MMC-HVDC controller optimization method in combination with quasi-proportional resonance-nonlinear active disturbance rejection

Abstract

The invention discloses an MMC-HVDC controller optimization method in combination with quasi-proportional resonance-nonlinear active disturbance rejection, and belongs to the technical field of new energy power generation system grid connection. In order to solve the problem of poor grid connection stability of the existing back-to-back MMC-HVDC grid connection method of the direct-driven permanent magnet synchronous power generation system, the method comprises the following steps: establishing a back-to-back MMC-HVDC module grid connection simulation model; establishing a Spearman correlation analysis experiment, obtaining a correlation coefficient, and screening out a PI controller which has a great influence on the grid-connected stability; a QPR-NLADRC controller is adopted to replace the PI controller which is screened out; establishing a multi-objective optimization function, determining parameter constraint conditions, performing multi-objective processing on a single-objective egret optimization algorithm, introducing an improvement strategy, and optimizing an MMC-HVDC controller combined with QPR-NLADRC; and carrying out stability analysis and judgment on the optimized system. The invention is used for the direct-drive permanent magnet synchronous power generation system.

Description

Technical Field

[0001] The invention relates to an MMC-HVDC controller optimization method combining quasi-proportional resonance-nonlinear self-disturbance rejection, belonging to the technical field of grid connection of new energy power generation systems. Background Art

[0002] Modular multilevel converter (MMC) has the characteristics of low loss, high efficiency and strong adaptability, and has become an important part of high voltage direct current (HVDC) technology. Wind farms and MMC-HVDC grid-connected systems contain different power electronic structures and controller links, which have high requirements on the controller structure and accuracy. Among them, the MMC-HVDC controller is crucial in the flexible direct current transmission system, and whether its design is reasonable directly determines the stability and operation performance of the system. Therefore, in-depth research on the MMC-HVDC controller structure and further optimization of the control strategy can improve the stability and reliability of the system.

[0003] The existing direct-drive permanent magnet synchronous power generation system adopts PI control through the back-to-back MMC-HVDC grid-connected method. Although the PI controller has a simple adjustment method, its tracking performance for interference is poor, which will cause system oscillation and poor grid-connected stability during the process of new energy grid connection. Summary of the invention

[0004] The purpose of the present invention is to solve the problem of poor grid connection stability of the existing direct-drive permanent magnet synchronous power generation system via a back-to-back MMC-HVDC grid connection method, and to provide an MMC-HVDC controller optimization method combining quasi-proportional resonance-nonlinear self-disturbance rejection.

[0005] The MMC-HVDC controller optimization method combining quasi-proportional resonance and nonlinear auto-disturbance rejection described in the present invention specifically includes:

[0006] S1. Using the digital model of the permanent magnet direct-drive wind power generation system, a grid-connected simulation model is established through back-to-back MMC-HVDC modules;

[0007] S2, monitoring the voltage quality of the DC voltage of the back-to-back MMC-HVDC module grid-connected simulation model;

[0008] Monitor the current quality of the AC current at the common connection point between the permanent magnet direct-drive wind power generation system and the MMC-HVDC module;

[0009] S3. According to the voltage quality monitoring and current quality monitoring results, a Spearman correlation analysis experiment is established to obtain the correlation coefficient;

[0010] S4. According to the correlation coefficient, select the PI controller that has a greater impact on grid stability;

[0011] S5, using the QPR-NLADRC controller that combines the quasi-proportional resonance QPR and the nonlinear auto-disturbance rejection control NLADRC to replace the PI controller selected in S4,

[0012] Obtained MMC-HVDC controller combined with QPR-NLADRC;

[0013] S6. Use response surface method to establish multi-objective optimization function and determine parameter constraints.

[0014] The single-objective egret optimization algorithm is processed with multiple objectives, and an improvement strategy is introduced to improve it.

