MMC-BESS low-computational-load extended-level model predictive control method

Abstract

The application discloses a kind of MMC-BESS low-computing amount extended level model predictive control methods, adopt hierarchical architecture: output current control layer is determined theoretical optimal level by minimizing current envelope area, and take its adjacent two integer levels as candidate, only evaluate 2 candidates can obtain optimal output level, constant 2 with the number of sub-modules, irrelevant;Based on the result, the circulating current control layer obtains the optimal bridge arm input combination by direct calculation and rounding, without exhaustive search, the amount of calculation is zero.Simultaneously set bridge arm input combination boundary constraint mechanism to ensure physical realizability.This method completely eliminates the tuning problem of weight factor in traditional model predictive control, the total amount of calculation is extremely low and does not increase with the number of sub-modules, with excellent output current quality, circulating current suppression ability and sub-module state of charge balancing performance, especially suitable for large-scale modular multilevel energy storage system.

Description

Technical Field

[0001] This invention relates to a low computational cost extended level model predictive control method for MMC-BESS, belonging to the field of power electronic converters. Background Technology

[0002] Modular multilevel energy storage converters (MMC-BESS) are of significant value in enhancing the grid's ability to absorb renewable energy due to their excellent harmonic performance, modular scalability, and high reliability brought by integrated energy storage. The control objectives of MMC-BESS are multi-tasking, primarily including: output current tracking, circulating current suppression, and submodule state-of-charge balancing.

[0003] Model predictive control (MRC) has been widely used in the control of modular multilevel converters due to its simple structure, rapid dynamic response, and outstanding multi-objective coordinated optimization capabilities. However, traditional MRC methods have three main limitations: first, the computational load increases sharply with the number of sub-modules, requiring the evaluation of a large number of candidate switch combinations in each sampling period; second, the weighting factor is difficult to tune; and third, power quality is easily affected.

[0004] To reduce computational load, existing technologies have proposed various optimized level model predictive control methods. One type retains weight factors, which reduces computational load by screening candidate combinations or grouping and sorting, but the tuning process of weight factors is complex, and the computational burden is still positively correlated with the number of submodules. Another type is the weight factor-free method, which calculates candidate levels through discrete models, further reducing the computational load of output current control. However, its circulating current control performance deteriorates significantly when facing large-scale systems, and the computational load of circulating current control is still positively correlated with the number of submodules. In particular, existing extended level model predictive control methods, although improving output current harmonic performance, have not proven that the optimal level calculated in the output current stage is the globally optimal solution, and the problem of high computational load in circulating current control has not been fundamentally solved. Summary of the Invention

[0005] The purpose of this invention is to provide a low-computational-load model predictive control method for modular multilevel energy storage converters, which aims to eliminate the correlation between computational load and the number of sub-modules, determine the globally optimal solution for the output level, and improve circulating current control performance while maintaining excellent output current performance.

[0006] To achieve the above objectives, the present invention adopts a hierarchical control architecture to independently regulate the output current and circulating current. The specific technical solution includes an output current control layer and a circulating current control layer, where S1) to S7) is the output current control layer and S8) to S11) is the circulating current control layer.

[0007] S1) Obtain the current phase current sample value at the current moment. (Where j represents phase j, and all subsequent letters j refer to phase j) and predicted reference current value And judge and Size relationship;

[0008] S2) According to and Calculate the current reference value corresponding to minimizing the current enclosure area. ;

[0009] S3) According to The minimum area level was calculated. ;

[0010] S4) Perform rounding down and rounding up to obtain the adjacent low voltage. Peaceful high level ;

[0011] S5) According to The parity determines the corresponding bridge arm input combination. ;

[0012] S6) According to Parity derivation Corresponding bridge arm input combination ;

[0013] S7) respectively and Substitute the values ​​into the output current prediction model and calculate the corresponding current enclosing area cost function value. and The level with the smaller cost function value is selected as the optimal output level. ;

[0014] S8) Obtain the bridge arm input combination corresponding to the optimal output level. and circulation prediction reference values ;

