Impact of Manufacturing Stresses On Multiple-Rib Synchronous Reluctance Motor Performance

The effects of manufacturing stresses on the overall magnetic properties of electrical steel core applied in the stator and transversally-laminated rotor of Synchronous Reluctance motors are studied. In addition, it is investigated how the -manufacturing-degraded electrical steel core affects the electromagnetic performance of two different Synchronous Reluctance motors. The traditional transversally-laminated rotor of this kind of a machine requires the adoption of mechanical elements (inner and tangential ribs) to maintain its mechanical integrity. Especially, for high-speed applications, the number of rotor ribs must be increased to make the rotor withstand higher centrifugal forces. Lamination manufacturing typically affects the magnetic properties of electrical steel by reducing its permeability at the machined edge. This machining effect is greater when the overall core layer area surrounded by the machined edges is small i.e., the machined surface area ratio to the total core area is high. Considering this, the degradation of the core magnetic properties in the ribs could improve the performance of the machine, and this effect would be more visible when the machine is designed with a higher number of narrow ribs with the adoption of Topology Optimization. The paper aims to propose some guidelines for the simulation of these effects starting from the measurements on defined core samples and then based on the measured results the simulation model should be updated accordingly. The samples are manufactured from several assembled sheets in order to modify the width of the cut area while maintaining the same external geometry.


I. INTRODUCTION
I NTEREST in "green" products, such as hybrid and electric vehicles, is constantly growing due to the emission reduction demand imposed by the European Union for new cars [1], [2]. Electric vehicles can be considered as a possible solution to achieve zero Green House Gas (GHG) emissions as far their power supply comes from sustainable sources, i.e., solar or wind turbine plants and to reduce air pollution, especially in large cities, [3], [4]. In order to achieve the highest performance characteristics of the electric motors within limited space and weight (demanded by the transport sector), motor design solutions with Rare-Earth (RE) magnets are the most used ones in traction applications, [5], [6], but currently, researchers at academy and industry are seeking for RE free motor design solutions or designs with a reduced amount of them [7], [8], [9], both for the increasing trend in the sales of electric vehicles, [10], for the risk of having high and volatile prices of raw materials, and for the vulnerability of many countries regarding the supply of the RE materials, [11]. In this scenario, the Synchronous Reluctance (SynRel) motor could be a valid alternative in mass production of traction motors, especially for those transport segments (e.g., conventional passenger cars or electric trains) that do not require strict size/weight requirements, [12]. SynRel machines have a power density that is comparable to the one of induction motors, but they have lower rotor losses because of synchronous speed operation avoiding slip related losses; these aspects make this type of solution attractive, [13], [14]. Its drawbacks are a limited constant power speed ratio (CPSR), very critical for traction applications, and its low power factor resulting in increasing demands for current/voltage characteristics of the power electronics applied together with SynRel machines. The limited CPSR of this motor technology is a result of absence of permanent magnets inside the rotor leading to elliptic voltage limits with the center coinciding with the one of the d-q axis current. With increasing speed, the ellipse limit area is reduced with a consequent reduction of the maximum admitted current that limits the maximum applied torque, [15]. This effect can be mitigated by adopting a very highly salient rotor. The presence of the ribs (tangential and inner ones, shown in Fig. 1) affects the saliency ratio of the machine, especially in high-speeds operation. This is mainly due to the low flux inside the machine (in field weakening operation) which does not guarantee the saturation of these elements (ribs) thus increasing the quadrature inductance, [16], [17].
In order to reduce ribs impact on CPSR, several solutions have been proposed in the literature: the adoption of filled barriers with epoxy resin [18]; the use of a particular optimization algorithm (Topology Optimization), [19], the magneto-structural optimization, [20], and the multi-material rotor solution, [16], [21].
As shown in [19], the results of the Topology Optimization lead to a solution with multiple inner ribs, thinner if compared This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ to the other solution, but still manufacturable in mass production. The manufacturing process of magnetic materials (steel or magnets) typically affects their performance, [22], [23], [24]. Therefore, it is conceivable that the multiple-rib design would be more (especially inside the ribs) affected by the manufacturing, with a possibility of a slight improvement in the machine performance.
