France



1 Introduction&lt;br/&gt;High-intensity magnetic field around the system will be generated when the wireless charging system of electric vehicles(EV) is working, which will reduce the transmission efficiency of the system and pose risks to human health[1]. Ferrite core has been widely used to shield the magnetic flux leakage on the back side of the coil and improve the magnetic coupling between the primary and the secondary sides due to its high permeability and large resistivity. However, this shielding method is not very effective in shielding the magnetic field in the non-working area on the side of the system. As a supplement to the passive shielding method, the main idea of the active shielding method is to use the reverse magnetic field generated by the shielding coil to cancel the magnetic field leakage in the target area. The reactive shielding method based on resonance of a passive compensation loop has been proven to be an effective method to offset the magnetic leakage in the target area, but adjustment of the shielding circuit impedance is complicated[2]. The shielding coil with independent power supply can be controlled to produced a magnetic field that is completely opposite to the incident magnetic field, but it inevitably reduces the overall transmission efficiency of the system[3]. A good solution is to make active shielding coil be in series with the branch where the energy transfer coil is located[4]. It should be noted that the active shielding coil may be coupled with the main magnetic linkage. Under this condition, the equivalent magnetic coupling between the primary and the secondary sides becomes smaller, which will reduce the output power of the system.&lt;br/&gt;In this paper, a dual-loop active shielding coil directly connected in series with the primary side is designed to reduce the reverse magnetic coupling between the primary and the secondary sides and the magnetic flux leakage around the wireless charging system of EV. The parameters of the shielding coil are determined by multi-objective optimization. Finally, a prototype with an input power of 1.2kW is built. The proposed active shielding coil has superior performance in system transmission efficiency and shielding effect on the target surface comparing to conductor shielding.&lt;br/&gt;2 Active Shielding Coil Design&lt;br/&gt;Circuit theory is adopted to model the wireless power transmission system with double LCC compensation topology as shown in Fig. 1, where &lt;i&gt;L&lt;/i&gt;&lt;sub&gt;1&lt;/sub&gt; and &lt;i&gt;L&lt;/i&gt;&lt;sub&gt;2&lt;/sub&gt; are self-inductances, &lt;i&gt;L&lt;/i&gt;&lt;sub&gt;3&lt;/sub&gt; is the self-inductance of shielding coil. To prevent load power from being affected by the shielding coil, what we need to do is to minimize the mutual inductance &lt;i&gt;M&lt;/i&gt;&lt;sub&gt;23&lt;/sub&gt; between the active shielding coil and the secondary side.&lt;br/&gt;The proposed dual-loop active shielding coil shown in Fig. 2 is only arranged on the primary side, since the steel plate of the car chassis has a good shielding effect on the magnetic field around the secondary side. The inner diameters of the inner loop coil is set to a fixed value of 50cm, and the target surface is set at a distance of 55cm from the center of the coil. The number of turns and outer diameter of the shielding coil are optimized through ANSYS simulation.&lt;br/&gt;The outer diameter of the shielding coil is optimized in the range of 29cm to 44cm with a step length of 3cm. In addition, the number of turns of the inner loop and outer loop is optimized in the range of 1 to 3 and 1 to 4, respectively. The optimization targets are the mutual inductance between the active shielding coil and the secondary side, the maximum magnetic flux density on the target surface, and the length of the active shielding coil. To solve the multi-objective optimization problem, the constraints are set as follows:&lt;br/&gt;&lt;i&gt;M&lt;/i&gt;&lt;sub&gt;23&lt;/sub&gt;≤0.5μH&lt;br/&gt;&lt;i&gt;B&lt;/i&gt;&amp;lt;27μT&lt;br/&gt;&lt;i&gt;L&lt;/i&gt;≤13m&lt;br/&gt;To further determine the optimal solution, an evaluation function &lt;i&gt;F&lt;/i&gt;(d) is constructed and defined as the normalized objective function &lt;i&gt;f &lt;/i&gt;&lt;sub&gt;i &lt;/sub&gt;multiplied by their respective weight coefficients &lt;i&gt;ω&lt;/i&gt;&lt;sub&gt;i &lt;/sub&gt;as follows:&lt;br/&gt;&lt;i&gt;F&lt;/i&gt;(d)=Σ&lt;i&gt;ω&lt;/i&gt;&lt;sub&gt;i&lt;/sub&gt;|&lt;i&gt;f &lt;/i&gt;&lt;sub&gt;i&lt;/sub&gt;-&lt;i&gt;f &lt;/i&gt;&lt;sub&gt;i&lt;/sub&gt;&lt;i&gt;&lt;sup&gt;*&lt;/sup&gt;&lt;/i&gt; |/(&lt;i&gt;f&lt;/i&gt;&lt;sub&gt;set&lt;/sub&gt;-&lt;i&gt;f &lt;/i&gt;&lt;sub&gt;i&lt;/sub&gt;&lt;i&gt;&lt;sup&gt;*&lt;/sup&gt;&lt;/i&gt;) (i=1,2,3)&lt;br/&gt;where &lt;i&gt;f &lt;/i&gt;&lt;sub&gt;i&lt;/sub&gt;&lt;i&gt;&lt;sup&gt;*&lt;/sup&gt;&lt;/i&gt; is the ideal value with a value of 0, and &lt;i&gt;f&lt;/i&gt;&lt;sub&gt;set&lt;/sub&gt; is the constraint value of the objective function.The solution with the smallest evaluation function value is adopted for subsequent simulation and experimental verification.&lt;br/&gt;The magnetic field distribution on the target surface is compared in three condition through ANSYS simulation: no shielding, with aluminum shielding and active shielding. When the active shielding is applied, the maximum magnetic flux density of the target surface is reduced to 16.2uT comparing to the case of 33.05uT when only the ferrite core is applied. However, the maximum magnetic flux density is 20.01uT in the case of aluminum shielding.&lt;br/&gt;The proposed active shielding design is verified by a prototype with an output power of 1.2kW. To investigate the effect of shielding methods on transmission efficiency, the input voltage needs to be adjusted to keep the system input power constant. The results of the experiment will be presented in the final paper.&lt;br/&gt;3 Conclusion&lt;br/&gt;This paper introduced a dual-loop active shielding coil connected in series with the primary side to reduce the magnetic leakage around the system. An evaluation function has been proposed to optimize the number of turns and outer diameter of the shielding coil. The proposed shielding method obtains a superior shielding effect on the target surface than conductor shielding, and has less passive impact on the transmission efficiency of the system.

**References:**

[1] S. Park, IEEE Transactions on Magnetics., vol. 54, p.1-8 (2018).&lt;br/&gt;[2] H. Moon, S. Kim and H. H. Par, IEEE Transactions on Magnetics., vol. 51, p.1-4 (2015).&lt;br/&gt;[3] T. Campi, S. Cruciani and F. Maradei, IEEE Transactions on Electromagnetic Compatibility., vol. 62, p.1398-1405 (2020).&lt;br/&gt;[4] S. Y. Choi, B. W. Gu and S. W. Lee, IEEE Transactions on Power Electronics., vol. 29, p.5770-5783 (2014).

