Medium–Large induction machines

In the locked-rotor measurement, the current is measured at different supply voltage levels, and, based on that, the rated starting current is scaled to the rated voltage.. According to

Scaling procedures for starting current measurements

B Y W A Q A S M A R S H A D ,

S A M I K A N E R V A ,

S I L V I A B O N O ,

M A S S I M O M E N E S C A R D I ,

& H O L G E R P E R S S O N

I N THIS ARTICLE, THE ACCURACYlevels of different scaling procedures for

starting current measurements are addressed through the statistical analysis of locked-rotor tests involving hundreds of medium–large induction

machines This is performed to ascertain the confidence level

of the predicted full-voltage starting currents from the reduced-voltage factory measurements, a necessity for motors with a strict tolerance to the starting current level It is shown that, for traditional scaling methods employing only measurements, this accuracy level is 88–90% for scaling step from 0.6 to 1.0 p.u voltage This figure could be raised to

Digital Object Identifier 10.1109/MIAS.2010.938385

© CREATAS

28

95–97% through a novel method that

also uses finite element method (FEM)

simulations for each individual test

volt-age The differences between FEM and

voltage measurements are documented

and used for correcting the full-voltage

FEM simulation Further, the use of

FEM is shown to help in the

understand-ing of the causes and locations of

measured voltage-dependent saturation

uncertainties The region (end-core,

over-hang, or main core) that contributes

the most to saturation uncertainties

is shown to be identifiable through

origin/parametric dependencies, leading

to a better product understanding and

more reliable scaling methods in future

Starting Currents in

Induction Machines

Starting current levels for a direct-online,

single-cage induction motor are typically specified to be four

to seven times the rated motor currents (4–7 p.u.) [1] These

figures are for medium–large induction motors, whereas these

values can be as high as 10–12 p.u for smaller motors

Start-ing currents of the delivered motors must usually follow one

of the following:

n the International Electrotechnical Commission (IEC)

60034-1 [2], allowing a 20% tolerance margin

n the National Electrical Manufacturers Association

(NEMA) MG1 standard [3], specifying tolerances

according to kilovoltampere per horsepower

n the customer specifications, defining individual

tolerance or absolute figures, e.g., in chemical oil

and gas sector:

nthe Shell Design and Engineering Practice (DEP)

33.66.05.31 [4] or American Petroleum Institute

(API) 541 standard [5] that requires a starting

current without tolerances, e.g., below 6.5 p.u

for medium-voltage motors

The Trend for Low-Inrush Current Motors

Recently, the demand for low-inrush current motors has been

slowly increasing because the starting current is typically less

than 5 p.u., and no tolerance margin is accepted between the

specification and delivered motor (referred below as

zero-tolerance designs) Marine, offshore petrochemical, and, in

general, all industry connected to a weak grid are

increas-ingly demanding such direct online solutions Costs, weight/

space requirements, and complexity are some of the factors

that rule out the use of alternative solutions, such as

soft-starters, transformer-starting, and converter-start [6] The

zero tolerances impose stringent demands upon exact

knowl-edge of all steps of the motor manufacturing process

Factors Affecting Starting Currents

The inductance of a machine and its variation during

startup roughly define the starting current for a given

applied voltage The inductance is generally divided into

three distinct regions: subtransient, transient, and steady

state The inductances are influenced by a number of motor

geometrical details as well as materials The starting currents

during the design stage can thus be influenced [6]–[12] by the following geometrical and physical parameters:

n Stator design: number of turns, end-windings layout, core length, and slot geometry

n Rotor design: slot dimensions, bar profile, and material

n Frame design: end-ring shield-ing, etc

The trade-offs between other per-formance parameters such as starting and breakdown torque, rated efficiency, and power factor also need to be con-sidered when limiting the inrush cur-rent during the design stage The mechanical and thermal limits of the structure also need to be taken into account [10]–[14]

Design Parameters Accuracy

Since all the manufactured motors need to be tested before delivery, in addition to the design effort, the emphasis is on (especially for zero-tolerance designs):

n design parameters accuracy

n measurements accuracy

n scaling accuracy

The accuracy level of the design parameters primarily depends upon the production site history that has been achieved through years of experience of machine modeling (reluctance networks, equivalent circuits, and FEM analysis), statistically calibrated against measurements on prototypes and delivered products Skin effect, iron saturation, and inductances modeling are some of the key parameters that define the true starting performance of an induction motor

