Nine-phase eight-vector SVPWM (Space Vector Pulse Width Modulation) control method based on dynamic sliding mode control

Abstract

The invention discloses a nine-phase eight-vector SVPWM (Space Vector Pulse Width Modulation) control method based on dynamic sliding mode control in a nine-phase rectification system, and belongs to the field of multiphase motor rectification. The invention provides a solution for the problems of harmonic suppression, robustness and the like of the existing multi-phase rectification technology. According to the method, a mathematical model is obtained through analysis based on the topology of a nine-phase rectification circuit; in the modulation technology, on the basis of a multi-phase space decoupling transformation equation, a nine-phase eight-vector SVPWM control strategy is adopted, and the switching time of an MOSFET is calculated; in closed-loop control, a reactive component adopts sliding mode control, and a sliding mode surface is constructed to solve a control law of the reactive component; a second-order differential equation is constructed on an active component, a new sliding mode surface is constructed on the basis of a traditional sliding mode variable structure, and a discontinuous item is transferred to a first-order derivative of a control quantity to obtain a dynamic sliding mode control law which is essentially continuous in time. According to the invention, a nine-phase eight-vector and dynamic sliding mode combined control strategy is adopted, and the problems of overlarge low-order harmonics in a multi-phase rectification system and jittering in sliding mode control are solved.

Description

[0001] The invention relates to the field of multi-phase motor rectification, and in particular to a nine-phase eight-vector SVPWM control method based on dynamic sliding mode control. Background Art

[0002] As an important component of power electronics technology, multiphase rectifiers have been widely used in electric vehicles, aerospace, renewable energy generation, and high-power industrial fields in recent years. Compared with traditional single-phase or two-phase rectifiers, multiphase rectifiers not only significantly improve the efficiency of the system by increasing the number of phases, but also effectively reduce the distortion and harmonic content of the current waveform. Therefore, the use of multiphase rectifiers can achieve precise modulation of AC power and improve the power quality of the power grid. Especially under low-frequency conditions, it can effectively reduce the harmonic pollution of the power grid and improve the stability of the power system.

[0003] Although multi-phase rectification technology has obvious advantages in improving power quality, existing control methods still face some technical difficulties. First, the suppression of low-order harmonic currents remains a challenge. Due to the complexity of the voltage space vector of the multi-phase rectifier, low-order harmonics (such as third, fifth, and seventh harmonics) are often difficult to suppress effectively, which not only affects the system efficiency, but also increases the pollution of harmonics to the power grid. Secondly, due to its superior robustness and dynamic response speed, sliding mode control technology has been well applied in the nonlinear and uncertainty problems of multi-phase rectification systems. However, traditional sliding mode control methods often cause jittering, especially when the system undergoes high-frequency switching, which directly affects the stability of the system and reduces the output power quality. This jittering not only increases the volatility of the system, but may also cause additional electromagnetic interference, thereby affecting the long-term stability and reliability of the system.

[0004] In order to solve these problems, many researchers have proposed optimization control strategies, such as improving the harmonic suppression effect of the system through space vector decoupling technology, or quickly responding to current changes by using a current hysteresis regulator. However, the switching frequency of the current hysteresis regulator is unstable, which will lead to increased electromagnetic interference. In addition, although methods such as precise feedback linearization control and model predictive control can provide higher control accuracy, their computational complexity is high and they are not suitable for actual real-time control scenarios.

[0005] Therefore, how to combine different control strategies in a multi-phase rectifier system to effectively suppress low-order harmonic currents and improve the response speed and robustness of the control system is of great significance in the research on improving the power quality of the current power system. Summary of the invention

[0006] The technical solution adopted by the present invention to solve its technical problems is: a nine-phase eight-vector SVPWM control method based on dynamic sliding mode control. It includes the following steps: 1) constructing a rectifier circuit topology according to design requirements and calculating device parameters; 2) calculating the conduction time of each switch tube in the nine-phase eight-vector SVPWM algorithm; 3) constructing a rectifier circuit mathematical model and the required second-order circuit equation according to the topological structure; 4) solving the control law of sliding mode control through the sliding mode control mathematical model and differential equation; 5) constructing a simulation model; 6) comparative analysis to obtain the advantages of dynamic sliding mode control and nine-phase eight-vector SVPWM control;

[0007] According to the present invention, a nine-phase eight-vector SVPWM control method based on dynamic sliding mode control is provided, which may also have such technical features, wherein step 1) builds a rectifier circuit topology according to design requirements, and calculates device parameters, including the following steps: selects a bridge rectifier structure according to design requirements, and calculates the parameters of the inductance and resistance on the AC side and the capacitance and resistance on the output side.