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/ba4c9472340bfaa611896a8f76f1e87a.jpg)
Figure. 1: Equivalent circuit of the system with active shielding coil

![](https://assets.underline.io/uploads/markdown_image/1/image/d4a13eecc734961c4e8612dfaa91b4a3.jpg)
Figure. 2: Design of dual-loop active shielding coil

INTERMAG 2021

 Active Shielding Coil Design for Wireless Charging System of Electric Vehicle

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 High-intensity magnetic field around the system will be generated when the wireless charging system of electric vehicles(EV) is working, which will reduce the transmission efficiency of the system and pose risks to human health[1]. Ferrite core has been widely used to shield the magnetic flux leakage on the back side of the coil and improve the magnetic coupling between the primary and the secondary sides due to its high permeability and large resistivity. However, this shielding method is not very effective in shielding the magnetic field in the non-working area on the side of the system. As a supplement to the passive shielding method, the main idea of the active shielding method is to use the reverse magnetic field generated by the shielding coil to cancel the magnetic field leakage in the target area. The reactive shielding method based on resonance of a passive compensation loop has been proven to be an effective method to offset the magnetic leakage in the target area, but adjustment of the shielding circuit impedance is complicated[2]. The shielding coil with independent power supply can be controlled to produced a magnetic field that is completely opposite to the incident magnetic field, but it inevitably reduces the overall transmission efficiency of the system[3]. A good solution is to make active shielding coil be in series with the branch where the energy transfer coil is located[4]. It should be noted that the active shielding coil may be coupled with the main magnetic linkage. Under this condition, the equivalent magnetic coupling between the primary and the secondary sides becomes smaller, which will reduce the output power of the system. In this paper, a dual-loop active shielding coil directly connected in series with the primary side is designed to reduce the reverse magnetic coupling between the primary and the secondary sides and the magnetic flux leakage around the wireless charging system of EV. The parameters of the shielding coil are determined by multi-objective optimization. Finally, a prototype with an input power of 1.2kW is built. The proposed active shielding coil has superior performance in system transmission efficiency and shielding effect on the target surface comparing to conductor shielding. 2 Active Shielding Coil Design Circuit theory is adopted to model the wireless power transmission system with double LCC compensation topology as shown in Fig. 1, where L1 and L2 are self-inductances, L3 is the self-inductance of shielding coil. To prevent load power from being affected by the shielding coil, what we need to do is to minimize the mutual inductance M23 between the active shielding coil and the secondary side. The proposed dual-loop active shielding coil shown in Fig. 2 is only arranged on the primary side, since the steel plate of the car chassis has a good shielding effect on the magnetic field around the secondary side. The inner diameters of the inner loop coil is set to a fixed value of 50cm, and the target surface is set at a distance of 55cm from the center of the coil. The number of turns and outer diameter of the shielding coil are optimized through ANSYS simulation. The outer diameter of the shielding coil is optimized in the range of 29cm to 44cm with a step length of 3cm. In addition, the number of turns of the inner loop and outer loop is optimized in the range of 1 to 3 and 1 to 4, respectively. The optimization targets are the mutual inductance between the active shielding coil and the secondary side, the maximum magnetic flux density on the target surface, and the length of the active shielding coil. To solve the multi-objective optimization problem, the constraints are set as follows: M23≤0.5μH B&lt;27μT L≤13m To further determine the optimal solution, an evaluation function F(d) is constructed and defined as the normalized objective function f i multiplied by their respective weight coefficients ωi as follows: F(d)=Σωi|f i-f i* |/(fset-f i*) (i=1,2,3) where f i* is the ideal value with a value of 0, and fset is the constraint value of the objective function.The solution with the smallest evaluation function value is adopted for subsequent simulation and experimental verification. The magnetic field distribution on the target surface is compared in three condition through ANSYS simulation: no shielding, with aluminum shielding and active shielding. When the active shielding is applied, the maximum magnetic flux density of the target surface is reduced to 16.2uT comparing to the case of 33.05uT when only the ferrite core is applied. However, the maximum magnetic flux density is 20.01uT in the case of aluminum shielding. The proposed active shielding design is verified by a prototype with an output power of 1.2kW. To investigate the effect of shielding methods on transmission efficiency, the input voltage needs to be adjusted to keep the system input power constant. The results of the experiment will be presented in the final paper. 3 Conclusion This paper introduced a dual-loop active shielding coil connected in series with the primary side to reduce the magnetic leakage around the system. An evaluation function has been proposed to optimize the number of turns and outer diameter of the shielding coil. The proposed shielding method obtains a superior shielding effect on the target surface than conductor shielding, and has less passive impact on the transmission efficiency of the system.

**References:**

[1] S. Park, IEEE Transactions on Magnetics., vol. 54, p.1-8 (2018). [2] H. Moon, S. Kim and H. H. Par, IEEE Transactions on Magnetics., vol. 51, p.1-4 (2015). [3] T. Campi, S. Cruciani and F. Maradei, IEEE Transactions on Electromagnetic Compatibility., vol. 62, p.1398-1405 (2020). [4] S. Y. Choi, B. W. Gu and S. W. Lee, IEEE Transactions on Power Electronics., vol. 29, p.5770-5783 (2014).

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/ba4c9472340bfaa611896a8f76f1e87a.jpg)
Figure. 1: Equivalent circuit of the system with active shielding coil

![](https://assets.underline.io/uploads/markdown_image/1/image/d4a13eecc734961c4e8612dfaa91b4a3.jpg)
Figure. 2: Design of dual-loop active shielding coil

poster

I. INTRODUCTION This paper proposed a new flux-concentrating rotor for PM machine, taking advantages of spoke-type and Halbach array PM flux-concentrating. As shown in Fig. 1 (a), it is the evolution of the proposed rotor. The upper part of tangentially magnetized magnets still remain the spoke-type flux-concentrating property, while the bottom tangentially magnetized magnets and the radially magnetized magnets form a primitive Halbach array [1] [2]. Since the ratio match tangentially/radially magnetized magnets directly affects the effect of flux-concentrating, as well as the saliency ratio and torque performance, this paper focuses on the optimization of the size of permanent magnets on the rotor on the basis of keeping the same amount of PMs and the polar arc coefficient. II. OPTIMAL DESIGN OF NEW FLUX-CONCENTRATING ROTOR Response surface method (RSM) is a set of statistical and mathematical techniques to find the “best fitted” response of the physical system through experiment or simulation [3]. In RSM, a polynomial model, called a fitted model, is generally constructed to represent the relationship between the performance and design parameters. As shown in Fig. 1 (a), it is the evolution of the proposed rotor. Ensuring the same volume of PM in three rotors, the new rotor was optimized. The geometric parameters of the new rotor are shown in Fig. 1 (b). In order to quantitatively compare with the motor whose pole-arc coefficient is 0.8, including in various electromagnetic characteristics. b is determined as 8 mm. The volume of PM is defined as (1). where p is pole pairs, l is stack length, m is the volume of PMs. The optimized objective is formulated to maximize the air-gap flux density under the constraint of keeping the PM volume. Tab. I shows electromagnetic analysis results under various combinations of lx and hy. The relationship between air-gap flux density and lx, hy is shown in Fig. 2 (a). Two optimized points are found, where the analysis result is better than the predicted value. The regression equations of air-gap flux density obtained by RSM analysis is as (2). As shown in Fig. 2 (b), the predicted value and the actual result maintain a high consistency. It can be concluded that the regression equation obtained by the RMS is reliable. As shown in Fig. 3, The result of optimization is to maximize the air-gap magnetic density. . lx is determined as 16.5 mm, and hy is determined as 3.75 mm. Fig. 4 shows the comparison of two rotors. The final optimized model is built by 2-D FEA, as shown in Fig.5. III. FEA AND EXPERIMENTAL VERIFICATION Tab. II gives the main parameters of optimized rotor. Fig. 6 (a) compares the air-gap flux density of the motors with optimized rotor and the original rotor. The air-gap flux density in the motor with optimized rotor is higher than that in the initial one. As shown in Fig. 6 (b), the line back electromotive force (EMF) of the motor with optimized rotor increases slightly due to the higher air gap flux density. Fig. 6 (c) shows the C-T relationship under maximum torque per ampere (MTPA) control and id=0 control. The results show that under MTPA control and id=0 control, the output torque of the optimized rotor is higher than the initial model. In Fig. 6 (d), the saliency ratio is higher in the optimized rotor than in the original one. so the flux-weakening ability is stronger. Considering the manufacturing process, the actual rotor structure is shown in Fig. 7 the shape of the radially magnetized PMs are rectangular. The stator, rotor and test platforms are shown in Fig. 8. Fig. 9 compares the experimental and FEA results. Fig. 9 (a) is the comparison between the experimental and analytical result about back EMF without load, Fig. 9 (b) shows the C-T relationship of the motors with original rotor and optimized rotor under MTPA control and the results present a high consistency and error is less than 2%. But the error is large in the relationship between experiment and analysis under id=0 control due to the reluctance torque is fully utilized with the large current. IV. CONCLUSION In this paper, a new flux-concentrating rotor for PM machines is designed to combing spoke-type IPM rotor and Halbach PM array The size parameters of PM of the new rotor are optimized by the response surface method for the maximum air-gap flux density and torque capacity considering the constant pole-arc coefficient and the volume of PM.The 72-slot/16-pole PM motor with the optimized rotor is developed by 2-D FEA and experiments. The experimental result the air-gap flux density is increased by 6.7%and the torque is increased by 7.5%.

