France



&lt;br/&gt;1 Introduction&lt;br/&gt;In aviation field, direct drive permanent magnet (PM) rectifier generator is often used in Multi Electric Airplane (MEA). Multiphase AC windings and rectifier technology make DC generator an excellent DC source with low torque ripple and low voltage ripple. And by regulating the engine speed within specified range, the DC-link voltage can be maintained constant with various electric loading. The key point of designing such a kind of generator, is to predict the electromagnetic and thermal field accurately.&lt;br/&gt;To solve the problem above, most research focused on electromagnetic thermal coupling analysis based on computational fluid dynamics (CFD), which can accurately predict temperature distribution [1]. Obviously, CFD-based coupling analysis method is not fit for optimization of direct drive PM rectifier generator for CFD process is time-consuming.&lt;br/&gt;In this paper, we aim to adopt lumped parameter thermal network (LPTN) as a substitute to replace the CFD process, implementing an electrical circuit-electromagnetic field-thermal network coupling analysis algorithm, enabling a much faster solution to analyze and optimize a PM rectifier generator.&lt;br/&gt;2 Coupling Analysis&lt;br/&gt;2.1 Electrical Circuit and Electromagnetic Field Coupling Analysis&lt;br/&gt;Electrical circuit and electromagnetic field are coupled by electromotive force(EMF). Electromagnetic field analysis takes winding current as the input, and outputs EMF. Electrical circuit analysis takes EMF as the input, and outputs winding current. Electromagnetic field is usually solved by 2D time stepping FEM method. Electrical circuit is solved by branch current method. The coupling relation is so strong that magnetic vector potential and winding current can only be analyzed simultaneously.&lt;br/&gt;2.2 Electromagnetic Field and Thermal Network Coupling Analysis&lt;br/&gt;Electromagnetic field analysis takes winding current as input, and outputs iron loss, copper loss, solid loss, etc. Thermal networks analysis takes loss as input, and output temperature of generator parts. The above is just one way coupling (designated as unidirectional) from electromagnetic field to thermal network, which is carried out by transferring and mapping loss data.&lt;br/&gt;In this paper, we have also considered the coupling from thermal network to electromagnetic field, involving material property, which will be described in Section 2.3. Here this is called bidirectional coupling method.&lt;br/&gt;As for the LPTN model of generator, thermal resistance of each part refers to Mellor’s T-shaped model [2]. The structure of each part is equivalent to a hollow cylindrical structure. Conduction, natural and forced convection are considered in the LPTN model. Model parameters and assumptions will be detailed in the full paper.&lt;br/&gt;2.3 Material Property&lt;br/&gt;The temperature dependent material property is measured in the lab. The result is as follows:&lt;br/&gt;(1)For permanent magnet, coercivity and remanence decrease with temperature.&lt;br/&gt;(2)For silicon steel, the iron loss decreases with temperature.&lt;br/&gt;(3)For copper and permanent magnet, the resistivity increase with temperature.&lt;br/&gt;2.4 Implementation&lt;br/&gt;Thermal network of generator is modelled and solved in Matlab, and electrical circuit and electromagnetic field are solved in ANSYS/Maxwell. Because the electromagnetic field and the temperature field have different time constant, during each iteration, several periods of electromagnetic field-circuit calculation are need to reach a steady state, followed by LPTN calculation. A control program is coded in Matlab to call thermal solver and Maxwell solver, and to transfer and map loss and temperature dependent material property. The convergence criterion is met if the temperature difference between adjacent iterations is less than 1°C.&lt;br/&gt;2.5 Experiment Validation&lt;br/&gt;This prototype machine is an inner rotor PM generator with an integrated fan on the same shaft as rotor. The designed output power is 4.5kW at 7250rpm, and the DC link voltage is 50V. The generator is of 10 poles and 12 slots, filled with six phase double layer winding. The surface mounted PM is sintered SmCo30H.&lt;br/&gt;After 5 iterations, the coupled field analysis quickly reached convergence. The PM and winding temperature at each iteration and the corresponding loss is shown in Fig 2 .&lt;br/&gt;As shown in Fig 1, a load test has been carried out to validate the accuracy of the bidirectional coupling method. The servo motor drives the prototype at a constant speed and torque meter measures the mechanical torque. The rectifier circuit is connected to the terminal of the prototype. Two thermocouples are embedded in the end windings.&lt;br/&gt;Fig 2 shows the measured and simulated results. To illustrate the effect of temperature on electromagnetic field, the simulated result by unidirectional method is provided where the initial temperature is set to 25°C. Due to the consideration of thermal effect on material properties, the simulated result by bidirectional method is more accurate than unidirectional method. The prediction error of the bidirectional coupling method is within 5%.&lt;br/&gt;This paper proposed a much faster method to predict the electromagnetic and thermal field of direct drive PM rectifier generator. Lumped parameter thermal network (LPTN) is used as a substitute to replace the CFD process, greatly shortens the solution time. Furthermore, temperature effect on material property is considered, enabling a bidirectional coupling analysis in the paper. The experimental validation shows that the proposed method is enough accurate. The method can be applied to analyze and optimize a PM rectifier generator, as well as other typed motor.

**References:**

&lt;br/&gt;[1] Yapeng Jiang , Dong Wang, Junquan Chen, &quot;Electromagnetic-Thermal-Fluidic Analysis of Permanent Magnet Synchronous Machine by Bidirectional Method,&quot; &lt;i&gt;IEEE Transactions on Magnetics&lt;/i&gt;, vol.3, no.54, pp. 1-5, 2018.&lt;br/&gt;[2] Mellor P.H, Roberts D, Turner D.R, &quot;Lumped parameter thermal model for electrical machines of TEFC design,&quot;&lt;i&gt; IEEE PROCEEDINGS-B&lt;/i&gt;, vol.138, no.5, pp.205-218, 1991.

**Images:**
![](https://assets.underline.io/uploads/markdown_image/1/image/55f231196743cfde6ee8b35ae9c9d637.jpg)

![](https://assets.underline.io/uploads/markdown_image/1/image/9e3d5b7f8437466378caee97ef19984c.jpg)




INTERMAG 2021

 Coupled Electromagnetic and Thermal Analysis of Permanent Magnet Rectifier Generator Based on LPTN

INTERMAG 2021 Welcome Video



	
<figure class="video_container">
 <iframe src="https://screencast-o-matic.com/player/crflqFVnUpZ?v=6&ff=1&title=0&controls=1" width=640 height=360 frameborder="0" allowfullscreen="true"> </iframe>
</figure>

Want to buy a ticket? 

Register for INTERMAG 2021 here

[![](https://assets.underline.io/uploads/markdown_image/1/image/17ce76e7960db633258024eef8d6da7f.png)](https://intermag.org/register)

Got your ticket, but still can’t access the event?

Contact us and our team will look into it.