[0008] According to the present invention, a nine-phase eight-vector SVPWM control method based on dynamic sliding mode control is provided, which may also have such a technical feature, wherein, in step 2), calculating the conduction time of each switch tube in the nine-phase eight-vector SVPWM algorithm includes the following steps: according to the nine-phase voltage vector diagram and the 3rd, 5th, and 7th harmonic plane voltage position vector diagram, the action time of eight vectors in each sector is calculated, and the action time of the zero working state is added, and the switching time of 9 groups of MOSFET switch states can be calculated through the MOSFET switch sequence diagram.

[0009] According to the present invention, a nine-phase eight-vector SVPWM control method based on dynamic sliding mode control is provided, which may also have such technical features, wherein, in step 3), constructing a rectifier circuit mathematical model and a second-order circuit differential equation according to the topological structure includes the following steps: by establishing a basic mathematical model of a multi-phase rectifier circuit, the current equation and the DC voltage equation under the rotating coordinates d and q axes are obtained, and by a quasi-linearization processing method, a single-input single-output second-order system with the DC voltage and d-axis current components as state variables is obtained.

[0010] According to the present invention, a nine-phase eight-vector SVPWM control method based on dynamic sliding mode control is provided, which may also have such technical features, wherein the solution of the control law of sliding mode control by sliding mode control mathematical model and differential equation in step 4) includes the following steps: in the reactive component, the sliding mode surface is constructed by using the error, and the control law of the reactive component is obtained by rotating the current equation under the coordinate q axis. In the active component, the second-order differential equation in step 3 is introduced into the mathematical model of dynamic sliding mode, and the derivative of the active component control law is obtained by constructing the sliding mode surface twice, and the active component is obtained by integrating it.

[0011] According to the present invention, a nine-phase eight-vector SVPWM control method based on dynamic sliding mode control is provided, which may also have such a technical feature, wherein the construction of the simulation model in step 5) includes the following steps: a nine-phase eight-vector SVPWM control rectifier model based on dynamic sliding mode control is constructed according to the calculations of steps 1 to 4. In addition, in order to verify the suppression effect of the nine-phase eight-vector method on harmonic current, a three-three phase shift decoupling control rectifier model based on dynamic sliding mode control is constructed. In order to verify the robustness and dynamic response of the dynamic sliding mode control, PI control and sliding mode control models are constructed on the basis of the nine-phase eight-vector control for comparison.

[0012] According to the present invention, a nine-phase eight-vector SVPWM control method based on dynamic sliding mode control is provided, which may also have such technical features, wherein the comparative analysis in step 6) to obtain the advantages of dynamic sliding mode control and nine-phase eight-vector SVPWM control includes the following steps: through the simulation model constructed in step 5, the harmonic space current trajectory circle, the DC side output voltage, the output voltage waveform when the load and the reference voltage suddenly change, and the output side THD are analyzed, and at the same time, the DC output voltage is subjected to parameter statistics when the load suddenly changes.

[0013] The beneficial effects of the present invention are as follows: a nine-phase space vector decoupling matrix and a second-order differential equation based on a dynamic sliding mode mathematical model are constructed, thereby proposing a nine-phase eight-vector SVPWM control method based on dynamic sliding mode control. First, the size and position of the nine-phase voltage vector are obtained, and the voltage of other harmonic subspaces is ensured to be 0 by controlling the action time of each vector, and the time of the switching point is calculated according to the switching sequence of each sector to generate a MOSFET pulse signal; secondly, a dynamic sliding mode control strategy is proposed, and a new sliding mode surface is constructed on the basis of the traditional sliding mode surface by constructing a second-order differential equation, and the discontinuous term is transferred to the first-order derivative of the control quantity, thereby eliminating the chattering of the sliding mode control and improving the reliability and stability of the system.

[0014] The features and advantages of the present invention will be described in detail through embodiments in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is the topology diagram of the nine-phase rectifier circuit.

[0016] Figure 2 It is the structural diagram of the rectifier system.

[0017] Figure 3 It is the voltage vector diagram in different spaces.