**References:**

[1]Igor P. Wiltuschnig, Paulo R. Eckert, David G. Dorrell ,“A Study of the influence of quasi-Halbach arrays on a torus machine,” IEEE Trans. Magn.,.vol. 52, no. 7, pp. 1325-1328. July 2016. [2]M. M. Rahman, K.-T. Kim, and J. Hur, “Design and optimization of neodymium-free spoke-type motor with segmented wing-shaped PM,” IEEE Trans. Magn., vol. 50, no. 2, Feb. 2014, Art. ID 7021404. [3]Byeong-Hwa Lee;Jung-Pyo Hong;Jung-Ho Lee ,“Optimum Design Criteria for Maximum Torque and Efficiency of a Line-Start Permanent-Magnet Motor Using Response Surface Methodology and Finite Element Method,” IEEE Trans. Magn. .vol. 48, no. 2, pp. 863-866. Feb. 2012.

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/71f9b4e9aeefd5ad0fbdc3e6fb75fafa.jpg)

![](https://assets.underline.io/uploads/markdown_image/1/image/d80d0be36065b68335586a8a185383f6.jpg)

 Design and Optimization of New Flux-Concentrating Rotors combining Halbach PM array and spoke-type IPM for PMSM

I. Introduction There has been a growing demand for permanent magnet electric machines with high torque and power density per unit current. Furthermore, electric machines are increasingly required to have high efficiency by regulation of energy efficiency. Many optimization algorithms are being studied to achieve high efficiency of IPMSM. Practically, manufactured electric machines can’t expect high performance for all products due to material deviations and dimensional tolerances. Therefore, we must consider the problems that occur while manufacturing to perform an optimization algorithm. However, considering the manufacturing tolerances along with many of the design variables of IPMSM requires a lot of time and cost. Therefore, we propose to apply multiple regression method to improve the performance of the conventional particle swarm optimization (PSO) algorithm. In addition, we can estimate the function formula for the design variables and derive the global solution by using estimated function formula. The convergence speed of the conventional PSO algorithm is improved by using the information of the solution obtained by performing the multiple regression. In addition, we will ensure the robustness of the design by estimating the appropriate regression equation for each design space. II. Particle Swarm Optimization with Multiple Regression Method Since the PSO algorithm finds the optimal solution using the information of each individual, the optimal solution can be found only when the particle moves. In this paper, multiple regression analysis is performed on the accumulation data obtained through the PSO algorithm. Then, an approximate solution can be obtained by using multiple regression. PSO with multiple regression algorithm has the same main loop as conventional PSO algorithm. First, initial particles are created with random locations, and update the velocity and position to find the global solution every iteration. If the amount of data for multiple regression is considered to be sufficient, then regression analysis is performed. We iterate this process until we reach the defined objective function. We selected seven numerical test functions to validate the proposed algorithm. The data obtained from performing the PSO algorithm is used to construct the Multiple Regression Model. Multiple regression is performed within the overall design space, and the result is a rough model. However, it is not a problem because it is used to improve the convergence speed. III. Robust Design Optimization of IPMSM A robust solution is the solution that has the lowest variability of the objective function in the design space that meets the requirements of the objective function. Constraints are needed to solve the problem of finding a robust solution. Constraints are obtained through multiple regression analysis. In particular, in order to consider the variability of design variables, the surrogate model must be configured in all solutions to find a robust solution. However, multiple regression analysis over the entire design space can improve the convergence of the algorithm, but it is challenging to construct an accurate surrogate model. Therefore, we used a method of reducing the design space into limited space that can be explained by multiple regression. A detailed procedure to find a robust solution is as follows. First, perform the modified PSO algorithm to save the data of the particles and find the first solution. When finding a solution, convergence speed is improved by performing multiple regression in the total design space. Second, Construct a multiple regression model around the first solution. The construct region of the multiple regression model limits the region to the adjusted coefficient of determination above 0.95. Multiple regression models verify the variability of the objective function. Third, the first and second process is repeated except for the multiple regression areas obtained in the second section. This procedure is repeated to the extent that it does not violate the constraints of the objective function. We have applied the proposed algorithm to the robust design optimization of IPMSM. We considered the variability of the objective function according to manufacturing tolerances. Target model was selected as IPMSM with eight poles and 48 slots as the target model for driving a vehicle. The constraint condition is over 165 Nm of average torque at 3000 RPM and objective function is minimization of the torque ripple. The central case torque and central case torque ripple are the average torque and torque ripple at the reference point. The worst case torque and worst case torque ripple are the average torque and torque ripple at the design variable values that represent the largest variation within the manufacturing tolerances. As a result of optimization using proposed algorithm, global solution and robust solution were found as shown in Fig. 1, 2. The robust solution has smaller torque ripple variability in Fig. 1. however, global solutions are large variability when considering manufacturing tolerance in Fig. 2. The engineers have the problem of selecting a global solution with higher torque at a similar torque ripple level without considering manufacturing tolerances. However, considering manufacturing tolerances, engineers can choose a robust solution with low variability.

**References:**

B. D. S. G. Vidanalage, M. S. Toulabi, S. Filizadeh, IEEE Transactions on Energy Conversion., vol. 33, p.1547 (2018) C. Ma, L. Qu, IEEE Transactions on Energy Conversion., vol. 30, p.1144 (2015) J. Ou, Y. Liu, R. Qu, IEEE Transactions on Industrial Electronics., vol. 65, p.3772 (2018) J. H. Lee, J. Song, D. Kim, IEEE Transactions on Industrial Electronics., vol. 65, p.1791 (2018) B. Ma, G. Lei, C. Liu, IEEE Transactions on Magnetics., vol. 54, p.8102404 (2018)

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/908113e9a2e7b790164fee363a61922d.jpg)
Average torque variability of robust solution

![](https://assets.underline.io/uploads/markdown_image/1/image/65a881435d964a839699329924c44f98.jpg)
Average torque variability of global solution