1 Introduction In aviation field, direct drive permanent magnet (PM) rectifier generator is often used in Multi Electric Airplane (MEA). Multiphase AC windings and rectifier technology make DC generator an excellent DC source with low torque ripple and low voltage ripple. And by regulating the engine speed within specified range, the DC-link voltage can be maintained constant with various electric loading. The key point of designing such a kind of generator, is to predict the electromagnetic and thermal field accurately. To solve the problem above, most research focused on electromagnetic thermal coupling analysis based on computational fluid dynamics (CFD), which can accurately predict temperature distribution [1]. Obviously, CFD-based coupling analysis method is not fit for optimization of direct drive PM rectifier generator for CFD process is time-consuming. In this paper, we aim to adopt lumped parameter thermal network (LPTN) as a substitute to replace the CFD process, implementing an electrical circuit-electromagnetic field-thermal network coupling analysis algorithm, enabling a much faster solution to analyze and optimize a PM rectifier generator. 2 Coupling Analysis 2.1 Electrical Circuit and Electromagnetic Field Coupling Analysis Electrical circuit and electromagnetic field are coupled by electromotive force(EMF). Electromagnetic field analysis takes winding current as the input, and outputs EMF. Electrical circuit analysis takes EMF as the input, and outputs winding current. Electromagnetic field is usually solved by 2D time stepping FEM method. Electrical circuit is solved by branch current method. The coupling relation is so strong that magnetic vector potential and winding current can only be analyzed simultaneously. 2.2 Electromagnetic Field and Thermal Network Coupling Analysis Electromagnetic field analysis takes winding current as input, and outputs iron loss, copper loss, solid loss, etc. Thermal networks analysis takes loss as input, and output temperature of generator parts. The above is just one way coupling (designated as unidirectional) from electromagnetic field to thermal network, which is carried out by transferring and mapping loss data. In this paper, we have also considered the coupling from thermal network to electromagnetic field, involving material property, which will be described in Section 2.3. Here this is called bidirectional coupling method. As for the LPTN model of generator, thermal resistance of each part refers to Mellor’s T-shaped model [2]. The structure of each part is equivalent to a hollow cylindrical structure. Conduction, natural and forced convection are considered in the LPTN model. Model parameters and assumptions will be detailed in the full paper. 2.3 Material Property The temperature dependent material property is measured in the lab. The result is as follows: (1)For permanent magnet, coercivity and remanence decrease with temperature. (2)For silicon steel, the iron loss decreases with temperature. (3)For copper and permanent magnet, the resistivity increase with temperature. 2.4 Implementation Thermal network of generator is modelled and solved in Matlab, and electrical circuit and electromagnetic field are solved in ANSYS/Maxwell. Because the electromagnetic field and the temperature field have different time constant, during each iteration, several periods of electromagnetic field-circuit calculation are need to reach a steady state, followed by LPTN calculation. A control program is coded in Matlab to call thermal solver and Maxwell solver, and to transfer and map loss and temperature dependent material property. The convergence criterion is met if the temperature difference between adjacent iterations is less than 1°C. 2.5 Experiment Validation This prototype machine is an inner rotor PM generator with an integrated fan on the same shaft as rotor. The designed output power is 4.5kW at 7250rpm, and the DC link voltage is 50V. The generator is of 10 poles and 12 slots, filled with six phase double layer winding. The surface mounted PM is sintered SmCo30H. After 5 iterations, the coupled field analysis quickly reached convergence. The PM and winding temperature at each iteration and the corresponding loss is shown in Fig 2 . As shown in Fig 1, a load test has been carried out to validate the accuracy of the bidirectional coupling method. The servo motor drives the prototype at a constant speed and torque meter measures the mechanical torque. The rectifier circuit is connected to the terminal of the prototype. Two thermocouples are embedded in the end windings. Fig 2 shows the measured and simulated results. To illustrate the effect of temperature on electromagnetic field, the simulated result by unidirectional method is provided where the initial temperature is set to 25°C. Due to the consideration of thermal effect on material properties, the simulated result by bidirectional method is more accurate than unidirectional method. The prediction error of the bidirectional coupling method is within 5%. This paper proposed a much faster method to predict the electromagnetic and thermal field of direct drive PM rectifier generator. Lumped parameter thermal network (LPTN) is used as a substitute to replace the CFD process, greatly shortens the solution time. Furthermore, temperature effect on material property is considered, enabling a bidirectional coupling analysis in the paper. The experimental validation shows that the proposed method is enough accurate. The method can be applied to analyze and optimize a PM rectifier generator, as well as other typed motor.

**References:**

 [1] Yapeng Jiang , Dong Wang, Junquan Chen, "Electromagnetic-Thermal-Fluidic Analysis of Permanent Magnet Synchronous Machine by Bidirectional Method," IEEE Transactions on Magnetics, vol.3, no.54, pp. 1-5, 2018. [2] Mellor P.H, Roberts D, Turner D.R, "Lumped parameter thermal model for electrical machines of TEFC design," IEEE PROCEEDINGS-B, vol.138, no.5, pp.205-218, 1991.

**Images:**
![](https://assets.underline.io/uploads/markdown_image/1/image/55f231196743cfde6ee8b35ae9c9d637.jpg)

![](https://assets.underline.io/uploads/markdown_image/1/image/9e3d5b7f8437466378caee97ef19984c.jpg)

technical paper

I.Introduction The spoke-type permanent magnet (PM) machines are widely used in many industrial fields because of their excellent performance of concentrating flux from the PM. Accurate calculation of the magnetic field is the premise for predicting the machine performances [1]-[3]. The analytical calculation methods of the magnetic field commonly used for spoke-type machines include the lumped magnetic circuit (LMC) model [1], subdomain model [2], and the method combining the two above [3]. The rotor magnetic potential (RMP) model is another analytical method to accurately calculate the magnetic field [4], which has been widely used in interior radial PM machines but has not been applied to spoke-type machines. In this paper, the RMP models of the spoke-type machine are proposed. Furthermore, a method is provided by combining the proposed RMP model and the winding function, which can realize a fast and accurate calculation of the magnetic field in spoke-type machines. The modeling processes of the RMP model proposed in this paper are straightforward, and it has fewer variables to be solved. Moreover, the proposed method does not need to consider the change of the relative position between the stator and rotor caused by rotor rotation. II.The Proposed RMP Model The RMP model is used to reflect the rotor surface magnetic potential distribution. In this paper, the magnetic field is divided into d-axis magnetic field (the magnetic field generated by the interaction of PM and d-axis current component) and q-axis magnetic field (the magnetic field generated by the q-axis current component). When there is only the d-axis magnetic field, it can be seen from Fig. 1(a) that the pole cap region is not saturated. Therefore, the RMP of the pole cap region is equal to that of the corresponding PM surface. Due to symmetry, the magnetic potentials on the left and right PM surfaces are opposite to each other. Hence the zero-crossing point of the RMP is located in the middle of the bridge (point B in Fig. 1(a)). According to the flux lines distribution in the bridge region, the RMP in this region changes linearly to both sides as the distance from point B increases. Therefore, the d-axis RMP model is obtained, as shown in Fig. 1(c). When there is only the q-axis magnetic field, it can be seen from Fig. 1(b) that the flux lines do not pass through the PMs; hence the RMP in the pole cap region is equal to zero. In the bridge region, the flux lines are separated in the middle of the bridge. Since the d-axis magnetic field is larger than the q-axis magnetic field, the influence of the d-axis magnetic field should be taken into account in the q-axis RMP model. Therefore, the q-axis RMP reaches the maximum value in the middle of the bridge (point B in Fig. 1(b)) and decreases to both sides along the flux lines. Therefore, the q-axis RMP model is obtained, as shown in Fig. 1(d). It should be noted that the position of points A and B in Fig. 1(a) and Fig. 1(b) are same. III.The Solution of the Proposed RMP Model The d-axis and q-axis LMC models are established to solve the unknown parameters (Urd1 and Urq1), as shown in Fig. 1(e) and Fig. 1(f). Rau and Rad are nonlinear reluctances corresponding to the upper and lower bridge regions, respectively. Rgd is d-axis air-gap reluctance. Rm is PM reluctance. Fc is PM magnetomotive force (MMF). Fd is the equivalent stator MMF, and its value is Fd=(Fd1+Fd2+Fd3-Fd4-Fd5-Fd6)/3. Fd1-Fd6 are the MMF at the corresponding stator teeth generated by the d-axis current component. Rgq is q-axis air-gap reluctance. Fq is the MMF generated by the q-axis current component. It should be noted that when there is only q-axis magnetic field, the bridge region is not saturated. Therefore, the reluctance Rau is only affected by the d-axis magnetic field. Moreover, there is a saturation overflow phenomenon in the bridge region (the width of the saturation region is wider than the actual bridge width). To reflect this phenomenon, the flux density value of point A in Fig. 1(a) is set as 1.6T (the magnetic potential drop in the core with flux density value less than 1.6T is ignored). Therefore, the length lro of the saturation overflow region (as shown in Fig. 1(a)) changes with different excitation conditions. The value of lro can be solved by using the continuity principle of flux, and then the value of abscissa x1 of point A can be obtained. Since the dq-axis cross-coupling effect has been taken into account in the above models, the total RMP model can be obtained by adding the d-axis and q-axis RMP models directly. Then the magnetic field and related parameters can be calculated by combining winding function with the proposed RMP model. IV.FEA Verification The basic parameters of the machine are listed in Table I. The validity of the proposed method (ANA) is verified by FEA. The comparison results of no-load air-gap flux density and back EMF are shown in Fig. 2 (a) and Table II. Under different current effective value I (20A to 200A) and different current angle φ (0° to -90°), the corresponding results of the air-gap flux density fundamental amplitude (FA) and average torque are shown in Fig. 2 (b) and Fig. 2 (c), respectively. As shown in Fig. 2, the maximum error of air-gap flux density FA does not exceed 8%, and the maximum error of average torque does not exceed 2%.