[0015] Obtained the multi-objective improved Heron Eagle optimization algorithm;

[0016] S7, using the multi-objective improved Heron Eagle optimization algorithm to optimize the MMC-HVDC controller combined with QPR-NLADRC;

[0017] S8. Establish the impedance model of the permanent magnet direct-drive wind power generation system and the back-to-back MMC-HVDC module, and conduct stability analysis and judgment on the optimized system.

[0018] Preferably, the specific method of selecting the PI controller having a greater impact on grid-connected stability according to the correlation coefficient in S4 includes:

[0019] The closer the correlation coefficient is to 1, the greater the positive correlation is, and the greater the impact on grid stability is;

[0020] The closer the correlation coefficient is to -1, the greater the negative correlation is, and the greater the impact on grid stability is;

[0021] According to the grid-connected requirements, the optimal number of PI controller groups is selected, and based on the optimal number of groups, the PI controller with a greater impact on the grid-connected stability is screened out.

[0022] Preferably, the specific method for establishing the Spearman correlation analysis experiment in S3 includes:

[0023] Suppose there are two data sets X and Y, the number of elements in each data set is N;

[0024] The i-th value in the dataset X is X i , the i-th value in the data set Y is Y i , the value range of i is [1,N];

[0025] Sort X and Y in ascending or descending order to obtain two sorted sets x and y respectively;

[0026] The set of PI parameters in the Spearman correlation analysis experiment is the data set X, and the set of objective functions in the Spearman correlation analysis experiment is the data set Y;

[0027] Take the absolute average value of the correlation coefficient r between each parameter and the target, and select the parameters with the largest absolute average value:

[0028]

[0029] Among them, i represents the i-th target;

[0030] j represents the jth element;

[0031] x ij Represents the jth element of the i-th target in the dataset X;

[0032] y ij Represents the jth element of the i-th target in the dataset Y;

[0033] Indicates the average value of all elements in the data set X;

[0034] It means to find the average value of all elements in the data set Y.

[0035] Preferably, the QPR-NLADRC controller described in S5 specifically includes:

[0036] The core component of the nonlinear active disturbance rejection control NLADRC is NLESO, which adds quasi-proportional resonance QPR to the disturbance feedforward channel to form an improved quasi-resonant nonlinear extended state observer;

[0037] Transfer function G of improved quasi-resonant nonlinear extended state observer QPR (s) is:

[0038]

[0039] Where K1 represents the basic gain of the QPR controller for the error signal,

[0040] K R represents the gain value of the QPR controller at the resonance point,

[0041] ω c represents the gain bandwidth of the QPR controller at the resonance point.

[0042] ω0 represents the set resonant frequency of the QPR controller,

[0043] s represents a complex variable.

[0044] Preferably, the improvement strategy introduced in S6 includes: SPM chaos mapping initialization, improved nonlinear perturbation factor, improved Levy flight strategy and improved attenuation factor strategy.

[0045] Preferably, the optimization objective of the multi-objective optimization function in S6 is:

[0046] Back-to-back MMC DC voltage overshoot U m Minimum;

[0047] The DC voltage stabilization time point t of the back-to-back MMC is the shortest;

[0048] Phase A current at the common connection point i A The harmonic distortion rate THD is minimal.

[0049] Preferably, the specific method of establishing the impedance model of the permanent magnet direct-drive wind power generation system and the back-to-back MMC-HVDC module in S8 includes:

[0050] The ideal current source I of the wind power system connected in parallel is PMSG and impedance Z PMSG , MMC is equivalent to an ideal voltage source V in series s and impedance Z MMC , and the current I at the common connection point is obtained PCC (s) and voltage V PCC They are:

[0051]

[0052] Preferably, in S8, stability analysis is performed on the optimized system, and the conditions for stability analysis are:

[0053] Under the ideal voltage source, when the impedance Z of the two cascade systems is MMC and Z PWMG When the amplitude curve has an intersection and the phase difference at the intersection is close to or greater than 180°, the system is considered unstable.