 Robust Design Optimization Based on Separated Surrogate Model Method for IPMSM

In this paper, a novel global search algorithm tornado optimization (TO) and its hybridization with pattern search method (PSM) are proposed to solve the multimodal optimization problem. The TO enhances capability of global searching with two strategies, intensification strategy and diversification strategy. The hybridization with PSM supports the exact and fast convergence on adjacent optima from the points that are discovered in global search using TO. Outstanding performance of the proposed algorithm is verified by comparison with niching genetic algorithm at the test function and the algorithm is applied to the optimal design of the interior permanent magnet synchronous motor for electric vehicle. I. Introduction To analyze the accurate characteristics of the electric motors for electric vehicle (EV), finite element method (FEM) analysis, which takes a lot of time, is required because of the nonlinear magnetic saturation of the motors [1]. Computational problem of FEM becomes worse when it is applied to the multimodal optimization problem, such as an optimal design of motor. To solve this problem, several algorithms such as niching genetic algorithm (NGA) have been studied. The NGA is an improved algorithm of GA for the multimodal function [2]. However, the NGA still have a problem with calculation time because it takes a lot of function calls to converge to the accurate peak point [3]. To reduce the number of function calls and to enhance the convergence characteristic, this paper proposes the tornado optimization (TO) hybridized with pattern search method (PSM). The TO is an algorithm to search overall area and to designate some points as candidates for several optima. The performance of the algorithm is maximized with the usage of the PSM, which is a kind of deterministic algorithm. II. TORNADO OPTIMIZATION WITH PSM In hybrid algorithm for the multimodal function, adopting strategy and role sharing of each search algorithm are important. Global search algorithm adopts diversification strategy to search overall area and to find undiscovered candidates. On the other hand, local search algorithm converges candidates to each adjacent optimum quickly and precisely by adopting intensification strategy [4]. In TO, the processes of being sucked into tornadoes and searching the opposite side by Zone 1 are employed as intensification strategy. On the other hand, thunder strike of Zone 2 is used as a diversification strategy to prevent particles from being trapped in a local optimum. First, the superior initial particles are generated by Latin hypercube sampling (LHS), which scatters samples evenly, and search each location [5]. With their results, tornadoes occur with their own zones. The rest of particles are sucked into tornado. When they enter the zone of tornado, they will be moved in two ways. As seen in Fig. 1, if they enter in Zone 1 which is the half of the zone, they are swept away by tornado and fly to the opposite side of the tornado. On the other hand, if they enter in Zone 2 which is the inside of the zone, they will be scattered outside the tornado, just as thunder strikes from inside to outside of the real tornado. They will continue to be sucked into the tornado at the relocated area. And then, the searched tornado extends its zone to the location of the absorbed particle. At the same time, the proportion of Zone 2 increases. Zone 1 reduces the number of function calls in advance before the PSM to be executed and helps to find optima hiding beyond the tornado. Zone 2 prevents meaningless re-discovery of discovered tornado and it helps to find new candidates that has not been found yet. In addition, expansion of Zone 2 increases the weight of thunder strikes instead of being swept away. It means that the proportion of diversification strategy increases with each iteration of the search. After repeating this process, new tornadoes occur in potentially optimal candidates recorded during the process. After enough tornadoes are created by TO, those candidates are converged to adjacent optimum by the PSM. The hybridization with the PSM makes TO to find the accurate peak point without being trapped to the local optima [3]. III. VERIFICATION OF PROPOSED ALGORITHM To verify the performance of the proposed algorithm, TO and NGA are applied to the mathematical test function. Test was carried out 100 times with each algorithm. As shown in Fig. 2, there is no doubt that the TO shows outstanding performance than the NGA. The Dstavg means the average distance from the found peak point to the actual peak point. When the function calls of TO and NGA are 1108 and 1380, respectively, the Dstavg of the TO is 72% less than the result of NGA. The TO consumed 19% less function calls than the NGA, even though the accuracy was higher. The application of the TO to the optimal design of the interior permanent magnet synchronous motor (IPMSM) will be covered in full paper. IV. CONCLUSION This paper introduces a novel hybrid algorithm on the purpose of both reducing function call and increasing accuracy of multimodal problem. The outstanding performance of the TO was confirmed through a test function, compared with NGA. The TO was applied to the optimal design for minimizing the cogging torque of IPMSM for EV and successfully derived the optimum model. Acknowledgement This work was supported by a National Research Foundation of Korea (NRF) grant funded by the Korean government (Ministry of Science and ICT) (No. 2019R1F1A1061132).

**References:**

[1] M. Caruso, A. O. Di Tommaso, L. Ferraris, 2017 Twelfth International Conference on Ecological Vehicles and Renewable Energies (EVER)(2017) [2] B.-K. Son, G.-J. Park, J.-W. Kim, IEEE Trans. Magnetics., vol.52, No.3 (2016) [3] Emanuele Dilettoso, Nunzio Salerno, IEEE Trans. Magnetics., vol.42, No.4, pp.1203-1206 (2006) [4] K. Lei, F. Wang, Y. Qiu, Proc. Auto. Decentralized Syst. (ISADS), pp. 513–519 (2005) [5] Y.-R. Kang, J.-C. Son, D.-K. Lim, IEEE Access., vol.8, pp. 117405-117412 (2020)

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/64e0812025167577a83918e641bb885b.jpg)
Fig 1. Concept of Tornado Optimization

![](https://assets.underline.io/uploads/markdown_image/1/image/eab6d8aa0606079f102d34d7340fde75.jpg)
Fig 2. Result of test function. (a) TO, (b) NGA

 Tornado Optimization with Pattern Search Method for Optimal Design of IPMSM

This paper focuses on designing a optimal novel structure for high power densitiy of a concentric winding integer slots permanent magnet synchronous motor(PMSM). This research shows the motor design possibility with a higher power density than a typical distributed winding model, within the same size. If the number of slots per phase per pole(NSPP) of PMSM is larger than 1, distributed windings are applied mostly. The distributed winding is a winding method that the number of turns is distributed into each slot according to NSPP. The distributed winding can make a sinusoidal magnetomotive force(MMF) waveform. Also, harmonics can be reduced. Distributed Windings can be classified into lap winding and concentric winding according to the number of pitch of each coil. In lap winding, the number of pitch of each coil is same. On the other hand, the number of pitch of each coil is different maintaining same center of each coil in concentric winding. The distinction between the winding methods is shown in Fig. 1. The concentric winding is divided into identical concentric winding and nonidentical concentric winding according to the number of each coil. Identical concentric winding is that the number of turns in each coil is identical. Nonidentical concentric winding, maintaining a sinusoidal waveform even if the number of turns in each coil is nonidentical, has the advantage of adjusting the number of turns in each coil can obtain a higher winding factor than the lap winding[1-3]. The winding factor is the determinant by the MMF at the air gap and is proportional to the fundamental amplitude of back electromotive force (BEMF)[4-5]. Consequently, the nonidentical concentric winding has potential to enhance a power density of motors according to winding factor. In general, all slots in the stator are the identical shape, so if each coil has a different number of turns, slot fill factors vary from each slot. As a result of slot fill factor differences, some slots are inefficient in utilizing space. It affects the BEMF's fundamental wave amplitude as well, which the output performance of the nonidentical concentric winding motors could be worse than the identical concentric winding one. To improve this problem, we need to design the structure of stator with optimization algorithm to make each slot shape efficiently. In this paper, we study about nonidentical slotted concentric winding in the following ways. A concentric winding integer slots model is selected with 3-phase 4-pole 24-slots PMSM that NSPP is 2. the winding factors of identical concentric winding type and various nonidentical concentric winding types are calculated and compared in case of the number of turns distribution for nonidentical one. nonidentical concentric winding is selected to the winding factor higher than identical concentric winding. For improving the above mentioned problem, a new structure of the stator, nonidentical slotted concentric winding, is optimized and designed. the winding factors of identical concentric winding type and various nonidentical concentric winding types are calculated and compared in case of the number of turns distribution for nonidentical one. Variable shown in Fig. 2. for algorithm is composed of factors which control a slot area. The purpose of optimization is power density and performance. Therefore, objective functions are average torque and torque ripple, and they are minimized as the following two objective function equations. obj1= 1/Tavg obj2= Te, ripple = (Te,max-Te,min)/Te * 100 Te, avg is the average torque and Te, ripple is the torque ripple ratio. Te, max and Te, min are the maximum and minimum of the instantaneous torque, respectively. Through a finite element analysis(FEA) of designed models, we ultimately compared the performance of motor models according to identical concentric winding, nonidentical concentric winding and nonidentical slotted concentric winding. Finally, we propose a novel structure of stator with each different slot areas and nonidentical concentric winding. it is called nonidentical slotted concentric winding. This research shows the motor design possibility with a higher power density than a typical distributed winding model, within the same size.

