part a- lab 2 draft

Upload: omar-abdelrahman

Post on 03-Apr-2018

225 views

Category:

Documents


0 download

TRANSCRIPT

  • 7/28/2019 Part a- Lab 2 Draft

    1/16

    Date Submitted: 02/15/2013

    Report Submitted: Instructor: Prof. David R. Haas

    Omar Khalaf Course Title: Electrical Engr. Lab III

    Garrett Dicken Course #: ECE 494

    Anas Kabashi Lab Report Section #: 102

    Spring 2013

    Experiment Title

    Separation of Eddy Current & Hysteresis Losses

    Comments by Professor:

    Corrected By: Grade:

  • 7/28/2019 Part a- Lab 2 Draft

    2/16

    Table of Contents

    1.1. Abstract of Synopsis...............................................................................................................1

  • 7/28/2019 Part a- Lab 2 Draft

    3/16

    1. Abstract of Synopsis

    Modeling and defining of a procedure leading to a good estimation of power losses in

    electrical machines from materials is challenging. The standardized measurement device

    for measuring the magnetic properties of soft magnetic materials is theEpstein Frame.

    The sample used in this experiment is a Grain Oriented electrical steel which in contrast

    to non-Grain provides more increased magnetic flux and also a decreased magnetic

    saturation. The power losses are measured by means of a wattmetermethod in which the

    primary current and secondary voltage are used. During the measurement the Epsteinframe behaves as an unloaded transformer

    Page 1 of 14

    http://en.wikipedia.org/wiki/Wattmeterhttp://en.wikipedia.org/wiki/Transformerhttp://en.wikipedia.org/wiki/Wattmeterhttp://en.wikipedia.org/wiki/Transformer
  • 7/28/2019 Part a- Lab 2 Draft

    4/16

  • 7/28/2019 Part a- Lab 2 Draft

    5/16

    3. Final Connection Diagram

    Figure 1.1

    Page 3 of 14

  • 7/28/2019 Part a- Lab 2 Draft

    6/16

    4. Data Sheet

    Table 1.1 Magmatic Density (Wb/m)

    0.4 0.6 0.8 1.0 1.2

    Freq. =

    30 (Hz)

    Es Calculated

    (V)20.76 31.14 41.52 51.9 62.26

    Es Measured

    (V)20.89 31.34 41.88 52.0 62.5

    % Error 0.63 0.64 0.87 0.19 0.38

    Power

    Measured

    (Watts)

    1.5 3 4.8 7 10

    Freq. =

    40 (Hz)

    Es Calculated 27.68 41.52 55.36 69.2 83.04

    Es Measured 27.47 41.56 55.6 69.7 83.9

    % Error 0.76 0.096 0.433 0.72 1.03

    Power

    Measured

    (Watts)

    2 3.9 6 9 13

    Freq. =

    50 (Hz)

    Es Calculated 34.6 51.9 69.2 86.5 103.8

    Es Measured 34.51 52.00 69.6 86.5 103.9

    % Error 0.26 0.19 0.57 0.0 0.096

    Power

    Measured

    (Watts)

    2.6 5 8 12 17

    Freq. =

    60 (Hz)

    Es Calculated 41.52 62.28 83.04 103.8 124.56

    Es Measured 41.56 62.8 83.7 104.2 125.1

    % Error 0.096 0.83 0.79 0.38 0.43

    Power

    Measured

    (Watts)

    3.2 6 10 15 20

    Freq. =

    70 (Hz)

    Es Calculated 48.44 72.66 96.88 121.1 145.32

    Es Measured 49.29 72.7 96.7 121.4 144.9

    % Error 1.75 0.055 0.18 0.25 0.29

    Power

    Measured

    (Watts)

    4 7 12 17 24

    Page 4 of 14

  • 7/28/2019 Part a- Lab 2 Draft

    7/16

    5. Computations and Results

    1) Kg Core Loss vs. frequency at different flux densities

    Figure 1.1

    2)

    Table 1.2Frequency

    (Hz)

    W/f at Bm= 0.4 W/f at Bm= 0.6 W/f at Bm= 0.8 W/f at Bm= 1.0 W/f at Bm= 1.2

    30 0.005 0.01 0.016 0.023333 0.03333

    40 0.005 0.00975 0.015 0.0225 0.0325

    50 0.0052 0.01 0.016 0.024 0.034

    60 0.005333 0.01 0.016667 0.025 0.0333370 0.005714 0.01 0.017143 0.024286 0.034286