[0015] S9) According to Calculate the upper and lower bridge arm voltages required to achieve circulating current tracking. ;

[0016] S10) Calculate the number of submodules for circulation control in each bridge arm. ;

[0017] S11) The optimal bridge arm combination is determined directly through the rounding function. ;

[0018] S12) Number of bridge arms input in the final output or Apply boundary value constraints;

[0019] In step S2), the current reference value corresponding to minimizing the current enclosing area is calculated. The formula is shown below:

[0020]

[0021] In step S3), the output current discrete model formula is used. , among them use The voltage difference between the upper and lower bridge arms can be directly calculated by substitution, thus obtaining the minimum area level. :

[0022]

[0023]

[0024] in This is the voltage of the lower bridge arm. For the upper bridge arm voltage, This is the reference voltage difference between the lower and upper bridge arms. This represents the average voltage across the capacitors of the phase sub-module. For bridge arm inductance, For load inductance, For load resistance, To control the cycle;

[0025] The rounding function in step S4) is: , ;

[0026] In step S5) Corresponding bridge arm input combination The constraint rule is: if If it is even, then

[0027]

[0028]

[0029] like If it is an odd number, then

[0030]

[0031]

[0032] Where N is the total number of submodules in a single bridge arm;

[0033] In step S6) Bridge arm input combination The constraint rule is: if If it is even, then

[0034]

[0035]

[0036] like If it is an odd number, then

[0037]

[0038]

[0039] The current-enclosed area cost function value in step S7) and The formulas for calculating the area of ​​a trapezoid are divided into the formula for calculating the area of ​​a triangle and the formula for calculating the area of ​​a trapezoid. The formula for calculating the area of ​​a trapezoid is: The formula for calculating the area of ​​a triangle is: The constraint method for which area calculation formula to use is as follows: If > , Using the formula for calculating the area of ​​a triangle, when corresponding > If the triangle is large, then the triangle calculation formula is used; otherwise, the trapezoid calculation formula is used. < , Using the formula for calculating the area of ​​a triangle, when corresponding > If the value is large, then the trapezoidal formula is used; otherwise, the triangle formula is used.

[0040] In step S9), the following formula is used based on the discrete circulation model: , among them use The replacement can directly calculate the voltages of the upper and lower bridge arms. :

[0041]

[0042] in This is the DC bus voltage. This represents the current circulating current sample value.

[0043] In step S10), the number of sub-modules for circulation control in each bridge arm is calculated. The formula is:

[0044]

[0045]

[0046] In step S11), the optimal bridge arm deployment combination is obtained by the following rounding function. :

[0047]

[0048]

[0049] in The function is a rounding function. This combination is the global optimal solution that minimizes the value of the circulation cost function. No traversal calculation is required during the circulation control stage.

[0050] The constraint rule in step S12) is that if the total number of sub-modules invested... or If the number of inputs exceeds the limit but there are redundant combinations, then the number of inputs to the submodules of the two bridge arms is increased or decreased synchronously to bring them back to the normal range. Range; for special boundary crossing situations, including , , or ,in If the integer is a positive integer exceeding the specified range, then replace it with... , , or Output the results.

[0051] Compared with the prior art, the present invention has the following beneficial effects:

[0052] 1. Extremely low computational load: Only 2 candidate levels need to be evaluated in the output current control stage, and no traversal search is required in the circulating current control stage, with a computational load of 0. The total computational load for each control cycle is constant at 2, which is independent of the number of sub-modules, making it particularly suitable for large-scale MMC-BESS systems.

