i
Estimating the magnetic characteristics of a salient pole
synchronous machine using ampere turns distribution method
By
Jayaram Subramanian
B.E., Anna University, 2012
A thesis submitted in partial fulfillment of the requirements for the degree of
Master of Applied Science
In the department of Electrical and Computer Engineering
©Jayaram Subramanian 2015
University of Victoria
All rights reserved. This dissertation may not be reproduced in whole or in part, by
photocopying or by other means, without the permission of the author
ii
Supervisory Committee
Estimating the magnetic characteristics of a salient pole
synchronous machine using ampere turns distribution method
By
Jayaram Subramanian
B.E., Anna University, 2012
Supervisory Committee
Dr.Subhasis Nandi,
Supervisor (Department of Electrical and Computer Engineering)
Dr.Nikitas Dimopoulos,
iii
Abstract
Supervisory Committee
Dr.Subhasis Nandi,
Supervisor (Department of Electrical and Computer Engineering)
Dr.Nikitas Dimopoulos,
Department Member (Department of Electrical and Computer Engineering)
Modeling plays a very important role in a variety of applications such as performance analysis, characterization, fault diagnosis, condition monitoring and stress analysis of electrical machines. With the importance of modeling of electrical machines increasing day by day, researchers are striving for better methods to solve the problem. One of the widely used techniques for modeling electrical machine is the finite element method. As computational power continues to be less and less expensive, the finite element method is becoming a widely used technique for modeling of electrical machines because of its advantages in terms of accuracy and efficiency. Many commercial finite element software packages are now available for this purpose. One such software, the Ansys Maxwell is used extensively for the modeling of electrical machines. It is the top of the line finite element package used by many motor manufacturers for industrial motor design and performance analysis. Ansys Maxwell has specific features such as the field calculator and RMxprt which facilitates the modeling of electrical machines. One of the important parameters while modeling electrical machine is the magnetic
iv characteristics of the core material. This plays a significant role in the performance characteristics and the analysis of electrical machines. This research work addresses this problem and provides a simple yet effective solution to determine the average magnetic characteristics of a salient pole synchronous machine that uses a material for the rotor with unknown magnetic characteristics. Existing techniques available to determine the magnetic characteristics of a material are mainly Epstein and single sheet tester. These two tests require a separate sheet of material and they are destructive. Therefore a non-invasive and non-destructive technique had to be designed to solve this problem as the manufacturers could not provide the data for the magnetic material used in the rotor.
In this work, an FE model of the salient pole synchronous machine was developed to closely emulate the characteristics of the experimental machine. This FE model was first subjected to magnetostatic simulation under different field currents using a known magnetic material. By comparing the result with the experimental machine and by performing a technique named ampere turn distribution technique, a new magnetic material characteristic was developed to follow the average characteristics of the rotor and the stator. Following the determination of the new material, this material was used in the simulation of the salient pole synchronous machine running as a motor and a generator under varying load condition and field currents. These results were then compared with the real machine to determine the effectiveness of the developed scheme.
The pursuit of research in this topic led to the following publication:
1. Subramanian, J.; Nandi, S.; Ilamparithi, T.; Winter, O., "Estimating the magnetic characteristics of a salient pole synchronous machine using ampere turns distribution method," Electrical Machines (ICEM), 2014 International
v
Table of Contents
Supervisory Committee ... ii
Abstract ... iii
List of Figures ... viii
List of Tables ... xi
List of Abbreviations ... xiii
List of symbols ... xiv
Chapter 1 Introduction to magnetic measurement techniques in rotating machines ... 1
1.1 Introduction to modeling of electrical machines ... 1
1.2 Introduction to measurement techniques of magnetic characteristics of materials in Electrical machines ... 2
1.2.1 Ring Test ... 3
1.2.2 Epstein Test ... 4
1.2.3 Single Sheet Test (SST) ... 6
1.2.4 Other Techniques ... 11
1.2.5 Comparison of Epstein and SST techniques ... 16
1.3 Motivation of the present research work ... 16
1.4 Thesis outline ... 17
Chapter 2 Finite Element Modeling ... 19
2.1 Electromagnetic analysis ... 19
2.2 Finite Element Analysis (FEA): ... 20
2.3 FE in Electrical Machines ... 21
2.4 Commercial FE packages available for Electrical Machines: ... 22
vi 2.5.1 Ansys RMxprt ... 25 2.5.2 Field Calculator:... 27 2.5.3 Solvers... 28 2.5.4 Material Properties ... 31 2.6 Conclusion ... 33
Chapter 3 Ampere Turn Distribution Scheme ... 34
3.1 Magnetic Materials used in Electrical Machines ... 34
3.1.1 Non-oriented Steel ... 34
3.1.2 Grain-oriented Steel ... 35
3.2 Finite element modeling of Salient Pole Synchronous Machine ... 35
3.2.1 Salient Pole Synchronous Motor ... 36
3.2.2 Salient pole synchronous generator ... 39
3.3 AT (ampere turn) Distribution Scheme ... 42
3.3.1 Steps of AT distribution scheme ... 42
3.3.2 Open circuit test ... 43
3.3.3 Magnetostatic simulation in Ansys Maxwell ... 44
3.3.4 Calculation of ampere turns for different parts of the motor: ... 50
3.3.5 Calculation of magnetic flux density from the experiment data ... 53
3.4 Comparison of Steel1008, M27 and the new material ... 63
3.5 Conclusion ... 67
Chapter 4 Comparison of results with new magnetic material and real SPSM ... 68
4.1 SPSM as a Motor ... 68
4.2 Comparison of Experiment and FE at different load conditions with SPSM as a Motor 70 4.2.1 Full Load condition ... 70
vii
4.2.2 75% Full Load condition ... 72
4.2.3 66% Full Load condition ... 74
4.2.4 50% Full Load condition ... 76
4.2.5 33% Full Load condition ... 78
4.2.6 No Load condition ... 80
4.3 Harmonic analysis of the stator current in SPSM as a motor... 86
4.4 SPSM as a generator ... 89
4.4.1 Generator with Resistive Load ... 90
4.4.2 Generator with Resistive-Inductive (RL) Load ... 95
4.4.3 Generator with Resistive-Capacitive (RC) Load: ... 100
Chapter 5 Conclusion ... 107
5.1 Conclusion ... 107
5.2 Advantages and Disadvantages of the ampere turn distribution scheme ... 107
5.3 Contributions ... 108
5.4 Future Scope ... 109
References: ... 110
