A method for neutral point potential balance control of a T-type three-level inverter based on VSVPWM

By using VSVPWM for sector-specific vector coordinated adjustment and dynamic adjustment factor allocation, a virtual space vector is constructed, which solves the problems of attenuation of the midpoint potential balance capability and large output voltage harmonics in T-type three-level inverters, and achieves rapid stabilization of the midpoint potential and efficient control of the output voltage.

CN123122282APending Publication Date: 2026-11-17NAVAL UNIV OF ENG PLA
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611474279.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-21
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

The existing neutral point potential balance control method for T-type three-level inverters suffers from reduced balancing capability under high regulation, large output voltage harmonics, and high control complexity, failing to meet the requirements of high dynamic and high power quality application scenarios.

Method used

By employing sector-specific vector collaborative regulation and dynamic adjustment factor allocation based on VSVPWM, a virtual space vector is constructed to dynamically adjust the action time of redundant small vectors. This is combined with compensation from non-midpoint influence vectors to achieve rapid and stable equilibrium of the midpoint potential and reduce output voltage harmonics.

Benefits of technology

It achieves stable midpoint potential balance across the entire modulation ratio range, effectively reduces output voltage harmonics, avoids waveform distortion, requires minimal computation, is suitable for real-time operation on DSPs and FPGAs, and is compatible with existing T-type three-level topologies.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123122282A_ABST
    Figure CN123122282A_ABST
Patent Text Reader

Abstract

This invention belongs to the field of inverter control, specifically relating to a midpoint potential balance control method for a T-type three-level inverter based on VSVPWM. The method includes the following steps: standardizing the space voltage vector and rotating it to the first large sector of a 60° coordinate system; dividing each large sector into five smaller sectors based on the voltage vector component boundary conditions; and calculating the vector action time of each smaller sector; based on the feedback obtained midpoint voltage and load current, designing a coordinated adjustment strategy for the midpoint potential balance of redundant vectors and non-midpoint influence vectors for different smaller sectors; introducing an adjustment factor k to dynamically allocate the action time of redundant small vectors and zero or large vectors to achieve dynamic midpoint potential balance; and generating a drive signal based on the adjusted vector action time to control the operation of the power switching transistors of the T-type three-level inverter. This invention provides a method that can quickly suppress midpoint potential shift and maintain stable balancing capability across the entire modulation ratio range.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention belongs to the field of inverter control, specifically relating to a method for controlling the midpoint potential balance of a T-type three-level inverter based on VSVPWM. Background Technology

[0002] T-type three-level inverters employ a series capacitor voltage divider structure on the DC side. This structure is susceptible to midpoint potential drift due to load fluctuations, component parameter variations, and switching delays. Existing midpoint potential balance control methods rely on Space Vector Pulse Width Modulation (SVPWM) to adjust the midpoint current using redundant small vectors to achieve potential balance. However, this existing SVPWM method has significant drawbacks: achieving balance solely through adjusting the duration of redundant small vectors easily leads to output voltage distortion; and under high-modulation conditions, the midpoint potential balance capability significantly degrades, failing to meet the demands of high-dynamic, high-power-quality applications. Summary of the Invention

[0003] The present invention provides a VSVPWM-based midpoint potential balance control method for T-type three-level inverters. This method addresses the problems of large output voltage harmonics, attenuation of balancing capability under high modulation, and high control complexity in existing midpoint potential balance control methods for T-type three-level inverters. By using sector-based vector collaborative adjustment and dynamic adjustment factor allocation, the method achieves rapid and stable balance of the midpoint potential, while reducing output voltage harmonics and expanding the balance control range.

[0004] The present invention provides a method for controlling the midpoint potential balance of a T-type three-level inverter based on VSVPWM, comprising: firstly, constructing a topology model of a T-type three-level inverter; characterized in that it further comprises the following steps:

[0005] S1: The reference voltage space vector is normalized. The large sector where the reference voltage space vector is located is determined according to the boundary conditions of the voltage vector components. The reference voltage space vectors in different large sectors are uniformly mapped to the 60° coordinate system corresponding to the first large sector through rotation coordinate transformation. According to the relationship of the voltage vector components after mapping, each large sector is divided into 5 small sectors. A virtual space vector is constructed based on the virtual space vector pulse width modulation method. The action time of each virtual space vector is calculated according to the volt-second balance principle.

