MMC Model Predictive Power Control System Based on Extended Power

Abstract

The present invention discloses an MMC model predictive power control system based on extended power, which includes modules such as an output power prediction module, a power selection and objective function calculation module, and a bridge arm energy prediction and objective function calculation module, etc. Aiming at the problem of power pulsation in MMC when a fault occurs in the AC power grid, the present invention uses the extended power calculation method, formulates the design principle of the power reference value of MMC, and realizes the tracking of the power reference value through model predictive control, effectively suppressing the power pulsation during the fault. Compared with the existing MMC control strategies, this method can suppress the active power pulsation or reactive power pulsation of MMC without the need for positive and negative sequence separation of voltage and current, and will not introduce additional current harmonic components. The structure of the MMC model predictive power control system based on extended power proposed in the present invention is simple and has strong engineering practical value.

Description

Technical Field

[0001] The invention belongs to the technical field of power electronics, and particularly relates to an MMC model predictive power control system based on extended power. Background Art

[0002] The flexible DC transmission technology based on the topology of the modular multilevel converter (MMC) replaces the direct series connection of switching devices by means of sub-module cascading, and has the advantages of low manufacturing difficulty, low switching loss, high waveform quality, etc., so it has been widely used. However, when a fault occurs in the AC power grid, the flexible DC transmission system will also be severely affected. The AC current and voltage will both contain negative sequence components, resulting in double-frequency pulsation of power, which seriously affects the control performance of the MMC.

[0003] As an advanced control theory, model predictive control has received extensive attention in the field of power electronics in recent years. In the model predictive control strategy, first, an objective function needs to be defined according to the expected control target, and then by calculating and comparing the objective function for different switching actions in each control period, the optimal solution is selected to control the system. Due to the great flexibility in the selection of the objective function, the objectives that can be achieved by this control strategy are very diverse. However, it should be noted that the model predictive control method based on the traversal method has a large amount of computation. To address this problem, a simplified MMC model predictive control strategy under an ideal power grid has been proposed in existing research. The input priorities of different sub-modules in the upper and lower bridge arms are sorted according to the sub-module capacitor voltage. When the number of sub-modules to be input is determined, the specific sub-module input method can also be determined as one. Therefore, it only needs to compare the objective functions when different numbers of sub-modules are input in the upper and lower bridge arms. This simplified method greatly optimizes the computational efficiency of the MMC model predictive control strategy. However, since this method does not consider the power pulsation suppression method of the MMC under grid faults, the safe operation of the system will be severely threatened when a grid fault occurs. In fact, compared with the MMC control strategy based on a linear PI controller, the model predictive control strategy is not limited by the control bandwidth and is more suitable for the optimal control of the MMC under a faulty power grid. Some scholars have proposed an extended power calculation method and applied it to the control strategy of a doubly-fed wind turbine under grid faults. If the extended power can be combined with the model predictive control strategy and applied to the control system of the MMC, it will be able to significantly improve the control performance of the MMC under grid faults and simplify its control system structure. Therefore, studying the MMC model predictive power control method based on extended power has important theoretical value and practical significance. Summary of the Invention

[0004] The object of the present invention is to overcome the deficiency that the output power of the MMC has pulsation under an unbalanced power grid in the above-mentioned existing technologies, combine the MMC model predictive control strategy with the extended power, and provide an MMC model predictive power control system based on the extended power.

[0005] In order to achieve the above object of the invention, the following technical solutions are adopted:

[0006] An MMC model predictive power control system based on extended power, characterized in that it includes a sampling module, a Clark transformation module, a voltage prediction module, an output current prediction module, an output power prediction module, a power selection and objective function calculation module, a bridge arm energy prediction and objective function calculation module, an internal circulating current prediction and objective function calculation module, a final objective function calculation module, and an objective function comparison and control instruction output module;

[0007] In the sampling module, it includes:

[0008] A voltage sampling module that samples the three-phase voltage U sabc of the AC side of the MMC converter, the DC bus voltage U dc , and the upper and lower bridge arm sub-module voltages U pabc (i) and U nabc (i), where i = 1 to N, and N is the number of sub-modules, and both the upper and lower bridge arms have N sub-modules;

[0009] A current sampling module that samples the three-phase current I vabc of the AC side of the MMC converter, the upper and lower bridge arm currents I pabc and I nabc ;