[0054] Preferably, in the digital model of the permanent magnet direct-drive wind power generation system described in S1, a grid-connected simulation model is established through back-to-back MMC-HVDC modules, in which the MMC on the wind turbine side adopts outer loop power and inner loop current control, and the MMC on the grid side adopts outer loop DC voltage control and inner loop current control.

[0055] Advantages of the present invention: The MMC-HVDC controller optimization method combining quasi-proportional resonance and nonlinear anti-disturbance control described in the present invention replaces PI controllers at different positions through Spearman correlation analysis, thereby improving the grid-connected stability of the direct-drive permanent magnet synchronous power generation system via back-to-back MMC-HVDC.

[0056] Considering the QPR-NLADRC parameter optimization problem, a multi-objective improved Heron Eagle optimization algorithm was proposed. After establishing a mathematical surrogate model through response surface analysis, the parameters were optimized.

[0057] Finally, the feasibility and effectiveness of the proposed method are analyzed by establishing the back-to-back MMC-HVDC and power generation system impedance model. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 This is a topological structure diagram of the permanent magnet direct-drive wind power generation system of the present invention connected to the grid via back-to-back MMC-HVDC;

[0059] Figure 2 This is a back-to-back MMC topology diagram;

[0060] Figure 3 This is the back-to-back MMC control structure diagram;

[0061] Figure 4 It is a structural diagram of the improved quasi-resonant nonlinear extended state observer of the present invention;

[0062] Figure 5 It is a flow chart of the multi-objective improved egret optimization algorithm of the present invention;

[0063] Figure 6 This is an impedance analysis comparison chart. DETAILED DESCRIPTION

[0064] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0065] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other.

[0066] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments, but they are not intended to limit the present invention.

[0067] Embodiment 1:

[0068] The MMC-HVDC controller optimization method combined with quasi-proportional resonance-nonlinear auto-disturbance rejection described in this embodiment includes:

[0069] S1. Using the digital model of the permanent magnet direct-drive wind power generation system, a grid-connected simulation model is established through back-to-back MMC-HVDC modules;

[0070] S2, monitoring the voltage quality of the DC voltage of the back-to-back MMC-HVDC module grid-connected simulation model;

[0071] Monitor the current quality of the AC current at the common connection point between the permanent magnet direct-drive wind power generation system and the MMC-HVDC module;

[0072] S3. According to the voltage quality monitoring and current quality monitoring results, a Spearman correlation analysis experiment is established to obtain the correlation coefficient;

[0073] S4. According to the correlation coefficient, select the PI controller that has a greater impact on grid stability;

[0074] S5, using the QPR-NLADRC controller that combines the quasi-proportional resonance QPR and the nonlinear auto-disturbance rejection control NLADRC to replace the PI controller selected in S4,

[0075] Obtained MMC-HVDC controller combined with QPR-NLADRC;

[0076] S6. Use response surface method to establish multi-objective optimization function and determine parameter constraints.

[0077] The single-objective egret optimization algorithm is processed with multiple objectives, and an improvement strategy is introduced to improve it.

[0078] Obtained the multi-objective improved Heron Eagle optimization algorithm;

[0079] S7, using the multi-objective improved Heron Eagle optimization algorithm to optimize the MMC-HVDC controller combined with QPR-NLADRC;

[0080] S8. Establish the impedance model of the permanent magnet direct-drive wind power generation system and the back-to-back MMC-HVDC module, and conduct stability analysis and judgment on the optimized system.

[0081] Furthermore, the specific method of selecting the PI controller having a greater impact on the grid stability according to the correlation coefficient described in S4 includes:

[0082] The closer the correlation coefficient is to 1, the greater the positive correlation is, and the greater the impact on grid stability is;

[0083] The closer the correlation coefficient is to -1, the greater the negative correlation is, and the greater the impact on grid stability is;

[0084] According to the grid-connected requirements, the optimal number of PI controller groups is selected, and based on the optimal number of groups, the PI controller with a greater impact on the grid-connected stability is screened out.

[0085] Furthermore, the specific method for establishing the Spearman correlation analysis experiment described in S3 includes:

[0086] Suppose there are two data sets X and Y.