[0006] S2: Based on the small sector where the reference voltage space vector is located, the midpoint voltage deviation, and the load current direction associated with the voltage vector, a dynamic adjustment factor k is introduced to dynamically adjust the action time of the redundant small vectors participating in the virtual space vector synthesis. The change in the action time of the redundant small vectors is compensated by the corresponding non-midpoint influence vector so that the virtual space vectors before and after adjustment meet the volt-second balance constraint.

[0007] S3: Generates a drive signal based on the adjusted vector action time to control the operation of the power switch of the T-type three-level inverter.

[0008] As a further optimization of the present invention, in step S1, the 27 basic space voltage vectors of the three-level inverter are divided into zero vector, small vector, medium vector and large vector according to the magnitude of their amplitudes. Among them, the small vectors include P-type redundant small vectors and N-type redundant small vectors that have the same spatial position but have opposite effects on the midpoint potential. A virtual space vector is formed by combining multiple basic space voltage vectors, so that the amplitude and phase of the virtual space vector are fixed.

[0009] As a further optimization of the present invention, step S2 specifically includes:

[0010] Sector 1: Sector 1 is divided into a first half sector and a second half sector. In the first half sector and the second half sector, according to the corresponding load current direction, a P-type redundant small vector or an N-type redundant small vector that can generate an adjustment effect opposite to the midpoint voltage deviation is selected as the adjustment vector. The action time of the selected redundant small vector is adjusted using the dynamic adjustment factor k. At the same time, the action time of the corresponding zero vector is adjusted to maintain the volt-second balance of the virtual space vector.

[0011] Small sector 2: Divide small sector 2 into a first half sector and a second half sector; in the first half sector and the second half sector, select a P-type redundant small vector or an N-type redundant small vector to participate in the midpoint potential regulation according to the corresponding load current direction, and use the dynamic adjustment factor k to adjust the action time of the selected redundant small vector. At the same time, introduce a PPP zero vector to compensate for the change in action time, so as to keep the amplitude and phase of the virtual space vector unchanged before and after the adjustment.

[0012] Small sectors 3 and 4: Do not divide the small sector into first half and second half sectors. According to the load current direction, select a P-type redundant small vector or N-type redundant small vector that can generate the required midpoint current direction. Use the dynamic adjustment factor k to adjust the action time of the selected redundant small vector. At the same time, adjust the action time of the large vector that participates in the virtual space vector synthesis with it accordingly, so that the virtual space vector before and after adjustment satisfies the volt-second balance constraint.

[0013] Small sector 5: Divide small sector 5 into a first half sector and a second half sector; in the first half sector and the second half sector, select P-type redundant small vectors or N-type redundant small vectors to participate in the midpoint potential regulation according to the corresponding load current direction, and use dynamic adjustment factor k to adjust the action time of the selected redundant small vectors. At the same time, adjust the action time of the large vectors in the same direction accordingly, and introduce PPP zero vectors to compensate for the change in action time, so as to maintain the volt-second balance of the virtual space vector.

[0014] As a further optimization of the present invention, the DC bus voltage is set to... , the midpoint of the DC bus is O, and the midpoint voltage offset is defined , where is the potential of the midpoint of the DC bus relative to the negative terminal of the DC bus; the value range of the dynamic adjustment factor k is −1<k<1, k has the same sign as Δu, and |k| increases monotonically as |Δu| increases; when Δu>0, according to the current load current direction, among the redundant small vectors participating in virtual space vector synthesis, the midpoint current is increased the action time of the redundant small vector, and the midpoint current is reduced the action time of the redundant small vector; the opposite is true when Δu<0; wherein, the midpoint current takes the direction flowing out from the midpoint O as the positive direction, the increment and decrement of the action time are determined by k, and through the compensation of non-midpoint influencing vectors, the virtual space vector before and after adjustment satisfies the volt-second balance constraint, and the amplitude and phase remain unchanged.

