A method and device for adaptive suppression of sub / super-synchronous oscillations based on MSA-WOA

Abstract

The application relates to a MSA-WOA-based sub / super-synchronous oscillation adaptive suppression method and device, which comprises the following steps: setting a multi-channel adaptive damping controller MC-ADC in a control link of a voltage source type converter; determining input and output signals of the multi-channel adaptive damping controller, obtaining a transfer function of a sub-synchronous channel controller and a super-synchronous channel controller based on the signals; obtaining an equivalent damping coefficient term provided by a wind power grid-connected system model after an additional adaptive damping controller based on the transfer function; establishing a target function and a constraint condition based on the equivalent damping coefficient term, optimizing proportional phase-shifting link parameters based on a multi-strategy adaptive improved whale algorithm, and obtaining optimal proportional phase-shifting link parameters; and setting the optimal proportional phase-shifting link parameters as proportional phase-shifting link parameters of the sub-synchronous channel controller and the super-synchronous channel controller. Compared with the prior art, the application has the advantages of realizing fast and accurate identification of oscillation characteristics.

Description

Technical Field

[0001] This invention relates to the technical field of damping suppression, and in particular to an adaptive suppression method and apparatus for sub / supersynchronous oscillations based on MSA-WOA. Background Technology

[0002] With the application of high-proportion power electronic equipment in high-proportion wind power generation, the power system exhibits a new "dual-high" characteristic. In light of this, the "mechanical-electrical-magnetic" interactions between power electronic devices and between them and the power grid cause periodic fluctuations in voltage, current, and power information over time, easily leading to new types of oscillations with large-scale frequency variations.

[0003] The occurrence of this new type of oscillation is closely related to factors such as wind speed, the number of wind turbines, and system impedance. Furthermore, existing methods for suppressing subsynchronous / supersynchronous oscillations mostly focus on adjusting the internal control strategy of the turbine or introducing specific filters and damping controllers on the grid-connected side. However, these methods are often based on specific oscillation mechanism assumptions or single-mode oscillation characteristics, making it difficult to simultaneously consider the dynamic characteristics of multiple frequency bands and modes under different operating conditions. In addition, some damping control schemes only adjust the power frequency component, failing to consider the coupling relationship between power frequency power and oscillation component power from the overall perspective of power / energy flow, resulting in limited oscillation suppression effects, especially under conditions of large grid impedance variations and complex operating environments. Summary of the Invention

[0004] The purpose of this invention is to realize an oscillation suppression method that takes into account different oscillation modes and adapts to multiple operating conditions. Starting from the power transmission mechanism, it provides a sub / supersynchronous oscillation adaptive suppression method based on MSA-WOA by comprehensively analyzing the power frequency power and oscillation component power and realizing rapid and accurate identification of oscillation characteristics.

[0005] The objective of this invention can be achieved through the following technical solutions:

[0006] An adaptive suppression method for subsynchronous / supersynchronous oscillations based on MSA-WOA, comprising the following steps:

[0007] S1: A voltage source converter is connected in parallel at the grid connection point of the wind turbine in the wind power grid connection system model, and a multi-channel adaptive damping controller MC-ADC is set in the control loop of the voltage source converter. The controller includes a sub-synchronous channel controller and a super-synchronous channel controller.

[0008] S2: Determine the input and output signals of the multi-channel adaptive damping controller, and obtain the transfer functions of the subsynchronous channel controller and the supersynchronous channel controller based on the signals;

[0009] S3: The equivalent damping coefficient term provided for the wind power grid-connected system model after obtaining the additional adaptive damping controller based on the transfer function;

[0010] S4: Based on the equivalent damping coefficient, establish the objective function and constraints, and optimize the proportional phase shifter parameters based on the multi-strategy adaptive improved whale algorithm to obtain the optimal proportional phase shifter parameters.

[0011] S5: Set the optimal proportional phase shifter parameters to the proportional phase shifter parameters of the subsynchronous channel controller and the supersynchronous channel controller.

