Multiphase SVPWM Strategy Analysis and Implementation of Seven-Phase Permanent Magnet Synchronous Motor

+e multiphase motor drive systems have become a focus in many application areas such as ship electric propulsion, urban mass transit, aerospace, and weapon equipment, as they are characterized by high power density, low torque pulsation as small torque ripple, large output power, strong fault tolerance, and high reliability. However, with the increase of the phase number of the motor, the current harmonic component increases correspondingly, which leads to the decrease of the control performance compared with the three-phase system. In order to overcome this challenge, implementation method of driving control technology for seven-phase permanent magnet synchronous motor (PMSM) based on SVPWM algorithm is discussed thoroughly in this paper. Simulink and experiments have been developed to check its practical feasibility. +e results show that the near-six vector SVPWM algorithm (NSV-SVPWM) achieves better performance than other methods.


Introduction
Multiphase PMSM is widely used in special occasions such as high current and high voltage, and it works in the case of low voltage and high power drive without any restriction of current or voltage. Compared with three-phase permanent magnet motor, PMSM has many advantages, such as simple structure, small volume, high power output density, high efficiency, low torque ripple, and low vibration noise, and it has been widely used in the field of micro electric vehicle [1][2][3][4][5]. As the number of motor phases increases, the pulse frequency of the output torque of the multiphase motor drive system increases, and the amplitude of the mechanical torque ripple decreases; thus, the static and dynamic performance of the multiphase motor drive system can be improved, and the reliability of the whole drive system is improved. For the dimension of the multiphase motor drive system, the threedimensional system of the traditional three-phase motor is transformed into a multidimensional system. e increase of the dimension of the motor makes the control strategy of the multiphase motor more abundant, thus making the more complex motor control algorithm more easily realized.
In view of the above advantages of multiphase PMSM, the seven-phase PMSM is considered. However, due to the increase of the phases of motors, the current harmonic contents of multiphase motor system are larger, which leads to the decrease of the control performance compared with the three-phase system [6,7]. e inverter nonlinearity is the key source of current harmonics. To solve this problem, domestic and foreign experts and scholars have conducted a lot of research [6,[8][9][10], such as starting from the design improvement of motor, changing the distribution of motor winding and distribution of groove to reduce the harmonic current, and starting from the angle of control strategy, current decoupling to different harmonic subspaces to reduce the harmonic content [6].
With regard to the PWM of seven-phase system, the SVPWM theory can be still employed to represent the behavior of seven-phase systems as a natural extension of the traditional three-phase SVPWM transformation [11,12]. In this paper, the space vector modulation (SVM) has been extended to a seven-phase voltage source inverter, considering reference space vectors in all the three d-q planes [13]. Especially, the proposed SVM strategy univocally selects the inverter switch configurations among the 128 possibility by privileging the space vector on the first subspace plane, d1-q1, the one responsible of balanced sinusoidal output voltage waveforms [14]. en, vector analysis of the other two planes that produce the harmonic current is made to find out whether the vector that generates harmonics can be counteracted as much as possible. So, the implementation method of driving control algorithm based on SVPWM for low-order current harmonics suppression is discussed in this paper. e goal of this paper is design a framework that addresses the above challenges. First, the mathematical model of the static coordinate system of the seven-phase PMSM is analyzed in detail. en, a novel synchronous rotating coordinate transformation matrix was presented, and a mathematical model of the rotating coordinate system is obtained by matrix converter. And then, the 3rd harmonic subspace and the 5th harmonic subspace are derived based on the spatial voltage vector distribution of seven-phase bridge inverter. After analyzing the performance of these methods, namely, the near-four vectors SVPWM (NFV-SVPWM) algorithm and NSV-SVPWM algorithm, NSV-SVPWM algorithm is carried out. Finally, a driving control platform based on STM32F407 primary control chip is designed for seven-phase PMSM.
Preliminary results have been announced in the conference paper [15]. A more complete framework of control technology of seven-phase PMSM for low-order current harmonics suppression is proposed. We also describe in detail the experiments analysis. e rest of the paper is organized as follows. Section 2 gives mathematic model of seven-phase PMSM. e seven-phase SVPWM algorithm is described in Section 3. Simulations and experiments analysis of PMSM and the detailed discussion are described in Section 4. Finally, a summary is provided in Section 5.