[0015] As a further optimization of the present invention, the redundant vectors are redundant small vectors that exert opposite influences on the midpoint potential; the non-midpoint influencing vectors are zero vectors and large vectors that have no effect on the midpoint potential.

[0016] The VSVPWM-based midpoint potential balance control method for T-type three-level inverters provided by the present invention can quickly suppress midpoint potential offset by designing a coordinated adjustment strategy by sectors, combined with the dynamic adjustment factor k, and the balance capability is stable in the full modulation ratio range; the present invention adopts the virtual vector with fixed amplitude and phase, cooperates with the volt-second balance constraint, effectively reduces output voltage harmonics and avoids waveform distortion caused by adjustment; the present invention is implemented by a pure software algorithm, no additional hardware circuit is required, the calculation amount is small, it can run in real time based on DSP and FPGA, and is compatible with existing T-type three-level topologies. Brief Description of the Drawings

[0017] Figure 1 is a topology diagram of a T-type three-level inverter;

[0018] Figure 2 is Figure 1 the charging and discharging diagram of the DC bus;

[0019] Figure 3 is a midpoint current diagram under the action of the small vector POO;

[0020] Figure 4 is a space voltage vector diagram in the α-β stationary coordinate system;

[0021] Figure 5 is a space vector diagram based on VSVPWM;

[0022] Figure 6 is a small sector distribution diagram in sector I. Detailed Description of the Embodiments

[0023] like Figure 1 , 2 As shown, in the topology of the currently used T-type three-level inverter, O is the DC bus voltage, and O is the midpoint of the DC bus. , These are the capacitors on the upper and lower sides of the DC bus, respectively. , These are the voltages of the upper and lower capacitors, respectively. , These are the currents on the upper and lower sides of the DC bus, respectively.

[0024] Assuming current , The flow passes through the series capacitors from top to bottom. and When the current is positive, the midpoint current is positive. The direction of flow out from neutral point O is taken as the positive direction; The capacitance values ​​are equal, denoted by C. And defined as follows: Indicates the output state of the corresponding phase. This corresponds to state P. This corresponds to state O. This corresponds to state N. The relationship with the potential at point O can be expressed as:

[0025]

[0026] Integrating the above differential equation yields the neutral point voltage. The final expression:

[0027]

[0028] In the formula, This is the initial value of the midpoint voltage.

[0029] Each phase arm consists of four power switches with anti-parallel diodes. The structure allows each phase arm to independently output three voltage levels by controlling the specified on / off combinations of four switching transistors: positive (P), zero (O), and negative (N).

[0030] The combined working states of the three-phase bridge arms generate These 27 space voltage vectors are mapped onto the α-β stationary coordinate plane, such as... Figure 4 As shown, based on the magnitude of each vector, these 27 vectors can be categorized into four types of space voltage vectors, namely:

[0031] 3 types of zero vectors That is, three phases , , Simultaneously outputting the same level has no effect on the midpoint potential;

[0032] 12 small vectors That is, three phases , , Two phases are at the same level, and the other phase is at the neutral point 0 level. Each small vector corresponds to two redundant switching states, which can be further subdivided into N-type small vectors and P-type small vectors. N-type small vectors are also called negative small vectors, and P-type small vectors are also called positive small vectors. The two have the same amplitude, but they have opposite effects on the DC side neutral point capacitor. One charges the neutral point potential, and the other discharges the neutral point potential. The two have complementary regulation effects on the neutral point potential that are equal in magnitude and opposite in direction.

[0033] 6 types of vectors That is, three phases , , With one phase positive, one phase negative, and one phase at 0 level, the effect on the neutral point potential is not adjustable.

[0034] 6 types of large vectors That is, three phases , , One phase outputs a positive level, and two phases output a negative level, or two phases output positive and one phase outputs negative, which has no effect on the neutral point potential.