[0087] The number of elements in the data set is N;

[0088] The i-th value in the dataset X is X i ,

[0089] The i-th value in the data set Y is Y i ,

[0090] The value range of i is [1, N];

[0091] Sort X and Y in ascending or descending order to obtain two sorted sets x and y respectively;

[0092] The set of PI parameters in the Spearman correlation analysis experiment is the data set X, and the set of objective functions in the Spearman correlation analysis experiment is the data set Y;

[0093] Take the absolute average value of the correlation coefficient r between each parameter and the target, and select the parameters with the largest absolute average value:

[0094]

[0095] Among them, i represents the i-th target;

[0096] j represents the jth element;

[0097] x ij Represents the jth element of the i-th target in the dataset X;

[0098] y ij Represents the jth element of the i-th target in the dataset Y;

[0099] Indicates the average value of all elements in the data set X;

[0100] It means to find the average value of all elements in the data set Y.

[0101] Furthermore, the QPR-NLADRC controller described in S5 specifically includes:

[0102] The core component of the nonlinear active disturbance rejection control NLADRC is NLESO, which adds quasi-proportional resonance QPR to the disturbance feedforward channel to form an improved quasi-resonant nonlinear extended state observer;

[0103] Transfer function G of improved quasi-resonant nonlinear extended state observer QPR (s) is:

[0104]

[0105] Where K1 represents the basic gain of the QPR controller for the error signal,

[0106] K R represents the gain value of the QPR controller at the resonance point,

[0107] ω c represents the gain bandwidth of the QPR controller at the resonance point.

[0108] ω0 represents the set resonant frequency of the QPR controller,

[0109] s represents a complex variable.

[0110] Furthermore, the improvement strategies introduced in S6 include: SPM chaos mapping initialization, improved nonlinear perturbation factor, improved Levy flight strategy and improved attenuation factor strategy.

[0111] Furthermore, the optimization objective of the multi-objective optimization function in S6 is:

[0112] Back-to-back MMC DC voltage overshoot U m Minimum;

[0113] The DC voltage stabilization time point t of the back-to-back MMC is the shortest;

[0114] Phase A current at the common connection point i A The harmonic distortion rate THD is minimal.

[0115] Furthermore, the specific method of establishing the impedance model of the permanent magnet direct-drive wind power generation system and the back-to-back MMC-HVDC module in S8 includes:

[0116] The ideal current source I of the wind power system connected in parallel is PMSG and impedance Z PMSG , MMC is equivalent to an ideal voltage source V in series s and impedance Z MMC , and the current I at the common connection point is obtained PCC (s) and voltage V PCC They are:

[0117]

[0118] Furthermore, in step S8, the stability analysis and determination is performed on the optimized system, and the conditions for the stability analysis and determination are:

[0119] Under the ideal voltage source, when the impedance Z of the two cascade systems is MMC and Z PWMG When the amplitude curve has an intersection and the phase difference at the intersection is close to or greater than 180°, the system is considered unstable.

[0120] Furthermore, in the digital model of the permanent magnet direct-drive wind power generation system described in S1, a grid-connected simulation model is established through back-to-back MMC-HVDC modules, in which the MMC on the wind turbine side adopts outer loop power and inner loop current control, and the MMC on the grid side adopts outer loop DC voltage control and inner loop current control.

[0121] In the present invention, the PI controller with a greater impact on system stability is selected in combination with the correlation coefficient obtained by Spearman correlation analysis, and is improved into a QPR-NLADRC controller, and the parameters are adjusted using an improved intelligent optimization algorithm.

[0122] Through mathematical model derivation, a simulation model of permanent magnet direct-drive wind power generation system connected to the grid via back-to-back MMC-HVDC is built in the simulation software as shown in the attached figure. Figure 1 , back-to-back MMC topology as shown in the attached Figure 2 As shown, the back-to-back MMC control block diagram is shown in the attached Figure 3 shown.