**References:**

[1] T.A.Lipo, "Introduction to AC Machine Design", University of Wisconsin, pp. 53-82, 2004. [2] M. M. Liwschitz, "Distribution Factors and Pitch Factors of the Harmonics of a Fractional-Slot Winding," in Transactions of the American Institute of Electrical Engineers, vol. 62, no. 10, pp. 664-666, Oct. 1943. [3] G. R. Slemon, A. Straughen “Electric Machines” University of Toronto, pp. 369-383, 1980. [4] Wen Ouyang, A. El-Antably, Surong Huang and T. A. Lipo, "Analysis of optimal stator concentric winding patterns design," 2005 International Conference on Electrical Machines and Systems, Nanjing, vol. 1, pp. 94-98, 2005. [5] J.R. Hendershot, T.J.E.Miller, “Design of Brushless Permanent-Magnet Machines” Magna Physics Publishing & Oxford University Press, pp. 180-181, 2010 [6] J. H. Lee, J. Kim, J. Song, Y. Kim and S. Jung, "A Novel Memetic Algorithm Using Modified Particle Swarm Optimization and Mesh Adaptive Direct Search for PMSM Design," in IEEE Transactions on Magnetics, vol. 52, no. 3, pp. 1-4, March, 2016.

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/3cc92d2bebd6cacb0fb22b847ccaf7b1.jpg)

![](https://assets.underline.io/uploads/markdown_image/1/image/3b059057e1fd0dc89bf8ef5b24adb7fe.jpg)

 A Novel Structure Design about Nonidentical Slotted Concentric Winding

In this paper, subdivided kriging multi objective optimization (SKMOO) is proposed for the optimal design of interior permanent magnet synchronous motor (IPMSM). The SKMOO with inverse searching can obtain uniform and accurate pareto front set with reduced computation cost compared to conventional algorithms. In the SKMOO algorithm, Fill blank method is applied to prevent premature convergence. In addition, the subdivided kriging grid is proposed to make well distributed and precise pareto front set. Superior performance of the SKMOO is confirmed the by compared conventional MOO algorithms with the test function and applied it to the optimal design of IPMSM for electric vehicle. I. Introduction The interior permanent magnet synchronous motor (IPMSM) has the torque caused by the magnet and reluctance torque. Therefore, the IPMSM torque greater than surface mounted PMSM [1]. However, such a permanent magnet type motor has a disadvantage of high cogging torque and torque ripple which causes in the motor’s noise, vibration, and efficiency reduction. This is important to electric vehicle for comfortable driving [2]. Therefore, we aim to design the IPMSM with reduced cogging torque and torque ripple. Finite element method (FEM) analysis is required since the accurate analysis is difficult by using simple mathematical models due to the nonlinear characteristic of the IPMSM [3]. However, the actual motor design using FEM analysis should meet several requirements. It takes a lot of time and effort to simultaneously consider multi objective functions in motor design that require FEM interpretation [4],[5]. To solve this problem, a lot of studies are in progress [6],[7]. The most widely used algorithms are multi objective particle swarm optimization (MOPSO) and non-dominated sorting genetic algorithm-II (NSGA-II). However, the MOPSO and NSGA-II still need many function calls to get the accurate and well distributed pareto front set. Therefore, subdivided kriging multi objective optimization (SKMOO) is proposed which applies the surrogate model to reduce function calls. Outstanding performance of the proposed algorithm is verifyverified through comparison of results with the MOPSO and NSGA-II. And the optimal design was derived by applying the SKMOO to design of the IPMSM with reduced cogging torque and torque ripple. II. Proposed Algorithm MOO performance can be determined based on the accuracy, uniform distribution, and number of function calls while consisting the SKMOO uses the following methods to improve algorithm performance. Pareto front set are extracted from samples using non-dominated sorting and sorted one objective function. The Euclidean distance of adjacent solutions is calculated to obtain a uniform pareto front set and predicted solution is added directly between the two solutions with the longest distance. The spline interpolation is used to increase the accuracy of the solution as shown in Fig. 1. Design variables are inversely searched to generate a real solution. The most similar combination of design variables is found by comparing the objective function value of the surrogate model with the predicted value and the function call is executed to the explored design variables. Fill blank method is used to prevent premature convergence. This method is greatly efficient because it does not overlap the area already explored. When the algorithm converges after the set number of iterations, subdivided kriging grid is proposed to improve performance. The same methods are used to adds solutions, but more accurate solutions can be added because the kriging grid is finely divided. Also, the solution with the closest distance is deleted, if the distance is less than a set distance and if it is more than a set distance, a new solution is added as the intermediate point of before and after solution to obtain more uniform distribution. Fig. 1 shows that the solution is directly added in objective function area. However, the solution is added only at the grid points. Therefore, there is a difference between predicted and real solution. In this case, solutions are uneven and inaccurate. Subdivided grid is a method to solve this problem by dividing kriging grid finely as shown in right of Fig. 1. III. verification of proposed algorithm The SKMOO, the NSGA-II and MOPSO are applied to the test functions. GD indicates how similar the pareto set to the real pareto set and SP indicates how uniformly the solution is distributed. GD and SP value mean better performance if values close to zero. Table. I summarize the results and shows that the SKMOO has superior performance on GD, SP, and the number of function calls than NSGA-II and MOPSO. IV. Conclusion In this paper, SKMOO based on surrogate model is presented to solve MOO problems. The SKMOO can reduce computation cost and show superior performances in the MOO problems. The proposed algorithm is validated by comparing with MOPSO and NSGA-II. Finally, we applied the SKMOO to the optimal design of IPMSM. Taguchi method is used to select variables among several variables that are sensitive to the objective function. The application of the SKMOO to optimal design of the IPMSM will be covered in full paper. V. Acknowledgment This research was supported by Korea Electrotechnology Research Institute (KERI) Primary research program through the National Research Council of Science & Technology (NST) funded by the Ministry of Science and ICT (MSIT) (No. 21A01082).

**References:**

[1] Guo Hong, Tian Wei, and Xiaofeng Ding, IEEE Access, Vol. 6, pp. 23568-23581, (2018). [2] Young-Rok Kang, Ji-Chang Son, and Dong-Kuk Lim, IEEE Access, Vol. 8, (2020). [3] M. Caruso, A. O. Di Tommaso, L. Ferraris, R. Miceli, F. Viola,2017 Twelfth International Conference on Ecological Vehicles and Renewable Energies (EVER)., (2017) [4] Min Li, Fabien Gabriel, Maria Alkadri, IEEE Transactions on Magnetics, Vol. 52, (2016) [5] Siguang An, Shiyou Yang, Osama A. Mohammed, IEEE Transactions on Magnetics, Vol. 54, (2018) [6] Khalil Alkebsi, Wenli Du, IEEE Access., Vol. 8, (2020) [7] Vikas Palakonda, Rammohan Mallipeddi, IEEE Access, Vol. 8, (2020)

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/de072cbbc28eead8bddd09747917a1f9.jpg)
Fig.1 Conceptual diagram of SKMOO

![](https://assets.underline.io/uploads/markdown_image/1/image/51d55f48cade1a8cf8b2cc9ec7346562.jpg)
Table. I Test function results

 Optimal Design of IPMSM for EV using Subdivided Kriging Multi Objective Optimization