[0035] The midpoint current and its effect on the midpoint voltage corresponding to each vector are shown in the table below:

[0036]

[0037] Taking a small vector POO as an example, such as Figure 4 As shown, the current flowing through the neutral point ,when hour, The midpoint potential increases. Similarly, the effects of other vector actions on the midpoint current and its influence on the neutral point can be obtained. The action of different vectors will change the direction and magnitude of the midpoint current, thus affecting the midpoint voltage balance.

[0038] To address this issue, this embodiment employs Virtual Space Vector Pulse Width Modulation (VSVPWM), combining multiple basic space voltage vectors according to specific rules to construct a new, equivalent virtual space vector. This ensures that the amplitude and phase of the virtual vector remain fixed, and that the amplitude and phase of the virtual vector do not change with fluctuations in the midpoint potential. Thus, the isolation and elimination of the midpoint potential's influence are achieved at the vector synthesis level. The specific rules for virtual vector synthesis are shown in the table below:

[0039]

[0040] Based on the constructed virtual vector set, the VSVPWM modulation algorithm uses conventional vector synthesis calculation. Specifically, the modulation process includes three main steps: sector positioning, vector action time calculation, and switch sequence generation.

[0041] The sector positioning process involves dividing the spatial vector map into six large sectors, N=Ⅰ~Ⅵ, as detailed below. Figure 5 As shown, each large sector is further divided into 5 smaller sectors, n=1 to 5, as follows: Figure 6 As shown. Then determine the reference voltage space vector. The location of the large sector, and through the formula Standardize the vector. , For the α-axis and β-axis components obtained after per-unit processing, the location of the reference voltage space vector within the large sector is determined by the following table:

[0042]

[0043] In the vector action time calculation step, to simplify the calculation, a rotation transformation is used to uniformly rotate and map the reference voltage space vectors in other large sectors to the 60° coordinate system corresponding to the first large sector. The mathematical expression for this transformation is:

[0044]

[0045] in , For the standard unit , The conditions for determining small sectors based on axis voltage are shown in the table below:

[0046]

[0047] Then, the action time of the vector in each small sector is obtained according to the volt-second balance principle, as shown in the table below:

[0048]

[0049] After obtaining the duration of each virtual voltage vector, the actual duration of each basic voltage vector can be further deduced based on the synthetic mapping relationship between the virtual vector and the basic spatial voltage vector. Taking the third sub-sector of sector I as an example:

[0050]

[0051] in For vectors The duration of action.

[0052] Taking large sector I as an example, the specific implementation is as follows:

[0053] In the midpoint potential balancing process of small sector 1, a cooperative adjustment scheme of redundant vectors and non-midpoint influencing vectors is adopted to construct a virtual vector, and the resultant vector expression is:

[0054]

[0055] where, is the changed vector action time. Projecting the vectors onto the 60° coordinate system, we can obtain:

[0056]

[0057] where , are the upper and lower capacitor voltages, and an adjustment factor is introduced to change the adjustment time of the small vectors.

[0058] Define the midpoint voltage offset , where is the potential of the midpoint of the DC bus relative to the negative terminal of the DC bus. The value range of the adjustment factor k is −1<k<1, which is positively correlated with the actual midpoint voltage offset; when the midpoint voltage is higher than the balance value, the proportion of the action time of the vector whose corresponding midpoint current is less than 0 is increased; when the midpoint voltage is lower than the balance value, the proportion of the action time of the vector whose corresponding midpoint current is greater than 0 is increased.

[0059] Specifically, k has the same sign as Δu, and |k| increases monotonically as |Δu| increases; when Δu>0, according to the current load current direction, among the redundant small vectors participating in the synthesis of the virtual space vector, the action time of the redundant small vector with the midpoint current is increased, and the action time of the redundant small vector with the midpoint current is decreased; the opposite is true when Δu<0; wherein, the midpoint current takes the direction flowing out from the midpoint O as the positive direction, the increment and decrement of the action time are determined by k, and through the compensation of the non-midpoint influencing vectors, the virtual space vector before and after the adjustment satisfies the volt-second balance constraint, and the amplitude and phase remain unchanged.