Usually, in the literature, the evaluation of the manufacturing process is abridged with the measurements of the magnetic properties of the core samples affected by manufacturing and with some consideration (analysis) from the material and different manufacturing points of view. The research work proposed in this paper aims for the evaluation of the impact of these effects on a performance characteristics of a specific motor typology (SynRel) at different load points which could suffer or even benefit from the degradation of the magnetic properties of the material when it is located in the rotor ribs.
Based on core samples measurement results, it is possible to update the material properties in the critical motor core elements or assign an extra magnetic region with different magnetic properties on the core surface in the simulated model to improve the accuracy of the motor simulation results. In order to carry out these properties, experimental tests on samples with different widths (differently affected by the manufacturing process) are going to be carried out.
This study has been carried out on a motor designed for Electric Vehicle applications, in which the manufacturing effects are more visible due to its high power and torque density. However, the same conclusion is valid for other applications.
II. DESIGN CRITERIA FOR HIGH-SPEED SYNCHRONOUS RELUCTANCE MOTORS FOR TRACTION APPLICATIONS As shown in [17] and [20], the optimization of the machine geometry can assist in improving the performance of the machine with a fixed topology (number and thickness of inner ribs for each barrier). The mechanical analysis is included in the optimization process to compute the mechanical stress inside the rotor used as an additional constraint of the optimization (together with the electromagnetic ones). The combination of electromagnetic and mechanical computation in high-speed machines is a necessary approach to ensure satisfactory robustness of the rotor structure to withstand high centrifugal force at a high rotational speed.
The first solution analyzed in this paper is the positioning of the ribs along the q-axis, which is the most frequently used approach in literature. In this case, the thickness of the tangential ribs has to be kept quite large due to the limited mechanical support of the radial rib.
The optimization solution for this kind of a motor is shown in Fig. 1 (left).
Another type of study that has been analyzed in this work regards topology optimization. The investigated algorithm is the Solid Isotropic Microstructural (or Material) with Penalization for intermediate densities (SIMP). This method uses the finite elements created by the mesh in FEM analysis, and the properties of each element are modified with an adequately defined density ρ.
As shown and discussed in [19], the topology optimization is used to identify the best combination of the number of internal ribs for each barrier and their preliminary position and thickness. After that, a magneto-structural optimization can refine the topology optimization results. The motor is designed with four barriers per pole and a notch; from the topology optimization, the best combination is the adoption of four inner ribs in the first barrier (the closest to the shaft), four internal ribs in the second barrier, two inner ribs in the third barrier, and only one radial rib in the fourth barrier. In the topology optimization, the value of the thickness of the tangential ribs is selected to the minimum one according to the manufacturing indications. The solution is shown in Fig. 1 (right).
The results of the radial-rib, Fig. 1 (left), and multi-inner rib, Fig. 1 (right), SynRels and their comparison in terms of performance are presented and discussed in Section IV. In Section III the lamination manufacturing (magnetic) effects are analyzed based on the measurement results.

