figs. 9(b) and 11(b) are more to improving the precision of the rotor circuit parameters.

Authorized licensed use limited to: MALAVIYA NATIONAL INSTITUTE OF TECHNOLOGY. Downloaded on February 03,2021 at 14:07:03 UTC from IEEE Xplore. Restrictions apply.
 YAMAMOTO et al.: SIMPLE METHOD TO DETERMINE DOUBLE-CAGE ROTOR EQUIVALENT CIRCUIT PARAMETERS OF INDUCTION MOTORS 281

impedance is frequency dependent and magnetic saturation of

the closed-slot bridge affects rotor leakage reactance. The results
are summarized as follows.
1) A new formula to calculate the double-cage rotor equiva-
lent circuit parameters was proposed.
2) The formula in 1) was very simple so that it can be calcu-
lated by a calculator but correctly determines the double-
cage rotor equivalent circuit parameters from only no-load
and locked-rotor tests.
Fig. 13. Terminal connections in standstill test using an adjustable single-
phase ac voltage source. (a) ab-c connection. (b) a-b connection. 3) An improved equivalent circuit model taking account of
magnetic saturation of the closed-slot rotor bridge was
shown.
F. Standstill Test Using Adjustable Single-Phase AC 4) Parameters of the improved equivalent circuit model were
Voltage Sources also determined by no-load and locked-rotor tests.
Using an adjustable single-phase ac voltage source (including 5) The proposed method was implemented on four 5.5-kW
an inverter by applying pulsewidth modulation (PWM) on a IMs: double- and single-cage IMs with semi-closed- and
single voltage vector) is useful because the torque is zero and closed-slot rotors. Comparison between measured and
the test is easy to make. In general, there are two kinds of test calculated results (torque–slip and current–slip curves)
circuits shown in Fig. 13. demonstrated the validity of the proposed method.
In [23], these standstill tests using a pure sine wave have
been implemented and compared with the locked-rotor test.
The following equation can be used to convert the standstill test
REFERENCES
currents (Iab−c and Ia−b in Fig. 13) into the stator current IS
in the IM equivalent circuit by calculation. These equations are [1] G. Kron, Equivalent Circuit of Electrical Machinery. Hoboken, NJ, USA:
Wiley, 1951.
derived so that the magnetomotive force of the standstill test [2] P. Jacobs, “Defining the equivalent circuit of the double-cage motor,”
corresponds to that of the locked-rotor test Trans. Amer. Inst. Elect. Eng. III, Power App. Syst., vol. 72, no. 2, pp. 651–
657, Aug. 1953.
IS = Iab−c (15) [3] IEEE Standard Test Procedure for Polyphase Induction Motors and Gen-
erators, IEEE Standard 112, 2004.
2 [4] Rotating Electrical Machines - Part 2-1: Standard Methods for Determin-
IS = √ Ia−b . (16) ing Losses and Efficiency From Tests, IEC 60034-2-1, 2014.
3 [5] K. Kato, T. Yano, and W. Hoshi, “Performance calculation on double
Using (15) or (16), the input impedance for each stator current squirrel-cage induction motors,” (in Japanese), J. Inst. Elect. Eng. Jpn.,
vol. 73, no. 780, pp. 990–996, 1953.
in the IM equivalent circuit can be obtained from the standstill [6] B. J. Chalmers and A. S. Mulki, “Design synthesis of double-cage in-
test [23]. duction motors,” Proc. Inst. Elect. Eng., vol. 117, no. 7, pp. 1257–1263,
