A modular multilevel converter thermal balance control method and system based on a dynamic holding factor

Abstract

The application discloses a kind of modular multilevel converter thermal balance control method and system based on dynamic retention factor, for the thermal imbalance problem of flexible direct current transmission modular multilevel converter, propose thermal-electricity collaborative control method based on dynamic retention factor (DRF), break through the limitation of traditional capacitor voltage balancing strategy.Capacitor voltage tolerance boundary (rated value ±5%~10%) is set, combined with real-time temperature feedback to build dynamic weight adjustment mechanism, collect IGBT junction temperature to calculate normalized temperature deviation value, generate dynamic retention coefficient, and then obtain the corrected capacitor voltage, in charging mode, low-temperature module is preferentially put into in ascending order, in discharging mode, high-temperature module is preferentially put into in descending order;Further integrate model predictive control (MPC), based on thermal model rolling solution minimization temperature difference objective function, adjust dynamic adjustment coefficient ΔK.The scheme can effectively solve the problem of small-capacitor module thermal fatigue early decline, provide high reliability technical support for high-voltage high-power new energy grid-connected equipment.

Description

Technical Field

[0001] This invention relates to the field of flexible DC transmission systems, specifically a method and system for thermal balance control of modular multilevel converters for flexible DC transmission based on a dynamic holding factor. Background Technology

[0002] Modular multilevel converters (MMCs), as core equipment for flexible DC transmission, rely on a cascaded submodule structure and intelligent voltage balancing control to achieve efficient conversion of high-voltage power. Because the parameters of each submodule differ, without additional capacitor voltage control strategies, uneven capacitor voltages within the bridge arm will occur, leading to deterioration of power loss distribution, uneven heat distribution, and a surge in switching frequency among the submodules. This exacerbates IGBT switching losses and thermal stress accumulation, seriously threatening system reliability.

[0003] Therefore, capacitor voltage balance control is a necessary means to ensure the stable operation of MMC. However, traditional balance control strategies have the following drawbacks: 1. Severe losses: One-sided pursuit of absolute balance of capacitor voltage in all sub-modules deviates from actual engineering requirements, resulting in excessively high switching frequencies and increased device losses; 2. Uneven heat distribution caused by differences in capacitor values ​​in sub-modules: Small capacitor sub-modules are frequently switched on and off due to their fast charging and discharging speeds, resulting in significantly higher temperatures than large capacitor sub-modules and accelerating device aging; 3. High risk: Thermal cycling stress multiplies the risk of IGBT failure. For example, the probability of failure doubles for every 10°C increase in temperature. Summary of the Invention

[0004] This invention addresses the various problems existing in the prior art by providing a modular multilevel converter thermal balance control method and system based on a dynamic holding factor.

[0005] According to a first aspect of the present invention, a method for thermal balance control of a modular multilevel converter based on a dynamic holding factor is provided, comprising the following steps:

[0006] Step S1: Monitor the capacitor voltage of each sub-module of the modular multilevel converter in real time. and IGBT real-time junction temperature ;

[0007] Step S2: Based on the real-time junction temperature of the IGBT Calculate the normalized temperature deviation value of each submodule. ;

[0008] Step S3: Based on the normalized temperature deviation value Generate the dynamic retention coefficients of each submodule. ;

[0009] Step S4: Utilize the dynamic retention coefficient Correct the capacitor voltage of each submodule The corrected capacitor voltage is obtained. ;

[0010] Step S5: Based on the current direction of the current in the current arm and the corrected capacitor voltage... The sub-modules are sorted to determine their switching priority, thereby achieving thermal-electric coordinated control.

[0011] Where i∈N, N is a natural number greater than 1, representing the total number of the sub-modules.

[0012] Preferably, in step S2 above, the step of adjusting the IGBT junction temperature in real time... Calculate the normalized temperature deviation value of each submodule. Specifically:

[0013]

[0014] in, It is the real-time junction temperature of the IGBT in the i-th submodule; It is the average junction temperature of the IGBTs in all submodules of the current bridge arm; It is the maximum allowable junction temperature of the IGBT device.

[0015] Preferably, in step S3 above, the step based on the normalized temperature deviation value... Generate the dynamic retention coefficients of each submodule. Specifically:

[0016]

[0017] in, This is a dynamic adjustment coefficient used to control the dynamic holding coefficient. The adjustment range.