[0123] By monitoring the voltage and current quality of back-to-back MMC DC voltage and the AC current at the common connection point of the wind power generation system and MMC-HVDC, a Spearman correlation analysis experiment is established to analyze the influence of PI parameters in the back-to-back MMC controller on the grid-connected stability. The closer the correlation coefficient is to 1, the greater the positive correlation is, and the closer the correlation coefficient is to -1, the greater the negative correlation is. Based on this, the PI controller that has a significant impact on system stability is screened.

[0124] A controller structure combining quasi-proportional resonance (QPR) and nonlinear active disturbance rejection control (NLADRC) is proposed, and the improved QPR-NLADRC controller replaces the PI controller (although the traditional PI controller is simple to adjust, its interference tracking performance is poor. The improved controller can solve the system oscillation problem caused by the grid connection of new energy).

[0125] Three multi-objective optimization functions are established through response surface methodology (controller parameters are about the internal DC voltage overshoot f1 of the back-to-back MMC, the stable time point f2, and the AC harmonic distortion rate f3 of phase A at the PCC), and the constraint conditions of the parameters are determined (constrained according to the actual parameters of the model, f1∈[500,2000], f2∈[0.05,0.5], f3∈[10,35]). The single-objective Egret Eagle optimization algorithm is subjected to multi-objective processing, and a variety of improvement strategies are introduced (SPM chaos mapping initialization, improved nonlinear perturbation factor, improved Levy flight strategy, improved attenuation factor), and the multi-objective improved Egret Eagle optimization algorithm is obtained to optimize the controller parameters.

[0126] The impedance model of the power generation system and the back-to-back MMC-HVDC is established, and the Bode diagrams of the MMC-HVDC system with PI controller and improved QPR-NLADRC configuration are compared to verify the effectiveness of the proposed structure in improving system stability.

[0127] A controller structure combining quasi-proportional resonance (QPR) and nonlinear active disturbance rejection control (NLADRC). NLESO is the core component of NLADRC. Compared with proportional resonance (PR), quasi-proportional resonance (QPR) controller can improve stability and expand bandwidth when the system has frequency deviation. In order to eliminate harmonic disturbances, a quasi-resonant controller is added to the disturbance feedforward channel to form an improved quasi-resonant nonlinear extended state observer (the specific structure is shown in the attached figure). Figure 4 This structure:

[0128] 1. It can enhance the robustness caused by frequency changes and better adapt to frequency changes;

[0129] 2. It can also better suppress the oscillation of the system near the resonant frequency, thereby improving the system stability;

[0130] 3. In addition, the control bandwidth of the NLADRC system can be expanded, enabling faster response to frequency instability and external disturbances.

[0131] The improved Heron Eagle optimization algorithm introduces SPM chaos map initialization, improved nonlinear perturbation factor, improved Levy flight strategy, and improved attenuation factor strategy to solve the problems of uneven initial population distribution, uneven exploration results, falling into local optimal solution, falling into local optimal solution too early, or not being able to make full use of subsequent iterations. The optimization algorithm is based on the back-to-back MMC DC voltage overshoot U m Minimum, back-to-back MMC DC voltage stable time point t is the shortest, and the A phase current i at the common connection point A Multi-objective optimization is carried out with the minimum harmonic distortion rate THD as three objectives.

[0132] The mathematical expression of the objective function is determined by the response surface method. Replacing the original simulation model with a mathematical proxy model can not only effectively avoid various coupling problems caused by the complex structure of the simulation model, but also accurately reflect the mathematical relationship between the parameters and the target quantity. The first five parameters with the largest correlation with the target value are screened in advance through the Spearman correlation to optimize the parameter variables of the objective function.

[0133] The multi-objective improved Heron Eagle optimization algorithm needs to be performance tested and compared with other algorithms.

[0134] In the present invention, Figure 1 As shown in the figure, it is the topological structure diagram of the permanent magnet direct-drive wind power generation system connected to the grid via back-to-back MMC-HVDC, where U dc is the internal DC voltage of the back-to-back MMC, i A It is the A-phase current emitted by the wind power generation system to the grid side.

