Multi-stage model predictive control method for modular multilevel converter

Abstract

The application provides a multi-stage model predictive control method suitable for a modular multilevel converter, and the method comprises: a first-stage alternating current control link, which comprises: obtaining a reference control option for output current tracking by using an alternating current index function, so as to realize alternating current control; a second-stage circulating current control link, which comprises: calculating the optimal number of submodules that should be put into the upper and lower bridge arms of the modular multilevel converter (MMC) in the first-stage control by introducing two circulating current factors; and a third-stage submodule capacitor voltage control link: in each control period, the capacitor voltages of each bridge arm submodule are grouped and sorted. The application improves the performance of circulating current suppression and reduces the calculation amount of the controller.

Description

Technical Field

[0001] This invention belongs to the field of power transmission technology, and specifically relates to a multi-stage model predictive control method applicable to modular multilevel converters. Background Technology

[0002] Modularity and scalability are key characteristics of Modular Multilevel Converters (MMCs). These features enable MMCs to be widely used in high-voltage direct current (HVDC) systems, power electronic transformers, and other applications. Regardless of the application, MMCs require a control method to achieve multiple control objectives, such as output current and submodule capacitor voltage balancing, minimizing circulating current, and reducing ripple in submodule capacitor voltage and DC-side current. The basic control objectives of MMCs consist of three parts: output current control, circulating current control, and capacitor voltage control. One scholar proposed an open-loop classical control strategy, in which the applied modulation index is calculated based on the required input and output voltages. Subsequently, stationary automatic control methods and closed-loop classical control methods in the synchronous equation coordinate system were proposed.

[0003] In recent years, Model Predictive Control (MPC) has gradually emerged in the field of power electronics control. MPC is a nonlinear optimization control method with advantages such as fast dynamic response, simple control principle, ease of implementation, and ability to perform multi-objective optimization. Some scholars have proposed applying MPC to Multi-Objective Control (MMC).

[0004] The performance of classical control methods depends on the design and adjustment of the PI controller, the type of modulation scheme, and the switching frequency. Furthermore, the control process of this method has a long settling time, and control delay can severely impact system performance. In addition, the system is highly sensitive to controller parameters, and the parameter tuning process is complex; the quality of the parameter design directly affects system performance.

[0005] To address the control objective of MPC for MMC, a weighted objective function is constructed. However, this requires iterative prediction and optimization over all switching states, leading to significant computational complexity and the need to adjust the weighting factors. Some researchers have proposed a Finite Control Set-Model Predictive Control (FCS-MPC) method, which effectively reduces the controller's computational load by selecting the optimal control scheme that minimizes the cost function in each control cycle, but this introduces substantial circulating harmonic components. Other researchers have proposed a simplified computational method that ensures control performance by selecting the optimal number of inputs rather than switching states, but this does not consider circulating current suppression strategies.

[0006] In summary, none of the above methods consider circulating current suppression strategies, resulting in low output power quality. Furthermore, the controller's weighting factor design is complex, requiring further optimization.

[0007] Therefore, it is necessary to design a multi-stage model predictive control method suitable for modular multilevel converters to solve the above-mentioned technical problems. Summary of the Invention

[0008] To address the aforementioned technical problems, this invention provides a multi-stage model predictive control method suitable for modular multilevel converters, the method comprising:

[0009] The first stage of AC current control includes: obtaining reference control options for output current tracking using AC current index functions to achieve AC current control;

[0010] The second stage of the circulating current control includes: calculating the optimal number of sub-modules that should be put into the upper and lower arms of the modular multilevel converter (MMC) in the first stage of control by introducing two circulating current factors.

[0011] The third stage is the submodule capacitor voltage control process: In each control cycle, the capacitor voltage of each bridge arm submodule is grouped and sorted.

[0012] Furthermore, the alternating current control circuit specifically includes:

[0013] By using the relationship between the arm voltage and the DC side voltage, the nearest-level modulation method (NLM) is used to obtain the reference control options for output current tracking. The number of submodules turned on within a small range is then substituted into the current tracking error index function for optimization, and the number of modules that minimizes the index function value is selected.

[0014] Furthermore, the relationship between the arm voltage and the DC side voltage is determined by the following expression:

[0015]

[0016] In the formula d pj and d nj The modulation indices of the upper and lower arms are d and d, respectively. pj and d nj The sum equals 1, u pj with u nj These are the output voltages of the upper and lower bridge arms, respectively.