[0012] Furthermore, the objective function and constraints are as follows:

[0013] ;

[0014] in, The lead time constant of this synchronization channel. The time constant of this synchronization channel. It is the lead time constant of the supersynchronous channel. It is the lag time constant of the supersynchronous channel. This represents the equivalent damping control coefficient term.

[0015] Furthermore, the equivalent damping control coefficient term is:

[0016]

[0017] in, s For differential operators, Represents angular frequency. For subsynchronous channel controllers or supersynchronous channel controllers, the transfer function is... When the transfer function is the subsynchronous channel controller, the equivalent damping control coefficient term is the subsynchronous equivalent damping control coefficient. When the transfer function is the supersynchronous channel controller, the equivalent damping control coefficient term is the supersynchronous equivalent damping control coefficient.

[0018] Furthermore, the transfer function of the controller is:

[0019] ;

[0020] in, c2 The angular frequency of the power frequency component. c1 This is another oscillating component that is complementary to the oscillating component. Indicates the lead time constant. This represents the time lag constant when the controller is a subsynchronous channel controller. and These are the lead time constant and lag time constant of the subsynchronous channel, respectively. When the controller is a supersynchronous channel controller... and These are the lead time constant and lag time constant of the supersynchronous channel, respectively. Describes the differential operator. Indicates the damping ratio. This represents the proportionality coefficient.

[0021] Furthermore, the specific steps for optimizing the proportional phase-shifting element parameters based on the multi-strategy adaptive improved whale algorithm to obtain the optimal proportional phase-shifting element parameters are as follows:

[0022] Define supersynchronous decision variables and subsynchronous decision variables, wherein the subsynchronous decision variables are determined by the lead time constant of the subsynchronous channel. The lag time constant of the subsynchronous channel Composition, the lead time constant of the supersynchronous channel and the lag time constant of the supersynchronous channel Composition; Perform the following steps for both hypersynchronous and subsynchronous decision variables:

[0023] A1. Initial population individuals are generated using Sobol low-difference sequences;

[0024] A2. Take the equivalent damping control coefficient term corresponding to the decision variable as the fitness, calculate the fitness of each initial individual, take the individual with the largest fitness as the initial global optimal solution, and take the initial global optimal solution as the current global optimal solution.

[0025] A3. Calculate the control parameters and generate a random number p for strategy switching. If p < 0.5, perform a shrinking encirclement operation based on adaptive weights, calculate the fitness of the new current individual, and take the new current individual with the highest fitness as the current global optimum. Then, update the current global optimum again using the optimal neighborhood perturbation strategy. Otherwise, perform a spiral update, calculate the fitness of the new current individual, and take the new current individual with the highest fitness as the current global optimum. The control parameters include a position weight factor. A and distance scaling factor C ;

[0026] A4. Determine if the maximum number of iterations has been reached. If yes, output the current global optimal solution; otherwise, update the control parameters and return to A3.

[0027] Furthermore, the specific steps for performing the adaptive weight-based shrinking and wrapping operation are as follows:

[0028] When | AWhen | > 1, a new current individual is calculated by randomly selecting an agent based on adaptive weights; otherwise, a new current individual is calculated based on the current global optimal solution and adaptive weights.

[0029] Furthermore, the new current individual obtained by randomly selecting an agent based on adaptive weights is:

[0030]

[0031] in, Indicates the new current individual, For adaptive weights, To randomly select agents, that is, to randomly select individuals. Let t represent the current individual and t represent the current iteration number.

[0032] The new current individual, calculated based on the current global optimal solution and adaptive weights, is:

[0033]

[0034] in, This is the current globally optimal solution.

[0035] Furthermore, the new current individual obtained by spiral update is:

[0036]

[0037] in, This represents the distance between the current individual and the current global optimal solution. l It is a random number. b It is the bubble mesh shape factor. Update the parameters for the exponent.

[0038] Furthermore, the index update parameters are:

[0039]

[0040] in, This indicates the maximum number of iterations.