[0010] The Clark transformation module performs Clark transformation on the three-phase voltage U sabc , the three-phase current I vabc , the upper and lower bridge arm currents I pabc and I nabc , the upper and lower bridge arm sub-module voltages U pabc (i) and U nabc (i) to obtain the corresponding voltage vectors U sαβ , current vectors I vαβ , the upper and lower bridge arm currents I pαβ and I nαβ , the upper and lower bridge arm sub-module voltages U pαβ (i) and U nαβ (i) in the α-β coordinate system;

[0011] The voltage prediction module calculates the voltage vector U of the next sampling period according to the grid voltage U sαβ obtained in the current sampling periodsαβ(next) ;

[0012] The output current prediction module calculates the current vectors I of the next sampling period when different sub-module input methods are adopted for the upper and lower bridge arms in the current sampling period, according to the grid voltage U sαβ and current I vαβ obtained in this sampling period; among them, there are N + 1 sub-module input methods in this sampling period, and I needs to be calculated N + 1 times vαβ(next) ; vαβ(next) to obtain I vαβ(next) (m), m = 1, 2,... N + 1;

[0013] The output power prediction module calculates the traditional active power P sαβ(next) and reactive power Q vαβ(next) (m), extended active power P g(next) (m), and extended reactive power Q g(next) (m) of the next sampling period according to the voltage vector U gnew(next) and current vector I gnew(next) (m) of the next sampling period, where m = 1, 2,... N + 1;

[0014] In the power selection and objective function calculation module, it includes:

[0015] The control objective selection module selects the corresponding traditional power and extended power as the power feedback values P gsel (m) and Q gsel (m) according to different control objectives;

[0016] The output power objective function calculation module calculates the output current objective function J gsel (m) and Q gsel (m) and the current reference value I vαβref ; 1m ;

[0017] The arm energy prediction and objective function calculation module includes:

[0018] The arm energy prediction module calculates the voltage values U pαβ (i), U nαβ (i), U pαβ and I nαβ of different sub-modules of the upper and lower bridge arms in the next sampling period when N + 1 sub-module input methods are adopted for the upper and lower bridge arms in the current sampling period, according to the voltages U αmp(next) (i), U αmn(next) (i), U βmp(next) (i) of different sub-modules of the upper and lower bridge arms, the currents Iβmn(next) (i), and further calculate the total energy E of the upper and lower arm sub - modules in the next sampling period p(next) (m) and E n(next) (m);

[0019] The arm energy objective function calculation module, according to the obtained total energy E of the upper and lower arm sub - modules p(next) (m), E n(next) (m) and the upper and lower arm energy reference values, calculate the arm energy objective function J 2m , m = 1, 2,... N + 1, and the upper and lower arm energy reference values are given as NCU dc 2 ; where C is the capacitance of a single sub - module, CU dc 2 is the energy of a single sub - module, and NCU dc 2 is the total energy reference value of N sub - modules;

[0020] The internal circulating current prediction and objective function calculation module includes:

[0021] The internal circulating current prediction module, according to the upper and lower arm currents I in this sampling period pαβ and I nαβ , calculate the internal circulating current I cαβ , and then according to the DC bus voltage U dc , respectively calculate the internal circulating current I in the next sampling period when the upper and lower arms adopt the N + 1 sub - module input method cαβ(next) (m), m = 1, 2,... N + 1;

[0022] The internal circulating current objective function calculation module, according to the predicted internal circulating current I cαβ(next) (m) and the internal circulating current reference value, calculate the internal circulating current objective function J 3m , m = 1, 2,... N + 1; the internal circulating current reference value is given as 0;

[0023] The final objective function calculation module calculates the final objective function J according to the following method m : J m = p1J 1m + p2J 2m + p3J 3m , m = 1, 2,... N + 1, where p1, p2, p3 are weight factors;

[0024] The objective function comparison and control instruction output module compares the final objective function J when the N + 1 sub - module input method is adopted m, where \(m = 1, 2, \cdots, N + 1\), select the sub-module input method with the minimum objective function as the control instruction for this sampling period to achieve the control of the MMC converter.

[0025] Furthermore: In the output power prediction module, calculate the traditional active power \(P\) g(next) (m), traditional reactive power \(Q\) g(next) (m), extended active power \(P\) gnew(next) (m), and extended reactive power \(Q\) gnew(next) (m) according to the following method:

[0026]

[0027] where \(U'\) sα(next) and \(U'\) sβ(next) are the \(\alpha\)-axis and \(\beta\)-axis components of the voltage vector \(U\) sαβ lagged by 90 degrees.