[0017] Furthermore, the alternating current index function J1 is determined by the following expression:

[0018] J1=|i * oj (k+1)-ioj (k+1)

[0019] In the formula i * oj (k+1) is the reference value for AC current, i oj (k+1) is the predicted value of AC current.

[0020] Furthermore, the predicted value of the alternating current is determined by the following expression:

[0021]

[0022] In the formula, R0 is the load resistance; L f L0 is the bridge arm inductance; u is the bridge arm inductance. pj with u nj These are the output voltages of the upper and lower bridge arms, respectively; i oj (k) is the detected value of the alternating current, i oj (k+1) is the predicted value of the alternating current; T s Sampling time.

[0023] Furthermore, the circulation control process specifically includes:

[0024] By introducing two circulation factors, Δi is set. cirj1 and Δi cirj2 The control arm and overall capacitor voltage, and the reference value of the circulating current are ultimately expressed as i. dc / 3+(Δi cirj1 +Δi cirj2 );

[0025] Circulation prediction expression:

[0026]

[0027] The bridge arm voltage difference and voltage reference values ​​are:

[0028]

[0029]

[0030] In the formula L eq =L f +2L, i cirj (k+1) represents the reference value of the circulation at time k+1, where the predicted circulation value i cirj (k+1) is determined by the following expression:

[0031]

[0032] Furthermore, the expressions for the output voltage and circulating current factor are:

[0033]

[0034]

[0035] In the formula, M = T s / C, u*Δj is the reference value for the voltage difference between the bridge arms.

[0036] Furthermore, the submodule capacitor voltage control stage specifically includes:

[0037] When grouping and sorting, when Δn p,nj When >0, select the Δn corresponding to the minimum value of J2 from the j-phase upper / lower bridge arm resection group. p,nj If a submodule is deployed, then Δn is selected from the deployment group when J2 takes the minimum value. p,nj Each submodule was removed;

[0038] The number of submodules that need to be deployed in the upper / lower bridge arm at the next moment is determined by the following expression:

[0039] Δn p,nj =n p,nj (k+1)-n p,nj (k)

[0040] In the formula, n p,nj (k+1) and n p,nj (k) represents the number of sub-modules deployed in the upper / lower arm at times k+1 and k, respectively.

[0041] Furthermore, the submodule capacitor voltage index function:

[0042]

[0043] On the other hand, the present invention also provides a multi-stage model predictive control system suitable for modular multilevel converters, the system comprising:

[0044] The control module is used in the first stage of AC current control to obtain the reference control option for output current tracking using the AC current index function, thereby realizing AC current control.

[0045] The calculation module is used in the second-stage circulating current control loop to calculate the optimal number of sub-modules that should be deployed in the upper and lower arms of the modular multilevel converter (MMC) in the first-stage control by introducing two circulating current factors.

[0046] The grouping and sorting module is used in the third stage of submodule capacitor voltage control: in each control cycle, the capacitor voltage of each bridge arm submodule is grouped and sorted.

[0047] This invention provides a multi-stage model predictive control method suitable for modular multilevel converters, which avoids weight factor design, improves the performance of circulating current suppression, and reduces the computational load of the controller.

[0048] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures pointed out in the description and the drawings. Attached Figure Description

[0049] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0050] Figure 1 A topology diagram of a three-phase modular multilevel converter according to an embodiment of the present invention is shown.

[0051] Figure 2 The simulation waveform of a conventional model prediction method according to an embodiment of the present invention is shown.

[0052] Figure 3 Simulation waveforms of a multi-stage model prediction method according to an embodiment of the present invention are shown.

[0053] Figure 4 A block diagram of multi-stage model predictive control according to an embodiment of the present invention is shown. Detailed Implementation

[0054] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0055] Existing model predictive control (MMC) schemes mainly focus on reducing computational load and lowering switching frequency, but they still have shortcomings in circulating current suppression and can lead to problems such as complex weight factor design and large harmonic components in the bridge arm circulating current. This invention addresses the current state of research on MMC-based model predictive control methods by proposing a multi-stage model predictive control method suitable for modular multilevel converters. This method avoids weight factor design, improves circulating current suppression performance, and reduces the computational load of the controller. The following is a description of this method.

[0056] The circuit topology of the three-phase modular multilevel converter (MMC) is as follows: Figure 1 As shown. Each phase has two bridge arms, upper and lower, and each bridge arm consists of N cascaded half-bridge sub-modules and a bridge arm inductor L. f Each SM (submodule) consists of a half-bridge structure and a submodule capacitor C. dc and i dc These represent the DC-side voltage and current of the MMC, respectively. pj and u nj (j = a, b, c) represent the voltages of the upper and lower arms, respectively. pj and i nj (j=a,b,c) represent the currents in the upper and lower arms, respectively.