**References:**

[1] Q. Chen, G. Liu, and W. Zhao, IEEE Transactions on Magnetics., vol. 49(9), pp. 5150–5157 (2013). [2] P. Liang, F. Chai, and Y. Li, IEEE Transactions on Industrial Electronics., vol. 64(5), pp. 3479–3488 (2017). [3] P. Liang, F. Chai, and Y. Bi, Journal of Magnetism and Magnetic Materials., vol. 417, pp. 389-396 (2016). [4] Q. Li, T. Fan, and X. Wen, IEEE Transactions on Magnetics., vol. 49(3), pp. 1193-1201 (2013).

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/3220580be1e7460840ec14da0e2fa0f4.jpg)
Fig. 1. (a) d-axis magnetic field (b) q-axis magnetic field (c) d-axis RMP model (d) q-axis RMP model (e) d-axis LMC model (f) q-axis LMC model

![](https://assets.underline.io/uploads/markdown_image/1/image/b959b7b8f4307f450d5a59f2a751f4a3.jpg)
Fig. 2. The comparison results (a) no-load (b) air-gap flux density at load (c) average torque

 Analytical Calculation for Magnetic Field in Spoke-Type Permanent Magnet Machines Based on a Rotor Magnetic Potential Model

I. INTRODUCTION Coreless axial-flux permanent-magnet synchronous machines (AFPMSMs) have many advantages such as high power density, simple structure, high overload capability, no cogging torque and so on. It has been used more and more widely in many applications, such as flywheel energy storage systems, wind energy power generation, electric vehicles, and ultra-precision positioning and processing areas [1-3]. However, it is difficulty to accurate modeling of AFPMSM as 3D electromagnetic problems, such as the curvature effect and fringing effect. The electromagnetic characteristics, such as air-gap flux density, back electromotive force (EMF), and electromagnetic torque, are critical in motor design. Precise design methods and fast performance estimations for coreless AFPMSM would be necessary. The most accurate method to design the AFIPMM is the 3D finite-element analysis, which is exceedingly time-consuming [4-5]. To solve this problem, a three-dimensional analytical approach, based on the volume integral method and second order vector potential (SOVP), is developed to calculate the coreless AFPMSM’s steady-state and transient performances. II. ANALYTICAL MODEL The 3-D geometric model of the coreless AFPMSM is shown in Fig. 1 (a). This coreless AFPMSM consisting of one stator with ironless winding and two rotors with Halbach magnet array mounted on nonmagnetic supports. Such a Halbach rotor can be modeled using equivalent current sheets, thereby enabling the field at any point in 3-D space to be calculated using the Biot–Savart law [3]. For instance, the axial magnetized magnet can be created by the radial and circumference current filament model, as shown in Fig. 1(b), and then a volume integral approach is proposed to obtain the field for the current sheet shown as in Fig. 1(c) and ironless winding. The calculation of the magnetic field using the integral equations is obtained in very short times thanks to the use of particular volumetric basis elements for which the solution is obtained analytically. The field solution from volume integral approach is then used as the magnetic charge source term for an equivalent magnetic charge model. The replace function of the field for wingding and Halbach rotor were given by Fourier series, thus the program difficulty (not shown here due to space limits) for three-dimensional distribution of the magnetic field will be reduced. The 3-D torque equations are derived using a magnetic charge sheet approach that enables the final torque equation to be expressed by a single equation without the need for a large number of summations. III. RESULTS AND VALIDATION The accuracy of the approach was validated 3D finite element analysis shown as in Fig 1 (d). The flux density calculated with the proposed approach are in good accordance with that obtained by 3D FEM simulation.Currently, we have designed and established a coreless AFPMSM prototype in laboratory as shown in Fig.2. Performance of the AFPMSM such as the flux density and torques are measured with the prototype. The experiment results are validated by FEA and proposed approach results and they will be included in the full paper. IV. CONCLUSIONS A three-dimensional analytical approach based on the volume integral method and magnetic charge sheet for the flux density and torque calculations in a coreless AFPMSM. These results will be compared to the 3D FEA and experiment results in the full paper.

**References:**

[1] Z. Kohari, Z. Nadudvari, L. Szlama, M. Keresztesi, and I. Csaki, IEEE Trans. Appl. Supercond., vol. 21, no. 3,pp. 1497–1501. (2011). [2] T. F. Chan, L. L. Lai, IEEE Trans. Energy Convers., vol. 22, no. 1, pp. 86–94. (2007). [3] N. Taran, V. Rallabandi, G. Heins, and D. M. Ionel, IEEE Trans. Ind. Appl., vol. 54, no. 6, pp. 5907–5917. (2018). [4] Xu L, Xu Y, Gong J. . IEEE Transactions on Magnetics,vol. 54, no. 11, pp. 1-5.(2018). [5] Daghigh A, Javadi H, Torkaman H. .IEEE Transactions on Magnetics,vol.52,no.9.(2016)．

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/9f6c252054a481f964baaafc996e4475.jpg)
Fig.1（a）Structure of coreless AFPMM（b）Equivalent current sheet model for a axial magnet piece ( c）Contour plot for the magnetic field by proposed method (d)Comparison of FEA and calculated results for flux density

![](https://assets.underline.io/uploads/markdown_image/1/image/74e017f083f724d22d670552f199140e.jpg)
Fig.2 Experimental setup to test the coreless AFPMM

 3-D Analytical Modeling and Optimization of Axial Flux Coreless PM Motor

Introduction In recent years, SMPMSM is quietly famous compared with other conventional motors in industrial applications due to higher efficiency, higher reliability, low cost, and smooth torque [1]–[3]. In general, several objective optimizations need to be consider, which required an optimization technique for an electrical motor. This paper deals with an optimal design of a surface-mounted permanent magnet synchronous machine (SMPMSM) with an exact analytical subdomain model by using a genetic algorithm method. To analyze the characteristic of permanent magnet (PM) motors, the classical optimization method, such as the finite element method (FEM), is intensively used. However, FEM has several time problems that require a longer computational time to evaluate the performance of PM motors. This problem can be overcome by using a genetic algorithm (GA) method combined with a subdomain model (SD), which developed an improved performance of SMPMSM, for instance, total harmonic distortion (THDv) and phase back-EMF. In this design, a three-phase 12-slot/8-pole PM motor is established with an exact SD model with radial magnetization patterns. Then, the GA ensemble with SD model to search the optimality of SMPMSM machine design. In the final analysis, the optimal new design of SMPMSM demonstrated by comparing with the initial design that is investigated by FEM. The result of induced back-EMF, and magnetic flux density of optimal design is compared with the initial design to show the advantages of GA optimization method. Analytical Model, Genetic Algorithm and Parameter of optimization The purpose of this analysis is to investigate the optimization design of a three-phase 12-slot/8-pole SMPMSM with a radial magnetization pattern. This optimization problem has been solved by obtains the minimum THDv as a fitness function. In order to process the optimization design of SMPMSM, several parameters need to be optimized i.e magnet pole-pitch, αp and slot-opening angle, boa. The roulette method is used to select the objective function of optimization with 50 maximum generations per iteration. The cross and mutation probability need to be set which requires linear and uniform recombination which is 0.89 and 0.025 respectively. The parameters of 12-slot/8-pole SMPM motor as follows: radius of the rotor, Rr = 23mm, the radius of PM surface Rm = 26mm, the inner radius of the stator Rsb = 42.5mm, axial length, la = 50mm, air-gap length, lg = 1mm, remanence of the magnet, Br, 1.12T, relative coil permeability, µr = 1.05 with the number of pole pairs, Np is 8 and number of stator slots, Ns is 12. The result, Discussion and Conclusion The phase back electromotive force has a significant relationship with flux-linkage because phase back electromotive force is derived from the rate of change of flux-linkage. In Fig. 1. show the waveform of phase back electromotive force against rotor position of initial and optimal design in radial magnetization patterns. The initial design of phase back electromotive force waveform shows more trapezoidal compare with the optimal design which has a sinusoidal waveform. The comparison of total harmonic distortion between the initial and optimal design of SMPMSM are demonstrated in Fig. 2 below. The result from Fig.2. shows the highest total harmonic distortion of 12-slot/8-pole motor is an initial design which is 11.34% compared with the optimal design where 1.3%. The magnitude of harmonic in phase back electromotive force |En| of initial design shows 22.24 V compare to optimal design which is 17.4 V. This shows a good agreement that the optimal motor design gives lowest total harmonic distortion compare with the initial design. The optimizes of SMPMSM motor with the following parameters are given in Table 1.