[0028] Furthermore: In the power selection and objective function calculation module, select the power feedback values \(P\) gsel (m) and \(Q\) gsel (m) according to the following method:

[0029] When the control objective is set to suppress the active power pulsation, the power feedback values are designed as follows:

[0030]

[0031] When the control objective is set to suppress the reactive power pulsation, the power feedback values are designed as follows:

[0032]

[0033] The calculation method of the power objective function \(J\) 1m is as follows:

[0034] \(J\) 1m = |\(P\) gref - \(P\) gsel (m)| + |\(Q\) gref - \(Q\) gsel (m)|

[0035] where \(P\) gref and \(Q\) gref are the reference values of active power and reactive power respectively, and \(J\) 1m is the power objective function using the \(m\)-th sub-module input method.

[0036] Furthermore: In the per-unit system, the weight factors \(p1\), \(p2\), \(p3\) are generally given as 1.

[0037] The present invention uses extended power calculation to obtain the power reference value of the MMC, and realizes the tracking of the power reference value through the model predictive control algorithm. It can suppress the active power pulsation or reactive power pulsation of the MMC under an unbalanced power grid without introducing additional current harmonic components. This method can be realized in the stationary coordinate system without the need for positive and negative sequence separation of voltage and current, and the control system structure is simple. Description of the Drawings

[0038] Figure 1 It is a specific example structure diagram of the MMC converter station.

[0039] Figure 2 It is a schematic diagram of the system principle of a specific example of the present invention, and the names of each module are as follows:

[0040] 1 - Voltage sensor, 2 - Current sensor, 3 - Clark transformation module, 4 - Voltage prediction module, 5 - Output current prediction module, 6 - Output power prediction module, 7 - Power selection and objective function calculation module, 8 - Arm energy prediction module, 9 - Arm energy objective function calculation module, 10 - Internal circulating current prediction module, 11 - Internal circulating current objective function calculation module, 12 - Final objective function calculation module, 13 - Objective function comparison and control command output module.

[0041] Figure 3 It is a schematic diagram of the method for putting N + 1 seed modules in the present invention. Detailed Embodiment

[0042] In order to describe the present invention more specifically, the technical solutions of the present invention will be described in detail below with reference to the drawings and specific embodiments.

[0043] The system implementation of the MMC model predictive power control method based on extended power of the present invention is as Figure 2 shown. The system includes a voltage sensor 1, a current sensor 2, a Clark transformation module 3, a voltage prediction module 4, an output current prediction module 5, an output power prediction module 6, a power selection and objective function calculation module 7, an arm energy prediction module 8, an arm energy objective function calculation module 9, an internal circulating current prediction module 10, an internal circulating current objective function calculation module 11, a final objective function calculation module 12, and an objective function comparison and control command output module 13.

[0044] As Figure 1 、 2 、3 shows, the MMC model predictive power control method based on extended power of the present invention includes the following steps:

[0045] Using the voltage sensor 1 to measure the three-phase voltage U on the AC side of the MMC sabcCollect; use the current sensor 2 to collect the three-phase current I on the AC side of the MMC vabc , the upper and lower arm currents I of the MMC pabc and I nabc for collection, and use the voltage sensor 1 to collect the DC bus voltage U dc ; use the voltage sensor 1 to collect the voltages U of the N sub-modules on the upper and lower three-phase arms pabc (i) and U nabc (i) (i = 1 to N) for collection, where N is the number of sub-modules, and both the upper and lower arms have N sub-modules.

[0046] Adopt the Clark transformation module 3 to perform Clark transformation on the three-phase voltage U sabc , three-phase current I vabc , upper and lower arm currents I pabc and I nabc , upper and lower arm sub-module voltages U pabc (i) and U nabc (i), i = 1 to N, to obtain the corresponding voltage vectors U sαβ , current vectors I vαβ , upper and lower arm currents I pαβ and I nαβ , upper and lower arm sub-module voltages U pαβ (i) and U nαβ (i), i = 1 to N;

[0047] Use the voltage prediction module 4 to calculate the voltage vector U sαβ of the next sampling period according to the grid voltage U sαβ(next) obtained in the current sampling period. The specific calculation method is as follows:

[0048] U sa(next) = U sα cos(ω s T s ) - U sβ sin(ω s T s )

[0049] U sβ(next) = U sβ cos(ω s T s ) + U sα sin(ω s T s )

[0050] where ω s is the grid voltage angular frequency, and T s is the control period.