[0018] Figure 4 It is a closed-loop control structure diagram.

[0019] Figure 5It is the waveform of active component of sliding mode control.

[0020] Figure 6 It is the output side DC voltage waveform and corresponding THD analysis.

[0021] Figure 7 This is a comparison chart of DC voltage waveforms when the load changes suddenly.

[0022] Figure 8 This is a comparison chart of DC voltage waveform when the reference voltage changes suddenly.

[0023] Fig. 9 It is the circular waveform diagram of each space current trajectory. DETAILED DESCRIPTION

[0024] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0025] In this embodiment, a nine-phase eight-vector SVPWM control method based on dynamic sliding mode control is specifically implemented on the platform: the operating system is Windows 10, and the simulation software is Simulink.

[0026] Embodiment 1:

[0027] like Figure 1 The topological structure of the nine-phase rectifier circuit is shown in Figure 1. The axes of the generator stator windings are spaced 2π / 9 electrical degrees apart. There are 9 bridge arms in the nine-phase rectifier circuit. The upper and lower MOSFETs of each bridge arm are alternately turned on, that is, there can be 2 9 =512 switch states. To control the voltage vectors of other harmonic planes to 0, eight voltage vectors are selected in each sector. According to the switch state, the 512 working states are divided into 4 groups, namely {1-8}, {2-7}, {3-6}, {4-5}. The numbers in brackets are the number of bridge arms that are turned on or off by the MOEFET. According to the principle of maximum amplitude of the synthetic vector of adjacent vectors, the vector with the largest amplitude is selected in the above four groups of working states, namely: {1-8}max, {2-7}max, {3-6}max, {4-5}max. Each sector selects two vectors in each of the four groups of switch states, that is, a sector has a total of eight voltage vectors. Since two adjacent sectors will share four identical vectors, a total of 72 voltage vectors are selected for eighteen sectors.

[0028] Embodiment 2:

[0029] like Figure 2 The following are the steps in building the rectifier system model:

[0030] Step 1: Select the appropriate topology and device type according to the design requirements;

[0031] Step 2: The current of each phase on the AC side and the voltage on the DC side are taken through the sampling circuit, and the current i under the rotating coordinate d-axis and q-axis is generated through coordinate transformation d 、i q ;

[0032] Step 3: In the reactive component, replace i d 、i q As the controller input, the sliding surface is constructed based on the mathematical model to solve the sliding mode control law u gq In the active component, i d 、i q with u dc As the controller input, based on the established second-order differential equation, u dref -u dc As the error, a sliding surface is constructed, and on this basis, a new sliding surface is constructed to obtain the sliding mode control law. Will Integration gives the active component.

[0033] Step 4: The active and reactive components are transformed by Park inverse transformation to obtain the α-β coordinate components, and the pulse signal of the MOSFET is generated by sector judgment and switching time calculation.

[0034] Embodiment 3: When the voltage vector in the fundamental wave space is mapped to other subspaces, the amplitude and phase of the vector are different. Figure 3 (a) It can be seen that the relative magnitudes of the amplitudes of the four groups of voltage vectors selected in the fundamental wave space in other subspaces. Figure 3 (b) lists the specific phases of the voltage vector in each subspace in the first sector. The amplitudes of the voltage vectors in each space of the four working states are shown in Table 1.

[0035] Table 1 The amplitude of the voltage vector corresponding to the four working states in each α-β subspace

[0036]

[0037] Among them: K1=sin(π / 9), K2=sin(2π / 9), K3=sin(3π / 9), K4=sin(4π / 9). Let θ v (ν=1,3,5,7) is α v -β v The expected voltage vector u in the subspace vref The angle with the starting edge of the first sector.

[0038] According to the sine theorem, the projection value of the desired voltage vector in each plane at the two boundaries of the sector in which it is located is:

[0039]

[0040] Where: N is the sector number (N = 1, 2..., 18). 3ref =u 5ref =u 7ref =0, that is, when the synthetic voltage vector in the harmonic space is 0, according to the volt-second balance principle, equations (2) and (3) can be obtained, that is, the action time of the eight vectors in the first sector and the zero vector.