[0018] Preferably, in step S4 above, the use of the dynamic retention coefficient Correct the capacitor voltage of each submodule The corrected capacitor voltage is obtained. Specifically:

[0019] .

[0020] Preferably, in step S5 above, the step of determining the current direction of the current in the current arm and the corrected capacitor voltage is... The sub-modules are sorted as follows:

[0021] When the current of the current bridge arm is greater than zero and it is in charging mode, the sub-modules with lower corrected capacitor voltages are prioritized for power-on according to the ascending order of the corrected capacitor voltages, thereby reducing switching.

[0022] When the current of the current bridge arm is less than zero and it is in discharge mode, the sub-modules with higher corrected capacitor voltages are prioritized for switching according to the descending order of the corrected capacitor voltages, thereby increasing the switching frequency.

[0023] Preferably, the dynamic adjustment coefficient Online rolling optimization is performed using the Model Predictive Control (MPC) algorithm.

[0024] Preferably, the dynamic adjustment coefficient Online rolling optimization is performed using the Model Predictive Control (MPC) algorithm, specifically as follows:

[0025] A1: Establish a discrete state-space model with the junction temperature, heat capacity and total power loss of each sub-module at any time as state variables, to predict the temperature change in the next M steps;

[0026] A2: Define an objective function that aims to minimize the temperature difference between each sub-module in the current bridge arm within the next M steps;

[0027] A3: Set constraint conditions with capacitor voltage boundary, junction temperature safety limit and temperature change rate limit as constraints;

[0028] A4: In each control cycle, the optimization problem under the objective function is solved online on a rolling basis to determine the dynamic adjustment coefficient. ;

[0029] Where M is a natural number greater than 1, representing the total number of possible future times.

[0030] Preferably, in step A4 above, a sequence of dynamic adjustment coefficients for the next M steps is determined, and one of the dynamic adjustment coefficients is selected from the sequence as the dynamic adjustment coefficient for the current moment based on historical data of the dynamic adjustment coefficients.

[0031] Preferably, the capacitor voltage of each submodule It fluctuates within a tolerance range of ±5% to ±10% of the rated value.

[0032] According to a second aspect of the present invention, a modular multilevel converter thermal balance control system based on a dynamic holding factor is provided, the system comprising:

[0033] The real-time monitoring module is configured to collect the capacitor voltages of each sub-module of the modular multilevel converter in real time via sensors. and IGBT real-time junction temperature ;

[0034] The DRF calculation module, which is communicatively connected to the real-time monitoring module, is configured to calculate the normalized temperature deviation values ​​of each submodule. and dynamic retention coefficient ;

[0035] The voltage correction module is communicatively connected to the DRF calculation module and configured to receive the dynamic holding coefficient. and the capacitor voltage And calculate the corrected capacitor voltage. ;

[0036] The switching decision module, which is communicatively connected to the voltage correction module, is configured to receive the current direction of the current arm and the corrected capacitor voltage. Based on the corrected capacitor voltage Sort and generate PWM control signals;

[0037] The drive execution module is communicatively connected to the switching decision module and is configured to drive the IGBTs of the corresponding sub-modules to perform switching actions according to the PWM control signal.

[0038] A third aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of any of the aforementioned methods.

[0039] In a fourth aspect, the present invention provides a computer device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the steps of any of the aforementioned methods.

[0040] Compared with the prior art, the present invention has the following beneficial effects:

[0041] 1) Regarding the switching frequency, this invention introduces a voltage correction mechanism based on the Dynamic Hold Factor (DRF) to dynamically suppress the switching priority of non-limit-crossing sub-modules. This not only reduces the switching and conduction losses of power devices, but also reduces the power consumption of IGBTs at the switching frequency, while ensuring the thermal balance of the MMC module.

[0042] 2) Regarding the thermal effects of submodules, the Dynamic Holding Factor (DRF) dynamically adjusts the switching tendency through temperature-voltage coupling weights, solving the problem of premature aging of capacitor modules due to differences in capacitance values ​​and avoiding avalanche failure caused by local overheating.