[0051] Using the output current prediction module 5, according to the voltage U obtained in this sampling period sαβ , the current I vαβ , calculate the valve-side current vector I in the next sampling period when different sub-module input methods are adopted for the upper and lower bridge arms in this sampling period vαβ(next) , where there are N + 1 sub-module input methods in this sampling period, and I needs to be calculated N + 1 times vαβ(next) , to obtain I vαβ(next) (m), (m = 1, 2,... N + 1), and then use the output current objective function calculation module 8 to calculate the output current objective function J vαβref according to the current reference value I 1m , (m = 1, 2,... N + 1);

[0052] The above-mentioned N + 1 sub-module input methods are as shown in the appendix Figure 3 .

[0053] The current value I in the next sampling period vαβ(next) is calculated according to the following method:

[0054]

[0055] where L is the equivalent inductance including the commutation transformer and the arm reactor, T s is the sampling period, U sα and U sβ are the α-axis and β-axis components of the voltage vector U sαβ respectively, I vα and I vβ are the α-axis and β-axis components of the current vector I vαβ respectively, I vα(next) (m) and I vβ(next) (m) are the α-axis and β-axis components of the current vector I vαβ(next) (m) respectively; U nα (m), U nβ (m) are the α-axis and β-axis components of the lower bridge arm voltage when the m-th sub-module input method is adopted, and the calculation method is as follows.

[0056]

[0057] where l na , l nb and l nc correspond to the number of sub-modules input in the lower bridge arm of the corresponding phase.

[0058] U pα (m), U pβ (m) are the α-axis and β-axis components of the upper bridge arm voltage when the m-th sub-module input method is adopted, and their calculation methods are the same as the corresponding components of the lower bridge arm.

[0059] Using the output power prediction module 6, calculate the traditional active power P g(next) (m), the traditional reactive power Q g(next) (m), and the extended active power P gnew(next) (m), the extended reactive power Q gnew(next) (m) according to the following method:

[0060]

[0061] where U' sα(next) and U' sβ(next) are the α-axis and β-axis components of the voltage vector U sαβ lagged by 90 degrees.

[0062] Using the power selection and objective function calculation module 7, select the power feedback values P gsel (m) and Q gsel (m) according to the following method:

[0063] When the control objective is set to suppress the active power pulsation, the power feedback value is designed as follows:

[0064]

[0065] When the control objective is set to suppress the reactive power pulsation, the power feedback value is designed as follows:

[0066]

[0067] The power objective function J 1m is calculated according to the following method:

[0068] J 1m =|P gref -P gsel (m)|+|Q gref -Q gsel (m)|

[0069] where P gref and Q gref are the reference values of the active power and reactive power respectively, and J 1m is the power objective function using the m-th sub-module input method.

[0070] Using the arm energy prediction module 8, according to the voltages U pαβ (i) and U nαβ (i) (i = 1~N) of different sub-modules on the upper and lower arms of this sampling period, the upper and lower arm currents I pαβ and I nαβ, calculate the voltage values U of different sub-modules in the next sampling period when different sub-module input methods are adopted for the upper and lower arms in this sampling period αmp(next) (i), U αmn(next) (i), U βmp(next) (i) and U βmn(next) (i) (i = 1 ~ N, m = 1, 2,... N + 1), and further calculate the total energies E of the upper and lower arm sub-modules in the next sampling period p(next) (m) and E n(next) (m) (m = 1, 2,... N + 1), then use the arm energy objective function to calculate module 9, and calculate the arm energy objective function J according to the upper and lower arm energy reference values E pref(next) and E nref(next) Calculate the arm energy objective function J 2m , (m = 1, 2,... N + 1).

[0071] The calculation method of the α-axis component U of the voltage value of the upper arm sub-module in the next sampling period is as follows: αmp(next) (i) is as follows:

[0072]

[0073] Where: C is the sub-module capacitor voltage.

[0074] The voltages U of other sub-modules βmp(next) (i), U αmn(next) (i) and U βmn(next) (i) are calculated in the same way as U αmp(next) (i), just replace the corresponding variables.