I. Introduction Due to the outstanding advantages, such as simple structure, high torque density and excellent dynamic performance, permanent magnet motors (PMMs) have been widely applied over the past four decades. Yet, the implementation of PMSM is limited in servo system with high demand for low torque ripple. In this background, the research on how to reduce torque ripple has become a hot spot for many scholars. In order to reduce the high torque ripple of PMM, different methods have been put forward, which can be concluded as rotor skewing, auxiliary slots or teeth, magnet shaping and flux barriers etc. Among these methods, magnet shaping has been developed as an effective method. A rotor consists of several different PM segments arranged in the axial direction is proposed in [1], which is validated by the effectiveness of reducing torque ripple. However, the shape of PM is random and lack of pertinence. By injecting the optimal third harmonic into PM shape, torque performance can be improved with analysis of airgap flux density in [2]. However, only one harmonic is considered into magnet design process. Hence, the influence of different harmonic injection methods on PM shape design is investigated comprehensive in [3], in which various magnet shape is explored to reduce torque ripple. However, the consequents of different initial phase angles of injected harmonics are neglected. In this paper, a magnet design method considering multi-harmonic injection is proposed, which selects an inset-permanent-magnet motor (Inset-PMM) as research subjective. The key of the study is to extensively consider the factor of harmonic phase angle into the design process. Then, based on different harmonic combination and phase angle consideration, two feasible rotor topologies are proposed. Based on the reasonable design, the torque ripple can be reduced effectively, while the motor can also offer a relatively high output torque. II. Motor features and Analysis Inset-PMMs with two optimal phase angle of injected harmonics and four optimal phase angle of injected harmonics are shown in Fig. 1. The principle of initial phase angle of injected harmonics into rotor design process is listed in Fig. 2, which confirms the effectiveness of proposed method. The relationships between torque ripple and different initial phase angle of injected harmonics are given in Fig. 3, meanwhile, the Inset-PMM without phase angle of injected harmonics, considering two optimal phase angle of injected harmonics and four optimal phase angle of injected harmonics are proposed. The comparisons of the initial phase angle of the proposed Inset-PMM are listed in Table I. The flux density, harmonic spectrum, magnetic field distributions and flux distributions of three Inset-PMM are shown in Fig. 4 and Fig. 5, which can be seen that the amplitudes of injected harmonics are almost same. The comparisons of torque performance of three Inset-PMM are presented in Fig. 6, which validates the effectiveness of the proposed method on reducing torque ripple. III. Conclusion Based on the analysis above, the initial phase angle of injected harmonics into rotor design process on torque performance in Inset-PMMs are investigated. Torque ripple is reduced from 37.9% to 13.7% when the factor of harmonic phase angle is considered, the results not only verify the effectiveness of the proposed method on reducing torque ripple, but also prove the validity of the Inset-PMM which can realize the low-cost design and enhance torque capability simultaneously, providing the new orientation for reducing the torque ripple. More detailed electromagnetic performance analysis and FEA results will be presented in the full paper.

**References:**

[1] W. Zhao, T. A. Lipo and B. Kwon, “Material-Efficient Permanent-Magnet Shape for Torque Pulsation Minimization in SPM Motors for Automotive Applications,” IEEE Transactions on Ind Electron., vol. 61, no. 10, pp. 5779-5787, Oct. 2014. [2] K. Wang, Z. Q. Zhu and G. Ombach, "Torque Enhancement of Surface-Mounted Permanent Magnet Machine Using Third-Order Harmonic," IEEE Trans Magn., vol. 50, no. 3, pp. 104-113, March 2014. [3] Y. Zeng, M. Cheng, G. Liu and W. Zhao, "Effects of Magnet Shape on Torque Capability of Surface-Mounted Permanent Magnet Machine for Servo Applications," IEEE Transactions on Ind Electron., vol. 67, no. 4, pp. 2977-2990, April 2020.

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/387bb906ffe71557d235a7258b6cb676.jpg)

![](https://assets.underline.io/uploads/markdown_image/1/image/d920978404cf62639df1c861c4abc923.jpg)

 Design and Analysis of a Low Torque Ripple Inset-Permanent-Magnet Motor Considering Multi-Harmonic Injection

In this paper, internal division point genetic algorithm (IDP-GA) was proposed to lessen the computational burden of multi-variable multi-objective optimization problem using finite element analysis such as optimal design of electric vehicles. The IDP-GA could consider various objectives with normalized weighted sum method and could reduce the number of function calls with novel crossover strategy and vector-based pattern search method. The superiority of the proposed algorithm was verified by comparing performances with conventional optimization method at the test functions. Finally, the applicability of the IDP-GA in practical electric machine design was verified by successfully deriving an improved design of electric bicycle propulsion motor. I. Introduction Electric bicycles (EBs) are getting more attention in many countries for their convenience, long travelling distance and environment friendly features, therefore, EB propulsion motor with low cost and high performance is needed [1]. When designing a motor for EBs, high torque density and improved riding impression is required [2]. Permanent magnet assisted synchronous reluctance motor (PMa-SynRM) has advantage of high torque density as it utilizes both magnet torque and reluctance torque. Also, with the usage of inexpensive ferrite magnet, PMa-SynRM is competitive for the manufacturing cost. Hence, PMa-SynRM is selected for the EB propulsion motor. To relieve the noise and vibration of the motor, various performances such as cogging torque, and torque ripple should be considered, and those characteristics can be adjusted with optimization of the variables related to the structure of the motor [3], [4]. Therefore, optimal design of the PMa-SynRM can be defined as multi-objective multi-variable (MVMO) problem. Finite element analysis (FEA) is required for accurate analysis of electric machines. However, the FEA has the disadvantage of huge computational cost, and such burden becomes more worse when FEA is applied to optimization of MVMO problems. Therefore, this paper proposes the novel optimization algorithm that reduces the number of function calls. II. Proposed Algorithm GA is one of the heuristic algorithms and can successfully converges to the global solution [5]. However, the GA has problem with requiring many function calls to enhance the converge rate and cannot consider various objective functions at once. The IDP-GA reduces function calls with novel crossover strategy and VPSM. Also, multi-objective can be considered with NWSM. The NWSM is one of the multi-objective methods and can simply reduce several objectives to one objective. The total fitness is calculated by multiplying each normalized fitness by assigned weight and adding them all together. The designer can perform optimization by giving the desired weight to each objective with the NWSM. Conventional GA converges to global solution by selecting superior entities as parents and generating crossover on the middle of selected two parents. To enhance the convergence speed, the proposed algorithm applies internal division point (IDP) concept. The IDP literally generates crossover not on the middle point, but on the point divided by objective values of two parents. Fig. 1 shows the difference of conventional approach and IDP. Conventional crossover, which is blue dot, is generated on the center of two parents. However, the location of the red dot is determined considering objective values of parents. With the usage of IDP, superior crossover can be obtained. At the end of the algorithm, if the populations are regarded as gather near the global solution, VPSM is executed on the estimated solution to find the actual global solution, preventing unnecessary function calls, and enhancing convergence characteristic. The VPSM has pattern move and exploration move. On the pattern move step, unit direction vector is determined using objective value of infinitesimal displacement of starting point. Fig. 2(a) shows an example of calculating direction vector. Assuming that the problem region is continuous, with two function calls on the point DX and DY, the values of DX' and DY' can be interpolated. Then, increasing direction of each variable is determined, and the unit direction vector can be calculated. Secondly, as shown in Fig. 2(b), the exploration move is performed on the best particle with preset initial moving distance to determined direction, until the objective value decreases. If the value of next point decreases, pattern move is repeated on the previous point and exploration move with reduced moving distance. To verify the performance, the IDP-GA and conventional GA was applied to two mathematical test functions. The number of function calls of IDP-GA was decreased as 56% and 64% of GA in two tests, and the convergence rate was improved as well. The detailed explain of the IDP-GA, test functions, and application to optimal design of PMa-SynRM for EB will be described in a full paper. III. Conclusion This paper introduces IDP-GA to relieve the computational burden of the MVMO optimization problem applying FEA. The proposed algorithm was validated by applying to mathematical test functions and comparing performance with conventional algorithm. The applicability of the optimization of actual electric machines was also verified by deriving design with improved performances, when the IDP-GA is applied to optimal design of PMa-SynRM for EBs.

**References:**

[1] Y. Li, J. Hu, F. Chen, S. Liu, Z. Yan, and Z. He, IEEE Trans. Power Electronics, vol. 33, no. 10, pp. 8226-8230 (2018) [2] D. Lim, Y. Cho, J. Ro, S. Jung and H. Jung, IEEE Trans. Magnetics, vol. 52, no. 3, pp. 1-4 (2016) [3] Ji-Chang Son, Young-Rok Kang, and Dong-Kuk Lim, Energies, 13, 3395 (2020) [4] Young-Rok Kang, Ji-Chang Son, and Dong-Kuk Lim, IEEE Access, vol. 8, pp. 117405-117412 (2020) [5] K. Lee, H. Oh, S. Jung and Y. Chung, IEEE Trans. Magnetics, vol. 55, no. 6, pp. 1-4 (2019)

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/b8e7214a83d0a219e82755f99840b3c0.jpg)
Fig. 1. Comparison of conventional crossover and IDP.