[0053] 2. No weighting factor: The hierarchical control architecture is adopted, and the output current and circulating current are controlled by independent cost functions, which completely eliminates the problem of weighting factor tuning in the traditional MPC method;

[0054] 3. Excellent circulating current control performance: The optimal bridge arm combination is obtained through direct calculation mechanism, which significantly reduces the steady-state error and fluctuation amplitude of the circulating current, and the circulating current can closely follow the reference value;

[0055] 4. Comprehensive control performance: While reducing the amount of computation, it maintains excellent output current THD performance and SOC balancing capability, and its overall performance is superior to existing methods;

[0056] 5. Guarantee of global optimal solution: For the first time, the global optimal level position for minimizing the current enclosure area is derived theoretically, and the selected level is ensured to be the output global optimal solution through a high-low level comparison mechanism. Attached Figure Description

[0057] Figure 1 This is a topology diagram of MMC-BESS in an embodiment of the present invention;

[0058] Figure 2 This is a flowchart illustrating the control method of the present invention;

[0059] Figure 3 This is a flowchart illustrating the bridge arm input combined boundary constraint method of the present invention;

[0060] Figure 4 This is a comparison chart of the present invention and traditional methods in terms of computational complexity;

[0061] Figure 5 This is a comparison chart of the steady-state performance of the present invention and the traditional method;

[0062] Figure 6 This is a comparison diagram of the dynamic response waveforms of the present invention and the traditional method;

[0063] Figure 7 This is a comparison chart of SOC equilibrium between the present invention and the traditional method. Detailed Implementation

[0064] The specific embodiments of the present invention will be further described below with reference to the accompanying drawings. It should be noted that the technical solution provided in this embodiment uses a modular multilevel energy storage converter (MMC-BESS) as the application object, but the present invention is also applicable to other energy storage or drive systems that adopt modular multilevel topologies. The following description is only illustrative and does not constitute a limitation on the scope of protection of the present invention.

[0065] like Figure 1 The diagram shown is a topology diagram of the MMC-BESS in an embodiment of the present invention. Each phase consists of an upper arm and a lower arm, and each arm contains N sub-modules (SMs) and an arm inductor. A load inductor is connected in series on the load side. and load resistance The DC bus voltage is This invention achieves AC / DC energy conversion and energy storage unit integration through a modular multilevel structure. The low-computational-load model predictive control method of this invention employs a hierarchical control architecture, independently regulating the output current and circulating current. Specifically, it includes an output current control layer and a circulating current control layer, and its control flow is as follows: Figure 2 As shown, Figure 3 The boundary constraint handling process for bridge arm input combination is given.

[0066] The method of the present invention will be described in detail below with reference to specific steps.

[0067] Output current control layer:

[0068] The output current control layer takes output current tracking as the control objective, determines the optimal level by minimizing the current enclosing area, and selects two candidate levels for cost function evaluation, thereby reducing the amount of computation.

[0069] S1) Obtain the current phase current sample value at the current moment. (Where j represents phase j, and all subsequent letters j refer to phase j) and predicted reference current value And judge and Size relationship;

[0070] S2) According to and Calculate the current reference value corresponding to minimizing the current enclosure area. ;

[0071] S3) According to The minimum area level was calculated. ;

[0072] S4) Perform rounding down and rounding up to obtain the adjacent low voltage. Peaceful high level ;

[0073] S5) According to The parity determines the corresponding bridge arm input combination. ;

[0074] S6) According to Parity derivation Corresponding bridge arm input combination ;

[0075] S7) respectively and Substitute the values ​​into the output current prediction model and calculate the corresponding current enclosing area cost function value. and The level with the smaller cost function value is selected as the optimal output level. ;

[0076] In step S2), the current reference value corresponding to minimizing the current enclosing area is calculated. The formula is shown below:

[0077]

[0078] In step S3), the output current discrete model formula is used. , among them use The voltage difference between the upper and lower bridge arms can be directly calculated by substitution, thus obtaining the minimum area level. :

[0079]

[0080]

[0081] in This is the voltage of the lower bridge arm. For the upper bridge arm voltage, This is the reference voltage difference between the lower and upper bridge arms. This represents the average voltage across the capacitors of the phase sub-module. For bridge arm inductance, For load inductance, For load resistance, To control the cycle;

[0082] The rounding function in step S4) is: , ;

[0083] In step S5) Corresponding bridge arm input combination The constraint rule is: if If it is even, then

[0084]

[0085]

[0086] like If it is an odd number, then

[0087]

[0088]

[0089] Where N is the total number of submodules in a single bridge arm;