[0075] The calculation of the energy sum of the upper and lower arm sub-modules in the next sampling period is as follows:

[0076]

[0077] The calculation of the arm energy objective function is as follows:

[0078]

[0079] Where J 2m is the arm energy objective function for adopting the m-th sub-module input method.

[0080] Use the internal circulating current prediction module 10 to calculate the internal circulating current I according to the upper and lower arm currents I pαβ and I nαβ in this sampling period, and then calculate the internal circulating current I in the next sampling period when different sub-module input methods are adopted for the upper and lower arms in this sampling period according to the DC bus voltage U cαβ and U dc respectively cαβ(next)(m) (m = 1, 2, ... N + 1), and then use the internal circulating current objective function calculation module 11 to calculate the internal circulating current objective function J according to the internal circulating current reference value. 3m , (m = 1, 2, ... N + 1).

[0081] Calculate the internal circulating current I in the next sampling period by the following method. cαβ(next) (m):

[0082]

[0083] Where: L0 is the inductance of the arm reactor, I cα and I cβ are the α-axis and β-axis components of the current vector I cαβ respectively, and I cα(next) (m) and I cβ(next) (m) are the α-axis and β-axis components of the current vector I cαβ(next) (m) respectively.

[0084] The internal circulating current objective function is calculated as follows:

[0085] J 3m = |I cαref - I cα(next) (m)| + |I cβref - I cβ(next) (m)|

[0086] Where, J 3m is the internal circulating current objective function using the m-th sub-module input method, and I cαref (m) and I cβref (m) are the α-axis and β-axis components of the current vector I cαβref (m) respectively, and are generally given as 0.

[0087] Use the final objective function calculation module 12 to calculate the final objective function J according to the following method. m : J m = p1J 1m + p2J 2m + p3J 3m , m = 1, 2,... N + 1. In the per-unit system, the weighting factors p1, p2, p3 are generally given as 1.

[0088] Use the objective function comparison and control instruction output module 13 to compare the objective functions J m (m = 1, 2,... N + 1) of the N + 1 sub-module input methods, and select the sub-module input method with the minimum objective function as the control instruction for this sampling period to achieve the control of the MMC converter.

[0089] In summary, by adopting the MMC model predictive power control method based on extended power in the present invention, power pulsations can be effectively suppressed when a fault occurs in the AC power grid, thereby ensuring the safe and stable operation of the flexible DC transmission system; the power pulsation suppression method in the present invention has a simple control structure and strong engineering practical value.

[0090] The above description of the embodiments is to enable those of ordinary skill in the art to understand and apply the present invention. It is obvious that those skilled in the art can easily make various modifications to the above embodiments and apply the general principles described herein to other embodiments without creative efforts. Therefore, the present invention is not limited to the above embodiments, and all improvements and modifications made by those skilled in the art based on the disclosure of the present invention should be within the protection scope of the present invention.