[0060] Since , the influence on the neutral point potential is associated with the current direction, while , the influence on the neutral point potential is associated with the current direction. In order to simplify the control process, small sector 1 is divided into two front and rear half sectors, and in the first half sector, , The duration of action remains unchanged, adjustment , , The duration of action means that the potential control at the center point of the first half of the sector is only related to the direction of the current. The direction is related, while the second half of the sector is the opposite, making... , The duration of action remains constant, adjustment , , The duration of action, in the latter half of the sector, the potential control at the center point is only related to the current direction. The directions are related, which simplifies control.

[0061] In the first half of sector 1, to strictly ensure that the output voltage remains constant, the constraint condition can be obtained according to the volt-second balance principle:

[0062]

[0063] definition ,make ,but

[0064]

[0065] Similarly, in the latter half of sector 1, the following equation holds:

[0066]

[0067] Right now:

[0068]

[0069] make ,but

[0070]

[0071] Similarly, in sector 2, it can be represented as

[0072]

[0073]

[0074]

[0075]

[0076] Similarly, maintain in the first half of the sector , The duration of action remains unchanged, adjustment , The duration of action means that the potential control at the center point of the first half of the sector is only related to the direction of the current. The direction is related, while the second half of the sector is the opposite, making... , The duration of action remains constant, adjustment , The duration of action, in the latter half of the sector, the potential control at the center point is only related to the current direction. The direction is related. To ensure a fixed vector magnitude, a zero vector PPP is introduced.

[0077] In the first half of sector 2, to strictly ensure that the output voltage amplitude remains constant, the volt-second balance characteristic must be followed, that is, the volt-second product of the vector action time before and after adjustment must be equal. Therefore, the constraint relationship can be obtained as follows:

[0078]

[0079] because Therefore, the second line can be obtained.

[0080] Substituting into the first line, we get

[0081]

[0082] make ,but

[0083]

[0084]

[0085] In the latter half of sector 2, the constraint relationship can be obtained as follows:

[0086]

[0087] because Therefore, the first line can be obtained.

[0088]

[0089]

[0090] make ,but

[0091]

[0092]

[0093] Similarly, in sector 3, maintain... , The duration of action remains unchanged, but adjustments are made. , , The duration of action ensures that the neutral point potential control of this sector is only related to the current direction. The directions are related. Similarly, the constraint relationship can be derived as follows:

[0094]

[0095]

[0096]

[0097]

[0098] because Therefore, the second line can be obtained.

[0099]

[0100] Therefore, we have the following formula:

[0101]

[0102]

[0103] make ,but

[0104]

[0105]

[0106] Similarly, in sector 4, maintain... The duration of action remains unchanged, but adjustments are made. , , The duration of action ensures that the potential control at the center point of this sector is only related to the direction of the current. The directions are related. Similarly, the constraint relationship can be derived as follows:

[0107]

[0108]

[0109]

[0110]

[0111] because Therefore, the first line can be obtained.

[0112]

[0113] Right now Therefore, we have the following formula:

[0114]

[0115] make ,but

[0116]

[0117]

[0118]

[0119] In sector 5, the volt-second balance principle applies.

[0120]

[0121]

[0122]

[0123]

[0124] Small sector 5 reduces the action time of small vectors with smaller amplitudes, changes the action time of large vectors in the same direction, introduces PPP vector-assisted modulation, and designs three adjustment modes based on the current flow direction:

[0125] (1) Divide sector 5 into two halves, the first half of sector remains unchanged. , The duration of action remains unchanged, adjustment , , The duration of action means that the potential control at the center point of the first half of the sector is only related to the direction of the current. The direction is related, while the second half of the sector is the opposite, making... , The duration of action remains constant, adjustment , , The duration of action, in the latter half of the sector, the potential control at the center point is only related to the current direction. Directionally related:

[0126] In the first half of sector 5:

[0127] because Therefore, the second line can be obtained.

[0128]

[0129] Right now Therefore, we have the following formula.

[0130]

[0131] make ,but

[0132]

[0133]

[0134] In the sector following sector 5:

[0135]

[0136] because Therefore, the first line can be obtained.