viii
List of Figures
Figure 1-1. Classification of magnetic measurement techniques ... 2
Figure 1-2. Different shapes of the sample for ring tester [2] ... 3
Figure 1-3. Ring test rig with control algorithm [3] © 2011, IEEE ... 4
Figure 1-4. Single Sheet Tester frame [6] © 1974, IEEE ... 7
Figure 1-5. Single Sheet Tester [6] © 1974, IEEE ... 8
Figure 1-6. 2H coil SST [7] © 2009, IEEE ... 9
Figure 1-7. 2H SST measurement system [7] © 2009, IEEE ... 9
Figure 1-8. Double excitation SST [8] © 1999, IEEE ... 10
Figure 1-9. Open Type SST - DC magnetization [9] © 2010, IEEE ... 11
Figure 1-10. 3D tester (cubic sensing box) [10] © 2010, IEEE ... 12
Figure 1-11. 3D tester for SMC [11] © 2003, IEEE ... 13
Figure 1-12. Epstein FEM [13] © 2005, IEEE ... 14
Figure 1-13. SST FEM [13] © 2005, IEEE ... 14
Figure 1-14. Geometry of induction motor [14] © 2012, IEEE ... 15
Figure 2-1. Electromagnetic Analysis Solution [17] ... 19
Figure 2-2. FE software process [18] ... 23
Figure 2-3. Ansys Maxwell - process flow [20] ... 24
Figure 2-4. Ansys Maxwell and related products [20] ... 25
Figure 2-5. RMxprt - DC motor [21] ... 26
Figure 2-6. Field calculator [22] ... 27
Figure 2-7. Magnetostatic Solution Process [17] ... 29
Figure 2-8. Eddy current solution process [17] ... 30
Figure 2-9. Transient Solution process [17] ... 31
Figure 2-10. Material Properties ... 32
Figure 2-11. M27 Core loss Model (Red curve – actual one and Black curve – inserted automatically by Maxwell to smoothen the characteristics) ... 33
Figure 3-1. Salient pole synchronous machine – FE model ... 36
Figure 3-2. Excitation voltage - winding 1 ... 37
Figure 3-3. Excitation voltage - winding 2 ... 37
ix
Figure 3-5. Field current ... 38
Figure 3-6. Load setting ... 39
Figure 3-7. Constant Speed - Prime mover setting ... 40
Figure 3-8. Constant speed of 1800 RPM ... 41
Figure 3-9. Circuit for generator with R Load ... 41
Figure 3-10. Steps in AT (ampere turn) distribution scheme ... 43
Figure 3-11. Optometrics setup... 45
Figure 3-12. Field winding (Pink) in SPSM ... 46
Figure 3-13. Plot of H for 0.5A Field Current ... 47
Figure 3-14. Plot of B for 0.5A Field Current ... 48
Figure 3-15. Modify attribute tab in magnetostatic simulation ... 49
Figure 3-16. Ampere turn distribution in SPSM ... 52
Figure 3-17. Ampere turn distribution steps shown in Table 3-8 ... 58
Figure 3-18. BH steel1008 vs New material ... 63
Figure 3-19. BH plot of different parts of SPSM based on Table 3-9, 3-11 and 3-12 ... 64
Figure 3-20. Comparison of generated voltage ... 66
Figure 3-21. Comparison of OCC difference - Steel1008, M27 and new Material from the experimental OCC ... 66
Figure 4-1. SPSM used in the experiments ... 68
Figure 4-2. Stator current of SPSM at FL at 0.7A field current ... 71
Figure 4-3. Stator current of SPSM at FL - zoomed in version of Figure 4-2 ... 72
Figure 4-4. Stator current of SPSM at 75% FL at 1A field current ... 73
Figure 4-5. Stator current of SPSM at 75% FL - zoomed in version of Figure 4-4 ... 74
Figure 4-6. Stator current of SPSM at 66% FL at 1.2A field current ... 75
Figure 4-7. Stator current of SPSM at 66% FL - zoomed in version of Figure 4-6 ... 76
Figure 4-8. Stator current of SPSM at 50% FL at 1A field current ... 77
Figure 4-9. Stator Current of SPSM at 50% FL - zoomed in version of Figure 4-8... 78
Figure 4-10. Stator current of SPSM at 33% FL at 0.7A field current ... 79
Figure 4-11. Stator current of SPSM at 33% FL - Zoomed in version ... 80
Figure 4-12. Stator current of SPSM at NL at 1.2A field current ... 81
x
Figure 4-14. Comparison of stator current at 0.7A field current ... 82
Figure 4-15. Comparison of stator current at 1A field current ... 83
Figure 4-16. Comparison of stator current at 1.2A field current ... 83
Figure 4-17. Comparison of power factor at 0.7A field current ... 84
Figure 4-18. Comparison of power factor at 1A field current ... 84
Figure 4-19. Comparison of power factor at 1.2A field current ... 85
Figure 4-20. FFT of stator current (a) Experimental SPSM (b) FE simulation of SPSM with new material (c) FE simulation of SPSM with M27... 87
Figure 4-21 – Experimental set up of the generator ... 89
Figure 4-22. Generated voltage of SPSM at 50% FL and 0.9A field current for R load... 92
Figure 4-23. Zoomed in version of Figure 4-22 ... 93
Figure 4-24. Comparison of phase voltage for R load at 0.72A field current ... 93
Figure 4-25. Comparison of phase voltage for R load at 0.9A field current ... 94
Figure 4-26. Comparison of phase voltage for R load at 1.08A field current ... 94
Figure 4-27. Generated voltage of SPSM at 50% FL and 0.9A field current for RL load ... 97
Figure 4-28. Zoomed in version of Figure 4-27 ... 98
Figure 4-29. Comparison of phase voltage for RL load at 0.72A field current ... 98
Figure 4-30. Comparison of phase voltage for RL load at 0.9A field current ... 99
Figure 4-31. Comparison of phase voltage for RL load at 1.08A field current ... 99
Figure 4-32 – Generated voltage of SPSM at 50% FL and 0.9A field current for RC load ... 102
Figure 4-33 – Zoomed in version of Figure 4-32 ... 103
Figure 4-34. Comparison of phase voltage for RC load at 0.72A field current ... 103
Figure 4-35. Comparison of phase voltage for RC load at 0.9A field current ... 104
Figure 4-36. Comparison of phase voltage for RC load at 1.08A field current ... 104
Figure A1 - SPSM wiring diagram ... 113
xi
List of Tables
Table 1-1. Comparison of Epstein and SST [16] ... 16
Table 2-1. Electrical machine design software packages ... 22
Table 3-1. Composition of silicon steel [23] ... 35
Table 3-2. Open circuit voltage - comparison of different materials with the experimental motor ... 44
Table 3-3. Field current - FE... 47
Table 3-4. Magnetic field intensity (AT/m) of Core, Yoke, Pole, Teeth of SPSM ... 48
Table 3-5. Length of different parts of the SPSM ... 50
Table 3-6. Ampere Turn for Field Current of 1A ... 50
Table 3-7. Ampere turns from FE simulation ... 56
Table 3-8 . Estimated distribution of ampere turns for the actual machine ... 57
Table 3-9. Magnetic field intensity for actual machine ... 58
Table 3-10. BH for core ... 59
Table 3-11. BH for Yoke ... 60
Table 3-12. BH for Teeth ... 61
Table 3-13. BH for Pole ... 62
Table 3-14. Comparison of open circuit voltage for different materials ... 65
Table 4-1. Motor experiments... 69
Table 4-2. Stator current of the motor at FL condition ... 70
Table 4-3. Power factor at FL condition ... 71
Table 4-4. Stator current of the motor at 75% FL condition ... 72
Table 4-5. Power factor at 75% FL condition... 73
Table 4-6. Stator current of the motor at 66% FL condition ... 74
Table 4-7. Power factor at 66% FL condition... 75
Table 4-8. Stator current of the motor at 50% FL condition ... 76
Table 4-9. Power factor at 50% FL condition... 77
Table 4-10. Stator current of the motor at 33% FL condition ... 78