III. EXPERIMENTAL VALIDATION OF THE BH CURVE ON DIFFERENT SAMPLES
The experimental tests have been regard three ring prototypes realized as follows: 1) Single assembled sheets with Outer Diameter (OD) = 60 mm -Inner Diameter (ID) = 50 mm 2) Two assembled sheets to compose a core with total dimensions OD = 60 mm, ID = 50 mm. The sheets are with OD = 60 mm -ID = 55 mm; OD = 55 mm -ID = 50 mm 3) Five assembled sheets to compose a core with total dimensions OD = 60 mm, ID = 50 mm. The sheets are with OD = 60 mm -ID = 58 mm; OD = 58 mm -ID = 56 mm; OD = 56 mm -ID = 54 mm; OD = 54 mm -ID = 52 mm; OD = 53 mm -ID = 50 mm. The three samples have the same external dimensions in order to maintain the same set-up during the test, but they are affected in a different way by the manufacturing process. The first sample considers a core width of manufacturing equal to 5 mm; the second one considers an equivalent width of 2.5 mm (2 × 2.5 mm = 5 mm) while the third one has an equivalent core width of 1 mm (5 × 1 mm = 5 mm). For this selection, it was assumed  that a width above 5 mm should not have a strong manufacturing effect which justifies a machine performance degradation. The minimum layer was selected equal to 1 mm considering that a similar width range is often applied in tooth tips of the stator and in the rotor ribs. In addition, in order to maintain the same final geometry for the samples, the last core has a single cut with a width of a layer 2.5 mm. Therefore, despite having the same overall total external and inner dimensions of the assembled sheets, the area of the machined surface layer is different. The selected material is the fully processed non-oriented steel M270-35A with 0.35 mm thickness. The steel cores are shown in Fig. 2.
Before testing, the steel core samples have to be prepared, creating stacked cores and inserting a measuring winding and an excitation winding. The number of turns of each winding has been selected according to the size of the samples, the maximum current of the power amplifier, the maximum current density of the excitation winding, and the maximum induced voltage in the measuring winding. The measuring winding is composed of 30 turns, while the excitation one of 100 turns. With this choice, it is possible to reach a maximum field strength of 10000 A/m in safe conditions using the selected instrumantation. Using the same number of turns for each prototype allows for a good comparison reducing possible measurement errors due to different winding parameters. The measuring systems perform the characterization of the magnetic properties of the prototype. The step of the magnetic field (H) has been fixed, and the hysteresis curve has been measured for each step of the magnetic field. After that, the normal magnetization curve was obtained by combining the vertex of the hysteresis curves. Figs. 2 and 3 show the steel core samples, the stacked core, the two windings on the steel samples and the experimental setup.
The BH curve of the different samples has been tested by the Permeameter AMH-50K-S by Laboratorio Elettrofisico, a company that produces measuring systems for magnetic materials. Thanks to the instrumentation, it is possible to measure the magnetic field at different flux densities. According to the size of the samples and the maximum current of the power amplifier, a maximum field strength of 10000 A/m has been used. The measurements are made starting from 0 Hz (DC) up to 1000 Hz; the experimental setup is shown in Fig 3.
The samples have been tested at several frequencies (from DC to 1000 Hz), but for direct comparison of the samples with the highest precision (considering instrumentation capabilities) a certain reference frequency has to be selected for the comparison of the samples. The selected frequency should be relatively low to reach the highest field strength (based on Permeameter specification). For these reasons, a set of tests focused on 50 Hz frequency. Several tests are made at the same frequency with the aim of averaging the results and obtaining the BH curves of the different samples. Fig. 4 shows the BH curves for the different steel core samples compared with the one found in the datasheet of the material. From the curves, it is clear that laser cutting strongly affects the magnetic properties of the material, and the degradation depends on the thickness of the sample under test. Despite the material being a non-oriented one, there is a difference between the rolling direction and the transversal one. As they are manufactured, the samples show the average curve between the two directions. The data obtained from the datasheet should regard the rolling direction in which the permeability of the material is higher. This can justify the high difference between the curves in the high saturation area. To verify it, a test with the Epstein frame should be arranged with the sample cut in the rolling direction. However, the results are consistent with the one presented in the literature, and no additional tests have been carried out [23], [24], [25].
The relative permeability of the material at around 1T is 7400 considering the datasheet's curve, 1280 for the first sample (5 mm), 720 for the second sample (2.5 mm), and 460 for the third one (1 mm). In the non-saturated area, the difference between the measured relative permeability is very high, and the effect of the laser cutting cannot be neglected, especially if the machine has areas with very low core thickness.
Considering the saturated area with a flux density of 1.5T, the relative permeability is 680 for the datasheet's curve, 380 for the first sample, 320 for the second, and 250 for the third one. The difference, in this case, is lower, and the manufacturing degradation of the material could be neglected in some motor analysis/design without a significant performance difference.