![](https://assets.underline.io/uploads/markdown_image/1/image/c333d6a9764d1938e0f61b865a45b939.jpg)
Fig. 2. Conceptual schematic of VPSM. (a) Pattern move. (b) Exploration move.

 Multi-Variable Multi-Objective Optimization Algorithm for Optimal Design of PMa-SynRM for Electric Bicycle

Introduction - Recently, as the demand for robots and the supply ratio of renewable energy have increased, the usage of electric motors and generators has increased. Therefore, sensors required for detecting the position of the rotor has also increased. Of these, the safety window motor for preventing accidents in which an automobile window is pinched is based on the principle that a ring type bond magnet is attached to the lower part of the rotor of the motor to detect the magnetic flux and detect the position. At this time, a dead zone, which is a section where the magnetic flux is very weak, is generated between the N pole and the S pole of the ring-type bond magnet and the hall sensor does not detect the magnetic flux in this section When the range of this dead zone is large, the region where the magnetic flux can be detected is limited, so that accurate position detection is impossible. In this paper, we provide the optimum model of a polar anisotropic molding yoke with a magnetic flux-concentrated structure for the dead zone reduction plan of ring type bond magnets for accurate control of electric motors and generators. Equation - In order to concentrate the magnetic flux generated from the permanent magnet of the molding yoke, a structure having a high magnetic flux concentration factor is required. This indicates that the more magnetic flux passing through the magnet than the air gap, the more the magnetic flux is concentrated. Therefore, if the magnetic flux-concentrated structure is used, when a magnetic field is applied, the air gap flux density can be higher than that of the conventional model, and the dead zone of the ring type bond magnet can be reduced. Analysis & Result – In the case of the conventional model, the structure is such that a magnetic field is directly applied to a very different ring type bond magnet with a permanent magnet. Therefore, the magnetic flux does not penetrate deeply into the magnet, and a large Dead Zone is generated. To reduce this dead zone, this paper designed a model that could be molded with a high air gap flux density by inserting a magnetic flux concentration core. The thickness of the inner and outer teeth and the direction of the magnetic field applied to the permanent magnet were used as the design variable. In this paper, the results were analyzed through finite element analysis. The design variable of inner and outer teeth thickness was set range of 0mm to 12mm and the angle degree of the magnetic field applied to the permanent magnet was set the range of 0° to 90°. a) Air gap flux density according to magnetization angle When the inner teeth thickness is 11mm and the outer teeth thickness is 0mm, the change in air gap flux density according to the magnetization direction was analyzed. When the angle of the magnetization direction is too large or too small, the magnetic flux leaks to the permanent magnet and the outside. So it can be confirmed that the maximum air gap flux density is generated around 40°. b) Air gap flux density according to the thickness of the inner teeth When the angle of the magnetization direction is 40° and the thickness of the outer teeth is 0mm, the change in air gap flux density according to the thickness of the inner teeth was analyzed. Analyzing the results at 12.5° where the dead zone occurs in a ring type bond magnet, the magnetic flux leaks to the adjacent core when the thickness of the inner teeth is too large. So it can be confirmed that the magnetic flux is most concentrated on the pole of the ring type bond magnet when the thickness of the inner teeth is 11mm. c) Air gap flux density according to the thickness of the outer teeth When the angle of the magnetization direction is 40° and the thickness of the inner teeth is 11mm, the change in air gap flux density according to the thickness of the outer teeth was analyzed. It is confirmed that when the thickness of the outer teeth is 0mm, the magnetic flux leaked to the outside is minimized and the maximum air gap flux density is generated. Therefore, an optimum model that can be molded to the highest air gap flux density was obtained. Comparing the waveforms of the air gap flux density of the proposed model and the conventional model, the proposed model increased the air gap magnetic flux density by 23% compared to the conventional model. As a result, dead zone decreased by about 22%. Conclusion – In this paper, a flux-concentrated core was inserted to reduce the dead zone range of the ring type bond magnet. As a result, it is possible to reach a higher air gap flux density than the conventional model. Finally, the validity of this paper was verified through the finite element analysis method. Based on this, it is expected that the market share of the safety window can be expanded, and further, the technological competitiveness of domestic sensor industries can be improved. Acknowledgment - This work was supported by the National Research Foundation of Korea(NRF) grant funded by the Korea government(MSIT). (No. 2020R1A2C1013724) NRF - 2020R1A2C1013724

**References:**

H. J. Kim et al, IEEE Trans. Magn., vol. 46, no. 6, pp. 2314-2317, Jun. 2010

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/9cab3636cee4a5e0db5c3e684990f0cb.jpg)
Magnetic Flux Density of Conventional & Proposed Model

![](https://assets.underline.io/uploads/markdown_image/1/image/eb542ad3264d6f8c7f1b0dc4dff9549a.jpg)
Waveform Comparison of Conventional & Proposed Model

 Research on Polar Anisotropic Molding Yoke Shape to Reduce Dead Zone of Ring Type Bond Magnets

Introduction In the field of wireless charging for electric vehicles, the limitation of vehicle-mounted space is a major constraint for the on-board device. It is impossible to ask the drivers to park their vehicle at the same exact position and drivers often park the car with more than a few centimetres of misalignment [1]. Currently, unipolar coils are widely used; these have an upper limit for the coupling coefficient k. With a certain secondary coil size and air gap, a maximum coupling coefficient of the coupler at the aligned position can be estimated using a Dagum distribution [2]. However, the horizontal misalignment has a significant impact on the coupling coefficient, which makes it difficult to tune the resonance of the compensation. With the decrease in k, the primary inverter output may operate in an overcurrent mode with series-series compensation. This is a challenge for the power semiconductors. At the same time, the misalignment will affect the power transfer capability and efficiency. The system performance, within reasonable misalignment, should be robust which will make the system usable for practical charging systems. Proposed Primary Coupler In order to enhance the power transfer capability with a relatively large misalignment, this work investigates a unipolar transmitter coil and its robust optimization design process. Several design parameters of the coupler are shown in Fig. 1, including the pitches, the length and width of the coil, the types of Litz wire. The coupler is analysed and verified with finite element (FE) modelling. Earlier research has shown that by reducing the main coil inductance and connecting with the assistive coil in series, a superposition effect in the magnetic field can be produced to a certain extent. By optimizing the design parameters, a higher coupling coefficient can be obtained under a specific primary inductance. For example, the analysis results show that by increasing gap B between the Litz wires, the self-inductance can be reduced which effectively enhances the lateral and longitudinal freedom of movement of the magnetic coupler. Reducing the centre distance G between the assistive coils and the main coil can increase the self-inductance linearly, but this will weaken the smoothness of the magnetic field. Analysis and Comparison The magnetic field of the receiver pad is more consistent within the controllable misalignment range, and comparing to [3], all the coils are series connected with one high frequency power source. Fig. 2 shows a comparison of the simulation results of the traditional unipolar coil, which is fully optimized using the Dagum distribution, and the proposed coil proposed in this work. These results are with the same offset condition, same receiver coil, and other external conditions. The X and Y directions are the driving and door-to-door direction respectively. The newly designed emitter coil produces a smoother coupling coefficient over a wider range. Inductive Charging Coupler Robust Design Optimization To achieve less performance variation in the face of misalignment variation caused by the vehicle driver, which is off the control in the design phase, a robust design can be introduced. The misalignment distribution is regarded as the uncertainty parameter in the robust design. The relationship between the coupler design parameter and the coupler circuit parameter is analysed through the FE simulation. The system performance parameters are derived from the circuit. The primary inverter output current, the output power and efficiency of the inductive power transfer system are combined with the distribution of the misalignment in the robust design optimization to form the robust design object. The coupler is optimized using the robust design object. Conclusion and Future work In the full paper, the impact of the additional coil on the magnetic field will be analysed extensively. A robust optimization design process based on the coupler design will be described. A prototype will verify the simulation results.