[0090] In step S6) Bridge arm input combination The constraint rule is: if If it is even, then

[0091]

[0092]

[0093] like If it is an odd number, then

[0094]

[0095]

[0096] The current-enclosed area cost function value in step S7) and The formulas for calculating the area of ​​a trapezoid are divided into the formula for calculating the area of ​​a triangle and the formula for calculating the area of ​​a trapezoid. The formula for calculating the area of ​​a trapezoid is: The formula for calculating the area of ​​a triangle is: The constraint method for which area calculation formula to use is as follows: If > , Using the formula for calculating the area of ​​a triangle, when corresponding > If the triangle is large, then the triangle calculation formula is used; otherwise, the trapezoid calculation formula is used. < , Using the formula for calculating the area of ​​a triangle, when corresponding > If the value is large, then the trapezoidal formula is used; otherwise, the triangle formula is used.

[0097] The controller first obtains the phase current at the current moment by sampling the current. The predicted reference current is calculated through the upper-level control algorithm. .by For example, according to Calculate the reference current value corresponding to minimizing the area enclosed by the current. Subsequently, the output current discrete model was used. (in and Calculate the minimum area level required to achieve this current, which is an inverse function of the bridge arm voltage. .right Round down and up to obtain two consecutive high and low integer levels. Based on the parity of the levels, use... , , , Derive the number of sub-modules that can be engaged in the upper and lower bridge arms corresponding to these two voltage levels.

[0098] Substitute the number of bridge arms engaged in the two sets into the output current prediction model. (in and For the number of bridge arm insertions, (For the average voltage of phase SM), calculate the corresponding current encirclement area. The calculation of the area enclosed by the current needs to be based on... and Based on the size relationship, select the area of ​​the trapezoid. Or the formula for the area of ​​a triangle Compare the two area values ​​and select the level with the smaller area as the optimal output level. and put its corresponding bridge arm into the combination. It is passed to the circulation control layer.

[0099] Circulation control layer:

[0100] The circulation control layer takes circulation tracking as the control objective and obtains the optimal bridge arm combination through direct calculation without traversal search, thereby reducing the amount of computation while ensuring the accuracy of circulation control.

[0101] S8) Obtain the bridge arm input combination corresponding to the optimal output level. and circulation prediction reference values ;

[0102] S9) According to Calculate the upper and lower bridge arm voltages required to achieve circulating current tracking. ;

[0103] S10) Calculate the number of submodules for circulation control in each bridge arm. ;

[0104] S11) The optimal bridge arm combination is determined directly through the rounding function. ;

[0105] In step S9), the following formula is used based on the discrete circulation model: , among them use The replacement can directly calculate the voltages of the upper and lower bridge arms. :

[0106]

[0107] in This is the DC bus voltage. This represents the current circulating current sample value.

[0108] In step S10), the number of sub-modules for circulation control in each bridge arm is calculated. The formula is:

[0109]

[0110]

[0111] In step S11), the optimal bridge arm deployment combination is obtained by the following rounding function. :

[0112]

[0113]

[0114] in The function is a rounding function. This combination is the global optimal solution that minimizes the value of the circulation cost function. No traversal calculation is required during the circulation control stage.

[0115] The circulating current control layer receives the optimal bridge arm input combination from the output current control layer. and circulation prediction reference values .

[0116] Based on the discrete circulation model Calculate the upper and lower bridge arm voltages required to achieve circulating current tracking. Next, the actual bridge arm voltages at the current optimal output level are calculated, and the difference between the two is the voltage used for circulating current control. Dividing this voltage by twice the average submodule voltage yields the number of submodules used for circulating current control in each bridge arm. .

[0117] The key step is to directly address... Round to the nearest integer. Add this integer to the number of upper and lower bridge arms engaged at the optimal output level, respectively, to obtain the final bridge arm engagement combination. (in (Belongs to the set of integers). This process guarantees the minimum value of the circulation cost function without any traversal calculations, as shown in the mathematical proof below:

[0118] The circulating control cost function is Obviously, Size and Proportional. Due to the... Rounding can make To minimize, the optimal bridge arm deployment combination is: .