This lamination manufacturing aspect strongly depends on the type of machine and the application (load, speed, specific power, tangential stress). When the selected machine is a SynRel, and the application is an industrial one characterized by constant torque and speed, except for some short transients, the machine works in the rated condition near by the knee of the saturation in which the difference in terms of permeability between the affected and non-affected materials is low; therefore, the manufacturing effect could be neglected. Otherwise, if the machine works with a different level of torque, when it is low, the impact of the manufacturing is more visible, reducing the permeability of the material and, therefore, the saliency and the performance.
When the selected machine is a PM motor, the level of saturation is guaranteed by the magnets reducing the effects of manufacturing. If the machine works in flux weakening mode (typically applied in traction applications) the effects of the manufacturing are important.

IV. THE EFFECT OF LASER CUTTING ON THE MOTOR PERFORMANCE
This Section presents a comparison between the solutions with radial ribs and multiple inner ribs. The comparison also includes cases using different material data 1. Material obtained directly from the datasheet 2. Material obtained from the measurementsdis applied to each part of the machine 3. Material obtained from the measurements is applied only to the ribs and material obtained directly from the datasheet is applied to the rest of the core. Since the BH curves have been obtained for different layer widths (1 mm, 2.5 mm, and 5 mm), it allowed to assign each part of the core of the machine with the BH curve of the measured sample that has the closest (single layer) width to the actual width of that core part; the same is made when material obtained from the measurements is applied only to the rotor ribs.
For the sake of clarity, Fig. 5 shows the areas in the rotor and in the stator in which the different curves are applied (Non affected, 5 mm, 2.5 mm, and 1 mm).
The comparison is performed using FEA; several values of current vector angle and amplitude have been tested and for every value the performance of the motor is computed. In this way the mapping of the machine has been obtained. After that with a post-processing (by interpolating the computed values in order to improve the quality of the results) the limit curves (maximum torque over the speed) have been obtained taking into account the limitation of the maximum applied current and voltage. In these points the other quantities (current amplitude,      power factor, saliency ratio, and inductance difference) have been computed.
Figs. 6 to 10 show the comparison of the maximum torque, the saliency ratio, the axis inductance difference, the maximum current, and the power factor, respectively, in the speed range of the radial ribs and the multi-inner-rib solution in three different conditions: 1) All areas of the core have the same BH curve, the one obtained from the datasheet of the material (Non Affected). 2) Different areas of the stator and rotor cores are affected according to the core measurement results. The applied BH curve depends on the core layer width (in correlation with the measured single layer thickness) of the part of the machine according to Fig. 5 (All Affected). 3) Only the rotor ribs are affected according to the core measurement results in the function of their width (Only Ribs Affected). Solution 3 is unusual because the manufacturing cannot affect only a certain area of the machine. But it is included to show how the manufacturing could improve the performance of the machine to find a new technological solution. If the solution can guarantee better performance, it is possible to realize the machine, reheating the electrical steel to guarantee the best material characteristic and after that damaging the ribs in order to affect only that area. This process is more expensive, but it could be justified if the performance improvement is significant. Another possibility during the laser cutting is to change the power of the laser during the cut, increasing it when the ribs are realized and reducing it in the other part of the machine.
From the torque comparison, it is clear that the solution with multiple inner ribs is better than the one with radial ribs, both considering the manufacturing effect and neglecting them. Table I shows the numerical comparison at the rated speed (6000 rpm) and at the maximum operating one (18000 rpm).
At the base speed, the difference between the two solutions is limited, and the multi-inner-rib one has a torque improvement compared with radial ribs of only 5.7% (364 Nm vs. 385 Nm) when the manufacturing effects are neglected. At this speed, the impact of manufacturing is visible when it is considered in all areas of the machine (stator and rotor core); this is mainly due to the decrement of the permeability of the stator flux path, which reduces the direct and quadrature axis inductance. The quadrature inductance is also affected by the reduction of the permeability in the ribs area, which could increase the saliency ratio (as shown in Fig. 7). These two effects are linked, and in the multi-inner-rib solution leads to an overall improvement in the saliency ratio. In contrast, in the radial ribs, the reduction of the d-axis inductance has a stronger effect. This could be due to the thickness of the ribs that do not allow a strong reduction of the permeability as happens in the thinner multi-inner-rib one. The decrement of the permeability of the stator flux path leads to a reduction of difference of the axes inductances and, consequently, the torque capability at the maximum current (as shown in Fig. 8).