Locked-Rotor and Run-Up Measurements

The starting current measurement is carried out by either locked-rotor or run-up methods The accuracy of the measure-ment is influenced by the test engineers’ experience and care, rotor position, and temperatures in the stator and rotor

In the locked-rotor measurement, the current is measured

at different supply voltage levels, and, based on that, the rated starting current is scaled to the rated voltage IEC 60034-1 [2] defines the locked-rotor current as the greatest steady-state, root-mean-square current taken from the supply with the motor held at rest, over all angular positions of its rotor,

at rated voltage and frequency According to IEEE 112-2004 [15], for this test, when possible, the measurements shall be taken at rated voltage and frequency since the current is not directly proportional to the voltage because of changes in reactance caused by saturation of the leakage paths

In the run-up measurement, the motor is started by reduced voltage, and the measured starting current is scaled to the rated voltage Usually, the motor is rotating slowly in a backward direction before startup (reverse-rota-tion start) to obtain the current at zero speed by the time the supplied electromagnetic transients have decayed

Scaling

When the full-voltage test is not possible on the factory test floors (e.g., for the largest induction machines) or when

THE DIFFERENCES BETWEEN FEM AND VOLTAGE MEASUREMENTS

ARE DOCUMENTED AND USED FOR CORRECTING THE FULL-VOLTAGE FEM SIMULATION.

29

the maximum allowable stall-time limits full-voltage tests,

the reduced-voltage test is applied In this case, the

start-ing current is extrapolated from the reduced-voltage

measurements, a method commonly known as voltage

scal-ing In such cases, not only the measurements themselves

but also the accuracy of scaling methods (full-voltage

cur-rents prediction from reduced-voltage measurements)

attains vital importance For run-up tests, the same scaling

methods may be used as in the locked-rotor test

Accuracy of Scaled Measurement Results

As will be discussed later, the traditional scaling methods,

including those presented in [15] and [16], have a degree

of uncertainty Sometimes, factory acceptance tests (FATs)

provide scaled starting current levels that are not a true

representation of the motors’ actual starting behavior

Documented experience from many years reveals that even

though there may be occasional complaints for scaled results

predicted by FATs, seldom have such complaints been raised

during commissioning or the normal operation of the

motors This indicates that scaling would only cause the dif-ferences between measured results and design values Figure 1 shows the authors’ perception regarding uncertainties asso-ciated with the starting currents A careful benchmarking procedure by the authors on a few hundred machines has singled out scaling as the biggest factor associated with the full-voltage starting currents accurate prediction

Outline and Scope of the Study

This section presents the details of the work conducted by the authors regarding scaling The study covers a vast range of direct-online induction motors (taken from the list

in Table 1) Five to eight locked-rotor measurement results for each individual machine, all at different voltage levels, are used The following are the topics addressed:

n accuracy evaluation of the traditional scaling

n presentation of the novel combined FEM-measure-ments scaling method

n confidence-level evaluation of the aforementioned method based on statistical analysis

n insight into starting currents saturation physics by studying parametric dependencies

Issues such as repeatability, reliability, and test-conditions influence (e.g., ambient temperature and hot or cold motor) are addressed This is done through new measurements on 50 motors and the utilization of earlier documented measurement results of a few hundreds of randomly selected motors This vast amount of measured data allows the separation of measurement uncertainties from those of scaling procedures In this study it

is revealed that no significant differences exist between the measured values of locked current and run-up starting currents Hence, the analysis in the following sections is restricted to locked-rotor tests only Since the present study deals with vol-tages at machine terminals, the role the network impedance plays is not significant and is omitted from the study

The FEM approach is introduced as an alternative to traditional scaling methods that employ only measure-ments Thus, two new scaling methods that, in addition to measurements, employ FEM simulations are studied Since the FEM models use the true hysteresis (BH) iron curve for

Design (1–2)%

Measurements (1–3%)

Slip

0

Starting Currents Uncertainty

Scaling (1–15%)

1

Different accuracy bands for the full-voltage starting

currents The arrows show distinct regions corresponding to

the design and measurements uncertainty.

TABLE 1 MACHINE DATA.

IC516

Data might differ according to machine type.