[0057] The multi-stage model predictive control method applicable to modular multilevel converters in this invention will be described in detail below, wherein the method includes:

[0058] The first stage of AC current control includes: obtaining reference control options for MMC output current tracking using AC current index functions to achieve AC current control;

[0059] The second stage of the circulating current control includes: calculating the optimal number of upper and lower arm sub-modules of the modular multilevel converter (MMC) to be deployed in the first stage of control by introducing two circulating current factors.

[0060] The third stage is the submodule capacitor voltage control process: In each control cycle (each cycle represents the first and second stage control cycles), the capacitor voltage of each bridge arm submodule is grouped and sorted.

[0061] In this invention, circulation: by Figure 1 (in, Figure 1 SM in y (1~N), the value range of y is (1~N), representing the analysis of the first submodule to the Nth submodule. When the current is transmitted to the load through the floating capacitor, it causes voltage fluctuations and interphase circulating current is generated between the three-phase bridge arms. If the interphase circulating current is not controlled, it will increase the bridge arm loss and cause the bridge arm current to be distorted.

[0062] The following is a detailed description of the multi-stage model predictive control method for modular multilevel converters provided by the present invention.

[0063] In some embodiments of the present invention, the alternating current control stage specifically includes:

[0064] The reference control options for MMC output current tracking are obtained using the AC current index function. The number of submodules turned on in a small range (-1, 1) is substituted into the current tracking error index function for optimization (i.e., the number of turned on is substituted into the index function for optimization at +0, +1 and -1 respectively). The number of modules that minimizes the index function value is selected.

[0065] The relationship between the arm voltage and the DC side voltage is determined by the following expression:

[0066]

[0067] In the formula d pj and d nj These are the modulation indices for the upper and lower arms, respectively. To maintain the stability of the DC-side voltage, d pj and d nj The sum equals 1.

[0068] In some embodiments of the present invention, the alternating current index function J1 is determined by the following expression:

[0069] J1=|i * oj (k+1)-i oj (k+1)

[0070] In the formula i * oj (k+1) is the reference value for AC current, i oj (k+1) is the predicted value of AC current. 。

[0071] The predicted value of the alternating current is determined by the following expression:

[0072]

[0073] In the formula, R0 is the load resistance; L f L0 is the bridge arm inductance; u is the bridge arm inductance. pj with u nj These are the output voltages of the upper and lower bridge arms, respectively; i oj (k) is the detected value of the alternating current, i oj (k+1) is the predicted value of the alternating current; T s Sampling time.

[0074] In some embodiments of the present invention, the circulating flow control element specifically includes:

[0075] By introducing two circulation factors, Δi is set. cirj1 and Δi cirj2 To control the voltage of the bridge arms and the overall capacitor ( Figure 1 (Neutron module capacitor voltage), so the reference value of the circulating current is ultimately expressed as i. dc / 3+(Δi cirj1 +Δi cirj2 ).

[0076] Circulation prediction expression:

[0077]

[0078] The bridge arm voltage difference and voltage reference values ​​are:

[0079]

[0080]

[0081] In the formula i oj (k) represents the detected value of the AC side current.

[0082] Ideal output voltage V on the AC side ol,j and circulation factor Δi cirj1 and Δi cirj2 The expression is:

[0083]

[0084]

[0085] In the formula, M = T s / C, N is the number of submodules for each bridge arm; u Σj Let i be the sum of the voltages of the upper and lower bridge arms. * oj (k) is the reference value for the AC side current, u * Σj (k) is the reference value for the bridge arm voltage and the sum of its values.

[0086] In some embodiments of the present invention, the submodule capacitor voltage control step specifically includes:

[0087] When grouping and sorting, the number of submodules Δn that need to be deployed in the upper / lower bridge arm. p,nj When (number of submodules to be deployed in the upper / lower arm) > 0, select the minimum value of J2 from the j-phase upper / lower arm cutoff group, corresponding to Δn. p,nj If a submodule is deployed, then Δn is selected from the deployment group when J2 takes the minimum value. p,njThe method involves the removal of individual sub-modules; the sorting computation is significantly reduced compared to traditional full sorting and heap sort, which can effectively alleviate the computational pressure of optimization.