Mathematic Model of Seven-Phase PMSM
e investigated motor is a hidden pole type seven-phase PMSM. Before studying the driving control system of the seven-phase PMSM, the mathematical model of the motor is established at first. erefore, for the convenience of analysis, the following assumptions are made for the seven-phase PMSM studied: (1) Neglect the influence of eddy current and hysteresis in iron core (2) e magnetic circuit of the motor is linear, and the saturation effect will not occur (3) e stator surface of the motor is smooth (4) Neglect influence of motor temperature on motor winding (5) e structure of the motor is implicit, and the reactance is equal (6) e seven-phase winding is "Y" type connection, and the spatial distribution is completely symmetrical In the stationary coordinate system, the flux equation, the stator voltage equation, the electromagnetic torque equation, and the mechanical motion equation are important mathematical model of the seven-phase PMSM.
In natural coordinate, the voltage equation for sevenphase PMSM is given: where T is interpreted as the stator flux matrix. e flux equation for seven-phase PMSM is given: where L v is the stator inductance matrix, and Ψ n represents the flux linkage of permanent magnet. It can be defined as where ψ, θ are the flux linkage amplitude of permanent magnet and the rotor position angle, respectively. α � 2π/7 is the angle between adjacent two-phase winding axes. e torque equation for seven-phase PMSM is given: where T e is motor electromagnetic torque, Μ eo is denoted as motor magnetic energy, and a is the polar logarithm of motor.
For the hidden pole rotor of motor, L v and θ are independent of each other, then the above function can be rewritten as e mechanical motion equation for seven-phase PMSM is where T l is denoted as motor load torque, and J is moment of inertia of mechanical and electrical system. Seven-phase 2 Complexity PMSM is a nonlinear system. e methods [16,17] can provide a reference for some parameters estimated and model constructed.
According to equations (2)-(6) of electromagnetic torque equation, it can be seen that the straight axis and quadrature axis currents of stator of seven-phase PSMS are coupled with each other, which brings inconvenience to the subsequent research on vector control theory. erefore, it is necessary to select appropriate spatial transformation and simplify the mathematical model of the seven-phase PMSM.
According to the principle of constant power, the transformation matrix from natural coordinate system to based on the matrix transformation of Clark and Park is given by If equation (7) is introduced into equations (2)-(6), through matrix transformation, the mathematical model of the motor in the rotating coordinate system can be obtained. e voltage equation for seven-phase PMSM is given by where In this way, other equations can be obtained. e flux equation for seven-phase PMSM is given by For the hidden pole rotor of motor, the torque equation can be calculated as where i q1 is the component of stator current on the q axis.

Seven-Phase Bridge
Inverter. e seven-phase motor drive systems are generally supplied by a seven-phase voltage source bridge inverter. Similar to three-phase bridge inverter, the topology of seven-phase bridge inverter is connected by 14 power switch devices, as shown in Figure 1.
In order to approach the circular magnetic field well, there should be as many regular polygon edges as possible; that is, more inverter switching state combinations should be generated. For seven-phase voltage inverter, the state variable S i (i � a, b, c, d, e, f, g) is introduced to represent the switching state of each bridge arm, and the turn-off and turn-on of the switch are represented by "0" and "1." en, the seven-phase voltage type inverter can be combined into 27 � 128 switching states [18,19].
In a multiphase voltage source inverter (VSI), switching states determine load equivalent circuit configurations. erefore, switching states can be divided into corresponding sets. In a seven-phase inverter, these basic equivalent circuit configurations can be categorized into three types, namely, C16 equivalent circuit, C25 equivalent circuit, and C34 equivalent circuit, as shown in Figure 2. erefore, the switching states can be represented by three sets: From the above, the space vector is obtained as Aiming to produce the desired fundamental (sinusoidal) component in the output phase voltage without low-order harmonics, SVPWM algorithm is used. Supposing that the DC bus voltage is considered as "1," Table 1 presents the magnitude relation of the voltage vector.

Current Harmonics Suppression
Method. Current harmonics (3rd, 5th) caused by the inverter nonlinearity in phase currents are considered. To increase efficiency and Complexity performance of multiphase power converters, the current total harmonic should be eliminated, which highly relies on their control strategy. In this section, two control strategies of SVPWM algorithm are presented in detail.