30

modeling saturation, at least theoretically, it was believed

that the FEM should be able to provide a better picture of the

saturation at the full-voltage conditions in the machine than

through one of the traditional empirical extrapolation

meth-ods that are based only on measurements Unfortunately,

experience showed that even FEM simulations required some

sort of calibration/tuning to achieve the acceptable

confi-dence level in absolute terms Two such calibration

approaches are presented and discussed

Traditional Scaling Methods

Description of the Methods

Traditional scaling methods rely only on measurements

These include the methods presented in literature [17] as

well as those recommended by standards such as IEEE

112-2004 [15] and IEEE 115-1995 [16] The locked-rotor tests

are made for maximum available test room voltages The

onset of saturation and its dependency upon the test voltage

is captured through various curve-fitting techniques of the

three to six available measurement results The full-voltage

current is obtained through an extrapolation procedure The

accuracy of the result is based on the employed curve-fitting

technique, the maximum test voltage, and the number of

tests Usually, reference to any physical dependencies and

design factors is not explicitly made Some of these methods

that have been studied by the authors are listed below:

1) The current is calculated as varying directly with

voltage (1), with the exponent k¼ 1 (IEEE 112-2004

[15]) and then the full-voltage current is corrected

based on a certain rule of thumb (experience-based)

correction factor

Inom¼ Imeasured

Unom

Umeasured

 k

, (1)

where Inom is the nominal current, Unom is the

nominal (rated) voltage, Imeasured is the measured

current, and Umeasured is the measured voltage

2) A more exact method (log–log) requires two

measure-ment points, and the exponent k in (1) is calculated

according to (2), IEEE 112-2004 [15]

k¼ log(Imeas1=Imeas2)= log(Emeas1=Emeas2), (2)

where E is the electromotive force that is

some-times approximated with the supply voltage U

3) IEEE 112-2004 [15] also states that, for better

accu-racy, a least-squares fit should be used when

deter-mining the scaling exponent k (2) This requires

multiple reduced-voltage test points

4) IEEE 115-1995 [16] instead suggests a semilog fit

Inom(Unom)¼ ef2(Unom), (3) where e is the base of natural logarithms (log) and

f2(Unom)¼ Unom Umeas1

Umeas2 Umeas1

3log Imeas2

Imeas1

 

þ log(Imeas1): (4)

5) A saturation factor approach is based on estimating the saturation factor (ksat) from tests and, conse-quently, predicting the full-voltage current from

Inom¼ Imeasksat(Unom=Umeas), (5)

ksat ¼ (Imeas2=Umeas2)=(Imeas1=Umeas1): (6)

6) An improved saturation factor approach uses a voltage dependence derived from the simple mag-netic circuit approach

NI¼ (Rairgapþ Riron)U, (7)

U / U; Riron/ 1=liron; liron/ 1=Uk: (8)

7) A second-degree polynomial fit of measured data is also considered

A complete onset of saturation is only possible at full voltages, which can be rather difficult to achieve in the tests rooms for many machines The accuracy of the tradi-tional scaling methods, i.e., those based only on measure-ments, is dependent upon the maximum test voltage that

is possible during the measurements (test room supply and motor stall time limitation), as well as on the num-ber of measurements (test room cost and test window limitation) In summary, the extrapolation accuracy of the employed curve-fitting technique is benchmarked in this study

Accuracy Evaluation Through Statistical Analysis

The accuracy of the different scaling methods is evaluated

by comparing the differences between the scaled results obtained from n 1 reduced-voltage measurements and the n:th measurement

1) The studied scaling method is applied on the

n 1 measurements (lowest test voltages), estimat-ing the startestimat-ing current value on the n:th measure-ment point (highest voltage)

2) The estimated value in n:th measurement point is compared with the actual measured value

This is followed by a detailed statistical analysis of these differences for certain sets of studied motors The analysis serves as the basis for defining the confidence level (error margin in terms of mean difference and standard deviation)

of the various studied scaling methods Classification of the motors based on different criteria such as power range, voltage level, and speed is also carried out

Comparative Analysis

The following methods are primarily investigated for benchmarking and also for improving their accuracy with the better curve-fitting techniques

Log–Log, Semilog, and Linear Scaling

The linear fit [the old practice, see 1) mentioned under the

“Description of the Methods” section], the semilog fit [see 4) above] and the log–log fit [see 2) and 3) above] are com-pared first It is found that, of these, the last is the most promising for predicting the full-voltage currents from the reduced-voltage tests Some of these results for synchro-nous motors have been provided in [17] 31