**References:**

[1] S. A. Birrell, D. Wilson, C. P. Yang, “How driver behaviour and parking alignment affects inductive charging systems for electric vehicles,” Transportation Research Part C: Emerging Technologies. [2] L. Li, Z. Wang, F. Gao, “FEA-Assisted Optimization Design of Asymmetric DD Type Structure Magnetic Coupler for Wireless Electric Vehicle Charger,” in 2020 IEEE ISIE, Delft, Netherlands. [3] S. Wang, D. G. Dorrell, Y. Guo, and M.-F. Hsieh, “Inductive Charging Coupler With Assistive Coils,” IEEE Trans. Magn., vol. 52, no. 7, pp. 1–4, 2016.

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/499ca208c0ff9238ebb1b3a17d7a856c.jpg)
Fig. 1 The proposed coupler design and optimization parameters

![](https://assets.underline.io/uploads/markdown_image/1/image/136654cbfc86dce9028b2ea4b405226e.jpg)
Fig. 2. The FEA simulation result comparison of misalignment tolerance a) Traditional unipolar coil, b) Proposed unipolar coil.

 The Electric Vehicle Wireless Charging Application Oriented Coupler Robust Optimization Design With Multiple Series Unipolar Coils

I.Introduction By reducing the over-dependence on current battery technologies, wireless power transfer (WPT) provides a brand-new power-supply solution and changes the traditional style of the energy utilization for electric-driven devices [1]. Accompanied by the increasing utilization of practical application, the diverse requirements of customers and various standards of manufacturers pose challenging issues on this technique, for example, Wireless Power Consortium (WPC) regulates the operating frequency of 87-205 kHz, while AirFuel (AF) alliance holds the operating frequency of 277-357 kHz. Thus, a number of efforts have been focused on the multiple frequencies multiple pickups based WPT, which can be divided into three aspects: 1) single-inverter single-transmitter [2-4], 2) multiple-inverters single-transmitter [5], and 3) single-inverter multiple-transmitters [6]. Due to the simplicity and low cost, single-inverter single-transmitter becomes the optimal choice. However, in [2], energies are transmitted by time division method, which fails to concurrently transfer multiple frequency power, in [3], the distribution of multiple frequency power cannot be adjusted arbitrarily, and in [4], the cross-coupling effect among the pickups will significantly deteriorate the transmission performance. Accordingly, how to randomly transmit multiple frequency power simultaneously without unnecessary coupling becomes a critical technical concern. In this paper, a dual-resonating-frequency compensation topology is adopted and implemented for exciting the decoupled-double D transmitting coils. Concretely speaking, two double D coils are located orthogonally to obtain the effect of magnetically decoupled in both primary side and pickup side. Meanwhile, the proposed system can provide two randomly predefined operating frequencies and the power distribution of two different frequencies can be arbitrarily regulated based on the users’ requirement by adopting the proposed frequency overlay modulation (FOM). Moreover, owing to the nonexistent cross-coupling between decoupled-double D pickup coils and the characteristic of compact integrated, each pickup can maintain the maximum efficiency respectively. II.Methodology Fig.1(a) depicts the WPT system with a single-transmitter for two pickups, where the configuration of dual-resonating-frequency compensation topology which consists of a series-connected LC (Csc Lsl1 Lsl2) tank and a parallel-connected LC (Cpc Lpl) tank is presented. Meanwhile, the parameters design procedure needs massive calculation based on equation (1), thus this paper specifies the operating frequency of 100kHz and 300kHz and each of them satisfies WPC and AF standards, respectively. Then, a simplified lookup table design method for randomly predefined operating frequency is shown in Fig.1(b). Fig. 1(c) testify the effectiveness of the proposed compensation topology. Besides, the proposed FOM can overlap randomly dual frequencies with various magnitudes, where Fig. 1(d) elaborates the proposed FOM with the same magnitude for 100kHz and 300kHz, which can be confirmed by Fig. 1(e). As shown in Fig.1(g), both simulation and practical experiments validate that the proposed system can maintain a high transmission efficiency of over 90% for each pickup of more than 10W. From the perspective of the magnetic, since the y-axis coils and x-axis coils are located orthogonally, the magnetic fluxes are symmetrical regarding the central axes. It indicates that when the current flows in y-axis coils as shown in fig.2 (b), the amount of magnetic flux from y-axis coils that passes through the x-axis equals to that passes out of it. Consequently, the coupling between y-axis coils and x-axis coils is theoretically zero which means the y-axis coils and x-axis coils are decoupled. In this paper, the y-axis coils and x-axis coils are orthogonally compactly placed and series-connected as the transmitting coils. While y-axis coils and x-axis coils are only orthogonally compactly placed in the pickup side and are utilized to energized pickups independently. Specifically speaking, y-axis coils in the transmitting side are fully coupled with the y-axis in the pickup side, and so do x-axis coils. This mechanism can boost the coupling thus promoting the efficiency for both pickups. Fig.1(g) quantitatively analyses the efficiency under the coupling of 0.08 (traditional circular coils) and 0.12 (decoupled-double D Coils), respectively, which indicates a maximum efficiency improvement of 25.2%. III.CONCLUSION This paper proposes an artfully designed dual-resonating-frequency compensation topology with a simplified parameter design procedure. Meanwhile, the proposed FOM regulation method can arbitrarily produce the compound excitation. By adopting decoupled-double D coils, the dual-frequency WPT system holds the merit of zero cross-coupling in the pickup side, thus can promote the transmission efficiency for each pickup. Theoretical analysis, simulation, and experimental results will be given in more detail to verify the feasibility. Acknowledgment This work was supported by a grant (Project No. T23-701/20-R) from the Hong Kong Research Grants Council, Hong Kong Special Administrative Region, China.

**References:**

[1] Z. Zhang, H. Pang, A. Georgiadis and C. Cecati, “Wireless power transfer – an overview,” IEEE Trans. Ind. Electron., vol. 66, no. 2, pp.1044-1058, 2019. [2] Z. Zhang, H. Pang, S. H. K. Eder, and R. Kennel, “Self-balancing virtual impedance for multiple-pickup wireless power transfer,” IEEE Trans. Power Electron., vol. 36, no. 1, pp. 958-967, 2021. [3] W. Liu, K. T. Chau, C. H. T. Lee, C. Jiang, W. Han, and W. H. Lam, “Multi-frequency multi-power one-to-many wireless power transfer system,” IEEE Trans. Magn., vol. 55, no. 7, pp. 8001609: 1-9, 2019. [4] Z. Zhang, Xing. Li, H. Pang, H. Komurcugil, Z. Liang, R. Kennel, “Multiple-frequency resonating compensation for multichannel transmission of wireless power transfer,” IEEE Trans. Power Electron. doi: 10.1109/TPEL.2020.3027916. [5] F. Liu, Y. Yang, Z. Ding, X. Chen, and R. Kennel, “A multifrequency superposition methodology to achieve high efficiency and targeted power distribution for a multiload MCR WPT system,” IEEE Trans. Power Electron., vol. 33, no. 10, pp. 9005-9016, 2018. [6] C. Zhao, D. Costinett. “GaN-based dual-mode wireless power transfer using multifrequency programmed pulse width modulation,” IEEE Trans. Ind. Electron., vol. 64, no. 11, pp. 9165-9176, 2017

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/7874a30861468e51bae13ac3af901dff.png)
Schematic of proposed dual-frequency WPT systems. (a) Configuration. (b) Simplified design scheme and key parameters. (c) Impedance characteristics. (d) Waveforms. (e) FFT analysis. (f)-(g)Characteristics of the transmission power and transmission efficiency.

![](https://assets.underline.io/uploads/markdown_image/1/image/0558ed582c372a01a5f6cd67e05b3733.png)
(a) Structure of decoupled double-D coils. (b) Magnetic field distributions of coils.

Next from INTERMAG 2021

Design and Optimization of New Flux-Concentrating Rotors combining Halbach PM array and spoke-type IPM for PMSM

Similar lecture

Comprehensive Analysis of Nanocrystalline Ribbon Cores in High-Power-Density WPT Pads for Electric Vehicles