**References:**

[1] S. Wang et al., “Multilevel optimization for surface mounted PM machine incorporating with FEM,” IEEE Trans. Magn., vol. 45, no. 10, pp. 4700–4703, 2009, doi: 10.1109/TMAG.2009.2024875. [2] K. Sindhya, A. Manninen, K. Miettinen, and J. Pippuri, “Generator Using Interactive,” vol. 64, no. 12, pp. 9776–9783, 2017. [3] H. Li, Z. Chen, and H. Polinder, “Optimization of multibrid permanent-magnet wind generator systems,” IEEE Trans. Energy Convers., vol. 24, no. 1, pp. 82–92, 2009, doi: 10.1109/TEC.2008.2005279.

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/26ad63fb00869f77cdf94d41e4a9e659.jpg)
The comparison of phase back electromotive force between initial and optimal design of 12-slot/8-pole SMPMSM under radial magnetization pattern.

![](https://assets.underline.io/uploads/markdown_image/1/image/c8181ba507618ea87f3870d3996c04bf.jpg)
The comparison of total harmonic distortion between initial and optimal design of 12-slot/8-pole SMPMSM under radial magnetization pattern.

 Genetic algorithm optimal design of SMPMSM using analytical subdomain model

Introduction MnZn ferrites are widely used in power electronics for high frequency transformers and inductors since they offer a good compromise between low core losses, permeability and saturation flux density. For properly designing magnetic components, the effective permeability and the core losses must be determined. Different phenomena, like magnetic hysteresis or eddy-currents, lead to core losses. State-of-the-art loss models are often based on the Steinmetz equation or extensions such as Generalized Steinmetz Equation [1]. These approaches aim to model the overall losses at once using a power law based on three (or more) coefficients. A given set of parameters is only accurate around the operating point used for determining these parameters, for the various phenomena scale differently with frequency and flux density. This results in reduced operating ranges in which the methods are accurate. In contrast, with the Bertotti method [2] hysteresis and classical eddy current losses as well as an extra term, called excess losses, are considered separately. The three terms are calculated as with the Steinmetz model but with specific coefficients for each loss term. Even if using three loss components widens the frequency and flux density ranges, the parameters remain frequency dependent and the excess losses lack a proper physical explanation. Therefore, this work presents a physical approach based on the combination of an electrical model of the material structure with the flux distribution resulting from solving Maxwell equations. The new model is used to explain and estimate the frequency dependency of complex permeability as well as the eddy-current and excess core losses that can be then combined with hysteresis losses predicted by classical methods. Equivalent electrical model of material structure The sintering process leads to a meta-material made of grains that are insulated one from each other. Such a structure reduces significantly eddy-current losses since the effective conductivity is small due to the insulating layer between grains. However, the apparent conductivity and permittivity of ferrite depend on frequency because the grain boundary interface acts like a capacitor (Fig 1.a). Its impedance reduces at higher frequencies what short-circuits the high resistivity of the boundary (cf. [3] for NiZn ferrites and [4] for MnZn ferrite). As proposed in [3], a two series cells model could be used to describe the grain and the boundary (Fig. 1.b) for calculating the overall conductivity and permittivity. Each cell is composed of a parallel conductance and capacitance, representing for one cell the admittance of the grains and for the other cell the admittance of the boundaries. By taking the limit of the admittances when the cross section dA and the length dl converges to zero, one can obtain the apparent conductivity and permittivity. These quantities are then a function of the frequency and of the material parameters, i.e.: conductivities and permittivities of the boundary and of the grain. Model of flux distribution vs. frequency The dielectric and ferromagnetic nature of ferrite results in standing waves in the core when the frequency increases. This affects the flux density distribution and induces eddy-currents. To model those effects, two approaches can be considered: lumped elements circuit [5] or full wave equation solution [6]. The first one assumes that induced currents flow only around the cross section in concentric loops, which is satisfied if only the first mode is excited. A more general approach is to solve the wave equations, as proposed in [6] for toroid and pot cores. The solution given in [6] converges quickly with respect to the number of considered spatial harmonics, reducing the computational cost. Proposed model Based on the electrical model of the material structure and on the solution for the flux distribution, a new core loss model is presented in this paper. This model allows to compute the complex permeance of a toroid core, even if the electrical material parameters of the material change with frequency. The effective complex permeability is derived by dividing the complex permeance by the ratio of the effective cross section over the effective magnetic length of the core. The overall core losses are calculated by adding the losses associated with the imaginary part of the permeance to the hysteresis losses. The latter are obtained from low frequency measurements and are then scaled linearly with frequency (similar to Bertotti model). Experimental validation The proposed approach is validated on a R36 toroid core made of N87. The complex effective permeability and the core losses are measured between 10 kHz and 1 MHz for various flux densities, at room temperature using an IWATSU SY-8219 BH Analyzer. The model is then fitted with complex permeability measurements using MATLAB curve fitting toolbox. Results and discussion As shown in Fig 2, the presented model captures accurately the variations of both the complex permeability and the core losses. Furthermore, it can be concluded that the so-called excess losses are due to additional eddy-current losses due to the excitation of wave modes in the core and to the increasing effective conductivity with frequency.

**References:**

[1] J. Li, T. Abdallah and C.R. Sullivan, Conference Record of the 2001 IEEE Industry Applications Conference. 36th IAS Annual Meeting, IEEE, (2002) [2] G. Bertotti, IEEE Transactions on Magnetics, Vol. 24, p. 621-630 (1988) [3] C. G. Koops, Physical Review, Vol. 83, p. 121-124 (1951) [4] M. Kacki, M.S. Rylko, J.G. Hayes and al., 2020 IEEE Applied Power Electronics Conference and Exposition (APEC) (2020) [5] C.R. Sullivan, PSMA 2019 Magnetics Workshop (2019) [6] W. Hauser, Modellbildung für strukturabhängige Effekte in Ferritkernen (2018)

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/7b6236946e32bac2c3dcdb7b998f4ba2.jpg)
Fig. 1: a) The granular microstructure of Ferrites leads to a capacitive behaviour that can be represented by a 1D model. b) Equivalent two cells model for the permittivity and conductivity of ferrites.

![](https://assets.underline.io/uploads/markdown_image/1/image/07e6234e666009d4a326fcb26db45258.jpg)
Fig. 2: Loss density against frequency for various flux density values.