As shown in Table I, the torque at high speed is higher when the manufacturing effect is considered, especially in the multi-inner-rib solution, but the inductance difference is lower; this is possible only at a higher current. The current, in this case, could be higher thanks to the higher saliency ratio and the reduction of the q-axis inductance. In the deep flux-weakening area, the current angle is very high (higher than 80°), and the control type is the Maximum Torque Per Voltage (MTPV). In this condition, the q-axis flux has a similar value to the d-axis one. Reducing the inductances allows an increase in the current satisfying the voltage requirements, thus obtaining a higher torque. This is also possible if the saliency ratio is lower and strongly depends on the values of the inductances. This is visible in the radial-rib rotor, in which the torque at 18000 rpm is equal to 20.5 Nm when the manufacturing effects are neglected and 22.2 Nm when they are considered. The torque improvement is not justified by an increase in the saliency ratio because it has similar values in all conditions, but there is an increase in the phase current (273.5 A vs. 264 A). The impact of manufacturing is very similar in the multiinner-rib rotor and the radial-rib one when the torque performance is considered. The difference in the base point is maintained at 5%, also considering the manufacturing effects. Considering the maximum operating speed (18000 rpm), the difference between the non-affected motors is around 90% (20.5 vs. 39.6) and maintains the same value when the manufacturing is considered (22.2 vs. 42.4).
The impact of the manufacturing on the saliency ratio of the machine depends on the type of the rotor and the speed. The difference in the saliency ratio between the two rotor solutions at base speed is 22% neglecting the manufacturing and becoming 21%, considering them. When the maximum speed is considered, this difference is more evident and is equal to 58% neglecting the manufacturing effects and becomes 59% considering them with an opposite trend.
The maximum current is the same at the base speed because all cases are able to satisfy the voltage requirements at this load, but at the maximum speed, the current values changes (below of the maximum supply current). Without manufacturing effect, the difference between the radial-rib rotor and the multi-inner-rib one is 68.7% and it becomes 62.8% considering the manufacturing effects in both cases.
Considering the power factor, the difference between the two solutions at base speed is quite low, around 11% neglecting the manufacturing effects and 10.5% considering them. At the maximum speed, the variation is slightly higher and equal to 14% neglecting the manufacturing effects and 16.5% considering them.
Both in the radial-rib rotor and in the multi-rib one, if the manufacturing effect is considered only in the ribs, the performance is improved, maintaining the same difference between the two rotor topologies.
Focusing on single rotor topology and analyzing the manufacturing effect, it is possible to see a positive impact in the high-speed area and a negative one in the low-speed area. Considering the radial-rib rotor, the torque at the base speed is reduced by 4.4% and increased by 8.3% at high speed. The power factor and the saliency ratio variations are quite small, less than 1% and 3%, respectively, at the base speed; and 1.6% and less than 1%, respectively, at high-speed. When the manufacturing effects are only considered in the rotor ribs, the maximum torque is similar to the non-affected one and at the maximum speed it is similar to the solution, which considers the manufacturing effects in the entire machine. Therefore, focusing the manufacturing effects only on the radial ribs does not justify an increase in the manufacturing cost as the performance improvement can be considered as a minor only at low speed loads. These effects are mainly due to the number of inner ribs for this solution: only one per pole per barrier, while the width is relatively large, reducing the manufacturing effects on it. Therefore, in the maximum torque point the ribs are strongly saturated and the difference between the permeability (affected curve vs non affected) is not so high to justify a performance difference. However, in the maximum speed operation (flux weakening) the saturation of these elements is not so high and the permeability between the different curves is quite high and the performance is affected by the adoption of the affected or non affected one.