[0041]

[0042] Taking the first sector as an example, 9 vector switching points t are introduced. jon (j=1,2,…,9):

[0043]

[0044] The vector switching point time is obtained from equation (4). According to the switching order of MOSFET in each sector, the switching point time can be brought into play to realize the driving of the rectifier system. The nine-phase eight-vector SVPWM control method mainly focuses on controlling the voltage vector in the main dq plane, while ensuring that the voltage vectors of other high-order dq planes are zero, thereby achieving a balanced and stable output. In addition, this control strategy can make full use of the DC voltage and maximize the efficiency of the system.

[0045] Embodiment 4:

[0046] Figure 4 The structure diagram of the closed-loop control of the present invention. Since the d-axis after the rotation coordinate change is usually aligned with the magnetic flux or excitation axis of the motor and is orthogonal to the rotor magnetic flux. Because it affects the torque and mechanical output power of the motor, the d-axis and q-axis voltage components are usually related to active power and reactive power. Therefore, the present invention adopts traditional sliding mode variable control in the reactive component, and focuses on constructing a second-order differential equation in the active component, and adopts dynamic sliding mode variable control to solve the active component output.

[0047] The traditional sliding mode variable structure design is adopted in the reactive component, and the formula of the current in the axis component is as follows:

[0048]

[0049] Define the error and switching functions:

[0050]

[0051] Derivative of the switching function:

[0052]

[0053] Using exponential reaching law control, let The reactive output component can be obtained:

[0054] u gq =-Ri q -ωLi d +e q +ε1Lsgn(s1)+k1Ls1 (8)

[0055] In order to better weaken the chattering, a single-input single-output second-order system is constructed with DC voltage and d-axis current components as state variables through quasi-linear processing method. A new sliding mode surface is constructed based on the traditional sliding mode variable structure, and the discontinuous term is transferred to the first-order derivative of the control quantity to obtain a dynamic sliding mode control law that is essentially continuous in time.

[0056] with u 2 dc and i d Construct a single-input, single-output second-order system for the state variables:

[0057]

[0058] where u 2 dc is the square of the DC voltage, i d 、i q are the current components of the rotating coordinate d-axis and q-axis, E m The peak value of the AC voltage is 220V, C is the DC capacitance, L is the equivalent inductance on the AC side, R L is the load resistance, and R is the equivalent resistance on the AC side.

[0059] The mathematical model of converting the single-input single-output system of active component structure into dynamic sliding mode is:

[0060]

[0061] Among them, [x1 x2] T for 2 dc i d ] T , f(x) and g(x) are known smooth functions as shown in equation (11), and η is the uncertainty in the system.

[0062]

[0063] The tracking error and switching function are defined as:

[0064]

[0065] After taking the derivative of s, a new dynamic switching function and its derivative are constructed:

[0066]

[0067] Where λ is a strictly positive constant, when σ=0, It is an asymptotically stable first-order dynamic system, and s approaches zero.

[0068] The control law of dynamic sliding mode control is taken as equation (15), and after integration, the active component u can be obtained gd .

[0069]

[0070] Stability analysis:

[0071] Assumption 1: The uncertainty satisfies the bounded condition, and there exists a bounded function B n (x), such that:

[0072]

[0073] Assumption 2: The derivative of the uncertain term is bounded

[0074]

[0075] Assumption 3 There exists a positive real number ε that satisfies

[0076]

[0077] Substituting the control law (15) into (14) we can obtain:

[0078]

[0079] According to the assumptions 1-3 given in the stability analysis, we can get:

[0080]

[0081] From formula (20), we can see that when and only when σ=0, According to the Lyapunov stability theorem, the system will reach and maintain the sliding state s = 0, and the linear sliding mode s will reach the sliding surface and maintain the sliding state in a finite time. After that, the system state x will gradually converge to zero, and the system will approach stability. The stability analysis proves that the sliding mode control law selected in this paper It can ensure that the system can reach and maintain a steady state.

[0082] Embodiment 5:

[0083] Figure 5The comparison of active component output using sliding mode control and dynamic sliding mode control shows that the most obvious difference between dynamic sliding mode control and sliding mode control is the u output by the dynamic sliding mode controller. gd The chattering is significantly reduced. The first-order dynamic sliding mode controller is actually designed The integrated control law u can be guaranteed gd Continuous, making the system have more stable power output performance.