Claims

1. A model predictive power control system for MMC based on extended power, characterized in that, It includes a sampling module, a Clark transformation module, a voltage prediction module, an output current prediction module, an output power prediction module, a power selection and objective function calculation module, a leg energy prediction and objective function calculation module, an internal circulating current prediction and objective function calculation module, a final objective function calculation module, and an objective function comparison and control instruction output module; In the sampling module, it includes: Voltage sampling module, sampling the three-phase voltages U of the AC side of the MMC converter sabc , the DC bus voltage U dc , the voltages of the upper and lower bridge arm sub-modules U pabc (i) and U nabc (i), where i = 1 to N, and N is the number of sub-modules, and both the upper and lower bridge arms have N sub-modules; Current sampling module, sampling the three-phase current I on the AC side of the MMC converter vabc , the upper and lower arm currents I pabc and I nabc ; The Clark transformation module performs Clark transformation on the three-phase voltage U sabc of the AC side of the MMC, the three-phase current I vabc , the upper and lower arm currents I pabc and I nabc , the upper and lower arm sub-module voltages U pabc (i) and U nabc (i), and obtains the corresponding voltage vector U sαβ , current vector I vαβ , upper and lower arm currents I pαβ and I nαβ , upper and lower arm sub-module voltages U pαβ (i) and U nαβ (i) in the α-β coordinate system; The voltage prediction module calculates the voltage vector U of the next sampling period based on the grid voltage U obtained in the current sampling period sαβ ; sαβ(next) ; The output current prediction module calculates, according to the grid voltage U sαβ and current I vαβ obtained in this sampling period, the current vectors I vαβ(next) in the next sampling period when different sub-module input methods are adopted for the upper and lower bridge arms in this sampling period; where there are N + 1 sub-module input methods in this sampling period, and I vαβ(next) needs to be calculated N + 1 times in total to obtain I vαβ(next) (m), where m = 1, 2,... N + 1; The output power prediction module calculates the traditional active power P sαβ(next) and the current vector I vαβ(next) (m) for the next sampling period to obtain the traditional active power P g(next) (m), the traditional reactive power Q g(next) (m), the extended active power P gnew(next) (m), and the extended reactive power Q gnew(next) (m), where m = 1, 2,... N + 1; In the power selection and objective function calculation module, it includes: The control target selection module selects the corresponding traditional power and extended power as the power feedback values P gsel (m) and Q gsel (m); Output power objective function calculation module, calculates the output current objective function J according to the selected power feedback values P gsel (m) and Q gsel (m) and the current reference value I vαβref ; 1m ; The leg energy prediction and objective function calculation module includes: The arm energy prediction module calculates, according to the voltages U pαβ (i) and U nαβ (i) of different sub-modules of the upper and lower arms in the current sampling period, and the currents I pαβ and I nαβ of the upper and lower arms, the voltage values U αmp(next) (i), U αmn(next) (i), U βmp(next) (i) and U βmn(next) (i) of different sub-modules in the next sampling period when the N + 1 seed module input method is adopted for the upper and lower arms in the current sampling period, and further calculates the total energies E p(next) (m) and E n(next) (m) of the upper and lower arm sub-modules in the next sampling period; The arm energy objective function calculation module calculates the arm energy objective function J based on the total energies E p(next) (m) and E n(next) (m) of the upper and lower arm sub-modules obtained and the upper and lower arm energy reference values, where m = 1, 2,... N + 1, and the upper and lower arm energy reference values are given as NCU 2m dc 2 ;​ The internal circulating current prediction and objective function calculation module includes: Internal circulating current prediction module, based on the upper and lower arm currents I pαβ and I nαβ in this sampling period, calculates the internal circulating current I cαβ . Then, based on the DC bus voltage U dc , respectively calculates the internal circulating current I cαβ(next) (m) in the next sampling period when the N + 1 seed modules are put into use for the upper and lower arms in this sampling period, where m = 1, 2,... N + 1; Internal circulation objective function calculation module, calculates the internal circulation objective function J according to the predicted internal circulation I cαβ(next) (m) and the internal circulation reference value 3m , m = 1, 2, ... N + 1; the internal circulation reference value is given as 0; The final objective function calculation module calculates the final objective function J according to the following method m : J m = p1J 1m + p2J 2m + p3J 3m , m = 1, 2, ... N+1, p1, p2, p3 are weight factors; The target function comparison and control instruction output module compares the final target function J of the N+1 sub-module input method. m where m = 1, 2,... N+1, and selects the sub-module input method with the minimum target function as the control instruction for this sampling period to achieve the control of the MMC converter.

2. The model predictive power control system for MMC based on extended power according to claim 1, characterized in that: In the output power prediction module, the traditional active power P g(next) (m), the traditional reactive power Q g(next) (m), the extended active power P gnew(next) (m), and the extended reactive power Q gnew(next) (m) are calculated according to the following method: Among them, U' sα(next) and U' sβ(next) are the α-axis and β-axis components of the voltage vector U sαβ after being lagged by 90 degrees.

3. The model predictive power control system for MMC based on extended power according to claim 1, characterized in that: In the power selection and objective function calculation module, the power feedback values P gsel (m) and Q gsel (m) are selected according to the following method: When the control objective is set to suppress the active power pulsation, the power feedback value is designed as follows: When the control objective is set to suppress the reactive power pulsation, the power feedback value is designed as follows: Power objective function J 1m is calculated as follows: J 1m = |P gref - P gsel (m)| + |Q gref - Q gsel (m)| Among them, P gref and Q gref are the reference values of active power and reactive power respectively, and J 1m is the power objective function using the m-th sub-module input method.

4. The model predictive power control system for MMC based on extended power according to claim 1, characterized in that: In the per-unit system, the weighting factors p1, p2, p3 are given as 1.

Images