Table 4-11. Power factor at 33% FL condition... 79
Table 4-12. Stator current of the motor at NL condition ... 80
xii Table 4-14. Percentage deviation of stator current from the experiment for M27 and new material
... 86
Table 4-15. Spectrum analysis of the stator current of SPSM ... 88
Table 4-16. Generator experiments ... 90
Table 4-17. Phase voltage at FL as a generator for R load ... 91
Table 4-18. Phase voltage at 75% FL as a generator for R load ... 91
Table 4-19 - Phase voltage at 66% FL as a generator for R load ... 91
Table 4-20. Phase voltage at 50% FL as a Generator for R load ... 92
Table 4-21. Phase voltage at 25% FL as a generator for R load ... 92
Table 4-22. Percentage deviation of phase voltage from the experiment for M27 and new material for R load ... 95
Table 4-23. Phase voltage at FL as a generator for RL load ... 96
Table 4-24. Phase voltage at 75% FL as a generator for RL load ... 96
Table 4-25. Phase voltage at 66% FL as a generator for RL load ... 96
Table 4-26. Phase voltage at 50% FL as a generator for RL load ... 96
Table 4-27. Phase voltage at 25% FL as a generator for RL load ... 97
Table 4-28. Percentage deviation of phase voltage from the experiment for M27 and new material for RL load ... 100
Table 4-29. Phase voltage at FL as a generator for RC load ... 101
Table 4-30. Phase voltage at 75% FL as a generator for RC load ... 101
Table 4-31. Phase voltage at 66% FL as a generator for RC load ... 101
Table 4-32. Phase voltage at 50% FL as a generator for RC load ... 101
Table 4-33. Phase voltage at 25% FL as a generator for RC load ... 102
Table 4-34. Percentage deviation of phase voltage from the experiment for M27 and new material for RC load... 105
xiii
List of Abbreviations
AC – Alternating Current AT – Ampere Turn DC – Direct Current FE – Finite ElementFEA – Finite Element Analysis FEM – Finite Element Model FFT – Fast Fourier Transform FL – Full Load
IEC – International Electrotechnical Commission NL – No Load
OCC – Open Circuit Characteristics PF – Power Factor
RD – Reverse Direction
SMC – Soft Magnetic Composite
SPSM – Salient Pole Synchronous Machine SST – Single Sheet Tester
xiv
List of symbols
– Area of the core (m2) – Area of the pole (m2)
– Area of the teeth (m2) – Area of the yoke (m2)
– Ampere turn of the core (AT)
– Ampere turn for the experiment (AT)
– Ampere turn for FE simulation (AT)
– Ampere turn of the air gap (AT) – Ampere turn of the iron (AT) – Ampere turn of the pole (AT) – Ampere turn of the teeth (AT) – Ampere turn of the yoke (AT) B – Magnetic flux density (T)
– Magnetic flux density of the air gap (T) - Magnetic flux density of the core (T) - Magnetic flux density of the yoke (T) - Magnetic flux density of the pole (T) - Magnetic flux density of the teeth (T) – Breadth of the pole (m)
xv – Width of the core (m)
– Width of the yoke (m)
-
Phase voltage (V)f – Frequency (Hz)
H – Magnetic field intensity (AT/m) IF - field current (A)
- Field form factor - Gap contraction factor
– Winding factor L – Inductance (H)
– Length of the core (m) – Length of the pole (m)
– Length of the teeth (m) – Length of the yoke (m) – Length of the air gap (m) N – Number of turns
NF – Total number of turns in the field winding – Pole arc (m)
R – Resistance (Ω)
– Width of the teeth (m)
xvi
- Flux per phase (wb)
– Pole pitch (m)
xvii
Acknowledgements
I would like to sincerely thank my supervisor Dr.Nandi for spending his valuable time in guiding me through this research work. I would also like to acknowledge him for the detailed supervision, advice, ideas and words of encouragement which helped me complete the thesis. I would like to thank the other member of my supervisory committee Dr.Nikitas Dimopolous for agreeing to be in my supervisory committee.
I would like to thank Mr.Rob Fichtner and Mr.Kevin Jones for helping me in the setup of the simulations and experiments.
I would like to thank Ilamparithi, Nagendrappa, Premkumar, Nethra and Komal for their continuous support and words of encouragement throughout my degree. I would also like to thank my friends Karthik, Raghavendran, Yashu, Jainish and Thejasvi for keeping me motivated through these years.
Last but not the least I would like to thank my family members without whom this work would have never been possible.
xviii
Dedication
1
Chapter 1
Introduction
to
magnetic
measurement
techniques in rotating machines
1.1 Introduction to modeling of electrical machines
Modeling of electrical machines is important for the following reasons: 1. Design and performance analysis of machines
2. Fault diagnosis and condition monitoring of the machines 3. Analyzing the characteristics of the electrical machines 4. Thermal and stress analysis under extreme conditions
There are a plethora of modeling techniques of electrical machines available for researchers such as finite element modeling and mathematical modeling. Of these, mathematical modeling is used for the parametric estimation of the electrical machines and the detailed mathematical modeling of motors is shown in [1]. Finite element (FE) method of electromagnetic analysis involves utilizing either own FE code or using commercial Finite software packages in the market. While modeling using FE method in these commercial software packages, one of the important parameters while defining the machine is the material used in the motor. The characteristics of the conductors and the core have to be specified in the motor model. Therefore knowledge of the magnetic characteristics of the material is highly important and usually the manufacturers can provide the details of the material used in the motor. Sometimes it is difficult to know the magnetic characteristics of the material since their characteristics might change due to aging, thermal and mechanical stresses on the machines. Therefore the determination of the magnetic characteristics is important not only when the motor is new but also periodically as long as the machine remains in service. This is to evaluate the performance of the motor continuously and check for faults which might develop in the machine while in operation. Some of the commonly used FE software packages in the modeling of electrical machines are Ansys, Magnet and Modelica. Detailed descriptions of these software packages and their characteristics are presented in chapter 2.
2
1.2 Introduction to measurement techniques of magnetic
characteristics of materials in Electrical machines
There are three main techniques in the determination of the magnetic characteristics of the material. They are Ring test, Epstein test and Single sheet test. Of these, Epstein test is widely used, followed by Single sheet test. Descriptions of these techniques are given below:
A classification of the current magnetic measurement techniques is shown below in Figure 1-1.