[0119] Bridge arm input combination boundary constraints

[0120] The rounding operation may cause the number of sub-modules connected to the bridge arm to exceed the allowable range. Furthermore, this invention sets up a boundary constraint mechanism to ensure the physical realizability of the output combination.

[0121] S12) Number of bridge arms input in the final output or Apply boundary value constraints;

[0122] The constraint rule in step S12) is that if the total number of sub-modules invested... or If the number of inputs exceeds the limit but there are redundant combinations, then the number of inputs to the submodules of the two bridge arms is increased or decreased synchronously to bring them back to the normal range. Range; for special boundary crossing situations, including , , or ,in If the integer is a positive integer exceeding the specified range, then replace it with... , , or Output;

[0123] To verify the effectiveness of the method of the present invention, this embodiment verifies the method through simulation and compares it with existing methods. The simulation parameters are as follows: control cycle. Number of submodules in a single bridge arm DC bus voltage Submodule capacitor Submodule rated voltage Bridge arm inductance Load inductance load resistance .

[0124] Figure 4 The diagram shows a comparison of the computational complexity of the present invention and traditional methods (Method 1 is the traditional EL-MPC, Method 2 is the EL-MPC based on optimized area, and Method 3 is the EL-MPC method of the present invention, and so on). The method of the present invention only needs to evaluate 2 candidate levels in the output current control stage and does not need to perform traversal calculations in the circulating current control stage. The total computational complexity is constant at 2, which is independent of the number of sub-modules and is significantly better than the traditional method.

[0125] Figure 5 The graph shows a comparison of the steady-state performance of the method of this invention and the traditional method. Under steady-state conditions with a modulation index of 0.9, the total harmonic distortion (THD) of the output current of the method of this invention is 0.8%, which is comparable to that of method 2 and better than that of method 1 (THD is 3.35%). In terms of circulating current control, the steady-state error of the circulating current of the method of this invention is 0.06 A, which is 68.4% lower than that of method 1 and 96.9% lower than that of method 2. The peak-to-peak value of the circulating current fluctuation is 6 A, which is between that of the two traditional methods, and the overall circulating current performance is significantly improved.

[0126] Figure 6 The diagram shows a comparison of the dynamic response waveforms of the present invention and the conventional method. In the output current dynamic response test, the reference current jumped from 45A to 22.5A. During the transient process, the output current quickly converged to the new steady-state value. The three methods exhibited similar dynamic characteristics and maintained good circulating current control throughout the process. Similarly, when the circulating current reference value jumped from 20A to 10A, all three methods achieved rapid tracking of the circulating current, exhibited consistent transient dynamic performance, and maintained excellent output current regulation capability throughout the process.

[0127] Figure 7 The diagram shows a comparison of SOC equalization between the present invention and the traditional method. It can be seen that the SOC equalization speed of the present invention and method 2 is consistent within the phase and in the SM, indicating that the present invention will not have a negative impact on SOC equalization.

[0128] Simulation results show that the proposed method achieves comparable output current THD performance to the traditional extended level model predictive control method based on current area optimization, while significantly improving circulating current ripple and steady-state error. The proposed method involves only two computations per control cycle, independent of the number of submodules. The output current control stage only requires evaluating two candidate levels, and the circulating current control does not require finite set traversal, allowing the optimal solution to be obtained directly through simple calculations. This significantly reduces the computational load compared to traditional methods.