[0084] Embodiment 6:

[0085] Figure 6 The output side DC voltage waveform and the corresponding THD analysis. It can be clearly seen from the comparison of the three methods that the PI control responds very quickly in the initial stage, but has obvious overshoot, and takes the slowest time to reach steady state. The overshoot of the sliding mode control model is much lower than that of the PI control, and the waveform is relatively stable. The dynamic sliding mode control model has almost no overshoot, takes the shortest time to reach steady state, and has the best stability. Figure 6 The right side shows the THD analysis of the output voltage of the three simulation models within one cycle when reaching steady state. It can be seen from the figure that the harmonic content of PI control is as high as 0.03%, while the harmonic content of the other two control methods is only 0.01%. The low-order harmonic content of dynamic sliding mode control is slightly higher than that of sliding mode control, but the high-order harmonics are lower than those of sliding mode control. Since this paper adopts nine-phase eight-vector SVPWM control, it can be clearly seen from the THD that odd harmonics are effectively suppressed.

[0086] Embodiment 7:

[0087] This is a comparison of the DC voltage waveform when the load changes suddenly. In order to test the robustness of the dynamic sliding mode method, the load resistance is changed at 0.2s to simulate the sudden change of load current in the circuit. Figure 7 (a) and Figure 7 (b) are the simulated waveforms of the current and voltage changes on the output side after a sudden increase and decrease in load at 0.2s.

[0088] After a sudden increase or decrease in load, the closed-loop control structure is adjusted. The DC side voltage regulation effect can be intuitively seen. The voltage fluctuation of PI control is the largest and the time to reach steady state is the longest. The voltage fluctuations of dynamic sliding mode and sliding mode control are not much different, but the dynamic sliding mode has a faster response speed. By comparing the three methods, dynamic sliding mode control performs best in dealing with sudden load changes, with faster response speed and smaller voltage fluctuations, proving its robustness and superiority in the face of load disturbances.

[0089] In the load mutation experiment, in order to detect the control effect of the control strategy in this paper on the transient response of the output side, the parameters such as the attenuation rate, overshoot and residual deviation of the output voltage are mathematically statistically analyzed, and the results are shown in Table 2. Because the load used in this study is a resistive load, the output voltage changes proportionally with the output current, and the statistical parameters of the output current are consistent with the voltage.

[0090] Table 2 Statistics of output performance parameters when load changes suddenly

[0091]

[0092] Embodiment 8:

[0093] Figure 8 This is a comparison of the DC voltage waveform when the reference voltage suddenly changes. To further test the robustness of the dynamic sliding mode control, the reference voltage is reduced from 600V to 500V at 0.1s. It can be seen from the figure that after the voltage suddenly drops, the PI control shows a more obvious voltage fluctuation, and the time to recover to the steady state is longer than the other two control methods. This reflects that the dynamic performance of PI control is poor when dealing with sudden changes, and the system is relatively slow to adjust the reference voltage. The response speed of the sliding mode control is slightly inferior to that of the dynamic sliding mode control. The voltage also returns to 500V after the sudden change, but there is a slightly larger voltage fluctuation. The dynamic sliding mode control shows a faster response speed and a smaller voltage fluctuation. After a short drop, the voltage quickly returns to 500V, and the voltage fluctuation is very small. The comparison of the three clearly shows the significant advantages of dynamic sliding mode control and sliding mode control in the face of a sudden drop in the reference voltage, especially the dynamic sliding mode control shows excellent stability and response speed.

[0094] Embodiment 9:

[0095] Fig. 9 It is the circular waveform diagram of each space current trajectory. Fig. 9 From top to bottom in (a) are the current trajectory circles of the fundamental space and the 3rd, 5th, and 7th subspaces corresponding to the PI control, sliding mode control, and dynamic sliding mode control based on nine-phase eight-vector control. After comparing the three, the current circle pulsation of the fundamental space controlled by PI is larger, indicating that the low-order harmonic content of the current is slightly higher than that of the other two methods. The harmonic currents of the sliding mode control and the dynamic sliding mode control are not much different, and the amplitudes are all within 0.5. Because the present invention adopts the nine-phase eight-vector decoupling method, a set of harmonic space current trajectory circles of three-three phase-shift decoupling control is added as a comparison, such as Fig. 9 As shown in (b), it is obvious from the comparison between the two methods in the figure that the harmonic current pulsation of the nine-phase eight-vector decoupling control is much smaller than that of the three-three decoupling control, which further proves that the nine-phase eight-vector has a significant effect on suppressing low-order harmonic currents.