3
1.2.1 Ring Test
Ring test is the most fundamental method of testing the magnetic properties of a material. Here annealing has to be done to reduce the effect of stresses in the material being used. The measurement of the core losses and hysteresis losses are done using the wattmeter method. The shape of the sample can be of different forms as shown in Figure 1-2 [2].
Figure 1-2. Different shapes of the sample for ring tester [2]
Once the sample is ready, the primary and secondary windings are wound around the sample to measure the core loss similar to a transformer experiment. This is a primitive and destructive method of testing and not widely used now.
Modifications were made to this technique by introducing a test bed incorporated with a new control algorithm to calculate all the measurement systems as shown in [3].This was followed by FE simulation of the stator yoke to verify the iron loss of the material. The test rig is shown in Figure 1-3.
4
Figure 1-3. Ring test rig with control algorithm [3] © 2011, IEEE
1.2.2 Epstein Test
A 25cm Epstein frame with double lapped joints has been the standardized procedure for characterization of magnetic characteristics of soft magnetic materials in the industries since 1936 [4]. Epstein tester follows the standard in IEC 402-2. This procedure has been tested by a lot of researchers and the results are highly reproducible. Hence the industry and researchers have widely preferred to use the Epstein tester for the determination of the magnetic properties of a given magnetic material. The setup of Epstein tester for magnetic measurement is as follows. Epstein test frame is designed using four strips of the magnetic material (or multiples of four) superposed at corners by double lapped joints. Each side of the square is provided with a secondary coil and external to it, a primary winding put together in a rigid rectangular frame. A total of 700 turns are used for the primary and secondary windings for DC and power frequency measurements (IEC 60404-2) and 200 turns for medium frequency testing (IEC60404-10). The sample to be tested must be 30mm wide and 280 – 305mm in length. The mean magnetic path
5 length is assumed to be 0.94m and this assumption has been tested in different methods in [5]. The Epstein frame can operate up to levels of 30KA/m and 1.5T with an accuracy of 1.5%. The power losses are measured by means of wattmeter method and during the experiment measurement, the Epstein frame behaves as an unloaded transformer. The magnetizing field intensity for individual test points in this procedure is calculated using the formula.
where
– Magnetic field intensity
– Number of magnetizing winding turns – Magnetizing current (Peak amperes) – Mean magnetic path length (0.94m)
With = 0.94m and 25cm frame, the equation reduces to The magnetic field is determined using the formula
where
– Measured Voltage – Frequency
– Area of the sample
Some of the advantages of the Epstein frame technique are:
1. Epstein tester is widely used and rigorously tested technique. Therefore reproducibility of results is easy.
2. The test sample can be placed in the test rig and can be removed after testing. Therefore easy replacement of the sample is possible. Furthermore since there is direct relationship between the current, H and Area, precalculated tables can be used for routine testing.
6 3. The test sample lies loosely in the test rig. Therefore no pressure, bending or strain is
subjected on the test sample/strips.
Some of the disadvantages of the Epstein frame technique are:
1. This is a destructive method of testing and there is a requirement of a large amount of samples for testing.
2. Epstein works only till B = 1.5T for non-oriented steel and 1.8T for oriented steel measurements. At high flux densities, digital control is necessary and the reproducibility and accuracy reduces.
3. The preparation of the specimen is time consuming and tedious.
1.2.3 Single Sheet Test (SST)
1.2.3.1
Single Excitation
SST is an alternate method for Epstein technique and tries to avoid some of its difficulties. The principle of SST is similar to an open circuit test in a transformer as shown in Figure 1-4. The specimen to be tested is placed between the yoke which employs a measuring coil to determine the B and H and a primary winding to apply the magnetizing field. In SST, the H and B values are acquired directly using a flux meter [7] and does not require calculations from the magnetizing current and mean magnetic path length like the Epstein technique. To achieve that, the flux must be uniform over the region being measured and such conditions can be achieved by using yokes as shown in [6]. Figure 1-5 shows the measuring instrument for the SST. It can be seen that the value of flux density can be calculated directly using the voltage divider and digital voltmeter. The core loss is measured using the equation shown below
where
7 – Magnetic flux density
– Period of the fundamental wave
The voltage induced in the H coil is amplified and the voltage induced in the B coil is integrated and amplified. These two voltages are multiplied and averaged over a single period.
Current international standard for SST are 50cm square sample, Single magnetizing coils and two yokes
Advantages:
1. It has similar measurement quantities like the Epstein frame. 2. It is easier to prepare the specimen.
3. It requires lower specimen mass and is easier to install compared to Epstein frame. 4. Easy to remove the samples and replace it with a new specimen.
Disadvantages:
1. It is an invasive method of testing i.e. it requires samples of a material in a specific shape. 2. It does not have good reproducibility.
8
Figure 1-5. Single Sheet Tester [6] © 1974, IEEE
One of the important aspects in the determination of magnetic characteristics is to determine the BH characteristics of the material at high flux densities i.e. above 1.5T. This aspect is one of the disadvantages with both Epstein and SST and has been tackled by researchers by introducing novel models of the SST.
1.2.3.2
Double Excitation:
In reference [7], a novel method of double excitation type SST in determining the magnetic properties is shown. Here, a novel 2H coil method is proposed for the H coils to be used in the SST. Figure 1-6 shows the 2H coil pair used in this method for the SST. The complete measurement system is shown in Figure 1-7. The two H coils help in increasing the accuracy while measuring the magnetic properties up to 2.1 T and H of 58000 A/m.
In reference [8], a double excitation type SST was introduced to help determine the magnetic properties at higher flux densities. The SST developed for this scheme is shown in Figure 1-8. Two magnetizing windings have been introduced – one for rolling direction (RD) and one for transverse direction (TD) as shown in Figure 1-8. These two windings helps in satisfying the
9 rotating flux condition exhibited in rotating machines. The way in which the TD winding is placed inside the RD helps in increasing the maximum flux density. In this type of excitation, closed path magnetic circuit was realized successfully to measure high flux densities. The results were in good agreement with the results from a normal SST at low flux densities proving the effectiveness of the double excitation type SST.
Figure 1-6. 2H coil SST[7] © 2009, IEEE
10
Figure 1-8. Double excitation SST[8] © 1999, IEEE
1.2.3.3
Open Type SST:
All the above SST and Epstein techniques are mainly used for AC excitation and determination of magnetic properties under AC excitation. Reference [9] shows the technique to determine magnetic properties under DC excitation using SST. This is particularly useful for determining iron losses for reactors in an inverter which works with DC excitation. An open type SST is designed with a help of Helmholtz coil as shown in Figure 1-9. The H is found by using a Hall probe since it is difficult to measure Hdc using normal H coil and to determine the change in B,
the output of B coil is integrated during change of current from zero to a specified value in the Helmholtz coil.
11
Figure 1-9. Open Type SST - DC magnetization [9] © 2010, IEEE
1.2.4 Other Techniques
1.2.4.1
3D Tester
Determination of magnetic characteristics is important for machine modeling and since the electrical machines experience 3D flux, the measurement technique must incorporate the effect of 3D flux in its calculations. Therefore few techniques were developed including these effects.