Focusing on single rotor topology and analyzing the manufacturing effect, it is possible to see a positive impact in the high-speed area and a negative one in the low-speed area. Considering the radial-rib rotor, the torque at the base speed is reduced by 4.4% and increased by 8.3% at high speed. The power factor and the saliency ratio variations are quite small, less than 1% and 3%, respectively, at the base speed; and 1.6% and less than 1%, respectively, at high-speed. When the manufacturing effects are only considered in the rotor ribs, the maximum torque is similar to the non-affected one and at the maximum speed it is similar to the solution, which considers the manufacturing effects in the entire machine. Therefore, focusing the manufacturing effects only on the radial ribs does not justify an increase in the manufacturing cost as the performance improvement can be considered as a minor only at low speed loads. These effects are mainly due to the number of inner ribs for this solution: only one per pole per barrier, while the width is relatively large, reducing the manufacturing effects on it. Therefore, in the maximum torque point the ribs are strongly saturated and the difference between the permeability (affected curve vs. non affected) is not so high to justify a performance difference. However, in the maximum speed operation (flux weakening) the saturation of these elements is not so high and the permeability between the different curves is quite high and the performance is affected by the adoption of the affected or non affected one.
Focusing on the multi-inner-rib solution, the results in terms of percentage variation of torque is quite similar to the previous solution when the manufacturing effects are considered or neglected. Instead, in terms of saliency ratio, the difference is quite low; in this case, at the base speed, it is about 1.4% decreasing when the manufacturing effects are considered and 1.2% increasing at high speed. Considering the power factor at base speed the difference is 1.3% decreasing while the one at high-speed is 3.6% increasing.
The difference in the radial-rib solution is 3% decreasing considering the manufacturing effects at low speed and 0.3% decreasing at maximum speed.

V. EXPERIMENTAL RESULTS
The results obtained by the simulation show that the performance of the SynRel is not significantly affected by the degradation of the material due to manufacturing process. The main effects are found in the maximum torque area in which the machine reaches very high saturation. In the flux weakening area, the presence of the degraded steel in the ribs allows for obtaining a better performance compared to the non-affected material. These results are encouraging and justify further experimental tests on the designed machine. From the torque ripple evaluation, it appears that the torque ripple of this solution (symmetric) is too high for the inteded application (Electric Vehicles), and a further refinement introducing an asymmetric solution is carried out [18]. This difference in the rotor does not affect the previous results in comparing the two machines and the material manufacturing effects.
The prototype is shown in Fig. 11; the rotor presents a combination of two asymmetries: asymmetries between two poles and asymmetry inside the single pole.
This allows to significantly reduce the torque ripple in the preferred direction. The stator and rotor cores of the prototype are made by laser cutting, and at the same company which has realized the steel core samples which were analyzed in Section III. This activity has been carried out within the European project ReFreeDrive [26], which aims to reduce the use of rare-earth materials in the next generation of electric drivetrains, facilitating the industrial feasibility for mass production at low manufacturing costs. Despite the particular rotor geometry (with  a high number of inner ribs), the possibility to realize punching to guarantee the mass production of the motor has been verified. The minimum thickness of the ribs in the rotor is around 0.7-0.8 mm, which is more than double the lamination thickness (0.35 mm).
The assembled motor and a suitable power electronics converter have been fully tested on a dedicated testbed courtesy of the IFP Energies Nouvelles (France) laboratory, shown in Fig. 12.
The tests have been limited to a maximum current of 675 A and a DC voltage of 750 V, therefore, the performance of the motor compared to the simulations presented in the previous Section is reduced.  In order to have a better comparison between experimental and simulation results, the simulation has been repeated with a new rotor design and the same limits as in the experimental tests. For the sake of clarity these new results are found only for the rib solution of the topology (the one that has been realized).