[0088] The number of submodules that need to be engaged (or disengaged) in the upper / lower bridge arm at the next moment is determined by the following expression:

[0089] Δn p,nj =n p,nj (k+1)-n p,nj (k)

[0090] In the formula, n p,nj (k+1) and n p,nj (k) represents the number of sub-modules deployed in the upper / lower arm at times k+1 and k, respectively.

[0091] The submodule capacitor voltage index function is used for optimization. During grouping and sorting, it determines the number of submodules Δn that need to be deployed in the upper / lower bridge arm. p,nj When >0, select the Δn corresponding to the minimum value of J2 from the j-phase upper / lower bridge arm resection group. p,nj If a submodule is deployed, then Δn is selected from the deployment group when J2 takes the minimum value. p,nj The specific function for determining the capacitor voltage index of a submodule after it is removed is as follows:

[0092]

[0093] In the formula, u cxjy (k+1) is the predicted value of the submodule capacitor voltage at time k+1.

[0094] In this invention, such as Figure 4 shown ( Figure 4 In the middle, S p (k) and S n (k) represents the number of sub-modules that should be put into operation in the upper and lower arms respectively. The specific implementation process of the multi-stage model predictive control method applicable to modular multilevel converters is as follows:

[0095] First, in the initial stage, the reference control options for output current tracking are obtained using the nearest-level modulation (NLM) method based on the relationship between the arm voltage and the DC-side voltage. Then, the number of submodules to be turned on within a small range is substituted into the current tracking error index function for optimization. The number of modules n that minimizes the index function value is selected. ﹡ pj n ﹡ nj Secondly, in the second stage, when capacitor voltage control is not considered, the reference value of the circulating current is set to i. dc / 3. Here, the circulation base is Δi cirj1 and Δi cirj2and set them to i respectively dc / 3+Δi cirj1 and i dc / 3+Δi cirj2 The circulating current reference value is controlled by a circulating current limiter, and the optimal sum of the reference values ​​u for the bridge arms is obtained through a voltage prediction model. ﹡ ∑j The total number of input SMs of the bridge arm is obtained by rounding using the nearest level approximation (NLC). ∑j To avoid affecting the control of AC side current measurement during circulating current control, the number of SMs in the input or bypass of the upper and lower bridge arms should be equal when circulating current control is performed. The final number of SMs put into use in the upper bridge arm is N. pj * =n ﹡ pj +(S∑j-(n ﹡ pj +n ﹡ nj )) / 2, the number of SMs (N) of the lower bridge arm finally put into use nj * Similarly, we can obtain the result.

[0096] Finally, in the third stage, during each control cycle, the capacitor voltage of each bridge arm submodule is grouped and sorted (input group and output group), when Δn p,nj When >0, select the Δn corresponding to the minimum value of J2 from the j-phase upper / lower bridge arm resection group. p,nj If a submodule is deployed, then Δn is selected from the deployment group when J2 takes the minimum value. p,nj Each submodule is removed. This method significantly reduces the computational cost of sorting compared to traditional full sorting and heap sort, effectively alleviating the computational pressure of optimization.

[0097] In this invention, a [system / facility] can be constructed in MATLAB / Simulink. Figure 1 The three-phase MMC system shown illustrates the advantages of the present invention. Given a system active power P* = 0.25MW and reactive power Q* = 0MW, at t = 0.5s, the system active power abruptly changes to P* = 0.125MW. The simulation waveform of a traditional MPC (Model Predictive Control) system is as follows... Figure 2 As shown, the simulation waveforms of the proposed multi-stage model predictive control method for modular multilevel converters are as follows. Figure 3 As shown. Detailed main circuit system parameters are listed in Table 1.

[0098] Table 1

[0099]

[0100] Figure 2 and Figure 3 The waveforms from top to bottom represent: the output current and setpoint of the three-phase MMC AC side, the three-phase output current on the grid side, the three-phase circulating current, and the capacitor voltages of all submodules in the upper and lower arms of phase a. (Comparison) Figure 2 (a) and Figure 3 (a) It can be seen that the AC side output current tracking accuracy is significantly improved, the harmonic component (THD) of the output current can be reduced to 1.11%, and it can respond to system step changes in a timely manner with a short adjustment time and good dynamic performance; the present invention uses two circulating current factors Δi cirj1 and Δi cirj2 Set them to 2A and 5A respectively, such as Figure 3 (c) and Figure 2 (c) In comparison, the three-phase circulating current has a more obvious control effect and solves the problem of energy imbalance between bridge arms. Figure 3 (d) and Figure 2 (d) represents the capacitor voltage of the submodule. Obviously, the traditional model predictive control method cannot maintain the capacitor voltage of the submodule at a stable value. In contrast, the method proposed in this invention achieves good balance control of the capacitor voltage of the submodule.