[0135] On this basis, U dc Overshoot U m and stable time point t, i A The harmonic distortion rate THD is used as the monitoring target to study the system stability.

[0136] Among them U m It is obtained by fitting the DC voltage waveform;

[0137] t is the moment when the fitted DC voltage amplitude is always within [1.48kV 1.52kV];

[0138] THD is A Obtained by FFT analysis.

[0139] like Figure 2 , is a back-to-back MMC topology diagram. The MMC on the wind turbine side adopts outer loop power and inner loop current control, and the MMC on the grid side adopts outer loop DC voltage control and inner loop current control.

[0140] Outer loop power control:

[0141] The three-phase AC current and voltage sent to the MMC on the fan side are converted through the coordinate system to obtain the current and voltage in the rotating coordinate system.

[0142] Then the power is obtained by calculation, and finally the dq axis current reference value in the MMC inner loop controller on the fan side is obtained:

[0143]

[0144] Where ΔP' is the difference between the power measurement value and the reference value;

[0145] K p1 , K i1 is the PI control parameter in the outer loop power control.

[0146] Inner loop power control:

[0147] Subtract the dq axis current reference value obtained by the outer loop control from the actual value.

[0148] After passing through the PI controller, the voltage reference value of the dq axis voltage can be obtained by subtracting the voltage drop caused by the equivalent leakage inductance of the converter transformer:

[0149]

[0150] Wherein, Δi' is the difference between the current measurement value and the reference value;

[0151] K p2 , K i2 is the PI control parameter in the inner loop current control;

[0152] L T is the equivalent leakage inductance.

[0153] Interphase circulating current suppression control:

[0154] Interphase circulating current suppression control can effectively reduce or eliminate circulating current and ensure safe and stable operation of the system.

[0155] Taking phase A of the MMC on the fan side as an example, the unbalanced voltage drop of phase A is obtained as follows:

[0156]

[0157] Among them, i' cir is the circulating current.

[0158] Outer loop DC voltage control:

[0159] The d-axis current reference value of the inner loop current control of the MMC on the grid side can be obtained by subtracting the DC voltage measurement value from the reference value and passing it through the PI controller. The q-axis reference value is set to 0.

[0160] The specific formula is:

[0161] i d_ref =(K p4 +K i4 ∫dt)ΔU dc ;

[0162] Among them, ΔU dc is the difference between the DC voltage measurement value and the reference value;

[0163] K p4 , K i4 is the PI control parameter in the outer loop DC voltage control.

[0164] like Figure 3 As shown in the figure, it is a back-to-back MMC control structure diagram. In order to eliminate harmonic disturbances, a quasi-resonant controller is added to the disturbance feedforward channel to form an improved quasi-resonant nonlinear extended state observer, as shown in Figure 4 As shown. Compared with proportional resonance, the quasi-proportional resonant controller can improve stability and expand the bandwidth when the system has frequency deviation. Its transfer function G QPR (s) is:

[0165]

[0166] Where K1 represents the basic gain of the QPR controller for the error signal;

[0167] K R Indicates the gain value (peak gain) of the QPR controller at the resonance point;

[0168] K R As it increases, the peak gain of the controller also increases, K R The parameter is proportional to the peak gain of the controller;

[0169] ω c Used to adjust the gain bandwidth of the QPR controller at the resonance point;

[0170] ω c The larger the QPR controller gain bandwidth, the larger the c The smaller it is, the closer the amplitude-frequency curve of the QPR controller at the resonance point is to that of the PR controller.

[0171] ω0 represents the set resonant frequency of the QPR controller;

[0172] s is a complex variable.