When the PWM inverter is used, to measure fundamental Jul. 1970.
[7] F. Corcoles, J. Pedra, M. Salichs, and L. Sainz, “Analysis of the induction
components of voltage and current by fast Fourier transforma- machine parameter identification,” IEEE Trans. Energy Convers., vol. 17,
tion would be recommended. no. 2, pp. 183–190, Jun. 2002.
[8] R. Wamkeue, I. Kamwa, and M. Chacha, “Unbalanced transients-based
maximum likelihood identification of induction machine parameters,”
G. Application to Deep Bar Rotor IMs IEEE Trans. Energy Convers., vol. 18, no. 1, pp. 33–40, Mar. 2003.
[9] J. Pedra and L. Sainz, “Parameter estimation of squirrel-cage induction
In the case of the deep bar rotor IMs, the rotor resistance motors without torque measurements,” Inst. Elect. Eng. Proc.—Elect.
increases at high slip frequency, accompanied by slot leakage Power Appl., vol. 153, no. 2, pp. 263–270, Mar. 2006.
reactance decrease, which leads to, both, a lower starting current [10] M. Gomez-Gonzalez, F. Jurado, and I. Perez, “Shuffled frog-leaping algo-
rithm for parameter estimation of a double-cage asynchronous machine,”
and a higher starting torque due to skin effects of the deep bar IET Elect. Power Appl., vol. 6, no. 8, pp. 484–490, Sep. 2012.
[24]. Such rotor circuit impedance variation with respect to the [11] J. R. Willis, G. J. Brock, and J. S. Edmonds, “Derivation of induced motor
slip frequency is similar to double-cage rotor IMs. In addition, models from standstill frequency response tests,” IEEE Trans. Energy
Convers., vol. EC-4, no. 4, pp. 608–615, Dec. 1989.
regardless of difference in single-cage, double-cage, and deep [12] S. A. Soliman and G. S. Christensen, “Modeling of induction motors from
bar rotors, it is thought that x3 is constant but x23 varies with standstill frequency response tests and a parameter estimation algorithm,”
rotor current IR . In fact, x3 is constant but x23 varies with IR in Elect. Mach. Power Syst., vol. 20, no. 2, pp. 123–136, 1992.
[13] R. Belmans, R. D. Findlay, and W. Geysen, “A circuit approach to finite
the tested single-cage rotor having small skin effect, as shown element analysis of a double squirrel cage induction motor,” IEEE Trans.
in Fig. 11(a). Thus, it is thought that the proposed method can Energy Convers., vol. 5, no. 4, pp. 719–724, Dec. 1990.
be applied to the deep bar rotor IMs. However, this requires [14] S. Williamson and D. R. Gersh, “Finite element calculation of double-
cage rotor equivalent circuit parameters,” IEEE Trans. Energy Convers.,
verification in a follow-up study. vol. 11, no. 1, pp. 41–48, Mar. 1996.
[15] A. Arkkio and T. P. Holopainen, “Space-vector models for torsional vi-
bration of cage induction motors,” IEEE Trans. Ind. Appl., vol. 52, no. 4,
V. CONCLUSION pp. 2988–2995, Jul./Aug. 2016.
A simple method to determine the double-cage rotor equiv- [16] H. K. Jafari, L. Monjo, F. Corcoles, and J. Pedra, “Using the instanta-
neous power of a free acceleration test for squirrel-cage motor parameters
alent circuit parameters was discussed. The method is also ap- estimation,” IEEE Trans. Energy Convers., vol. 30, no. 3, pp. 974–982,
plicable to single-cage IMs in which the rotor equivalent circuit Sep. 2015.