The log–log curve fitted

extrap-olation provides the lowest error

to the actual current at

maxi-mum tests voltage This error is

roughly two to ten times better

than the other two approaches

In this comparison, the highest

test-room voltage is treated as

the rated voltage

n A difference of up to 30% in

scaled values is observed when

comparing the log–log approach

[see 3) above] against the linear

approach [see 1) above]

n For the log–log approach, a

spread in scaled results of up to

20% is observed when k (2) is

derived from only two test

points [see 2) above]

correspond-ing to different test voltages

n There is also spread in log–log

estimations when successive highest voltages are

removed from the curve-fitting algorithm The

spread among scaled results is as high as around 5%

in certain cases when at least three test points are

used for the estimation Results with only two test

points have a much greater spread

Log–Log, Saturation Factor, and Second-Degree Fit Scaling

The best of the earlier discussed scaling methods, the log–log

method [see 3) discussed earlier] is consequently compared for

accuracy against a suggested second-degree polynomial fit

[see 7) above] as well as the traditional saturation factor

tech-niques [see 5) and 6) stated earlier] For the studied cases, it is

found that:

n All methods are equally good if the maximum test

voltage is above 0.95 p.u

n For scaled results using 0.6 p.u maximum test

voltage, the differences to actual measured values

are roughly of the same order for all the three

methods The saturation factor method exhibits

the lowest differences band, i.e., 5.5% The

sec-ond-degree fit has a difference band of 6%,

which is slightly below that of the log–log

method of 6.5%

n The differences for the second-degree fit are

ran-dom, whereas for the other two methods, they are

always in the same direction Thus, the log–log fit

is always predicting lower than the measured

values, whereas the saturation factor method always predicts higher than the measured values

Accuracy Levels

To provide generalized accuracy figures for the investigated scaling methods, a more detailed study was carried out for the log–log, the saturation factor, and the second-degree fit scaling methods Results are provided in Table 2 for a scaling step of 0.4 p.u., i.e., when the maximum test voltage used for scaling

of the results is 0.4 p.u below the maximum available test voltage The scaled results in Table 2 are compared against the maximum available test voltage, which may be lower than 1 p.u For example, when the maximum available test voltage is 0.6 p.u., the estimation at this voltage is made from tests that use a maximum test voltage of 0.2 p.u

The results in Table 2 show that this set of scaling meth-ods have an uncertainty margin of roughly 10% The simpler methods always provide an underestimation, whereas the more advanced ones can provide an overestimation as well

It is found that the error margin of the saturation factor approach (2) decreases more than that of the other ap-proaches, as the scaling voltage step is decreased Still, all these curve-fitting methods depend primarily on the quality

of information embedded in the test data (saturation), the full-voltage scaled results becoming poorer with a decreasing highest test voltage Hence, it can be concluded that the scal-ing methods that are based only on measurements are unac-ceptable for zero-tolerance inrush current design motors FEM Approach 1: Tuning of FEM Models

In the first variant, a two-dimensional (2-D) FEM simulation

is tuned against a reduced-voltage measurement with an expectation that the full-voltage simulation could provide correct results A common understanding is that the uncer-tainties/errors in 2-D FEM simulations are due to the wrong input data fed, such as the end-windings inductances and material properties, to 2-D-FEM models It was believed that such differences could be neutralized by a single calibration against the measurements, achieved through tuning of the input data, such as the end-windings inductances

One such example is provided in Figure 2 Here, the FEM model of a certain machine is tuned against a

reduced-voltage locked-rotor mea-surement through variation of end-winding inductances and material properties However, this approach did not yield the correct full-voltage starting current For a scaling step of 0.45 p.u., an underestimation

by 7% can be observed This error is believed to be due to the influence of three-dimensional (3-D) effects upon saturation (frame region and end-winding

TABLE 2 SCALING THROUGH MEASUREMENTS:

ACCURACY LEVELS FOR A SCALING STEP OF AT LEAST 0.4 PER UNIT.

Scaling Method

Error to

THE FEM APPROACH IS INTRODUCED AS

AN ALTERNATIVE

TO THE TRADITIONAL SCALING METHODS THAT EMPLOY ONLY MEASUREMENTS.