 New Core Loss Model for Ferrite Cores Based on a Meta-Material Approach

Closed core magnetic circuit components have long been the most difficult circuit component to miniaturize. Loss of cross sectional core area and available space to wrap windings trades off inductance as the component becomes smaller. For both circuit compactness and high frequency performance, there has been interest in the embedding of magnetic core materials within planar substrate (printed circuit board, PCB) and fabricating the windings lithographically [1-5]. This allows precise winding fabrication with control of capacitances, as well as being able to reclaim the circuit board area of the formerly surface mounted package. However, high turn numbers are difficult to achieve using PCB via fabrication methods. In standard PCB via fabrication, via diameters are limited by the aspect ratio of the hole that needs to be drilled. In-situ via fabrication in full 1.6mm thickness boards is limited to about a 1:8 aspect ratio. This translates to a ~200 micron via diameter, with a required via-to-via gap of about the same. This limits the number of turns that can be wrapped around the embedded core. Wire-wound magnetic components can use wire with diameters of 50 microns, and gaps equal to the varnish thickness of a couple of microns; wire-wound components can fit significantly more winding turns than embedded components using PCB via fabrication, and thus will have a much higher inductance for the same sized core. In this digest, an alternative ex-situ via forming method is presented, which uses embedded z-conductive via plugs, to fabricate embedded transformers. The effective aspect ratio is theoretically unlimited, but is 1:40 in the process presented here, competitive with wire-wound components. The general embedding process has six steps (Fig. 1): 1. Mill pocket for via plug, 2. Insert via plug and embed with epoxy, 3. Planarize both sides to reveal plug faces, 4. Mill pocket for magnetic core material, 5. Embed magnetic core with epoxy and planarize, 6. Double sided photolithography to form windings. The ex-situ fabricated via plug is the enabling part of the process; the plug is z-conductive in that it electrically conducts along the plug’s length, but is electrically isolating over the plug’s cross-section. When the plug is embedded in the PCB, the plane of the PCB is considered the xy plane, and the thickness of the PCB is in the z-direction. The major advantage of this process compared to alternative ex-situ via fabrication methods is how there is no precise pattern in the z-conductive plug. The plug, ideally, creates an xy continuum of z-direction conduction. Interlayer electrical connections occur wherever a circuit pattern on the top layer overlaps the circuit pattern on the bottom layer; precise alignment to the via plug is not required. Thus, the resolution of the lithography, not the via fabrication, is what determines the maximum number of turns on the winding. To fabricate the z-conductive plugs, commercial litz wire (660 strand bundle/ 46 American wire gauge strands) is epoxy cast into a bar, then cut into 1.6mm thick segments (“plugs”, fig. 2b and c). The 46 gauge strands effectively create 40µm diameter vias (fig. 2c), with a strand-to-strand gap of a few microns. A racetrack core [1, 6] size of 2.1mm outer diameter, 1.0mm inner diameter, leg length of 6.25mm, and a thickness of 1.0mm (fig. 2) is used for embedding. After step 5 (fig 1), the samples are polished with 1 micron slurry to reduce surface roughness and repair surface smearing of the copper litz wire. The transformer winding design has 14 turns per winding (200 µm lines, fig. 2). The lithography process developed is an etch process, where copper is deposited over the entire sample, after step 5, by e-beam evaporation or wet chemical plating. A photolithography process using microelectronics tools was developed, but is difficult to perform given the bulky board materials and the samples having a surface that was not precisely planar. One non-functioning 14-turn sample was completed using photolithography. As a novel alternative approach, spray paint was applied to a sample as an etch mask, and a 3D laser writer, which is less sensitive to surface non-uniformity, was used to ablate the negative image of the 14-turn winding pattern. The winding pattern was successfully etched using this robust approach, but windings were found to be consistently shorting. To help trace the path of the current shorting through the windings, a magneto-optical garnet was used, where a DC current was applied through the shorting path, and the garnet was placed on the windings and observed through a polarizing light microscope. The garnet indicates optically the magnetic fields emanated by the current, showing the shorting current path (Fig. 2d). To correct the shorting defect, high DC current was applied to thermally decompose the assumed high-resistance shorting anomaly. The approach effectively opened the shorting path, but collateral damage occurred destroying the device. In summary, we present this work as innovative both in approach and diagnostic technique. Support for this project has ended and work will not continue, but others working with the same miniature, high-frequency, high-inductance application requirements should consider this approach.

**References:**

References:: [1] D. Bowen, D. Basu, C. Krafft and I. Mayergoyz, "Fabrication and Evaluation of PCB-Embedded Broadband Signal Transformers With Custom Machined Racetrack-Shaped Ferrite Cores for Ethernet Applications," in IEEE Transactions on Magnetics, vol. 53, no. 11, pp. 1-5, Nov. 2017, Art no. 8401805 [2] D. Bowen, A. Lee, C. Krafft and I. D. Mayergoyz, "Design Control of Performance in Nested and Interleaved Winding Printed Circuit Board Transformers for Ethernet Applications," in IEEE Transactions on Magnetics, vol. 49, no. 7, pp. 4013-4016, July 2013. [3] D. Bowen, A. Lee, C. Krafft and I. D. Mayergoyz, “Zero-footprint Ethernet transformers using circuit-board embedded ferrites,” Journal of Applied Physics 115, 17E716 (2014) [4] J. Quilici, “Embedded magnetics technology overview,” see http://www.radial-e.com/?page_id=2for Radial Electronics Technology Brief. [5] “Planarmag Product Technology From TE Brings Automated Embedded Magnetic Products To The Ethernet Market,” Electronic Specifier, Feb. 2011, retrieved from https://www.electronicspecifier.com/communications/new-planarmag-product-technology-from-te-brings--automated-embedded-magnetic-products-to-the-ethernet-market Jan. 2020 [6] D. Bowen, C. Krafft and I. D. Mayergoyz, “Inductance optimization of miniature Broadband transformers with racetrack shaped ferrite cores for Ethernet applications,” AIP Advances 7, 056663 (2017)

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/d566179c3f82a8d154a69afc5295cd28.jpg)
Figure 1. Process for embedding a Z-conductive plug to fabricate a magnetic circuit component.

![](https://assets.underline.io/uploads/markdown_image/1/image/bea5a391e76230077b3160e1c07ae413.jpg)
Figure 2. 14-turn racetrack core transformer (a) Layout (b) Top Image (c) Image of turns overlapping via plug (d) Magneto-optic image of current flowing between primary and secondary windings

 Concept and Process Development using Ex-situ Fabricated High Density Interconnect Plugs for Circuit-Board Embedded Magnetic Components

I. INTRODUCTION Magnetoencephalographs (MEGs) require a magnetically shielded room (MSR) made of mu-metal to protect against very slow and small range of magnetic field (micromagnetic field) noise. As magnetic shields tend to be very expensive, the evaluation of the magnetic shielding factor SF of the MSR in advance is crucial. The permeability μ of the shielding material is the main factor affecting the magnetic properties, which influence the SF. Conventionally, an index of permeability at H = 4 mA/cm, which we denote as “μ4” in this paper, is used to evaluate the performance of a mu-metal. However, as we explained in a previous study [1], permeability at a lower range than 4 mA/m, such as the Rayleigh region [2], is important for magnetic shielding materials in MSRs used for MEGs. We also explained that the measurement system of permeability against a very low-frequency micromagnetic field fluctuation and heat treatment process can improve the permeabilities of a mu-metal [3]. The SF is also affected by the shape of the magnetic shield. As a first step for predicting the SF of an MSR model, we examined a cylindrical-shell magnetic shield (material thickness, t = 1 mm; length, L = 640 mm; diameter, D = 100 mm), which can be considered as infinitely long and can be applied with SF = 1+µt/D [4]. We propose a prediction method for the SF by using this formula and the estimated μ. The permeability index, μ, of a ring-shaped test piece (ring TP) was measured at a magnetic field strength of 0.001–0.675 A/m, which comprises the linear region in the μ–H curve, to detect the initial permeability, μi, and Rayleigh constant ν; Rayleigh’s law states the μ–H curve is present for μ = μi + νH. Before we apply two equations to calculate SF, we need to know the internal magnetic field of the material H = Hi, whose value is attenuated from applied magnetic field He due to the demagnetization effect of its cylindrical-shell shape. Although the demagnetizing factors of spheres and ellipsoids are known, it is difficult to calculate these values for other shapes. Therefore, we used magnetic field analysis to estimate the internal magnetic field Hi of the cylindrical shell. In this study, a prediction method of the SF for a cylindrical shell was demonstrated for three types of mu-metals, and its accuracy was verified by an experiment on the SF. II. METHOD We prepared three types of mu-metals (mu-metals 1, 2, and 3), which exhibited different properties. A ring TP and cylindrical shell were fabricated from mu-metals 1, 2, and 3. The ring TPs and cylindrical shells were simultaneously magnetically annealed in the same heat-treatment furnace. The prediction of the SFs was performed in three steps. First, ring TPs were measured in an alternating magnetic field of 1 Hz, and μi and ν were obtained via the least-squares method that was applied to the μ data of the linear region. Second, the Hi values of the cylindrical shells were estimated using magnetic simulation software ELF/MAGIC (ELF corporation). A cylindrical shell was modeled as a 3D object that was divided into 180 cells along its circumferential direction, 64 along the length, and using a single cell along the thickness. Hi was obtained as the average value of all dividing elements; He was varied as 0.01, 0.1, 1, and 10 A/m (0 to peak); and Hi was calculated for each level of He. Third, SF was predicted as SF = 1 + µt/D using the equation μ = μi + νHi. Fourth, the cylindrical shells fabricated from mu-metals 1, 2, and 3 were placed in a uniform magnetic field strength of He that varied at 1 Hz, and the magnetic field at the center of the shell Hp was measured. The SF was obtained as He/ Hp and compared with the predicted result. III. RESULT AND DISCUSSION The μ–H curve, μi and ν of the ring TPs are displayed in Fig. 1. The internal magnetic fields Hi of the cylindrical shell calculated using the magnetic simulation software were 1/1608 to 1/418 of He, and were in the Rayleigh region. Fig. 2 shows the predicted SF using Hi values of each He and the experimental results for the three types of cylindrical shells. It showed that mu-metal 3 has the highest μ in the Rayleigh region, mu-metal 2 has the lowest μ, and so as the order of the SF. Both results of the prediction and experiment show the tendency that the SF increases with an increase in He, and are in good agreement. IV. CONCLUSION To develop an accurate prediction method for the magnetic shielding factor in the design of an MSR, we proposed a prediction method for cylinders as a preliminary step and verified its accuracy via the comparison with the result of the experiment. The value of Hi indicate that only few of He can penetrate inside the material of cylindrical shell, and thus μ of micromagnetic field such as the Rayleigh region is important for a magnetic shield against even large magnetic field noise of 10 A/m. As shown in Fig. 2, it was confirmed that the value of μ calculated using Rayleigh’s law is linked to the magnetic shielding effect, and the result also indicated the possibility of applying it to predict SF of MSR.