e NFV-SVPWM Algorithm.
rough analysis of the near-four vectors SVPWM (NTV-SVPWM) algorithm [15], it is found that the medium and small vector groups of the space voltage vector in the seven-phase motor system are not fully utilized to remove the harmonic component of the phase current. erefore, we take into account the more abundant basic voltage vectors in the seven-phase inverter to reduce the harmonic component. Two near vectors, two large vectors, and zero vectors in the fundamental wave subspace are used to synthesize the reference voltage vector, that is, NFV-SVPWM algorithm [20], whose control strategy is the same as the three-phase SVPWM. e reference voltage in the sector is obtained by using the parallelogram rule synthesis. e maximum voltage vector and the middle vector correspond to the U g vector group and the U f vector group in Table 1, respectively. e NFV-SVPWM algorithm of seven-phase motor system follows the following principles: e selected basic voltage vectors are given the voltage reference vector in the synthetic fundamental wave space, and the voltage vector synthesized in the 3rd harmonic subspace should be guaranteed to be zero. As shown in Figure 3, in the basic wave space, voltage vectors U 71 , U 3 and U 67 , U 103 are in the same direction, while in the 3rd harmonic subspace, voltage vectors U 71 , U 3 and U 67 , U 103 are opposite to each other in direction. In order to make the voltage vector of the 3rd harmonic subspace zero, the action time of voltage vectors U 71 , U 3 and U 67 , U 103 is inversely proportional to the amplitude of the voltage vector in the 3rd harmonic subspace. In this way, the voltage vectors cancel each other without affecting the synthesis of voltage vectors in fundamental wave space.
For example, in the first sector, fundamental voltage vectors U 3 , U 67 , U 71 , U 103 and U 0 , U 127 two zero vectors are   Similar to the NTV-SVPWM analysis method, the following relations can be obtained: To reduce the switching losses, the switching signal of the voltage vector in sector one is shown in Figure 4.

e NSV-SVPWM Algorithm.
It is well known that 6 groups of voltage vector can synthesize the desired voltage vector. However, voltage vectors U a , U b , U c and U e with small amplitude are not continuously switched, which lead to desired voltage vector different direction. So, the voltage vector of the stator flux may offset. In NSV-SVPWM algorithm, the reference vector needs six voltage vectors to be synthesized in each sector of the fundamental wavelet subspace.
e previous analysis shows that only three vectors U d , U f , and U g of seven groups of voltage vectors in the fundamental subspace are used. erefore, NSV-SVPWM algorithm is used to synthesize the reference voltage vector by a total of three groups of six voltage vectors: U d , U f , and U g .
For example, in the first sector, select fundamental voltage vectors U 1 , U 3 , U 67 , U 71 , U 103 , U 111 and two zero vectors U 0 , U 127 to synthesize the reference vector U ref . e time of action is defined as T 1 , T 2 T 3 , T 4 , T 5 , T 6 and T 0 , respectively.
Aiming to cut down the harmonic content of phase current, the voltage vector synthesized in the α 3 − β 3 and α 5 − β 5 subspaces should be zero. As shown in Figure 5, it can be seen that voltage vectors U 1 , U 67 and U 103 are in the same direction in α 1 − β 1 subspace, while vector U 67 is opposite to the corresponding two vectors in the α 3 − β 3 and α 5 − β 5 subspaces. To make the synthesized voltage vectors be zero in the subspaces α 3 − β 3 and α 5 − β 5 , the reverse vector can be counteracted by the corresponding two other vectors. us, the resultant voltage of each direction is zero, and the current harmonic is suppressed.
In the α 1 − β 1 fundamental subspace, the following equations came up

Complexity 5
In the 3rd harmonic subspaces, the following equations are obtained: In the same way, the following equations came up in the 5th harmonic subspace According to the vector magnitude of Table 1, the substitution formulae (14)- (16), a set of equations with six variables and solving the equations can be expressed as 2.6638U dc sin(π/7) T s , 1.4773U dc sin(π/7) T s , In this way, the voltage equations are listed for each sector based on the fundamental subspace (α 1 − β 1 ), 3rd harmonic subspace (α 3 − β 3 ) and 5th harmonic subspace (α 5 − β 5 ), respectively. ese equations are simplified to obtain a new six-element first-order functions within each sector. For example, in sector m (1 ≤ m ≤ 14), the following relation holds  Figure 4: Switching signal of NFV-SVPWM algorithm in sector one. 6 Complexity In order to reduce the switching losses, the voltage vectors for each sector should be reasonably arranged. Sector one (m � 11) is taken as an example for analysis. e order of action of the space voltage vector is arranged as follows: e switching signal of the voltage vector in sector one is shown in Figure 6.