[0043] 3) Regarding capacitor voltage fluctuations, voltage boundary tolerance control (rated value ±5%~10%) is adopted to replace the traditional absolute voltage equalization target; the DRF mechanism avoids the chain reaction caused by voltage disturbances by suppressing the switching of non-critical modules, providing stronger voltage disturbance immunity for the grid with high penetration of new energy sources and reducing the capacity requirements of reactive power compensation devices.

[0044] 4) In terms of real-time dynamic adjustment, the integrated model predictive control (MPC) provides a "prediction-optimization" closed-loop control interface for the smart grid through online rolling optimization of the dynamic adjustment coefficient, which is suitable for the random fluctuation scenario of wind and solar power. Attached Figure Description

[0045] Figure 1 This is a structural diagram of the control algorithm for the modular multilevel converter (MMC) system of the present invention;

[0046] Figure 2 This is a hardware architecture diagram of the thermal balance control system of the present invention;

[0047] Figure 3 This is a flowchart of the control algorithm based on the dynamic retention factor of the present invention;

[0048] Figure 4 This is a flowchart of the Model Predictive Control (MPC) algorithm of this invention;

[0049] Figure 5 Simulation diagrams showing IGBT switching operation under different dynamic holding factors (DRF). Detailed implementation method.

[0050] The techniques described below can be modified in various ways and have multiple embodiments, which are described in detail below with reference to the accompanying drawings. However, this does not mean that the techniques described below are limited to the specific embodiments. It should be understood that the present invention includes all similar modifications, equivalents, and substitutions without departing from the spirit and scope of the techniques described below.

[0051] Figure 1 This diagram illustrates the control algorithm structure of the Modular Multilevel Converter (MMC) system of this invention. The core of the dynamic hold factor algorithm section in the diagram lies in DRF control based on the temperature coefficient. Here, T* represents the target temperature, and T... H and T L These are the temperatures of the submodules with the highest and lowest temperatures, respectively. T* is set to a range of [0, (T...). H -T L[2] If the traditional algorithm is used directly, the small capacitor submodule will be frequently switched on and off due to its fast charging and discharging speed, causing its temperature to rise continuously; while the large capacitor submodule will be switched on and off less often due to its slow charging and discharging speed, resulting in a relatively lower temperature. This uneven temperature distribution not only exacerbates the thermal stress of the small capacitor submodule, but may also cause local overheating problems, further affecting the thermal stability and reliability of the MMC system. With the help of DRF control, the traditional algorithm's "one-size-fits-all" control method for each submodule can be broken. Based on the real-time information of the submodule (such as capacitor voltage value, initial switching status of the submodule, etc.), the switching priority of each submodule can be dynamically adjusted according to the submodule's operating status.

[0052] Figure 2 This is a hardware architecture diagram of the thermal balance control system of the present invention. The control system includes the following units: a temperature monitoring unit, which uses a fiber optic temperature sensor mounted on the IGBT substrate of the submodule; and a data processing unit, which transmits the IGBT temperature signal to the data processing unit via a photoelectric conversion module. The data processing unit uses an FPGA main control chip to calculate the average junction temperature of the bridge arm in real time. Generate dynamic retention coefficient Perform capacitor voltage correction Data is acquired via an ADC circuit, and the calculation results are stored in a memory chip. In the switching execution unit, the driver chip receives the PWM signal from the FPGA main control chip via optical fiber, and transmits the IGBT gate-level drive signal to the sub-module power unit via a gate-level resistor network, thereby triggering the sub-module IGBT to switch.

[0053] Figure 3 This is a flowchart illustrating the thermal balance control of a flexible DC transmission modular multilevel converter based on a dynamic holding factor, as described in this invention. To optimize the switching frequency, a dynamic holding factor mechanism is introduced. Specifically, for submodules in a disconnected state but with voltages above the lower limit, their voltage values ​​are multiplied by a dynamic holding factor DRF (recommended value 1.02~1.05) slightly greater than 1 before being sorted. Similarly, the same process is applied to submodules in an engaged state or with voltages exceeding the upper limit. This mechanism, through a numerical amplification effect, substantially increases the probability that these submodules will maintain their current state during the next control trigger, thereby effectively reducing unnecessary switching operations. Theoretical analysis shows that the introduction of the dynamic holding factor DRF can reduce the switching frequency by approximately 15%~30%. Simulations also verify that this method significantly improves system efficiency while maintaining voltage fluctuation amplitude (typically controllable within ±2%).