    Page 5 of 14

  • 7/28/2019 Part a- Lab 2 Draft

    8/16

    Figure 1.3

    Table 1.3

    Bm (Wb/m) Slope = Ke* B2m Y-intercept = Kh*Bnm

    0.4 0.0002 0.00470.6 0.00002 0.0099

    0.8 0.0004 0.015

    1.0 0.0004 0.0225

    1.2 0.0003 0.0327

    Table 1.4

    f = 30 Hz f = 40

    Hz

    f = 50

    Hz

    f =

    60 Hz

    f = 70

    Hz

    Bm

    (Wb/m)Pe =Ke*

    Bmf(Watts)

    Ph=Kh*Bn

    m *f(Watts)

    Pe =Ke*

    Bmf(Watts)

    Ph=Kh*Bnm

    *f(Watts)

    Pe =Ke*Bm

    f(Watts)

    Ph=Kh*Bn

    m *f(Watts)

    Pe =Ke*Bm

    f(Watts)

    Ph=Kh*Bn

    m *f(Watts)

    Pe =Ke*Bm

    f(Watts)

    Ph=Kh*Bnm

    *f(Watts)

    0.4 0.18 0.141 0.32 0.188 0.5 0.235 0.72 0.282 0.98 0.329

    0.6 0.018 0.297 0.032 0.396 0.05 0.496 0.07

    2

    0.594 0.098 0.693

    0.8 0.36 0.45 0.64 0.6 1 0.75 1.44 0.9 1.96 1.05

    1.0 0.36 0.675 0.64 0.9 1 1.125 1.44 1.35 1.96 1.575

    1.2 0.27 0.981 0.48 1.308 0.75 1.635 1.08 1.962 1.47 2.289

    Page 6 of 14

  • 7/28/2019 Part a- Lab 2 Draft

    9/16

    3)

    Pe & Ph vs. frequency at different flux densities

    Figure 1.4

    Page 7 of 14

  • 7/28/2019 Part a- Lab 2 Draft

    10/16

    6. Discussion and Conclusion

    In this lab our main objective was to be able to separate between the eddy and hysteresis

    losses that take place using Epstein core loss testing equipment.

    To begin with, core losses arise in transformers and inductors as a result of alternating

    currents. They are usually present in the form of heat energy and result in increasing the

    temperature of the core. The total amount of core losses is dominated by the summation of the

    hysteresis and eddy current losses.

    The hysteresis loss is caused by continuous reversals in the alignment of the magnetic

    domains in the magnetic core. Succinctly put, the energy that is required to cause these reversals

    is the hysteresis loss.(Gonen) On the other hand, eddy current losses arise, Because iron is a

    conductor, time-varying magnetic fluxes induce opposing voltages and currents called eddy

    currents that circulate within the iron core. In the solid iron core, these undesirable circulating

    currents flow around the flux and are relatively large because they encounter very little

    resistance. Therefore, they produce power losses with associated heating effects and cause

    demagnetization. As a result of this demagnetization, the flux distribution in the core becomes

    non-uniform, since most of the flux is pushed toward the outer surface of the magnetic

    material.(Gonen) In order to decrease the magnitude of the eddy currents, the resistance of thecore has to be increased drastically. This is achieved by laminating the core material or using non

    conducting magnetic material for the core.

    The power lost as a result of hysteresis losses (Ph) is summarized in the following

    formula,

    Page 8 of 14

  • 7/28/2019 Part a- Lab 2 Draft

    11/16

    We can see that the factors that affect the hysteresis losses are the volume of the material,

    and hence the cross sectional area and the length of the core, the frequency and the maximum

    flux density through the core. All these factors are directly proportional to the power lost due to

    hysteresis; hence the increase of any of those factors will directly increase the hysteresis losses.

    Similarly, the eddy current losses are also measured through the following formula,

    The formula is very similar to that of the hysteresis power losses, except that as the

    frequency or the magnetic flux doubles, the eddy current losses is quadrupled. Also for the eddy

    current losses the lamination thickness has to be taken into consideration because as the

    thickness gets smaller, the power loss due to eddy currents diminishes.

    In this particular lab, we started off by following the procedure described earlier and by

    preparing a circuit similar to the one present in the final connection diagram. A circuit breaker

    was connected between the variable frequency supply and the variac which was later connected

    in parallel with the primary side of the Epstein test equipment. At the secondary side a double

    pole double throw (DPDT) switch was connected where one setting connected to a voltmeter andthe other connected to an LPF wattmeter. The wattmeter was connected because it has less

    resistance in its wires than the voltmeter; as a result a high proportion of the current is going to

    pass through the coils of the wattmeter than it would pass through the voltmeter. Consequently,

    readings obtained were more accurate through the wattmeter than they were through the

    voltmeter. The frequency was then altered and changed and at every setting the voltage input was

    changed to obtain higher flux densities. The values were obtained before the lab and the numbers

    were tested for errors afterwards. All the results obtained were tabulated in the datasheet portion

    of the lab.