**References:**

[1] M. Sakakibara, S. Muto, M. Okuda, Y. Adachi, and G. Uehara, IEEJ Trans. Magn. 140, 331 (2020). [2] R. M. Bozorth, IEEE Magnetics Society, 489 (1993). [3] M. Sakakibara, Y. Adachi, and G. Uehara, T. Meguro, IEEE Trans. Magn. (2020). [4] Wills, A.P., Phys. Rev. 9 pp. 193 213 (1899)

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/9310565b4f277b2dd9bdbf5c7589b14e.jpg)
Fig.1 Plots of μ–H. In (a) and (b), H denotes values in the range of 0 – peak. μ4 is an index of permeability at H = 4 mA/cm, which is conventionally used to evaluate the shielding factor.

![](https://assets.underline.io/uploads/markdown_image/1/image/11e87de63365dc520ed55ec59662c45a.jpg)
Fig.2 Estimated value of the shielding factor SF compared with the experimental results; (a) mu-metal 1, (b) mu-metal 2, and (c) mu-metal 3.

 Prediction of cylindrical magnetic shielding performance by considering the magnetic field strength inside the material

1.INTRODUCTION Wireless power transfer (WPT) has been widely used in various applications due to its great convenience of contactless energy delivery [1,2]. Generally, WPT systems can be categorized into single-phase type and multi-phase type. Compared to traditional single-phase WPT systems, the multi-phase type WPT systems can have more uniformed magnetic field, more compact coil structure and less output current ripple [3]. However, additional interphase mutual inductance is a problem in multi-phase WPT system, making the modeling and control more difficult [3,4]. Also, most multi-phase WPT systems are based on two-level inverters, and three-level inverters surpass traditional two-level inverters in many applications because of advantages like lower voltage stress on power switches and smaller electromagnetic interference (EMI). Among the various three-level topologies, the T-type neutral point clamp (TNPC) topology has high overall efficiency and equal conduction loss of all used power switches [5]. In this paper, a three-phase three-level WPT system based on TNPC inverter is proposed. First, a compact three-phase coil structure are designed to achieve symmetric interphase mutual inductance. Next, with a simplified modeling method, each phase of the three-phase system is simplified to a single-phase system. Further, a novel output regulation method with specific switching sequence is proposed to control the amplitude of the output voltage. Finally, the effectiveness of the proposed three-phase WPT system is verified by both the simulation and experimental results. 2.SYSTEM CONFIGURATION The main topology of the proposed WPT system with series-series compensation is shown in Fig. 1(a). The 3D coil structure is shown in Fig. 1(d) where every two of the six coils have a mutual inductance. The primary coils and the secondary coils have the identical shape. The three coils on the same side are in the same plane and occupy 120° of a circle. With this unique coil structure, the space is highly utilized, and the three-phase coils keep symmetric and has the same interphase mutual inductance. Mpp is the mutual inductance of the primary coils, Mss is the mutual inductance of the secondary coils, Msp1 is the mutual inductance of the aligned coils from the primary side and secondary side, and Msp2 is the mutual inductance of the coils which are not aligned from the primary side and secondary side. Based on the first harmonic approximate (FHA) method, the equivalent circuit is shown in Fig.1 (b), where RL=6Rdc/π2 [6]. When the three phases keep balanced, the voltage across o1 and n1, U1on=0, and the voltage across o2 and n2, U2on=0. Therefore, a simplified equivalent circuit can be derived, as shown in Fig.1 (c). In Fig.1 (c), each phase of the system is approximately equivalent to a single-phase system, where the inductance of the primary side is L1-Mpp, the inductance of the secondary side is L2-Mss, and the mutual inductance is Msp1-Msp2. 3.OUTPUT VOLTAGE REGULATION METHOD The inverter output voltage U1xo is subject to the switching states. Supposing the switching state of the phase x is S=[Sx1 Sx2 Sx3 Sx4], when S+1=[1 1 0 0], U1xo=+0.5Udc; when S0=[0 1 1 0], U1xo= -0.5Udc; when S-1=[0 0 1 1], U1xo=-0.5Udc. In order to reduce the state change of each controlled switches and make the control signal symmetric, a specific switching sequence is chosen for the proposed WPT system, that is S0-S+1-S0-S-1-S0. As shown in Fig. 2 (a) and (b), each of S+1 and S-1 account for 0.5aT in a period time T. The drive signals of each power switches are symmetric in half a period time, and the states of each controlled switches only change twice in a single period time, which makes the driver signals easy to produce. By using fourier decomposition method, the amplitude of the first harmonic can be expressed as m1=2Udcsin(aπ/2)/π (0&lt;=a&lt;=1). Also, the ratio of U2xn and U1xn in Fig.1 (c) can be obtain by U2xn/U1xn =[jw(Msp1-Msp2)]/[(R2+RL)R1+4π2fr2(Msp1-Msp2)2]. Therefore, by regulating the value of a, the transferred power can be controlled. 4.SIMULATION AND EXPERIMENTS In order to verify the effectiveness of the proposed WPT system, simulations are done in MATLAB/SIMULINK. Also, as shown in Fig.2(c), a 500W three-phase prototype is established. Fig.1(b) is the experiment results of the three-phase current and the voltage of U1xo at a=0.3. More analysis and experiment results will be given in full paper. 5.CONCLUSION In this paper, a three-phase three-level WPT system is proposed. Each phase of the system is approximately equivalent to a single-phase system which effectively simplify the system model. A novel output regulation method is proposed to control the output voltage. Simulation and experiment results show a good performance of the proposed WPT system and effectiveness of the control and modeling method, and indicate the proposed WPT system has a better output voltage spectrum and reduced voltage stress compared to traditional three-phase two-level WPT system. Acknowledge This work is supported in part by NSFC (52077186&51677159), in part by STIC Shenzhen Municipality (JCYJ20180307123918658), and in part by ITF (ITP/027/19AP), HK.

**References:**

[1]X. Gao, C. Liu, Y. Huang, "Design of An UAV-Oriented Wireless Power Transfer System with Energy-Efficient Receiver," IECON 2020, Singapore, 2020, pp. 2025-2030. [2]Y. Xiao and C. Liu, "Direct Load Voltage Control for Electrolytic Capacitor-less Wireless Power Transfer System Without DC/DC Converter," in IEEE Transactions on Industrial Electronics. doi: 10.1109/TIE.2020.3013780 [3]J. Pries et al., "A 50-kW Three-Phase Wireless Power Transfer System Using Bipolar Windings and Series Resonant Networks for Rotating Magnetic Fields," in IEEE Transactions on Power Electronics, vol. 35, no. 5, pp. 4500-4517, May 2020. [4]G. A. Covic, J. T. Boys, M. L. G. Kissin and H. G. Lu, "A Three-Phase Inductive Power Transfer System for Roadway-Powered Vehicles," in IEEE Transactions on Industrial Electronics, vol. 54, no. 6, pp. 3370-3378, Dec. 2007. [5]Z. Yuan et al., "Design and Evaluation of Laminated Busbar for Three-Level T-Type NPC Power Electronics Building Block With Enhanced Dynamic Current Sharing," in IEEE Journal of Emerging and Selected Topics in Power Electronics, vol. 8, no. 1, pp. 395-406, March 2020. [6]C. Wang, C. Zhu, G. Wei, J. Feng, J. Jiang and R. Lu, "Design of Compact Three-Phase Receiver for Meander-Type Dynamic Wireless Power Transfer System," in IEEE Transactions on Power Electronics, vol. 35, no. 7, pp. 6854-6866, July 2020.