[0054] The core idea of ​​this invention is to abandon absolute voltage balancing, allowing the capacitor voltage to fluctuate within ±5% to 10% of the rated value, while dynamically adjusting the switching priority. As mentioned above, by introducing the Dynamic Hold Factor (DRF) mechanism, temperature feedback is integrated into the capacitor voltage sorting process to correct the sorting weights, reduce unnecessary switching, and reduce the switching frequency and losses from the source.

[0055] To achieve the above objectives, a modular multilevel converter thermal balance control method based on a dynamic holding factor is provided, comprising the following steps:

[0056] Step S1: Real-time monitoring of capacitor voltages in each sub-module of the modular multilevel converter. and IGBT real-time junction temperature .

[0057] Step S2: Based on the real-time junction temperature of the IGBT Calculate the normalized temperature deviation value of each submodule. .

[0058] Normalized temperature deviation The calculation method is as follows:

[0059]

[0060] in, It is the real-time junction temperature of the IGBT in the i-th submodule; It is the average junction temperature of the IGBTs in all submodules of the current bridge arm; It is the maximum allowable junction temperature of the IGBT device.

[0061] Step S3: Based on normalized temperature deviation value Generate the dynamic retention coefficients of each submodule. .

[0062] Dynamic retention coefficient The calculation method is as follows:

[0063]

[0064] That is,

[0065]

[0066] in, This is a dynamic adjustment coefficient used to control the dynamic holding coefficient. The adjustment range is typically between 0.01 and 0.05, and can be dynamically adjusted according to the optimization algorithm model.

[0067] The dynamic retention coefficient This is considered as the dynamic hold factor (DRF) in the entire control process.

[0068] Step S4: Utilize the dynamic retention coefficient Correct the capacitor voltage of each submodule The corrected capacitor voltage is obtained. .

[0069] The corrected capacitor voltage The calculation method is as follows:

[0070] .

[0071] That is,

[0072]

[0073] Step S5: Based on the current direction of the current in the current arm and the corrected capacitor voltage... The sub-modules are sorted to determine their switching priority, thereby achieving thermal-electric coordinated control.

[0074] The specific sorting method for each submodule is as follows:

[0075] When the current of the current bridge arm is greater than zero and it is in charging mode, the sub-modules with lower corrected capacitor voltages are prioritized for power-on in ascending order according to the corrected capacitor voltage, thus reducing the switching.

[0076] When the current of the current bridge arm is less than zero and it is in discharge mode, the submodules with higher corrected capacitor voltages are prioritized for switching in descending order according to the corrected capacitor voltages.

[0077] That is, for the high-temperature small capacitor submodule ( >1): During charging, the charging sequence is shifted to the end, reducing the number of switching operations and thus lowering the temperature; for low-temperature large capacitor submodules ( <1): The order of discharge is shifted forward, increasing the number of switching operations, which in turn raises the temperature.

[0078] Where i ∈ N, N is a natural number greater than 1, representing the total number of submodules. Furthermore, the capacitor voltage of each submodule... It fluctuates within a tolerance range of ±5% to ±10% of the rated value.

[0079] In summary, when the submodule capacitor voltage is within the boundary range, DRF utilizes its equivalent dynamic holding factor. Amplifying the capacitor voltage of a submodule (high-temperature module) or decreasing the capacitor voltage of a submodule (low-temperature module) effectively extends the holding time of the current state of that submodule, reducing unnecessary triggering. The dynamic hold factor algorithm can improve system thermal management. Compared to traditional algorithms where frequent switching of submodules results in an average IGBT switching frequency above 300Hz, the dynamic hold factor algorithm reduces the switching frequency to ≤200Hz, a decrease of approximately 30%, when no switching is required.

[0080] Furthermore, the dynamic adjustment coefficient Online rolling optimization is performed using the Model Predictive Control (MPC) algorithm.

[0081] Figure 4 This is a flowchart of the Model Predictive Control (MPC) algorithm of this invention. Model Predictive Control (MPC) is an advanced control strategy based on a system model, with rolling time-domain optimization as its core, and explicitly handling constraints. It solves the finite-time-domain optimal control problem in real time within each sampling period through a closed-loop mechanism of "prediction-optimization-correction-rolling," and only implements the optimal control quantity at the current moment, subsequently repeating the process.