Reference [10] shows a technique of measuring the magnetic properties of grain oriented steel using a 3D tester model. The structure of the tester is shown in Figure 1-10. Using this technique, BH loci, core losses were calculated and validated with the experimental results. The disadvantage of this scheme was the difficulty in the production of strong fields while maintaining the field pattern.
12
Figure 1-10. 3D tester (cubic sensing box) [10] © 2010, IEEE
Reference [11] shows another 3D tester model for measuring the magnetic properties of the soft magnetic composite (SMC) material. This method also estimates the BH loci, power loss and core losses in the material. The structure of this tester is shown in Figure 1-11. Finite element model of the tester and the whole system was developed and studied and was followed by the implementation in an experimental test system.
All these techniques clearly shows that to identify a magnetic characteristics of a material, a separate specimen of the material in certain shapes is required which is followed by testing of that material in different conditions.
1.2.4.2
FE model
A two dimensional approach of finite element proposed in [12] by Enokizono was analyzed to determine the magnetic characteristics of a material. This approach attempted to determine the 2D magnetic properties at high flux density as this is important for a lot of applications of which electrical machine modeling and analysis is one. An extrapolation technique is used to determine the magnetic properties at high flux density above the saturation level of 2T. Bezier interpolation technique is used to determine the necessary co-efficient for Newton-Raphson iteration. This
13 showed the importance of Bezier interpolation technique in the identification of 2D magnetic properties.
Figure 1-11. 3D tester for SMC[11] © 2003, IEEE
A three dimensional approach for Finite Element modeling has been proposed in [13]. A 3D FEM was modeled for Epstein and SST and their results were compared. The 3D FEM modeling strategy for Epstein and SST are shown in Figure 1-12 and Figure 1-13 respectively. lFE in Figure 1-12 represents the length of the laminated magnetic core. It was found that Epstein showed more error compared to SST during FEM modeling.
14
Figure 1-12. Epstein FEM [13] © 2005, IEEE
Figure 1-13. SST FEM [13] © 2005, IEEE
15
1.2.4.3
Nondestructive Testing
Reference [14] describes a non-destructive method for the detection of BH characteristics of a material of a motor. It uses local and global magnetic measurements and objective functions to determine the B-H curve of the material through optimization techniques. Global measurements measure the excitation current and voltage which is used to determined the coupled magnetic flux. Local measurements measure the flux in a tooth by adding a search coil. Then using numerical inverse method and iteratively minimizing the quadratic difference between simulated and measured peak magnetic flux, the BH curve is obtained. The evaluation of these quadratic functions needs a lot of computations to minimize the error between simulated and measured quantities. Also a few more calculations have to be done during the computation of the magnetic flux of the material. Overall this method is computationally and memory requirement wise intensive. Besides, the machine has to be disassembled in order to put in search coils around stator teeth for measurement purposes. Also a hole has to be drilled to add these search coils to measure the flux. The geometry of the studied asynchronous motor is shown in Figure 1-14.
16 Reference [15] uses discrete evolutionary (DE) optimization technique to find the B-H characteristics of the material in a synchronous generator. In this method, permeability, current density, and bend adjustment co-efficient are optimized using DE optimization technique to reduce the error in the estimated no load voltages. While this method was very accurate in predicting the open circuit characteristics of the generator, integrating such optimization procedures with commercially available software would require considerable effort.
1.2.5 Comparison of Epstein and SST techniques
S.No Epstein Tester SST (82)
Acceptance Good Fair
Reproducibility Good Poor
Simplicity Good Poor
Applicability Poor Good
Calibration facility Poor Poor
Technique Invasive Invasive
Table 1-1. Comparison of Epstein and SST [16]
1.3 Motivation of the present research work
From the above techniques for magnetic measurements, it can be clearly seen that the existing techniques have limitations such as the requirement of specimen material, invasive technique and destruction of the tested specimens. Therefore a non-invasive method without the requirement of an additional specimen material needs to be developed for machines already in operation.
In this research work, an attempt has been made to develop a technique which is non-invasive and involves simple calculations to determine the BH (Magnetic flux density-Magnetic field intensity) characteristics of the material used in salient pole synchronous machine (SPSM). The stator of the real SPSM used for testing is made of M27 and the rotor is made of steel whose BH characteristic is unknown. FE modeling has been done using Ansys Maxwell for the SPSM. Experimental open circuit test results were compared with those obtained from the FE simulations and a novel ampere turn distribution technique was developed using a known magnetic material with characteristics expected to be similar to the actual one in order to match the BH characteristics of
17 the actual material used in the machine. Following this, simulations and experiments were performed to compare the simulated characteristics of the electrical machine with the newly determined material to the experimental motor and generator characteristics.
To the best knowledge of the author, ampere turn distribution technique is a completely new method to determine the BH characteristics of a material used in SPSM.
1.4 Thesis outline
The structure of the thesis is as follows:
In chapter 1, a discussion on the existing techniques to measure the magnetic properties such as Epstein tester and single sheet tester have been presented. This was followed by a comparison of the existing techniques and their shortcomings. Next, a survey on the latest techniques to determine the magnetic properties has been provided. Finally, the motivation of the thesis has been presented to stress the need for a non-invasive technique for the determination of the magnetic properties of the material in rotating machines.
In chapter 2, discussions have been provided on the FE software package, Ansys Maxwell. The advantages, disadvantages and the tools available in Ansys Maxwell have been presented in a detailed manner. The tools available in creating models for magnetostatic measurements have been provided.
In chapter 3, theoretical calculation in determination of the magnetic characteristics of the material has been provided followed by the modeling of the SPSM machine using Ansys Maxwell. Details have been provided on the creation of SPSM models using Ansys Maxwell for different conditions such as the motor and generator and for different loads. Further the ampere turn distribution technique to determine the magnetic characteristics has been described in detail. In chapter 4, simulations have been performed under various conditions for SPSM to test the newly determined material and to compare it with the experimental results. Open circuit test, motor and generator tests were performed at different load conditions and different field currents to give a comprehensive analysis on the accuracy of the newly derived material characteristics using the ampere turn distribution technique.
18 In chapter 5, the advantages and shortcomings of this ampere turn distribution scheme have been provided. Finally, the contributions of the research, future scope and conclusion have been discussed.
19
Chapter 2
Finite Element Modeling
2.1 Electromagnetic analysis
Solving electromagnetic problems has always been a difficult task because of the complications associated with varying and complex geometric shapes, materials along with complex mathematical operations involved in their solutions. Some of the techniques available for solving electromagnetic problems are shown in Figure 2-1.
Figure 2-1. Electromagnetic Analysis Solution [17]
Of all the techniques available for solving the electromagnetic problems, finite element method has emerged as one of the robust methods for the analysis and finding the best possible solution.