    Figure 1.1 in the computation and results shows the relationship between the density of

    power loss per kg of core material as a function of frequency. We observed how an increase in

    frequency resulted in an increase in the power loss. Moreover, as the flux density increased the

    magnitude of power loss also increased. These results match other work of literature that

    Page 9 of 14

  • 7/28/2019 Part a- Lab 2 Draft

    12/16

    indicates that the power loss is directly proportional to both the frequency and the magnetic flux

    passing through the core material.

    In table 1.2, the ratio between the power lost and the frequency was obtained as the

    frequency and the magnetic flux changed. Amazingly, the ratio was almost constant for a given

    magnetic flux, only slightly increased as the frequency increased. However, when the magnetic

    flux was increased from 0.4 to 1.2 Wb/m2, the resultant ratio of power lost to frequency was over

    six times larger. Hence, at a given frequency, the major factor affecting the power lost would be

    the magnetic flux density. All the various ratios of power losses and frequencies were plotted

    against frequency, for different values of maximum magnetic flux density, the slopes and the y-

    intercepts of these lines were viewed on the plot and were later tabulated in table 1.3. The slope

    when multiplied by the square of the frequency yields the power losses due to eddy currents.

    Whereas, the product of the y-intercept and the frequency yielded the magnitude of the power

    lost due to hysteresis. The hysteresis losses and the eddy currents losses were then calculated

    tabulated as they changed with frequency and magnetic flux density in table 1.4 and they were

    then plotted in figure 1.4. We concluded from the previous plot, that the hysteresis losses are

    usually much higher at lower frequencies and they only increase slightly as the frequency is

    increased. On the other hand, eddy losses are much smaller at lower frequencies but increase

    drastically as the frequency is increased relative to hysteresis power losses.

    As for the Kh and Ke calculations, we used traditional computation methods as the

    graphical approach did not prove helpful in our case. The following steps and measures were

    taken to figure out these values,

    To calculate n and Kh:

    For Bm = 0.4 T/m

    Log (0.047) = Log Kh + n Log (0.4)

    - 1.33 = Log Kh (0.398) n (1)

    For Bm = 0.6 T/m

    Log (0.0979) = Log Kh + n Log (0.6)

    Page 10 of 14

  • 7/28/2019 Part a- Lab 2 Draft

    13/16

    - 1.0092 = Log Kh (0.223) n (2)

    From Equation (1):

    Log Kh = (0.398) n 1.33 .. (3)

    Substitute (3) in (2):

    - 1.0092 = (0.398) n 1.33 (0.223) n

    n = 1.833

    and Kh = 0.251

    - To Calculate Ke:

    For Bm = 0.4 T/m

    Log (0.0018) = Log Ke + 2 Log (0.4)

    - 2.744 = Log Ke 0.796

    Ke = 0.0021

    Similarly:

    For Bm = 0.6 Ke = 0.0013

    For Bm = 0.8 Ke = 0.0064

    For Bm = 1.0 Ke = 0.0046

    For Bm = 1.2 Ke = 0.0043

    Keava. = (0.0021+0.0013+0.0064+0.0046+0.0043)/5 = 0.00374

    All in all, most of the outcomes of the lab were expected; the power loss increased

    drastically as frequency and the magnetic flux density consumed larger values. The plots and

    data from the tables shoed that. However, we faced a stumbling block when it came to measuring

    the proportionality constants of the material with respect to hysteresis and eddy current losses of

    the core material of the Epstein test equipment. Consequently, we had to use an unorthodox

    method of calculating the values Steinmetz exponent as well as the proportionality constants.

    Page 11 of 14

  • 7/28/2019 Part a- Lab 2 Draft

    14/16

    7. Bibliography

    M. I.T.Sta_, MagneticCircuitsandTransformers,pp.144-154,John Wiley andSons,1985.

    Vincent Del Toro, Basic Electric Machines , pp. 34-38, Prentice Hall, 1990

    Turan Gnen,Electrical Machines with MATLAB, pp. 17-41, 2nd Edition, CRC Press, BocaRaton, Fla, 2012

    Dr. Edwin Cohen, Dr. Sol Rosenstark,NJIT Electrical Engineering Laboratory IV, version 1.3,Revised by Dr. David R. Haas, 2013

    Page 12 of 14

  • 7/28/2019 Part a- Lab 2 Draft

    15/16

    8. Observed Data Sheet

    Page 13 of 14

  • 7/28/2019 Part a- Lab 2 Draft

    16/16

    Page 14 of 14