32

inductances), which are missed in a reduced-voltage 2-D

tuning endeavor

By considering the starting current (saturation) levels

for other machines, an error figure of 10% was

estab-lished for this method The error margin in this case is

still comparable with the traditional scaling methods

Further, the amount of effort required in the iterative

tuning procedure for each machine in this method makes

it unsuitable for daily factory scaling usage and is not

dis-cussed further

FEM Approach 2: Trends-Based Scaling

In an alternative investigated FEM, for a machine under

consideration, the FEM simulations are carried out

corre-sponding to each of the conducted tests It is found that,

for any given machine, the differences between

measure-ments and FEM simulations have a defined slope when

plotted against the test voltages even when no tuning is

performed Thus, it is possible to predict the full-voltage

starting currents using these differences predicting trends

The full-voltage FEM simulations are run and the

corre-sponding corrections made

Figure 3 shows the differences of FEM simulations to

measurements against test voltage for a certain set of

studied motors Results from both harmonic and

time-stepping FEM simulations are provided It can be seen that

the difference between the FEM simulations and

measure-ments in most cases is not constant but can still be

repre-sented through a straight line having a certain slope This

observation points to the possibility of using the

differen-ces between the FEM simulations and measurements at

reduced voltages to yield linear correction factors (slopes),

which in turn can be used to correct the full-voltage FEM

simulations to provide fairly accurate scaled results

The Analysis Method

The analysis method is investigated for accuracy on 205

different machines Table 3 provides a listing of the

machine types and frame sizes that have been used in the

analysis (full details are provided in Table 1) In the

presented analysis:

The FEM simulations are compared against three reduced-voltage measurements The maximum of these reduced test voltages is around 0.6 p.u

n The slope for the differences between the FEM simulations and measurements is obtained

n The full-voltage FEM simulation results are adjusted based on the slope of the differences and the scaling (voltage) step

n The scaled results are compared against the actual measurements at the highest test voltages

n The results are statistically analyzed according to the mean error and its standard deviation

Accuracy

Results for the differences are summarized in Figure 4 It can be seen that the mean difference is 1.4% and the

Starting Current: FEM to Measured (%)

–20 –15 –10 –5 0 5

0 0.2 0.4 0.6 0.8 1 1.2

Test Voltage (p.u.)

O, Solid Line: Harmonic [], Dashed Line: Time Stepping

Motor 4 Motor 3Motor 2 Motor 1

3

Motor 5 and 6

Starting currents: differences between the FEM and measurements.

TABLE 3 DESCRIPTION OF MACHINES USED IN THE TRENDS-BASED FEM STUDY.

Voltage

Current

0

1

2

3

4

5

6

7

Voltage (kV)

FEM Tuned Measured FEM Original

2

Tuning of the FEM results against the reduced-voltage

measurements.

0.96 0.97 0.98 0.99 1 1.01 1.02 1.03 1.04 1.05 1.06

Starting Current: Difference (p.u.) 0

20 40 60

Measured to Trends-Based FEM Predicted Starting Current Voltage Scaling Step for FEM: 0.4 p u (Mean: 1.01396, Standard Deviation: 0.0123)

4

Statistical summary of differences in starting currents between the highest-voltage measurements and

standard deviation is only 1.23% Hence, a 95%

confi-dence level is within the 5% error margin Figure 5 shows

these differences against the maximum available test

volt-age For a test voltage higher than 0.5 p.u., the differences

are within 3% Further, some of the cases could be clearly

identified to be outliers due to special test conditions, FEM

input data error, measurements errors, etc Hence, for all

practical cases, the 95% confidence level for the discussed

method is within a 3% error margin This error in scaling

is realistic even for zero-tolerance designs with a certain

built-in acceptable design margin

In the discussion earlier, the trends (slopes) used to

cor-rect the FEM simulations used only three test results To

benchmark an improvement in accuracy possible with

further tests, the aforementioned results are compared

against a case in which the trends (slopes) are determined

with five test points (Figure 6) From Figure 6, it can be seen

that the results for the three test-point cases are

underesti-mated, and a significant deviation exists between the

three-test-point and five-three-test-point cases Hence, it can be seen

that even this method is sensitive to the number of test

points, and a recommendation is to use as many test results

as possible From Figure 6, it is found that most results

belonging to the cases that are left and right of the center

column had either test points that were too close or some of

the results for the lowest test voltages were questionable

Hence, other recommendations are that the test should be as

separated as possible in terms of test voltages, and extreme

care is needed when making the lowest voltage tests

Comparison to the Traditional Methods

One of the aims of this work had been to determine the full-voltage starting current values when actual tests could only be possible with reduced voltages Figure 7 compares the differences in the full-voltage estimated results between the trends-based FEM and various traditional methods The newly suggested method has a maximum spread of 17% to traditional methods The overestimation, underestimation, or mixed behavior of the traditional scal-ing methods can also be witnessed