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/25b52455841eb9658cbdc6202cde8c47.jpg)
Fig. 1. Proposed three-phase three-level WPT system. (a) Main topology. (b)Equivalent circuit. (c) Simplified equivalent circuit. (d) 3D model of the primary coils and the secondary coils.

![](https://assets.underline.io/uploads/markdown_image/1/image/906d0bfa8e0f16fee8207cba065151e9.jpg)
Fig. 2 Primary voltage and experiment of the proposed WPT system. (a) Voltage waveform when a=0.37. (b) Voltage waveform when a=0.88. (c) Testbed. (d) Inverter output voltage in phase A and three-phase currents.

 A Novel Output Voltage Regulation Method for Three-phase Three-level Wireless Power Transfer Based on A Simplified System Model

I. INTRODUCTION Due to the ability to maintain system operation after electrical failure, fault-tolerant machines are increasingly used in safety-critical applications, for instance, automobiles, aviation and armament [1]-[2]. In this paper, based on the dual-winding fault-tolerant permanent magnet motor (DFPMM) which not only inherits permanent magnet synchronous motors’ advantages of high torque density and high efficiency but also has high-performance of fault-tolerance, a novel electric machine drive system is proposed and studied. The 12-slot/10-pole DFPMM, with two separate sets of three-phase single-layer concentrated windings wound on alternate teeth, is capable of restricting fault propagation. The drive system is equipped with two independent drive circuits, while only one DC power supply is needed. With the proposed drive system, fault-tolerant performance to cope with open-circuit faults can be implemented by applying corresponding control strategies. Especially when the one-phase open-circuit fault occurs, system utilization rate is larger compared with the traditional drive system. Simulation and experimental results incorporating current, torque and speed under conditions of one-phase open circuit fault are exhibited respectively. The results verify the effectiveness of the proposed fault-tolerant control strategies and the high reliability of the proposed DFPMM motor drive system. II. THE DFPMM SYSTEM CONFIGURATION AND DESIGN The structure of the proposed electric drive system is shown in Fig. 1(a), which contains a DFPMM, two sets of three-phase full-bridge drive circuits, two triacs, six fuses, and an independent 270V DC power supply. The two sets of windings of the DFPMM are independent of each other, and the ABC windings are controlled by T1 to T6, while the XYZ windings are controlled by T7 to T12. The triac TR1 is connected to the neutral point of the ABC windings and the midpoint of the capacitors, and TR2 is in a similar way. Fuses F1-F6 prevent a shoot-through fault caused by a simultaneous breakdown of power switches in the same phase. During normal operation, two independent three-phase bridge drive circuits control the two windings, each with the same power output which is 50% of the total output power, and the two triacs are kept off. III. FAULT TOLERANT CONTROL STRATEGIES Based on the principle of constant magnetomotive force, control strategies after faults occur will be introduced in the full paper in detail. With the double d-q mathematical model of DFPMM, it is a vector control system that consists of two separate speed controllers, current controllers and SVPWM modules for both two sets of windings. When an open-circuit fault occurs in one set of the windings, the corresponding triac will be turned on, forming a power conversion circuit with a neutral point, as the conductive triac will add a degree of freedom to control the system under the fault condition. Meanwhile, two capacitors are used to perform amplitude and phase modulation on the remaining two-phase currents. IV. SIMULATION AND EXPERIMENTAL VERIFICATION Simulation and experiments are carried out under the conditions of open-circuit faults occurring in one phase, to verify the effectiveness of the proposed electric drive system and its control strategies. The test platform is shown in Fig. 1(b). The results are shown in Fig. 2, including three-phase current of the fault set of windings, speed and torque waveforms. Simulation waveforms of current, torque and speed when an open-circuit fault occurs in phase-A with and without the fault-tolerant control strategies are shown in Fig. 2(a) and Fig. 2(b), respectively. And experimental waveforms of current, torque and speed when an open-circuit fault occurs in phase-A with and without the fault-tolerant control strategies are compared in Fig. 2(c) and Fig. 2(d). From these results, the motor with the proposed drive system maintains the normal performance after the fault occurs, by applying the corresponding control strategy. Thus, the damage caused by open-circuit faults can be avoided. V. CONCLUSION In this paper, a highly-reliable DFPMM drive system aiming at ensuring a normal operation after open-circuit faults is proposed. By applying the control strategies and corresponding actions to typical types of faults, both simulation and experiments have done, proving the effectiveness of the proposed highly-reliable electric drive system. It can cope with open-circuit faults in different situations, and it is suitable for safety-critical applications. More detailed theoretical analysis and experimental verification will be presented in the full paper. ACKNOWLEDGEMENT This work was supported in part by the National Natural Science Foundation of China under Project 51807094, and in part by the Project funded by China Postdoctoral Science Foundation under Project 2020M671499, and in part by Jiangsu Planned Projects for Postdoctoral Research Funds under Grant 2020Z145.

**References:**

[1] B. Wang, J. Wang and A. Griffo, "Stator Turn Fault Detection by Second Harmonic in Instantaneous Power for a Triple-Redundant Fault-Tolerant PM Drive", IEEE Transactions on Industrial Electronics, vol. 65, no. 9, pp. 7279-7289 (2018) [2] X. Jiang, D. Xu and L. Gu, "Short-Circuit Fault-Tolerant Operation of Dual-Winding Permanent-Magnet Motor Under the Four-Quadrant Condition", IEEE Transactions on Industrial Electronics, vol. 66, no. 9, pp. 6789-6798 (2019)

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/3a7fa4ccb4534a2d2dfda84bf2bd5503.jpg)
Fig. 1. The proposed DFPMM drive system. (a) Configuration topology. (b) Test platform.

![](https://assets.underline.io/uploads/markdown_image/1/image/f6b8eac2831d1c3d6d19a84a0d792b69.png)
Fig. 2. The waveforms of three-phase currents, torque, and speed under conditions of phase-A open-circuit fault. (a) Simulation waveforms when no fault-tolerant control strategy is adopted. (b) Simulation waveforms when the fault-tolerant control strategy is adopted. (c) Experimental waveforms when no fault-tolerant control strategy is adopted. (d) Experimental waveforms when the fault-tolerant control strategy is adopted.

 Analysis of Highly Reliable Electric Drive System Based on Dual-Winding Fault-Tolerant Permanent Magnet Motor

1. Introduction Due to the rapid growth in global energy consumption, it is imperative to improve the efficiency of electrical machines and other electrical devices. Iron loss is a crucial part of electrical machines, and has received increasing attention in the past few years. It is significant for both optimizing electrical machine design and improving thermal efficiency. Over the years, several iron loss models have been developed, such as the loss separation models [1] and various hysteresis models [2]. In order to evaluate the iron loss, the most traditional way is to analyze the machine using finite element method (FEM), and calculate the iron loss employing a loss separation model. However, it is deficient, especially when dealing with inverter-fed electrical machines, in which higher-order harmonics exist. In some cases, the estimation error exceeds one hundred percent [3], which is critical for iron loss prediction. On the other hand, in order to precisely describe the materials properties, several hysteresis models are developed. One benefit of hysteresis models is that the iron loss can be directly obtained if the B-H hysteresis loops are known. This paper introduces an iron loss estimation method based on the loss surface (LS) hysteresis model and has applied it on several specimens. The verification process is shown in Fig. 1. The LS hysteresis model is constructed based on the B-H hysteresis loops measured from a toroidal specimen laminated of non-oriented electrical steel sheets. A C-type specimen constructed by the same material is taken as the test object, and is analyzed using the finite element method. The obtained field results combined with the LS hysteresis model are used for iron loss calculation. The calculation results are compared with measured ones. The verification is carried on for both sinusoidal and non-sinusoidal cases. 2. The Proposed Procedure As the iron loss can be calculated directly from the integral of the hysteresis loop without the empirical values. Therefore, it is important to get the magnetic field intensity (H) from the flux density (B) and its derivative(dB/dt). The method begins with the measurement of the B-H hysteresis loops, which is performed on a toroidal specimen according to the International Standard IEC 60404-6 [4]. A primary winding is used for the excitation, and another winding is the measuring coil. The B-H hysteresis loops under different magnetic flux density magnitudes Bm have been measured to demonstrate the material behavior more precisely. The LS model, therefore, can be represented by the function between the magnetic field intensity(H) and the flux density (B) and its derivative (dB/dt) and its maximum value (Bm). The calculated object is first performed field analysis by using finite element method. The magnetic flux density distribution can be computed in all elements, which is also taken as the input of the LS model. Combined the derivative of magnetic flux density (dB/dt) and the magnitude Bm, the corresponding magnetic field intensity (H) can be obtained, therefore, the B-H hysteresis loops for arbitrary waveforms can be construed, from which iron loss can be calculated. 3. Verification First, a C-type specimen with the same material is manufactured to verify the proposed procedure, as shown in Fig. 1. In order to check the LS hysteresis modeling accuracy, a magnetic flux density with 0.8T amplitude and 50Hz fundamental frequency, combined with 35% of the third harmonic with a phase shift of 90°is taken as the input, and the magnetic field intensity and hysteresis loops are shown in Fig. 2. The modeled results match well with measured ones. The iron loss of the test C-type specimen is calculated by using both the proposed method and the loss separation model. The measured iron loss is 1.699W, and for the proposed method, the calculated iron loss is 1.530W, whereas for the loss separation models, the calculated iron loss is 1.282W. 4. Conclusion The proposed iron loss calculation procedure based on the LS hysteresis model has been developed and the verification of a C-type specimen has been done. Compared with actual measured results, the proposed iron loss calculation method is more accurate than loss separation model. Verification and analysis on more topologies will be discussed in the full paper.