[0129] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A low-computational-complexity extended-level model predictive control method for MMC-BESS, characterized in that, Includes the following steps: S1) Obtain the current phase current sample value at the current moment. (Where j represents phase j, and all subsequent letters j refer to phase j) and predicted reference current value And judge and Size relationship; S2) According to and Calculate the current reference value corresponding to minimizing the current enclosure area. ; S3) According to The minimum area level was calculated. ; S4) Perform rounding down and rounding up to obtain the adjacent low voltage. Peaceful high level ; S5) According to The parity determines the corresponding bridge arm input combination. ; S6) According to Parity derivation Corresponding bridge arm input combination ; S7) respectively and Substitute the values ​​into the output current prediction model and calculate the corresponding current enclosing area cost function value. and The level with the smaller cost function value is selected as the optimal output level. ; S8) Obtain the bridge arm input combination corresponding to the optimal output level. and circulation prediction reference values ; S9) According to Calculate the upper and lower bridge arm voltages required to achieve circulating current tracking. ; S10) Calculate the number of submodules for circulation control in each bridge arm. ; S11) The optimal bridge arm combination is determined directly through the rounding function. ; S12) Number of bridge arms input in the final output or Apply boundary value constraints.

2. The MMC-BESS low computational complexity extended level model predictive control method according to claim 1, characterized in that, In step S2), the current reference value corresponding to minimizing the current enclosing area is calculated. The formula is shown below:

3. The MMC-BESS low computational complexity extended level model predictive control method according to claim 1, characterized in that, In step S3), the output current discrete model formula is used. , among them use The voltage difference between the upper and lower bridge arms can be directly calculated by substitution, thus obtaining the minimum area level. : in This is the voltage of the lower bridge arm. For the upper bridge arm voltage, This represents the reference voltage difference between the lower and upper bridge arms. This represents the average voltage across the capacitors of the phase submodule. For bridge arm inductance, For load inductance, For load resistance, To control the cycle.

4. The MMC-BESS low computational complexity extended level model predictive control method according to claim 1, characterized in that, The rounding function in step S4) is: , .

5. The MMC-BESS low computational complexity extended level model predictive control method according to claim 1, characterized in that, In step S5) Corresponding bridge arm input combination The constraint rule is: if If it is even, then like If it is an odd number, then Where N is the total number of submodules in a single bridge arm.

6. The MMC-BESS low computational complexity extended level model predictive control method according to claim 1, characterized in that, In step S6) Bridge arm input combination The constraint rule is: if If it is even, then like If it is an odd number, then 7. The MMC-BESS low computational complexity extended level model predictive control method according to claim 1, characterized in that, The current-enclosed area cost function value in step S7) and The formulas for calculating the area of ​​a trapezoid are divided into the formula for calculating the area of ​​a triangle and the formula for calculating the area of ​​a trapezoid. The formula for calculating the area of ​​a trapezoid is: The formula for calculating the area of ​​a triangle is: The constraint method for which area calculation formula to use is as follows: If > , Using the formula for calculating the area of ​​a triangle, when corresponding > If the triangle is large, then the triangle calculation formula is used; otherwise, the trapezoid calculation formula is used. < , Using the formula for calculating the area of ​​a triangle, when corresponding > If the sum is large, the trapezoidal formula is used; otherwise, the triangle formula is used.

8. The MMC-BESS low computational complexity extended level model predictive control method according to claim 1, characterized in that, In step S9), the following formula is used based on the discrete circulation model: , among them use The replacement can directly calculate the voltages of the upper and lower bridge arms. : in This is the DC bus voltage. This represents the current circulating sample value.

9. The MMC-BESS low computational complexity extended level model predictive control method according to claim 1, characterized in that, In step S10), the number of sub-modules for circulation control in each bridge arm is calculated. The formula is:

10. The MMC-BESS low computational complexity extended level model predictive control method according to claim 1, characterized in that, In step S11), the optimal bridge arm deployment combination is obtained by the following rounding function. : in This is a rounding function, and this combination is the global optimal solution that minimizes the value of the circulation cost function. No traversal calculation is required during the circulation control stage.

11. The MMC-BESS low computational complexity extended level model predictive control method according to claim 1, characterized in that, The constraint rule in step S12) is that if the total number of sub-modules invested... or If the number of inputs exceeds the limit but there are redundant combinations, then the number of inputs for the submodules of the two bridge arms is increased or decreased synchronously to bring them back to the normal range. interval; For special boundary crossing situations, including , , or ,in If the integer is a positive integer exceeding the specified range, then replace it with... , , or Output the results.

Images