[0082] Dynamic adjustment coefficient The specific steps for online rolling optimization using the Model Predictive Control (MPC) algorithm are as follows:

[0083] A1: Establish a discrete state-space model with the junction temperature, heat capacity, and total power loss of each submodule at any time as state variables to predict the temperature change in the next M steps.

[0084] The discrete state equations are constructed as follows:

[0085]

[0086] in, Let i be the junction temperature of the i-th submodule at time k; Let i be the total power loss of the i-th submodule at time k; Let i be the heat capacity of the i-th submodule; To control the cycle.

[0087] Where k∈M, M is a natural number greater than 1, representing the total number of possible future times.

[0088] A2: Define an objective function that aims to minimize the temperature difference between each sub-module in the current bridge arm within the next M steps.

[0089] The objective function for minimizing the temperature difference between the bridge arms is constructed as follows:

[0090]

[0091] in, Let k be the average junction temperature of the IGBTs in all submodules of the bridge arm at time k.

[0092] A3: Set constraints based on capacitor voltage boundary, junction temperature safety limit, and temperature change rate limit.

[0093] Wherein, voltage boundary: U min with U max The maximum allowable voltage U of the IGBT device rated It fluctuates within the range of ±5% to 10% of the rated value.

[0094] Junction temperature safety threshold: .

[0095] Temperature change rate boundary: .

[0096] A4: In each control cycle, the optimization problem under the objective function is solved online on a rolling basis to determine the dynamic adjustment coefficient. .

[0097] Determining the dynamic adjustment coefficient First, the dynamic adjustment coefficient sequence for the next M steps can be determined, i.e., [△K0, △K1, ..., △K]. M-1 ]. In selecting the dynamic adjustment coefficient When k is in k, the first value ΔK0 in the sequence is usually taken as the optimal solution for the current time k.

[0098] However, this is not the only possibility. In some cases, other values ​​in the sequence can be chosen as the optimal dynamic adjustment coefficient for the current time k. For example, the dynamic retention coefficients of multiple sub-modules at the same time. All are greater than 1, or the dynamic retention coefficient of the same submodule is greater than 1 at different times. If all values ​​are greater than 1, then the value in the sequence that is larger than the first value ΔK0 is selected as the optimal dynamic adjustment coefficient.

[0099] Figure 5 The figures show simulation results of IGBT switching behavior under different Dynamic Hold Factors (DRF). The Dynamic Hold Factor algorithm improves system thermal management. As can be seen from the figures, compared to the traditional algorithm which results in a high average switching frequency of IGBT devices due to frequent submodule switching, the Dynamic Hold Factor algorithm effectively reduces the switching frequency without the need for switching. This not only reduces the switching and conduction losses of power devices and the power consumption of IGBTs at the switching frequency, but also ensures the thermal balance of the MMC module.

[0100] The present invention also provides a modular multilevel converter thermal balance control system based on a dynamic holding factor, the system comprising:

[0101] The real-time monitoring module is configured to acquire the capacitor voltage of each sub-module of the modular multilevel converter in real time via sensors. and IGBT real-time junction temperature .

[0102] The DRF calculation module, which communicates with the real-time monitoring module, is configured to calculate the normalized temperature deviation values ​​of each submodule. and dynamic retention coefficient .

[0103] The voltage correction module, which communicates with the DRF calculation module, is configured to receive the dynamic hold coefficient. and capacitor voltage And calculate the corrected capacitor voltage. .

[0104] The switching decision module, which communicates with the voltage correction module, is configured to receive the current direction of the current arm and the corrected capacitor voltage. Based on the corrected capacitor voltage The signals are sorted and PWM control signals are generated.

[0105] The drive execution module is connected in communication with the switching decision module and is configured to drive the IGBTs of the corresponding sub-modules to perform switching actions according to the PWM control signal.

[0106] As another aspect of the present invention, a computer-readable storage medium is provided having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of any of the aforementioned methods.

[0107] As another aspect of the present invention, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the steps of any of the aforementioned methods.

[0108] Although the present invention has been described in detail above with general descriptions and specific embodiments, some modifications or improvements can be made to it. The above descriptions are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Other changes and modifications made by those skilled in the art without departing from the spirit and scope of the present invention are still included within the scope of protection of the present invention.