20
2.2 Finite Element Analysis (FEA):
FEA is a numerical method for solving multiphysics problems. It is usually employed for problems with complicated geometries, loading and material properties. This method is usually used where an analytical solution may be difficult to handle and other modeling methods do not give accurate results. The model to be solved is first defined geometrically into smaller bodies interconnected by simple boundary lines or surfaces. These smaller bodies are solved using simpler equations and then followed by a calculus of variations method to minimize an associated error function.
Some of the advantages of FEA are:
Complex geometry can be included without difficulty and solved.
Different materials can be used with different parts of the geometry.
Local effects can be captured as the whole model is subdivided into simpler bodies
Total solution can be represented at the end for this complex geometry. Some of the disadvantages of FEA are:
Only an approximate solution is provided
FE method provides element dependent solution i.e. for irregular shaped elements, accuracy of the solution is less.
Using high quality numerical methods is restricted due to large number of meshes to be solved resulting in a very long solution time.
Applications of Finite Element method include:
Mechanical/Aerospace/Civil/Automobile
Structural Analysis
Electromagnetic problems
Thermal analysis/Fluid dynamics
21
2.3 FE in Electrical Machines
FEA is widely used for electromagnetic problems. It is widely used in the modeling of electrical machines as it helps in understanding the characteristics of the machine under different conditions without the need for an actual motor with reasonable accuracy. Furthermore from the advantages shown above for FEA, it can be clearly seen that FEA is a very useful scheme for electrical machine modeling as the machines exhibit complexities in their geometry and the usage of different materials in their construction. It is therefore clear that FEA is an important tool for designers of electrical machines for achieving low cost, high efficiency, reliability and minimum weight by optimizing the performance of the machine through proper design. Since FE simulation gives realistic results, different working conditions can be tested such as running the electrical machines as a:
1. Motor 2. Generator
3. OC/Short Circuit test
All these different characteristics can be studied under different loading conditions and different field currents. Also, with new types of electrical machines being developed, it becomes highly important to analyze these machines extensively to study their characteristics and functioning [18]. Since FE method is computationally intensive, the ability of the computers to run multiple processes followed by high CPU power becomes a necessity. Current availability of inexpensive multiprocessor capability, high CPU power and memory makes FE method a feasible method for the study of electrical machines.
In addition, different fault conditions can be incorporated easily to study the effects of the faults on the input and output characteristics of the machine. This has been suitably proved in [19]. Hence FE modeling plays an important role in the study of electrical machines.
22
2.4 Commercial FE packages available for Electrical
Machines:
Some of the commonly used software packages in electrical machine modeling are Ansys Maxwell, Comsol, Modelica, Flux and Speed. The field of application of these software packages is given in Table 2-1.
S.No Software License Field of application
1 Ansys Proprietary Ansys provides solution for different areas such as
Electronics, Electromagnetics, Multiphysics, Fluids and structures
2 Comsol Proprietary Comsol is widely used for MEMS, CFD and structural
analysis
3 MagNet Proprietary MagNet is a product of Infolytica corporation which is
dedicated for solving Electrical machine problems.
4 Flux Proprietary Flux is designed for electromagnetic and thermal
analysis. It is used for electromagnetic devices such as electrical machines, Transformers, HV devices, cables
and induction devices
5 Modelica Non Proprietary Modelica is used for solving electromagnetic problems by differential, algebraic and discrete equations
6 MEGA Proprietary MEGA was developed for 2D and 3D finite element
analysis of electromagnetic fields. It was designed by a group in the University of Bath
7 SPEED Proprietary It is a dedicated software for electromagnetic analysis of Electrical Motors
Table 2-1. Electrical machine design software packages
There are several other in house systems and open source software packages developed for electrical machines but the table above gives the widely used and popular software packages in the field of electrical machines.
23 1. Preprocessing (Model the required geometry, assign material, boundary conditions, loads
and constraints)
2. FEA solver (This is the work of the software package which assembles and solves the differential equations)
3. Postprocessing (Sort and display the results) These steps are shown in Figure 2-2
Figure 2-2. FE software process [18]
All the work in this thesis has been done using different products of Ansys software package such as Maxwell and RMxprt. The details of this software and its capabilities are discussed in this chapter. Some of the reasons for choosing Ansys software package are the following
1. Ability to perform magnetostatic, electrostatic and transient simulations 2. Ability to provide inbuilt electrical machine modeling packages like RMxprt 3. Ability to scale the problem to different ranges based on the needs.
2.5 Ansys:
Ansys Maxwell is an electromagnetic field simulation software used for 2D and 3D electromagnetic devices such as motors and transformers. It uses Finite element method to solve static, frequency domain and time varying electromagnetic and electric fields. This plays a major role in using Ansys for electrical machine design and analysis.
24 The process flow of the modeling in Ansys Maxwell is shown in Figure 2-3 [20].
Some of the advantages of Ansys Maxwell are 1. Automatic adaptive meshing
2. Dynamic link with Simplorer and Maxwell circuit editor 3. Transient in motion
4. Permanent magnet temperature dependency However some of the disadvantages are:
1. It takes a long time to obtain a steady state solution. It takes around 13 hours with an intel i7 core @3.40GHz, Windows 7 OS and 16GB RAM for 2s simulation for a 3 phase, 4 pole, 60Hz, 2KW SPSM.
25
Figure 2-4. Ansys Maxwell and related products [20]
Figure 2-4 shows the Ansys Maxwell and associated products available to the users for modeling of electromagnetic devices and machines. Some of the products can be interlinked as shown in Figure 2-4. For example, a model of electrical machine can be drawn in Ansys Maxwell and can be linked with Ansys PExprt for control of the machine since PExprt provides the option of power electronics such as bridges and converters. It can be linked to Simplorer to include a field circuit for powering the field windings/stator windings of the electrical machine. In this thesis, Maxwell circuit editor has been used for powering the field windings and stator windings of the SPSM. The following sections will discuss the important features in Ansys Maxwell which are useful in electrical machine modeling
2.5.1 Ansys RMxprt
Ansys RMxprt is a template based design tool for Electrical machines in Ansys Maxwell domain. It has templates for induction motor (single and three phase), synchronous motor (permanent magnet, salient and round rotor), DC motors (brushless and permanent magnet) and switched
26 reluctance motors. In this template, the outer diameter, inner diameter of the stator, rotor, length, height of yoke, teeth, number of slots, winding type, number of conductors and type of the material used have to be filled. Once these data are inserted into the model, they have to be converted to Maxwell 2D design and run as a Finite Element model. This is a very useful tool since it saves a lot of time in modeling or drawing the complete motor step by step. Figure 2-5 shows RMxprt modeling of PMDC motor.
RMxprt is particularly useful for designers of electrical machines since it gives the ability to optimize the machine to identify the best design. RMxprt can be used by clicking on the Project tab Insert RMxprt design and by selecting the appropriate machine model. Once a particular electrical machine is selected from the RMxprt template, the tool gives the general stator and rotor models. Here suitable details of its size, material and other parameters such as slots and teeth are filled and tested for its performance. Besides, there is also parametric and optimization capability built with the tool which automatically varies parameters within the template such as the diameter of the rotor or the length of the stator [21]. Using this tool, the designers can view the performance curves and choose the best design. This shows the capabilities of Ansys RMxprt in the modeling of electrical machines.