Origins of Uncertainty

In this section, the differences between the measurements and FEM are discussed in relation to the saturation behav-ior of the machines

Possible Reasons for Differences

The differences between FEM simulations and measure-ments are generally believed to be due to the following:

n the inaccuracies in the input data to FEM, e.g., end-windings inductances and material properties (BH curves and resistances)

n manufacturing tolerances and related deviations

n 3-D-effects approximations in the 2-D-FEM models

Of these, issues such as manufacturing tolerances are in general not voltage dependent, whereas others such as the saturation are known to be dependent upon voltages The iron saturation affects mainly the inductance related to the active length of the machine but does impact the end region/end-winding inductances as well (based on machine design, e.g., aspect ratio) Other parameters, such as the end-winding inductances and rotor resistances, that are sometimes treated as constants in the 2-D-FEM tools have also been found to vary with the voltage For example, the authors have found that the end-windings inductance dur-ing the start has been found to be considerably dependent upon the saturation of the end-core region

During a direct online start, the inrush current is limited by the stator leakage inductance An increasing applied voltage means a higher flux in the machine This

0.3 0.4 0.5 0.6 0.7 0.8 0.9

Maximum Test Voltage –10

–5 0 5 10

Method A Method B Method C Method D Method E Method F

Method G

7

Difference in scaled full-voltage currents: trends-based FEM against various traditional scaling methods (marked as different patterns).

0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65

Maximum Test Voltage –10

–5

0

5

10

5

Differences in starting currents between the highest-voltage

measurements and trends-based scaling against the

scaling (voltage) step.

–8 –6 –4 –2 0 2 4 6 8 10 Starting Currents Differences: Three Tests

Against Five Tests 0

50

100

6

Distribution of the difference between full-voltage starting

currents (in p.u.) derived from compensating slopes with three

and five tests (mean: 0.839, standard deviation: 1.312).

34

will cause certain iron regions in the

machine to saturate and will cause the

inductance to decrease Thus, the stator

current in the machine will increase more

than what would have been expected

through a linear correlation to maintain

the corresponding flux in the machine

Investigations show that the FEM

results are almost always lower than the

measured values, as illustrated in Figure 3

The following are the possible reasons

for lower currents in the FEM

simula-tions than those in measurements

n The lamination materials used

in the machines are found to

have generally lower losses than

those provided in the specification

and, hence, originally assumed

for calculations [18] Reduced

iron losses usually mean also poorer BH curve due

to more silicon and less iron

n The end-winding inductance that is input to the

FEM simulations usually represents the operating

conditions at rated slip At locked-rotor conditions,

high starting currents occur, which are known to

decrease the windings inductances due to

end-core saturation and, hence, higher than expected

currents result during measurements

Interpretation of Different Slopes

Figure 3 also shows that trends (slopes) for differences

between the FEM and measurements vary

A positive slope shows that the saturation model in the

FEM is tuned to fit the rated conditions and is

underesti-mated at lower voltages The input parameters to design

tools are generally calibrated for the full-voltage rated

con-ditions In case of the FEM, the modeled BH curves (for

this particular study) had been calibrated after the knee,

and hence, less saturation in the FEM simulations than in

reality is witnessed at lower voltages (operation in the

region below the knee) In such a case, the end-winding

inductance can be believed as independent of voltage, and

the change in the offset can be primarily due to the

satura-tion of the main core, the 2-D phenomena

A negative slope shows that the difference in the

cur-rents increases as test voltages are increased This means

that the measurements are showing a higher saturation of

the core than that modeled in the FEM with an increasing

voltage This is either because of the onset of saturation in

the end core region, which decreases the end-winding

inductance, or an increased saturation of the 2-D geometry,

which decreases the slot and differential leakage

inductan-ces The end result is an increase in the measured current

when compared with FEM

A zero slope indicates that the saturation model in the

FEM is acceptable, and the difference is due to a wrong

input parameter in the FEM In this case, the

end-wind-ings inductance can be said to be constant

2-D (Tooth-Tip/Airgap) or 3-D (End-Region) Saturation

As stated earlier, a higher saturation in reality than that

modeled by the FEM is believed to be the reason for the

negative slopes in Figure 3 This can

be either because of the saturation of the end core that affects the end-wind-ings inductance (3-D phenomena) or a higher than expected saturation of the teeth tips (due to flux paths in the air gap and surroundings) and effecting slot/differential leakage (2-D phenom-ena) It can also be the case that both phenomena are present with one being more dominant than the other It is found that this dominance can be observed by