Claims

1. A modular multilevel converter thermal balance control method based on a dynamic holding factor, characterized in that, Includes the following steps: Step S1: Monitor the capacitor voltage of each sub-module of the modular multilevel converter in real time. and IGBT real-time junction temperature ; Step S2: Based on the real-time junction temperature of the IGBT Calculate the normalized temperature deviation value of each submodule. Specifically: ； in, It is the real-time junction temperature of the IGBT in the i-th submodule; It is the average junction temperature of the IGBTs in all submodules of the current bridge arm; This is the maximum allowable junction temperature of the IGBT device; Step S3: Based on the normalized temperature deviation value Generate the dynamic retention coefficients of each submodule. Specifically: ； in, This is a dynamic adjustment coefficient used to control the dynamic holding coefficient. The adjustment range; Step S4: Utilize the dynamic retention coefficient Correct the capacitor voltage of each submodule The corrected capacitor voltage is obtained. Specifically: ； Step S5: Based on the current direction of the current in the current arm and the corrected capacitor voltage... The sub-modules are sorted to determine their switching priority, thereby achieving thermal-electric coordinated control. Where i∈N, N is a natural number greater than 1, representing the total number of the sub-modules.

2. The control method according to claim 1, characterized in that, In step S5 above, the current direction of the current in the current bridge arm and the corrected capacitor voltage are used as the basis for the calculation. The sub-modules are sorted as follows: When the current of the current bridge arm is greater than zero and it is in charging mode, the sub-modules with lower corrected capacitor voltages are prioritized for power-on according to the ascending order of the corrected capacitor voltages, thereby reducing switching. When the current of the current bridge arm is less than zero and it is in discharge mode, the sub-modules with higher corrected capacitor voltages are prioritized for switching according to the descending order of the corrected capacitor voltages.

3. The control method according to claim 2, characterized in that, The dynamic adjustment coefficient Online rolling optimization is performed using the Model Predictive Control (MPC) algorithm.

4. The control method according to claim 3, characterized in that, The dynamic adjustment coefficient Online rolling optimization is performed using the Model Predictive Control (MPC) algorithm, specifically as follows: A1: Establish a discrete state-space model with the junction temperature, heat capacity and total power loss of each sub-module at any time as state variables, to predict the temperature change in the next M steps; A2: Define an objective function that aims to minimize the temperature difference between each sub-module in the current bridge arm within the next M steps; A3: Set constraint conditions with capacitor voltage boundary, junction temperature safety limit and temperature change rate limit as constraints; A4: In each control cycle, the optimization problem under the objective function is solved online on a rolling basis to determine the dynamic adjustment coefficient. ; Where M is a natural number greater than 1, representing the total number of possible future times.

5. The control method according to claim 4, characterized in that, In step A4 above, the sequence of dynamic adjustment coefficients for the next M steps is determined, and one of the dynamic adjustment coefficients is selected from the sequence as the dynamic adjustment coefficient for the current moment based on the historical data of the dynamic adjustment coefficients.

6. The control method according to claim 1, characterized in that, The capacitor voltage of each submodule It fluctuates within a tolerance range of ±5% to ±10% of the rated value.

7. A modular multilevel converter thermal balance control system based on a dynamic holding factor, for performing the steps of the method as described in any one of claims 1 to 6, characterized in that, The system includes: The real-time monitoring module is configured to collect the capacitor voltages of each sub-module of the modular multilevel converter in real time via sensors. and IGBT real-time junction temperature ; The DRF calculation module, which is communicatively connected to the real-time monitoring module, is configured to calculate the normalized temperature deviation values ​​of each submodule. and dynamic retention coefficient ; The voltage correction module is communicatively connected to the DRF calculation module and configured to receive the dynamic holding coefficient. and the capacitor voltage And calculate the corrected capacitor voltage. ; The switching decision module, which is communicatively connected to the voltage correction module, is configured to receive the current direction of the current arm and the corrected capacitor voltage. Based on the corrected capacitor voltage Sort and generate PWM control signals; The drive execution module is communicatively connected to the switching decision module and is configured to drive the IGBTs of the corresponding sub-modules to perform switching actions according to the PWM control signal.

Images