Simulations and Analysis.
e proposed algorithm is implemented in Matlab/Simulink platform. e performance of the three control algorithms (NTV-SVPWM, NFV-SVPWM, and NSV-SVPWM) mentioned above section is analyzed. e parameters of the seven-phase PMSM are consistent with the mathematical model. e motor has a load of 1 N.m, the switching frequency is 20 kHz, and the DC bus voltage is 72V. e stator resistance is 1.78Ω, and d-axis inductance L d and q-axis inductance L q are 4.5 × 10 − 3 H and 4.5 × 10 − 3 H, respectively. e permanent magnet flux linkage ψ f is 0.175 Wb. e moment of inertia J is 0.4 × 10 − 3 kg · m 2 .
Simulation results of NTV-SVPWM are shown in Figure 7. In Figures 7(a) and 7(b), x-coordinate is times (s) and y-coordinate is phase current (A). It can be seen that the phase current is less sinusoidal, and the FFT harmonic analysis shows that the 3rd and 5th harmonic contents are as high as 36.65% and 17.22%, respectively, and the total harmonic THD is 40.61%. e results show that the sevenphase system extended from traditional three-phase SVPWM algorithm has the problem of harmonic phase current. Figure 8 presents the simulation results of NFV-SVPWM. Although the sinusoidal component of the phase current is slightly higher than the NTV-SVPWM algorithm, the sinusoidal phase current is still very poor as shown in Figure 8(b). rough the FFT harmonic analysis, it can be seen that the 3rd harmonic content is much less than that of the NTV-SVPWM algorithm, only 5.31%, while the 5th harmonic content is still large, 29.13%, and the overall THD is 16.01%. ough the NFV-SVPWM algorithm controls the influence of the harmonic subspace, only the 3rd harmonic subspace is considered, and the interference of the 5th harmonic subspace is not considered, which leads to the greater content of the 5th harmonic in the phase current. In Figures 8(a) and 8(b), x-coordinate is times (s) and y-coordinate is phase current (A). Figure 9 provides the simulation results of NSV-SVPWM. Using the same motor parameters, the simulation results show that the phase current of NSV-SVPWM appears with tiny distortion, as shown in Figure 9(b). At the same time, the phase voltage waveform is saddle waveform, as shown in Figure 9(c), where x-coordinate is times (s) and Figure 6: Switching signal of NSV-SVPWM algorithm in sector one.
Complexity y-coordinate is phase voltage (V). From the FFT harmonic analysis, it can be seen that the 3rd and 5th harmonic currents are much lower than those of the NTV-SVPWM algorithm, as shown in Figure 9(d). e harmonic contents of 3rd and 5th currents are only 4.29% and 1.76%, and the total THD is 3.51%.
According to the simulation results of above three algorithms, we can see that NSV-SVPWM achieves much better performance than the other methods.

Experiments and Analysis.
A prototype of the proposed algorithm was then built in the lab and tested. In order to ensure that the driving system has good real-time, economic and good motor control characteristics, the STM32F407VET6 chip of the ARM Cotex-M4 kernel as the main control chip of the drive system has been used. Combined with the rated parameters of seven-phase PMSM, a motor experimental platform with a rated voltage of 72 V is built, and its physical map is shown in Figure 10.      10 Complexity e phase voltage waveform of the NSV-SVPWM control mode is shown in Figure 11. It is not difficult to see that the phase voltage waveform is saddle wave, in which the phase difference between a phase voltage U a and b phase voltage U b is 2π/7. e difference between a phase voltage U a and c phase voltage U c is 4π/7, as shown in Figure 11(b). In a similar fashion, a phase voltage U a lags g phase voltage U g 2π/7, as shown in Figure 11(f). erefore, on the basis of the experimental results, the phase voltage waveform is consistent with the simulation results. Figure 12 presents the phase current experimental waveform i a . e sinusoidal component of phase current is very high and without current harmonics. It proves the feasibility of the NSV-SVPWM algorithm in the seven-phase PMSM drive control system. e experiment results are in agreement with the theoretical analysis and simulation results.

Conclusion
Multiphase PMSM is needed to produce the sinusoidal component in the output phase voltage, without low-order harmonics. In this paper, the rotating coordinate system of the seven-phase PMSM is obtained by matrix converter. To eliminate low-order harmonics, NTV-SVPWM algorithm, NFV-SVPWM algorithm, and NSV-SVPWM algorithm are discussed thoroughly based on simulations and experiments analysis. According to the simulation and experimental results, it can be concluded that the NSV-SVPWM algorithm not only suppresses the current high order harmonics, but also has excellent control performance.

Data Availability
No data were used to support this study.

Conflicts of Interest
e authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.