n the slopes’ dependencies upon the pole number

n the slopes’ dependencies upon the aspect ratio (core length/air gap diameter)

End-Region Saturation

End-region saturation is due to 3-D effects, which would be dominant for machines having a low aspect ratio (short machines with a high shaft height) and/or for machines having a low pole number (greater end-windings overhang)

Tooth-Tip Saturation

Tooth-tip saturation is due to 2-D effects, which would be dominant for machines having a high aspect ratio (long machines with a low shaft height) and/or for machines hav-ing a high pole number The 2-D leakage increases with the pole number

Findings

The following are the saturation uncertainties findings

n For all the studied groups of machines, the slope magnitudes always decreased with the pole num-ber This showed that the end-winding overhang is directly proportional to the saturation of the end region (frame, end-core, shaft, etc.) and, hence, the uncertainty in the 2-D FEM models

n The trend against the aspect ratio varies when the machines with only negative slopes are considered For the smaller machines, the end-core saturation (3-D leakage) is found to be predominant, whereas for the bigger machines, it is the main core satura-tion (2-D leakage) that is found to have the big-gest influence

Figure 8 shows the slopes of discussed differences against the aspect ratio (core length/air gap diameter) and pole number for certain particular groups of machines In Figure 8, (a) shows a group for which the core-end saturation is dominant (3-D/end-winding-leak-age), (b) shows another group for which the main core saturation is dominant (2-D/tooth tip, zigzag leakage) instead, and (c) shows a group for which the end-region (frame/end-plate) saturation is dominant (3-D/end-wind-ing leakage) In Figure 8(c), the decrease in slopes with the pole numbers reflect the fact that the end-windings overhang decreases with the pole number This group may also simultaneously exhibit the behavior of either Figure 8(a) or (b) as well

THE LOG–LOG CURVE FITTED EXTRAPOLATION PROVIDES THE LOWEST ERROR

TO THE ACTUAL CURRENT AT MAXIMUM TESTS VOLTAGE.

35

Uncertainties associated with the scaling of starting

cur-rents have been studied through a detailed statistical

analy-sis of hundreds of medium–large induction machines The

various traditional scaling methods using only

measure-ments are evaluated for their accuracy This figure is found

to be 88–90% for the individual studied methods Thus, it

can be stated that FATs that employ scaling to predict the

rated-voltage starting currents are not always accurate and

merely provide an indication of the starting current level

The introduced FEM and measurements-based novel

scaling method is shown to fulfill an accuracy level of up to

97% for a voltage scaling step of 0.35 p.u The value

decreases to 95% for a scaling step of 0.45 p.u The

differ-ences between the FEM simulations and the measurements

are found to be increasing, decreasing, or remaining

unchanged with an increasing test voltage The differences,

which are almost always negative, are due to the incorrect

prediction of saturation levels by the FEM By observing certain parametric dependencies, it is shown that it is pos-sible to locate the region that contributes the most to this saturation modeling uncertainty, i.e., tooth/air-gap region, end-region (frame/end-shield), and end-core region References

[1] C Sadarangani, Electrical Machines Stockholm, Sweden: Royal Insti-tute of Technology, 2000.

[2] Rotating Electrical Machines—Part 4: Methods for Determining Synchronous Machine Quantities from Tests, IEC Standard 60034-4, 2008.

[3] Motors and Generators, ANSI/NEMA MG Standard 1-2003.

[4] Electrical Machine—Cage Induction Types (amendments/supplements to IEC 60034-1 Rotating Electrical Machines, Part 1: Rating and Performances 2010), Shell DEP Standard 33.66.05.3-Gen, 1999.

[5] Form-Wound Squirrel-Cage Induction motor—250 Horsepower and Larger, API Standard 541, 1995.