[0173] Figure 5 The flowchart of the multi-objective improved egret optimization algorithm of the present invention is shown in Figure 1. The position formula of each egret in the population generated by SPM chaos mapping initialization is:

[0174]

[0175] After introducing the improved nonlinear perturbation factor, the position update formula in the exploration phase is:

[0176]

[0177] The improved Levy flight strategy introduces the dynamic factor s, the formula is:

[0178]

[0179] After introducing the improved attenuation factor, the position update formula in the development phase is:

[0180]

[0181] For different optimization algorithms (MOSBOA, multi-objective dragonfly algorithm (MODA) and multi-objective locust algorithm (MOGOA)), the same population size (100), maximum number of iterations (200) and dimension (3) were set, and 20 independent runs were performed using the true Pareto solution set of the DTLZ2 test function. The average value was taken as the final comparison result. The specific data is shown in Table 1.

[0182] Table 1

[0183] name GD IGD HV Spacing Spread MOSBOA 0.0200 0.1798 0.3028 0.6536 0.1016 MOISBOA 0.0199 0.1798 0.3031 0.6614 0.1046 MODA 0.0316 0.1897 0.2902 0.5002 0.0937 MOGOA 0.0305 0.1876 0.2933 0.5373 0.0877

[0184] like Figure 6 , is the impedance analysis comparison chart, according to the stability criterion:

[0185] Under the ideal voltage source, if the impedance Z of the two cascade systems is MMC and Z PWMG When the amplitude curve of has an intersection, and the phase difference at the intersection is close to or greater than 180°, the system is considered unstable. The conclusion is that the system is unstable before the improvement, and the system reaches a stable state after the improvement.

[0186] Although the present invention is described herein with reference to specific embodiments, it should be understood that these embodiments are merely examples of the principles and applications of the present invention. It should therefore be understood that many modifications may be made to the exemplary embodiments and that other arrangements may be devised without departing from the spirit and scope of the present invention as defined by the appended claims. It should be understood that the various dependent claims and features described herein may be combined in a manner different from that described in the original claims. It should also be understood that features described in conjunction with individual embodiments may be used in other described embodiments.

Claims

1. An optimization method for MMC-HVDC controller combining quasi-proportional resonance and nonlinear auto-disturbance rejection, characterized in that: It includes: S1. Using the digital model of the permanent magnet direct-drive wind power generation system, a grid-connected simulation model is established through back-to-back MMC-HVDC modules; S2, monitoring the voltage quality of the DC voltage of the back-to-back MMC-HVDC module grid-connected simulation model; Monitor the current quality of the AC current at the common connection point between the permanent magnet direct-drive wind power generation system and the MMC-HVDC module; S3. According to the voltage quality monitoring and current quality monitoring results, a Spearman correlation analysis experiment is established to obtain the correlation coefficient; S4. According to the correlation coefficient, select the PI controller that has a greater impact on grid stability; S5, using the QPR-NLADRC controller that combines the quasi-proportional resonance QPR and the nonlinear auto-disturbance rejection control NLADRC to replace the PI controller selected in S4, Obtained MMC-HVDC controller combined with QPR-NLADRC; S6. Use response surface method to establish multi-objective optimization function and determine parameter constraints. The single-objective egret optimization algorithm is processed with multiple objectives, and an improvement strategy is introduced to improve it. Obtained the multi-objective improved Heron Eagle optimization algorithm; S7, using the multi-objective improved Heron Eagle optimization algorithm to optimize the MMC-HVDC controller combined with QPR-NLADRC; S8. Establish the impedance model of the permanent magnet direct-drive wind power generation system and the back-to-back MMC-HVDC module, and conduct stability analysis and judgment on the optimized system.

2. The MMC-HVDC controller optimization method combining quasi-proportional resonance and nonlinear auto-disturbance rejection according to claim 1, characterized in that: The specific method of selecting the PI controller having a greater impact on grid-connected stability according to the correlation coefficient described in S4 includes: The closer the correlation coefficient is to 1, the greater the positive correlation is, and the greater the impact on grid stability is; The closer the correlation coefficient is to -1, the greater the negative correlation is, and the greater the impact on grid stability is; According to the grid-connected requirements, the optimal number of PI controller groups is selected, and based on the optimal number of groups, the PI controller with a greater impact on the grid-connected stability is screened out.