[0137]

[0138] Right now Therefore, we have the following formula.

[0139]

[0140] make ,but

[0141]

[0142]

[0143]

[0144] (2) Simultaneously adjust the P-type small vector and the N-type small vector, and adjust... , , Duration of action, maintain , The duration of action means that the potential control at the center point of the first half of the sector is only related to the direction of the current. The direction is related, while the second half of the sector is the opposite, making... , The duration of action remains constant, adjustment , , The duration of action, in the latter half of the sector, the potential control at the center point is only related to the current direction. Directionally related:

[0145] (3) Regulation , Duration of action, maintain , Since the duration of action remains constant, the potential control at the center point of the first half of the sector is only related to the direction of the current. The direction is related, while the second half of the sector is the opposite, making... , The duration of action remains constant, adjustment , The duration of action, in the latter half of the sector, the potential control at the center point is only related to the current direction. The directions are related.

[0146] Finally, based on the adjusted vector action time, a symmetrical PWM drive signal is generated to control the operation of the power switching transistors, achieving stable inverter operation and neutral point potential balance. To ensure the safe and reliable operation of the inverter and minimize the switching losses of the power devices, the arrangement of the switching sequence should adhere to the following two basic constraints:

[0147] 1. A direct jump from P to N is not allowed between any adjacent switching states, otherwise it will cause a momentary short circuit in the DC bus capacitor and cause the switching device to be damaged due to excessive voltage stress.

[0148] 2. Under the premise of satisfying the previous constraint, the total number of operations of each phase switching device in each switching cycle should be minimized in order to reduce switching losses and improve inverter efficiency.

[0149] Based on the above constraints, the vector action sequence of each small sector is optimized and arranged to obtain a nine-segment symmetrical switching sequence, as shown in the table below:

[0150] sector n Order of action of vectors n=1 ONN-OON-OOO-POO-PPO-POO-OOO-OON-ONN n=2 ONN-OON-PON-POO-PPO-POO-PON-OON-ONN n=3 ONN-PNN-PON-POO-PPO-POO-PON-PNN-ONN n=4 ONN-PNN-PON-PPN-PPO-PPN-PON-PNN-ONN n=5 ONN-OON-PON-PPN-PPO-PPN-PON-OON-ONN

[0151] This embodiment dynamically adjusts the action time of redundant vectors and non-midpoint influence vectors. The non-midpoint influence vector is either a zero vector or a large vector in the same direction. It precisely modulates the equivalent vector synthesized from virtual vectors in the target direction, and ultimately effectively suppresses output voltage distortion during the dynamic process of midpoint potential balance.

[0152] To further clarify the specific application of the above control strategy, taking the nine-segment modulation sequence of the first sub-sector of sector I as an example, the nine-segment modulation sequence and corresponding time of the first half of the first sub-sector of sector I are shown in the table below:

[0153]

[0154] The nine-segment modulation sequence and corresponding times for the second half of the first sub-sector of sector I are shown in the table below:

[0155]

[0156] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit the scope of protection of the present invention. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the essence and scope of the technical solutions of the present invention.

Claims

1. A method for balancing the neutral point potential of a T-type three-level inverter based on VSVPWM, first, a T-type three-level inverter topology model is constructed, characterized in that, It also includes the following steps: S1: The reference voltage space vector is normalized. The large sector where the reference voltage space vector is located is determined according to the boundary conditions of the voltage vector components. The reference voltage space vectors in different large sectors are uniformly mapped to the 60° coordinate system corresponding to the first large sector through rotation coordinate transformation. According to the relationship of the voltage vector components after mapping, each large sector is divided into 5 small sectors. A virtual space vector is constructed based on the virtual space vector pulse width modulation method. The action time of each virtual space vector is calculated according to the volt-second balance principle. S2: Based on the small sector where the reference voltage space vector is located, the midpoint voltage deviation, and the load current direction associated with the voltage vector, a dynamic adjustment factor k is introduced to dynamically adjust the action time of the redundant small vectors participating in the virtual space vector synthesis. The change in the action time of the redundant small vectors is compensated by the corresponding non-midpoint influence vector so that the virtual space vectors before and after adjustment meet the volt-second balance constraint. S3: Generates a drive signal based on the adjusted vector action time to control the operation of the power switch of the T-type three-level inverter.