Authorized licensed use limited to: MALAVIYA NATIONAL INSTITUTE OF TECHNOLOGY. Downloaded on February 03,2021 at 14:07:03 UTC from IEEE Xplore. Restrictions apply.
 282 IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 55, NO. 1, JANUARY/FEBRUARY 2019

[17] A. M. A. Mahmoud and R. W. Menzies, “A complete time domain model Hideaki Hirahara (M’13) received the B.E. degree
of the induction motor for efficiency evaluation,” IEEE Trans. Energy from the Polytechnic University, Tokyo, Japan, in
Convers., vol. EC-1, no. 1, pp. 68–76, Mar. 1986. 2000, the M.E. degree from the National Institution
[18] S. Yamamoto, H. Hirahara, A. Tanaka, and T. Ara, “A simple method for for Academic Degrees and University Evaluation,
determining equivalent circuit parameters of double-cage induction mo- Tokyo, in 2002, and the Ph.D. degree from Tokyo
tors from no-load and locked-rotor tests,” in Proc. IEEE Energy Convers. Metropolitan University, Tokyo, in 2016.
Congr. Expo., 2017, pp. 301–306. In 2002, he joined the Advanced Polytechnic Cen-
[19] S. Yamamoto, T. Ara, S. Oda, and K. Matsuse, “Prediction of starting ter, Chiba, Japan, as an Instructor. In 2004, he moved
performance of PM motors by DC decay testing method,” IEEE Trans. to the Polytechnic College Hamamatsu, Hamamatsu,
Ind. Appl., vol. 36, no. 4, pp. 1053–1060, Jul./Aug. 2000. Japan, as a Lecturer. In 2010, he joined the Polytech-
[20] J. R. Smith, Response Analysis of A.C. Electrical Machines: Computer nic Center Ehime, Japan, as an Instructor. In 2012, he
Models and Simulation. Taunton, U.K.: Res. Stud. Press, 1990, pp. 49–52. moved to the Polytechnic University, where he is currently an Assistant Profes-
[21] S. Yamamoto and T. Ara, “Prediction of characteristics of squirrel-cage sor. His research interests include electric machines and motor drives.
induction motor—Consideration of non-linearity of rotor winding leak- Dr. Hirahara is a member of the Institute of Electrical Engineers of Japan.
age reactance,” (in Japanese), in Proc. Nat. Conv. Rec. IEE Jpn., 1998,
pp. 51–52.
[22] T. Kano, H. Nakayama, T. Ara, and T. Matsumura, “A calculation method Akira Tanaka received the B.E. degree in applied
of equivalent circuit constants with mutual leakage reactance on syn- chemistry and M.E. and Ph.D. degrees in electrical
chronous machine with damper winding,” Elect. Eng. Jpn., vol. 167, no. 2, engineering from Toyo University, Tokyo, Japan, in
pp. 71–78, Apr. 2009. 1983, 1985, and 2017, respectively.
[23] S. Yamamoto and T. Ara, “Universal parameter measurement and speed From 1985 to 1988, he was with the Hitachi Tobu
sensorless vector control of induction and permanent magnet synchronous Semiconductor Co., Ltd., Saitama, Japan. In 1988, he
motors—Application to a closed-slot squirrel-cage induction motor,” (in joined the Faculty of Tokyo Vocational Training Col-
Japanese), in Proc. Nat. Conv. Rec. IEE Jpn., 2010, pp. 160–161. lege, Tokyo, Japan, as a Lecturer. In 1999, he moved
[24] I. Boldea and S. A. Nasar, The Induction Machines Design Handbook, to the Kyushu Polytechnic College, Fukuoka, Japan.
2nd ed. Boca Raton, FL, USA: CRC Press, 2010, pp. 221–222. Since 2012, he has been the Faculty at the Polytech-
nic University, Tokyo, Japan, where he is currently a
Professor. His current research focuses on electric equipment and maintenance.
Dr. Tanaka is a member of the Institute of Electrical Engineers of Japan
(IEEJ). He was the recipient of the IEEJ Industry Applications Society Distin-
guished Transaction Paper Award in 2016.

Takahiro Ara received the B.E. degree from the

Polytechnic University, Tokyo, Japan, in 1977, and
the Dr. degree from Meiji University, Tokyo, in 1991,
Shu Yamamoto (M’95) received the B.E., M.E., and both in electrical engineering.
Ph.D. degrees in electrical engineering from Meiji In 1977, he joined the Faculty of the Hamamatsu
University, Tokyo, Japan, in 1992, 1994, and 1997, Vocational Training Center, Hamamatsu, Japan, as an
respectively. Instructor. In 1980, he moved to the Polytechnic Uni-
In 1995, he joined, as an Assistant, the Faculty versity as an Assistant. From 2000 to 2017, he was a
at Polytechnic University, Tokyo, Japan, where he is Professor with the Polytechnic University, where he
currently a Professor. His research interests include is currently a Professor Emeritus. His research inter-
electric machines and motor drives. ests include electric machines and motor drives.
Dr. Yamamoto is a member of the Institute of Elec- Dr. Ara is a member of the Institute of Electrical Engineers of Japan (IEEJ).
trical Engineers of Japan (IEEJ). He was the recipient He was the recipient of the IEEE IAS Electric Machines Committee Prize Paper
of the IEEE IAS Electric Machines Committee Prize Award in 1997, and was also the recipient of the IEEJ IAS Outstanding Con-
Paper Award in 1997 and the IEEJ IAS Distinguished Transaction Paper Award tribution Award, the IEEJ IAS Distinguished Transaction Paper Award, and the
in 2016. IEEJ IAS Technical Achievement Award in 2012, 2016, and 2017, respectively.

Authorized licensed use limited to: MALAVIYA NATIONAL INSTITUTE OF TECHNOLOGY. Downloaded on February 03,2021 at 14:07:03 UTC from IEEE Xplore. Restrictions apply.