[0096] In summary, the multi-phase rectification control strategy combining nine-phase eight-vector and dynamic sliding mode proposed in the present invention effectively suppresses the current in the 3rd, 5th and 7th harmonic space, solves the jitter problem of traditional sliding mode control, and exhibits excellent dynamic response speed and stability when the load changes suddenly and the reference voltage changes, and has extremely strong robustness. The present invention effectively improves the power quality, which is of great significance for improving the stability of the power system and reducing grid pollution.

[0097] The above embodiments are only used to illustrate specific implementation modes of the present invention, and the present invention is not limited to the description scope of the above embodiments.

Claims

1. A nine-phase eight-vector SVPWM control method based on dynamic sliding mode control, characterized in that: The following steps are involved: Step 1: Build a nine-phase rectification model based on three-three phase shift decoupling control of PI control; Step 2, calculate the conduction time of each switch tube in the nine-phase eight-vector algorithm according to the vector diagram; Step 3: Build a nine-phase eight-vector control model based on PI control; Step 4: Solve the control law of dynamic sliding mode control according to the mathematical model of the rectifier circuit; Step 5: Build nine-phase eight-vector control models based on sliding mode control and dynamic sliding mode control respectively; Step 6: Compare and analyze the advantages of dynamic sliding mode control and nine-phase eight-vector SVPWM control.

2. The nine-phase eight-vector control method based on dynamic sliding mode control according to claim 1 is characterized in that: In the step 1, a nine-phase rectification model of three-three phase shift decoupling control based on PI control is constructed, and the harmonic current conditions of each harmonic subspace can be analyzed from the model.

3. The nine-phase eight-vector control method based on dynamic sliding mode control according to claim 1 is characterized in that: In the step 2, the conduction time of each switch tube in the nine-phase eight-vector algorithm is calculated according to the vector diagram, and the action time of eight vectors in each sector is calculated according to the nine-phase voltage vector diagram and the 3rd, 5th, and 7th harmonic plane voltage position vector diagram, plus the action time of the zero working state, and the switching time of 9 groups of MOSFET switch states is calculated through the MOSFET switch sequence diagram.

4. The nine-phase eight-vector control method based on dynamic sliding mode control according to claim 1 is characterized in that: In the step 3, a nine-phase eight-vector SVPWM control rectifier model based on PI control is built, and compared with the three-three phase-shift decoupling control model built in step 1, the nine-phase eight-vector SVPWM control algorithm is verified for the suppression effect of the 3rd, 5th and 7th harmonic space currents.

5. The nine-phase eight-vector control method based on dynamic sliding mode control according to claim 1 is characterized in that: In the step 4, the control law of dynamic sliding mode control is solved according to the mathematical model of the rectifier circuit. The mathematical model of the rectifier circuit is processed by a quasi-linearization method, and a single-input single-output second-order system is constructed with DC voltage and d-axis current components as state variables. A new sliding mode surface is constructed based on the traditional sliding mode variable structure, and the discontinuous term is transferred to the first-order derivative of the control quantity to obtain a dynamic sliding mode control law that is essentially continuous in time.

6. The nine-phase eight-vector control method based on dynamic sliding mode control according to claim 1 is characterized in that: In the step 5, nine-phase eight-vector control models based on sliding mode control and dynamic sliding mode control are respectively constructed. In terms of closed-loop control, the sliding mode control method uses the difference between the DC voltage and the reference voltage as input in the voltage outer loop, constructs a sliding surface to obtain the reference current, and the model realizes four-branch parallel current sharing by merging three mutual inductance coils. The branch current deviation caused by the inconsistency between the device parameters and the loop parasitic parameters generates an induced electromotive force in the magnetic core. According to Faraday's law of electromagnetic induction, the electromotive force will drive the unbalanced current to remain zero, so the current sharing of the deviation branch can be achieved. Active current sharing of the four branches is finally achieved through pairwise merging.

7. The nine-phase eight-vector control method based on dynamic sliding mode control according to claim 1 is characterized in that: The comparative analysis in step 6 shows the advantages of dynamic sliding mode control and nine-phase eight-vector SVPWM control. Through the simulation model built in step 5, the harmonic space current trajectory circle, DC side output voltage, output voltage waveform when the load and reference voltage suddenly change, and output side THD are analyzed. At the same time, parameter statistics of the DC output voltage are performed when the load suddenly changes.

Images