2. The method for neutral point potential balance control of a T-type three-level inverter based on VSVPWM according to claim 1, characterized in that, In step S1, the 27 basic space voltage vectors of the three-level inverter are divided into zero vector, small vector, medium vector and large vector according to their amplitude. Among them, the small vectors include P-type redundant small vectors and N-type redundant small vectors that have the same spatial position but have opposite effects on the midpoint potential. A virtual space vector is formed by combining multiple basic space voltage vectors, so that the amplitude and phase of the virtual space vector are fixed.

3. The method for neutral point potential balance control of a T-type three-level inverter based on VSVPWM according to claim 1, characterized in that, Step S2 specifically includes: Sector 1: Sector 1 is divided into a first half sector and a second half sector. In the first half sector and the second half sector, according to the corresponding load current direction, a P-type redundant small vector or an N-type redundant small vector that can generate an adjustment effect opposite to the midpoint voltage deviation is selected as the adjustment vector. The action time of the selected redundant small vector is adjusted using the dynamic adjustment factor k. At the same time, the action time of the corresponding zero vector is adjusted to maintain the volt-second balance of the virtual space vector. Small sector 2: Divide small sector 2 into a first half sector and a second half sector; in the first half sector and the second half sector, select a P-type redundant small vector or an N-type redundant small vector to participate in the midpoint potential regulation according to the corresponding load current direction, and use the dynamic adjustment factor k to adjust the action time of the selected redundant small vector. At the same time, introduce a PPP zero vector to compensate for the change in action time, so as to keep the amplitude and phase of the virtual space vector unchanged before and after the adjustment. Small sectors 3 and 4: Do not divide the small sector into first half and second half sectors. According to the load current direction, select a P-type redundant small vector or N-type redundant small vector that can generate the required midpoint current direction. Use the dynamic adjustment factor k to adjust the action time of the selected redundant small vector. At the same time, adjust the action time of the large vector that participates in the virtual space vector synthesis with it accordingly, so that the virtual space vector before and after adjustment satisfies the volt-second balance constraint. Small sector 5: Divide small sector 5 into a first half sector and a second half sector; in the first half sector and the second half sector, select P-type redundant small vectors or N-type redundant small vectors to participate in the midpoint potential regulation according to the corresponding load current direction, and use dynamic adjustment factor k to adjust the action time of the selected redundant small vectors. At the same time, adjust the action time of the large vectors in the same direction accordingly, and introduce PPP zero vectors to compensate for the change in action time, so as to maintain the volt-second balance of the virtual space vector.

4. The method for neutral point potential balance control of a T-type three-level inverter based on VSVPWM according to claim 1, characterized in that, setting the DC bus voltage as , the midpoint of the DC bus is O, and the midpoint voltage offset is defined , where is the potential of the DC bus midpoint relative to the negative terminal of the DC bus; the value range of the dynamic adjustment factor k is −1<k<1, k has the same sign as Δu, and |k| increases monotonically as |Δu| increases; when Δu>0, according to the current load current direction, among the redundant small vectors participating in virtual space vector synthesis, the action time of the redundant small vectors that increase the midpoint current is increased, and the action time of the redundant small vectors that reduce the midpoint current is decreased; the situation is opposite when Δu<0; wherein, the midpoint current takes the direction flowing out from the midpoint O as the positive direction, the increment or decrement of the action time is determined by k, and through the compensation of non-midpoint influencing vectors, the virtual space vector before and after adjustment satisfies the volt-second balance constraint, and the amplitude and phase remain unchanged.

5. The method for neutral point potential balance control of a T-type three-level inverter based on VSVPWM according to claim 1, characterized in that, Redundant vectors are small redundant vectors that have the opposite effect on the midpoint potential; non-midpoint influence vectors are zero and large vectors that have no effect on the midpoint potential.