[0101] On the other hand, the present invention also provides a multi-stage model predictive control system suitable for modular multilevel converters, the system comprising:

[0102] The control module is used in the first stage of AC current control to obtain the reference control option for output current tracking using the AC current index function, thereby realizing AC current control.

[0103] The calculation module is used in the second-stage circulating current control loop to calculate the optimal number of sub-modules SMs of the upper and lower arms of the modular multilevel converter (MMC) in the first-stage control by introducing two circulating current factors.

[0104] The grouping and sorting module is used in the third stage of submodule capacitor voltage control: in each control cycle, the capacitor voltage of each bridge arm submodule is grouped and sorted.

[0105] The functions and methods implemented by other modules of the multi-stage model predictive control system for modular multilevel converters in this invention correspond to the functions and methods implemented by other steps in the multi-stage model predictive control method for modular multilevel converters in this invention, and therefore will not be repeated here.

[0106] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the technical solution of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of the technical solution of the present invention.

Claims

1. A multi-stage model predictive control method applicable to modular multilevel converters, characterized in that, The method includes: The first stage of AC current control includes: obtaining reference control options for output current tracking using AC current index functions to achieve AC current control; The second stage of the circulating current control includes: calculating the optimal number of sub-modules that should be put into the upper and lower arms of the modular multilevel converter (MMC) in the first stage of control by introducing two circulating current factors. The third stage is the submodule capacitor voltage control process: In each control cycle, the capacitor voltage of each bridge arm submodule is grouped and sorted. The alternating current control circuit specifically includes: By using the relationship between the arm voltage and the DC side voltage, the nearest-level modulation method (NLM) is used to obtain the reference control options for output current tracking. The number of sub-modules turned on within a small range is substituted into the current tracking error index function for optimization, and the number of modules that minimizes the index function value is selected. The circulation control component specifically includes: By introducing two circulation factors, it was set to... and The reference value for the circulating current is ultimately expressed as the control arm and the overall capacitor voltage. ; Circulation prediction expression: The bridge arm voltage difference and voltage reference values ​​are: In the formula for Reference values ​​for the circulation at any given time, including predicted circulation values. Determined by the following expression: ； The expressions for the output voltage and the circulating current factor are as follows: In the formula , This is a reference value for the voltage difference between the bridge arms; The submodule capacitor voltage control section specifically includes: When grouping and sorting, At that time, from Selected from the upper / lower bridging arm resection group When taking the minimum value, the corresponding Individual submodules are deployed; conversely, select from the deployment group. When taking the minimum value Each submodule was removed; The number of submodules that need to be deployed in the upper / lower bridge arm at the next moment is determined by the following expression: In the formula, and They are respectively and The number of sub-modules deployed at any given time on the upper / lower bridge arm.

2. The multi-stage model predictive control method for modular multilevel converters according to claim 1, characterized in that, The relationship between the arm voltage and the DC side voltage is determined by the following expression: In the formula and These are the modulation indices of the upper and lower arms, respectively. and The sum equals 1. and These are the output voltages of the upper and lower bridge arms, respectively.

3. The multi-stage model predictive control method for modular multilevel converters according to claim 2, characterized in that, Alternating current index function Determined by the following expression: In the formula This is the reference value for alternating current. This is the predicted value of the alternating current.

4. The multi-stage model predictive control method for modular multilevel converters according to claim 3, characterized in that, The predicted value of the alternating current is determined by the following expression: In the formula For load resistance; For bridge arm inductance; For load inductance, and These are the output voltages of the upper and lower bridge arms, respectively. The measured value is the alternating current. This is the predicted value of the alternating current; Sampling time.

5. The multi-stage model predictive control method for modular multilevel converters according to claim 1, characterized in that, Submodule capacitor voltage index function: 。 6. A multi-stage model predictive control system applicable to modular multilevel converters, said system being used to implement the predictive control method of any one of claims 1-5, characterized in that, The system includes: The control module is used in the first stage of AC current control to obtain the reference control option for output current tracking using the AC current index function, thereby realizing AC current control. The calculation module is used in the second-stage circulating current control loop to calculate the optimal number of sub-modules that should be deployed in the upper and lower arms of the modular multilevel converter (MMC) in the first-stage control by introducing two circulating current factors. The grouping and sorting module is used in the third stage of submodule capacitor voltage control: in each control cycle, the capacitor voltage of each bridge arm submodule is grouped and sorted.

Images