27
2.5.2 Field Calculator:
Field calculator is another feature available in Ansys Maxwell which enables the users to build and postprocess the solutions obtained from the FE simulation. It has various operations attached to it such as vector operations, calculus operations, and algebraic operations. It can be applied over specific geometric shapes to perform field calculations, integrations and for exporting the determined results. Once the magnetostatic simulation is over, field calculator can be accessed from the results tab of the project. Figure 2-6 shows the Field calculator tool available in Ansys Maxwell.
Field Calculator is particularly useful while performing magnetostatic simulations. The three main purposes of field calculator are as follows:
1. Plot field quantities (Magnetic flux density - B, Magnetic field intensity - H, Current density - J) over different geometric entities in the Finite Element model
2. Perform integration (Line, surface, volume) over the geometric entities. 3. Export the field result over a specific location or a point.
28 Field Calculator is useful for finding the maximum and minimum value as well as the position of magnetic field intensity, magnetic flux density in a given region. The steps to find the maximum value and position are shown below.
To get a maximum value of Magnetic flux density (B) in a given volume: Input > Quantity >B
Vector > Mag
Input > Geometry > Volume (Volume of interest) Scalar > Max> Value
Output > Eval
To get the position of maximum value of Magnetic flux density (B) in a given volume: Input > Quantity >B
Vector > Mag
Input > Geometry > Volume (Volume of interest) Scalar > Max> Position
Output > Eval
The same procedure can be used to find the minimum value and position of B, H and J using field calculator. There is no direct method available in Ansys Maxwell to determine the average of B or H in a given region.
2.5.3 Solvers
There are several solvers available in Ansys Maxwell. They are as follows: 1. Magnetostatic
2. Eddy Current 3. Transient
These solvers can be accessed by clicking on tab Maxwell 2d Solution type. This gives the three solver types and any one of them can be chosen according to the requirement.
29
2.5.3.1
Magnetostatic Solvers
To perform magnetostatic analysis, magnetostatic solvers have to be used. They are usually used in inductors, motors, solenoids, actuators and many others and are used for objects that are stationary. The quantities which are computed through Magnetostatic solvers are magnetic field intensity (H), magnetic flux density (B) and current density (J). Figure 2-7 shows the steps involved in magnetostatic solution processes.
Figure 2-7. Magnetostatic Solution Process [17]
2.5.3.2
Eddy Current Solver
To perform magnetostatic analysis, eddy current solvers have to be used. They are usually used in inductors, motors, solenoids, stray field calculations and many others and are used for objects that are stationary. The quantities which are computed through eddy current solvers are magnetic field and magnetic scalar potential. The eddy current solvers are used for steady state, AC
30 magnetic fields at a given frequency. Figure 2-8 shows the steps involved in eddy current solution processes.
Figure 2-8. Eddy current solution process [17]
2.5.3.3
Transient Solver
To perform transient analysis, transient solvers have to be used. They are usually used in inductors, motors, solenoids, permanent magnets and many others and are used for objects that are moving. Figure 2-9 shows the steps involved in transient solution process.
31
Figure 2-9. Transient Solution process [17]
2.5.4 Material Properties
Ansys Maxwell has options for introducing the effect of core loss into the material characteristics using features available in the material properties tab. Figure 2-10 shows the options available in Ansys Maxwell for introducing different material characteristics. A specific BH curve of a material can be added by using the relative permeability tab and by importing the appropriate BH curve of the material. This feature is extensively used in this thesis to introduce different BH curves of material and testing the characteristics of the SPSM. Following this, there is also an option for introducing core loss model into the material. This can be achieved by adding a type of core loss model. They are of three types – electrical steel, power ferrite or hysteresis model. Electrical steel core loss model for M27 material is shown in Figure 2-11.
32 Other options available in Ansys Maxwell in the material properties are Bulk conductivity, Thermal expansion, Young’s Modulus, specific heat and many more. These properties can be utilized based on the need and the accuracy required for the FE model and the application.
33
Figure 2-11. M27 Core loss Model (Red curve – actual one and Black curve – inserted automatically by Maxwell to smoothen the characteristics)
2.6 Conclusion
This chapter has discussed in detail the need for FE modeling of Electrical machines. It has also provided comprehensive detail on the Ansys Maxwell software discussing mainly on the features such as RMxprt, solvers and core loss modeling necessary for the modeling of electrical machines. Other minor features involved in modeling the SPSM will be discussed in Chapter 3.
34
Chapter 3
Ampere Turn Distribution Scheme
The first part of the chapter deals with the magnetic materials commonly used in salient pole synchronous machine (SPSM). Following this, the modeling of SPSM as a motor and a generator using Ansys Maxwell has been discussed. Subsequently the novel scheme developed to determine the magnetic material used in SPSM has been provided. Furthermore comparison of the open circuit characteristics has been made with different magnetic materials and the experimental motor. Thus a complete overview of modeling of electrical machines using Ansys Maxwell and the ampere turn distribution scheme for determining the magnetic material in SPSM is provided in this chapter.
3.1 Magnetic Materials used in Electrical Machines
Silicon steel is the commonly used magnetic material for the electrical machines. It is widely used in transformers, inductors and electrical machines. One of the drawbacks of using steel is that with aging, the losses increase. The addition of silicon to the steel increases the electrical resistivity thereby reduces the eddy current losses and also improves the material stability and age. Silicon steel offers high saturation flux density, good permeability at high flux density and moderate losses compared to only steel.
Silicon steel contains iron and 0.3 – 4.5% silicon. The percentage of silicon varies with different grades of silicon steel. Usually low content of silicon is used for electrical machines and high content of silicon (4-5%) is used for transformers. This silicon steel is graded based on core losses and are classified into two main categories. They are
1. Non-oriented steel 2. Grain-oriented steel
3.1.1 Non-oriented Steel
Non oriented steel produces the same magnetic domains irrespective of the direction of magnetization in the plane of the material [23]. The name non oriented distinguishes it from oriented steel to show that the magnetic properties are the same in all directions.
35
3.1.2 Grain-oriented Steel
Grain oriented steel produces magnetic domains which depends on the direction of rolling [23]. The process of rolling and annealing used during the production of steel can be used to create steel with superior magnetic properties in one direction and inferior properties in another.
Silicon steel is split into different grades based on the core losses associated with it. The core losses will vary according to the content of Silicon in the Silicon steel. With increase in silicon, core loss decreases but it also lowers the induction permeability of the material.
In general silicon steel consists of iron mixed with silicon (Si), carbon (C), manganese (Mn), phosphorous (P) and sulphur(S). The composition of different types of silicon steel is shown in Table 3-1.