3. The MMC-HVDC controller optimization method combining quasi-proportional resonance and nonlinear auto-disturbance rejection according to claim 1, characterized in that: The specific method for establishing the Spearman correlation analysis experiment described in S3 includes: Suppose there are two data sets X and Y. The number of elements in the data set is N; The i-th value in the dataset X is X i , The i-th value in the data set Y is Y i , The value range of i is [1, N]; Sort X and Y in ascending or descending order to obtain two sorted sets x and y respectively; The set of PI parameters in the Spearman correlation analysis experiment is the data set X, and the set of objective functions in the Spearman correlation analysis experiment is the data set Y; Take the absolute average value of the correlation coefficient r between each parameter and the target, and select the parameters with the largest absolute average value: Among them, i represents the i-th target; j represents the j-th element; x ij Represents the jth element of the i-th target in the dataset X; y ij Represents the jth element of the i-th target in the dataset Y; Indicates the average value of all elements in the data set X; It means to find the average value of all elements in the data set Y.

4. The MMC-HVDC controller optimization method combining quasi-proportional resonance and nonlinear auto-disturbance rejection according to claim 1, characterized in that: The QPR-NLADRC controller described in S5 specifically includes: The core component of the nonlinear active disturbance rejection control NLADRC is NLESO. By adding quasi-proportional resonance (QPR) to the disturbance feed-forward path, Forming an improved quasi-resonant nonlinear extended state observer; Transfer function G of improved quasi-resonant nonlinear extended state observer QPR (s) is: Where K1 represents the basic gain of the QPR controller for the error signal; K R Represents the gain value of the QPR controller at the resonance point; ω c It represents the gain bandwidth of adjusting the QPR controller at the resonance point; ω0 represents the set resonant frequency of the QPR controller; s represents a complex variable.

5. The MMC-HVDC controller optimization method combining quasi-proportional resonance and nonlinear auto-disturbance rejection according to claim 1, characterized in that: The improved strategies introduced in S6 include: SPM chaos mapping initialization, improved nonlinear perturbation factor, improved Levy flight strategy and improved attenuation factor strategy.

6. The MMC-HVDC controller optimization method combining quasi-proportional resonance and nonlinear auto-disturbance rejection according to claim 1, characterized in that: The optimization objective of the multi-objective optimization function described in S6 is: Back-to-back MMC DC voltage overshoot U m Minimum; The DC voltage stabilization time point t of the back-to-back MMC is the shortest; Phase A current at the common connection point i A The harmonic distortion rate THD is minimal.

7. The MMC-HVDC controller optimization method combining quasi-proportional resonance and nonlinear auto-disturbance rejection according to claim 1, characterized in that: The specific method of establishing the impedance model of the permanent magnet direct-drive wind power generation system and the back-to-back MMC-HVDC module in S8 includes: The ideal current source I of the wind power system connected in parallel is PMSG and impedance Z PMSG , MMC is equivalent to an ideal voltage source V in series s and impedance Z MMC ; The current I at the common connection point is obtained PCC (s) and voltage V PCC They are:

8. The MMC-HVDC controller optimization method combining quasi-proportional resonance and nonlinear auto-disturbance rejection according to claim 1, characterized in that: The stability analysis and determination of the optimized system described in S8 is performed, and the conditions for the stability analysis and determination are: Under the ideal voltage source, when the impedance Z of the two cascade systems is MMC and Z PWMG When the amplitude curve has an intersection and the phase difference at the intersection is close to or greater than 180°, the system is considered unstable.

9. The MMC-HVDC controller optimization method combined with quasi-proportional resonance-nonlinear auto-disturbance rejection according to any one of claims 1 to 8, characterized in that: S1 uses a digital model of a permanent magnet direct-drive wind power generation system to establish a grid-connected simulation model through back-to-back MMC-HVDC modules. The fan side MMC adopts outer loop power and inner loop current control; The grid-side MMC adopts outer loop DC voltage control and inner loop current control.

Images