**References:**

[1] G. Bertotti, "General properties of power losses in soft ferromagnetic materials," in IEEE Transactions on Magnetics, vol. 24, no. 1, pp. 621-630, Jan. 1988. doi: 10.1109/20.43994. [2] Chevalier, T., Kedous-Lebouc, A. and Cornut, B. et al. A new dynamic hysteresis model for electrical steel sheet. Physica B: Condensed Matter, 275(1-3), 197-201. [3] Tekgun, B., Sozer, Y., and Tsukerman, et al. Core loss estimation in electric machines with flux-controlled core loss tester. IEEE Transactions on Industry Applications, 55(2), 1299-1308. [4] Methods of measurement of the magnetic properties of magnetically soft metallic and powder materials at frequencies in the range 20hz to 200khz by the use of ring specimens”, International Standard IEC 60404-6: 2003.

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/5ada5b095a3f2f7807d1aa9820381674.jpg)
Fig1TheLSmodelVerificationProcess

![](https://assets.underline.io/uploads/markdown_image/1/image/fb3c33aaca38a830099708fa21c2bbaf.jpg)
Fig2LSModelEstimatedAndMeasuredResults

 Iron Loss Calculation Based on Loss Surface Hysteresis Model and Its Verification

The aim of this paper is to investigate the feasibility of a hybrid ferromagnetic bonding in a stack used in electrical machines. In a context of growing energy needs and the absolute need to reduce greenhouse gases, the stakes are high. For instance, with the electrification of aircraft in full development, it is essential to improve the power density of on-board electric actuators. Beyond aeronautics, the advantage of improving machine performance through its intrinsic performance of its materials is that this increase in efficiency does not depend on the application and becomes possible in all sectors of activity. This study which is supported by the French ADEME (Agency of the Environment and the Control of Energy) focuses on steel sheet stack intrinsic performance improvement for electrical machine applications. Generally, a laminated stack is composed by electrical steel sheet. An epoxy-based bonding is often used to ensure a structural cohesion of the stack in order to be manipulated and integrated. It also helps to keep a proper electrical insulation between sheets, limiting eddy current circulation. This study proposes to add iron powder to the epoxy bonding as described on Fig. 1, in order to improve the permeability of the bonding, which is magnetically neutral, becoming then a SMC. However, SMCs are generally used as a magnetic yoke [1] and not as an adhesive layer. The powder load is only intended to exploit a magnetically inert volume. The magnetic flux which will circulate in the packet will do so essentially through the sheets, but the gain in permeability envisaged to make pass more flux under the same magnetic field. The powder load could also improve the thermal conduction along the z axis, improve the thermal performance of the sheet package. Nevertheless, the circulation of eddy currents must remain limited between the sheets and the composite joints. It is necessary to limit the powder load to ensure this. Actually, the method proposed here does not really reproduce the SMC manufacturing process (pressure at temperature of the dielectric mixture - magnetic powder). In the case of a conventional SMC, the process ensures inter-grain isolation [2], making these materials particularly efficient at high-frequency. In the composite joint proposed here, the inter-grain (and therefore inter-sheet) insulation relies only on the residue of the epoxy adhesive around the grains once the sheets have been bonded. The distribution and the quantity of the adhesive being difficult to control, it is essential to keep a sufficient margin, and therefore to limit the powder load. On the other hand, the addition of a magnetic filler has been shown to improve the permeability of a mixture, especially at high frequency [3]. This feasibility study is based on ring core characterization. In order to limit parameters, the thickness and the number of sheets are set, respectively at 0.2 mm and 10 per stack. The thickness of the composite joint is also set at 50 µm. Actually, the powder grains must be large enough to retain interesting magnetic properties. A thickness of 50 µm is a good compromise in the context of a feasibility study. Finally, the only variable parameter is the powder charge. It has been decided to set it at 25%, 50% and 75% of the mass of the composite joint. A joint of the same thickness but loaded at 0% will be used as a reference. The magnetic permeability of the joint is actually determined by the powder volume / joint volume ratio. However, it is much simpler to proceed by mass ratio and to use density to calculate the corresponding volume ratio (4.5%, 13% and 30%). Besides, two types of iron powder are used (AHC 300 and MH300, both 300 mesh for grain size), each with specific properties. Therefore, each ring core corresponds to electrical steel sheet stack with a determined load of specific type of iron powder. Then, ring cores are manufactured and their windings are designed to reach 2 T at 400 Hz and 1000 Hz, and 1 T at 5000 Hz. Specific iron losses (W/kg) are measured and gathered in Fig. 2. This study demonstrates that adding iron powder to adhesive bonding of steel sheet is technically possible. The gain observed is significant and appears promising from the perspective of a more developed study. However, some technical points must be improved. More measurement should be proceeded, including reproducibility measurement. It would appear interesting to charge the adhesive with a metal powder of the same thermal conductivity as iron (or similar) but which would be non-magnetic. It could then be determined whether the performance is improved via magnetic permeability or via thermal conduction. In order to fully characterize these samples, optimization by numerical simulation could be considered, in order to precisely determine the parameters essential to performance. Finally, it seems essential to perform tests on a machine and not on ring cores, in order to ensure the real potential gains.

**References:**

[1] A. Jakubas, P. Gebara, S. Seme, “Magnetic properties of SMC cores produced at a low compacting temperature”. Acta Phys. Pol. A, 131(5), 1289-1293. [2] Y. X. Pang, S. N. B. Hodgson, J. Koniarek, “The influence of the dielectric on the properties of dielectromagnetic soft magnetic composites. Investigations with silica and silica hybrid sol–gel derived model dielectric.” Journal of magnetism and magnetic materials, 310(1), 83-91. [3] Belyaev, B. A., & Tyurnev, V. V. (2018). “Electrodynamic calculation of effective electromagnetic parameters of a dielectric medium with metallic nanoparticles of a given size.” Journal of Experimental and Theoretical Physics, 127(4), 608-619.

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/398f5db97656d15612bd89dc54d799a1.jpg)
Fig. 1: Representation of the hybrid bonding in an electrical steel sheet stack.

![](https://assets.underline.io/uploads/markdown_image/1/image/8beea0f22ed5225ed70e5c7b610947ed.jpg)
Fig. 2: Results of specific iron losses measured in the samples for AHC300 and MH300 iron powder, at several frequencies and magnetic flux densities.

Downloads

Next from INTERMAG 2021

Analytical Calculation for Magnetic Field in Spoke-Type Permanent Magnet Machines Based on a Rotor Magnetic Potential Model

Similar lecture

Thermal Network Model of A SCB2500kVA Dry-type Transformer Coupled with Electromagnetic Loss

.css-70qvj9{display:-webkit-box;display:-webkit-flex;display:-ms-flexbox;display:flex;-webkit-align-items:center;-webkit-box-align:center;-ms-flex-align:center;align-items:center;}Downloads

Next from INTERMAG 2021

Analytical Calculation for Magnetic Field in Spoke-Type Permanent Magnet Machines Based on a Rotor Magnetic Potential Model

Similar lecture

Thermal Network Model of A SCB2500kVA Dry-type Transformer Coupled with Electromagnetic Loss

Downloads