Composition, %
Description of Material C Mn P S Si
Low Silicon Steel 0.003 0.5 0.03 0.001 1.6
Medium Silicon Steel 0.003 0.15 0.01 0.001 2.0
High Silicon Steel 0.003 0.15 0.01 0.001 2.7
Grain Oriented Silicon Steel 0.003 0.07 0.01 0.001 3.1
Table 3-1. Composition of silicon steel [23]
The experimental SPSM used in the testing and experiments as a motor and generator contains non-oriented steel in its stator. The stator of the SPSM is made up of M27 grade steel and the material of the rotor is unknown. M27 contains 3% silicon and remaining is mainly iron, apart from trace quantities of C, Mn, S and P. M27 belong to the group of non-oriented steel. Since the rotor material in SPSM is unknown, the novel scheme discussed in section 3.3 provides a suitable solution to determine the BH characteristics of the material in the stator and rotor of the SPSM.
3.2 Finite element modeling of Salient Pole Synchronous
Machine
As discussed in Chapter 2 on the capabilities of Ansys Maxwell, a salient pole synchronous machine (SPSM) model was drawn in Maxwell as shown in Figure 3-1. The model was drawn
36 according to the specifications given by the manufacturer. The name plate details, dimensions of the machine, wiring and connection diagram of the SPSM are provided in the appendix of this thesis. FE model of SPSM shown in Figure 3-1 can be used for the different purposes such as
1. Motor 2. Generator
3. Short/Open circuit test
These different functions can be achieved by incorporating suitable changes in the excitation and motion setup of the model. The modifications required to achieve these different conditions will be discussed in detail in this chapter.
Figure 3-1. Salient pole synchronous machine – FE model
3.2.1 Salient Pole Synchronous Motor
The model shown in Figure 3-1 is used for running the SPSM as a motor. In a motor, supply has to be given to the stator windings. Therefore in the excitation setup of the FE model, three phase ac supply of 120V RMS was provided directly to the windings. Figure 3-2 and Figure 3-3 show the settings given in Ansys Maxwell for the stator windings of the motor. Three phase AC supply is provided by supplying the voltage directly in the voltage tab with a phase shift of 120 degrees. The resistance and the inductance for the stator windings are given from the motor datasheet.
37 Type of the winding in the SPSM is stranded conductors and it can be selected in the type tab. In addition, the stator resistances and inductance can be included in the windings to create a realistic model of the stator windings. These values were given based on the manufacturer’s data. The current to the field windings can be provided using the circuit shown in Figure 3-4. The resistance and the inductance shown in Figure 3-4 are the field circuit parameters that need to be entered separately. The gradual buildup of field current is shown in Figure 3-5.
Figure 3-2. Excitation voltage - winding 1
38
Figure 3-4. Field circuit for motor
Figure 3-5. Field current
0 0.5 1 1.5 2 0 0.2 0.4 0.6 0.8 1 1.2 1.4 Time (s) C u rr e n t (A )
39
Figure 3-6. Load setting
To vary the load to different conditions, the motion set up tab is modified. This is shown in Figure 3-6. Different loads can be set up by modifying the load torque to different values. The full load is 10.6Nm (the negative sign in Figure 3-6 indicates that the externally applied load torque) and therefore different loading conditions can be given directly based on the percentage of this full load torque. The comparison of the simulation and the experimental results of SPSM with the newly determined material and real motor for different field currents and different loading conditions are shown in detail in Chapter 4.
3.2.2 Salient pole synchronous generator
The model shown in Figure 3-1 is used for running the SPSM as a generator. In a generator, the machine has to be rotated at a constant speed (synchronous speed) using a prime mover. To achieve this condition, in the motion setup of the FE model, the mechanical transients are removed and only a constant speed that is required can be given for the model. This way the model is made to rotate at a required constant speed. This is shown in Figure 3-7. Figure 3-8
40 shows the constant speed of 1800 RPM given to the generator which is the synchronous speed required to produce 60Hz sinusoidal three phase voltages. Since no voltage has to be given to the stator windings, they are left alone and can be modeled in the Maxwell circuit editor to incorporate loading of the generator. This condition is shown in Figure 3-9 where the parameters of the stator winding (winding resistance and inductance) are entered in the Maxwell circuit editor. The comparison of the simulation FE model and the experimental generator at different loading conditions, different field currents and different loading circuits is shown in detail in Chapter 4.
41
Figure 3-8. Constant speed of 1800 RPM
42
3.3 AT (ampere turn) Distribution Scheme
The flowchart in Figure 3-10 shows the step-by-step procedure of using AT (ampere turn) distribution scheme to determine the magnetic characteristics of the material used in the Salient pole synchronous motor. Ampere turn can also be expressed as magnetomotive force and can determined by multiplying the turns in the winding and the amount of current flowing in a given winding. This scheme is a novel non-destructive technique for identifying the magnetic characteristics without using any sample material and thus can be utilized by the already installed machines. FE modeling of the machine followed by performing experimental tests on the SPSM helps in determining the magnetic characteristics of the actual machine.
3.3.1 Steps of AT distribution scheme
1. Perform experimental open circuit test with the real motor for different field currents (0.1- 1.5A).
2. Perform open circuit test with the FE model of SPSM for different field currents (0.1- 1.5A) with steel1008 material –material with known BH characteristics.
3. Magnetostatic FE simulation with a known material- steel 1008 to determine the B (Magnetic flux density) and H (magnetic field intensity) in different parts of the iron in the SPSM for the different field current settings
4. Calculate AT and gap contraction factor, Kg using the open circuit test voltage from the
simulation results.
5. Use the experimental open circuit voltage and find AT for the air gap using the Kg
obtained from the simulations.
6. Compare the simulated AT for different parts of the machine with the experimental ampere turns and split the experimental ampere turns according to the same ratio. This gives new ampere turns for different parts of the motor. Repeat this process for all 15 field currents
7. Using the calculated AT, determine the magnetic field strength for the experimental motor and determine the new magnetic characteristics of the material.
The detailed procedure of the ampere turn distribution technique is discussed step by step in the following sections
43
Figure 3-10. Steps in AT (ampere turn) distribution scheme
3.3.2 Open circuit test
To determine the magnetic properties of the material, an open circuit test for fifteen different field currents from 0.1A to 1.5A, in steps of 0.1 A, was first done on the actual SPSM to obtain the experimental open circuit characteristics (OCC) of the machine. This was followed by performing OCC test with FE model using steel 1008 as the initial known material. An external circuit was used to give appropriate field current of 0.1 – 1.5A as required. Appropriate voltmeters were connected to the stator winding to measure the generated voltage. The simulation was run for two seconds and the rms value of the simulated phase voltage was noted for each field current. The two OCCs, one obtained from experiment and the other from the transient simulation, showed considerable difference in the generated voltages as shown in Table 3-2. Thus it shows that using steel1008 will not be very meaningful in simulating the SPSM. However, it was conjectured that for identical geometry, the magnetic field strength, H, required to set up a given flux density in the different parts of the iron will be proportional for different