Comparisons between simulation and experimental results are made in terms of maximum torque in the speed range (Fig. 13), maximum current (Fig. 14), and current angle control (Fig. 15). During the experimental tests, the current angle maximum value was limited to 80°(the same limitation has been imposed in the simulation results). This choice has been made because of some issues related to unexpected vibrations of the motor at higher values of the current angle The reason for this problem can be addressed to higher torque and torque ripple using the corresponding angle; in fact, during this test, the non-perfect balance of the rotor had caused vibrations at high speeds, even when the load torque was equal to zero.
There is a good correspondence between the experimental and simulation results leading to the verification of the model and correct evaluation of the manufacturing effects. The main difference between the experimental and simulation results is found at speeds over 16 krpm point in terms of torque and (especially) maximum current. The difference is easily understandable from the current control angle curves. In the experimental results, there is a reduction of the current angle control compared to the other speeds. This leads to a high reduction of the maximum current and a slight one in the torque. However, up to the rated operating point the measured results matched the simulation results (the one which includes the manufacturing steel effects) with the acceptable precision.
The mechanical integrity of the machine is verified up to 18 krpm with a test that provides a low acceleration time (4 s) from 0 to 18 krpm repeated several times; in addition, the maximum speed has been maintained for a long time. After these tests, no additional vibration or noise appeared in the machine, confirming that the rotor is able to manage that speed.

VI. CONCLUSION
Manufacturing affects the overall magnetic properties of electrical steel core material, thus influencing the electromagnetic performance of the machine. These effects can be visible when a SynRel motor works at different levels of saturation or not very significant if the magnetic condition of the core reach high saturated area or if the machine has relatively high equivalent air gap (PM motors). The permeability of the material can be 90% smaller than the original permeability in case of very thin core layers. This frequently happens in the design of transversallylaminated-rotor SynRel motors where thin rotor ribs are applied (for mechanical reasons).
A comparison of manufacturing effects on two solutions (radial-rib rotor and multi-innerrib one) is carried out and discussed. The results show that in a SynRel motor, manufacturing effects can impact the performance of the machine, depending on the applied load. The possible performance improvement of the SynRel by manufacturing effects is mainly due to the strong reduction of the q-axis inductance, because the ribs are typically thinner than the other parts of the machine core. Reducing the q-axis inductance (maintaining high values of the d-axis) allows better exploiting the MTPV curve, having a higher current and comparable (or higher) torque.
These considerations are true for both solutions with a slightly higher impact on the multi-inner-rib rotor. This last solution has a better performance when the manufacturing effects are neglected and has more advantages from the manufacturing effects because it presents a high number of ribs with thin thickness.
If the manufacturing effects are only considered in the ribs, the performance of the machine is improved both at the lowspeed area and at the high-speed one (with field weakening). The improvement is small considering the high effort in terms of manufacturing challenges and costs to obtain this solution, and it is not feasible realizing this kind of solution on purpose.
In the design of electrical machine for electric vehicles, stator and rotor yokes are hardly affected by the manyfacturing effects due to their relatively large thickness. The effect on stator tooth depends on their number, in order to reduce the impact of the manufacturing in the stator it is preferable to use the minimum number of teeth (which makes larger the width of a single tooth). In general, it is recommended to pay attention to these effects in machines with electrical excitation, as their power factor, magnetization current, and inductances strongly depend on the actual magnetic properties of the applied lamination core in every working points with a speed below the base one. The manufacturing effect of the stator lamination core usually leads to performance degradation of the machine since it leads to a higher magnetization current and a worse power factor. However, the manufacturing effect of the rotor lamination core in some cases can lead even to a performance improvement, e.g., by increasing the inductance ratio in a SynRel machine.
Further studies should investigate the realization of a model to better simulate this effect by assigning for each area a dedicated BH curve as the function of part width. Another possibility is to define in a general way several affected widths with a defined BH curve in a multi-layer model in order to automatically create the model without measuring the affected area by assigning the different curves. These models can be obtained only if a high number of core samples with several layer widths have been manufactured. Another research work to continue the investigation of the manufacturing effect is related to possible extra iron losses caused by it at different flux density amplitudes and operating frequencies.