[6] T A Gallant and P J Tavner, “Low starting current cage induction motors for the offshore petrochemical industry,” in Proc IEE Power Electronics Machines and Drives Conf., Bath, U.K., Apr 2002, pp 387–391 [7] J Bredtthauer and N Struck, “Starting of large medium voltage motors: Design, protection and safety,” IEEE Trans Ind Applicat., vol 31, no 5, pp 1167–1176, Sept./Oct 1995.

[8] M R Khan, I Husain, and M F Momen, “Lightly ferromagnetic rotor bars for three-phase squirrel-cage induction machines,” IEEE Trans Ind Applicat., vol 40, no 6, pp 1536–1540, Nov./Dec 2004 [9] O Weingartner, “Asynchronous rotor,” Patent DE3 502 697, June 5, 1986.

[10] D J T Siyambalapitiya, P G McLaren, and P J Tavner, “Transient thermal characteristics of induction machine rotor cage,” IEEE Trans Energy Conversion, vol 3, no 4, pp 849–854, Dec 1998.

[11] T A Gallant and J P Pestle, “The modern design of reliable squirrel cage rotors,” in Proc IEE Electrical Machines and Drives Conf., 1987,

pp 140–144.

[12] G E Parry and J J Middlemiss, “Design and construction to achieve low starting currents on hazardous area motors,” in IEE Colloquium Machines in Hazardous Areas (Dig No 1997/057), Feb 18, 1997,

pp 5/1–5/4.

[13] M R Feyzi and A M Parker, “Heating in deep-bar rotor cages,” IEE Proc Electr Power Applicat., vol 144, no 4, pp 271–276, July 1997 [14] M F Cabanas, J L Ruiz Gonzalez, J L B Sampayo, M G Melero,

C H Rojas, F Pedrayes, A Arguelles, and J Vina, “Analysis of the fatigue causes on the rotor bars of squirrel cage asynchronous motors: Experimental analysis and modeling of medium voltage motors,” in Proc 4th IEEE Int Symp Diagnostics for Electric Machines, Power Electron-ics and Drives (SDEMPED’03), Aug 24–26, 2003, pp 247–252 [15] IEEE Standard Test Procedure for Polyphase Induction Motors and Genera-tors, IEEE Standard 112-2004.

[16] IEEE Guide: Test Procedures for Synchronous Machines, IEEE Standard 115-1995.

[17] W Arshad, C Danielsson, H Persson, H Lendenmann, and J Ha¨gg,

“Rated starting performance of solid-pole salient synchronous motors from reduced voltage factory tests,” in Proc IEEE Industry Applications Society Petroleum and Chemical Industry Committee (PCIC’07), Calgary, Alta, 2007, pp 1–10.

[18] W Arshad, T Ryckebush, F Magnussen, H Lendenmann, and A Boddeflack, “Incorporating laminations processing and electrical machines manufacturing steps in design tools,” in Proc IEEE Industry Applications Society Annual Meeting, New Orleans, LA, Sept 2007,

pp 94–102.

Waqas M Arshad (waqas.arshad@se.abb.com) is with the ABB Corporate Research in Raleigh, North Carolina, and Va¨stera˚s, Sweden Sami Kanerva is with ABB Oy, Electrical Machines in Helsinki, Finland Silvia Bono and Massimo Menescardi are with the ABB SACE in Vittuone (Milan) Italy Holger Persson is with ABB Automation Products/ Machines in Va¨stera˚s, Sweden Arshad and Kanerva are Mem-bers of the IEEE This article first appeared as “Medium–Large Induction Machines Starting Currents: Scaling Accuracy & Sat-uration Uncertainties” at the 2007 IAS Annual Meeting

Aspect Ratio: Core Length/Airgap Diameter (Arbitrary Values)

Aspect Ratio: Core Length/Airgap Diameter (Arbitrary Values)

Pole Number (Arbitrary Values)

(a)

(b)

(c)

8

Finding saturation uncertainties origin from parametric

dependencies (x, y, and z are arbitrary numbers) Various

grouping of machines: (a) unexpected saturation dominant

in the core-end region (3-D/end-winding leakage); (b)

unexpected saturation is dominant in the main core region

(2-D/tooth tip and zigzag-leakage); (c) unexpected

saturation is dominant in the end (frame/end-plate) region

(another contributor to the 3-D/end winding-leakage).

36