[0041] In another aspect, the present invention proposes a sub / supersynchronous oscillation adaptive suppression device based on MSA-WOA, the device comprising a multi-channel adaptive damping controller setting module, a transfer function calculation module, an equivalent damping coefficient term calculation module, an optimal proportional phase shifting element parameter calculation module, and a parameter setting module;

[0042] Among them, the multi-channel adaptive damping controller setting module is used to connect a voltage source converter in parallel at the grid connection point of the wind turbine in the wind power grid connection system model, and to set a multi-channel adaptive damping controller MC-ADC in the control loop of the voltage source converter. The controller includes a sub-synchronous channel controller and a super-synchronous channel controller.

[0043] The transfer function calculation module is used to determine the input and output signals of the multi-channel adaptive damping controller, and obtain the transfer functions of the subsynchronous channel controller and the supersynchronous channel controller based on the signals;

[0044] The equivalent damping coefficient calculation module is used to provide the equivalent damping coefficient term for the wind power grid-connected system model after obtaining the additional adaptive damping controller based on the transfer function.

[0045] The optimal proportional phase shifter parameter calculation module is used to establish the objective function and constraints based on the equivalent damping coefficient term, and optimize the proportional phase shifter parameters based on the multi-strategy adaptive improved whale algorithm to obtain the optimal proportional phase shifter parameters.

[0046] The parameter setting module is used to set the optimal proportional phase shifting parameters to the proportional phase shifting parameters of the subsynchronous channel controller and the supersynchronous channel controller.

[0047] Compared with the prior art, the present invention has the following beneficial effects:

[0048] (1) This invention establishes a mathematical mapping relationship between oscillation signals and controller parameters by analyzing equivalent damping, and accurately identifies oscillation characteristics. This invention analyzes the equivalent damping control coefficient term after adding MC-ADC, transforming the oscillation suppression problem into a quantitative problem that can be optimized by algorithm, ensuring that the optimized proportional phase shifter parameters can accurately match the oscillation characteristics under the current operating conditions, providing stable and sufficient positive damping for the system under various operating conditions such as different wind speeds, different number of wind turbines, and different grid impedances, fundamentally destroying the self-sustaining condition of oscillation energy, and significantly improving the pertinence and effectiveness of oscillation suppression. When calculating the transfer function of the subsynchronous channel, the subsynchronous equivalent damping coefficient is obtained; when calculating the transfer function of the supersynchronous channel, the supersynchronous equivalent damping coefficient is obtained, realizing the quantitative distinction of damping effects of different modes, and realizing the accurate identification of oscillation characteristics.

[0049] (2) This invention also employs MSA-WOA optimization parameter tuning. The proportional phase shifting parameters of the MC-ADC directly affect the damping effect and dynamic speed. In the optimization algorithm, the Sobol sequence is used to initialize the population to improve the quality of the initial solution and strengthen the global search. An optimal neighborhood perturbation strategy is introduced to avoid the algorithm getting trapped in local optima and to solve the problem of premature convergence. Adaptive weights are added to dynamically adjust the influence of the optimal position, improve the convergence speed of the algorithm, and control the exploration and development intensity through a time-varying function. Multiple improvement strategies balance the optimization intensity and exploration capability, thereby obtaining the optimal solution of the function, improving the parameter optimization efficiency and working condition adaptability, and realizing the rapid identification of oscillation characteristics. Attached Figure Description

[0050] Figure 1 This is a flowchart of the present invention;

[0051] Figure 2 This is a flowchart of the solution model of the present invention;

[0052] Figure 3 This is a system structure diagram of the MC-ADC control of the present invention;

[0053] Figure 4 This is a structural diagram of the MC-ADC of the present invention;

[0054] Figure 5 This invention provides a comparison of the active power at the grid connection point of the wind power grid-connected system with and without MC-ADC and with optimization, wherein... Figure 5 (a) represents no oscillation suppression. Figure 5 (b) represents the MC-ADC without algorithm optimization. Figure 5 (c) is a case with MC-ADC and algorithm optimization;

[0055] Figure 6 This is the controller hardware structure of the present invention;

[0056] Figure 7 This is the grid connection point oscillation current spectrum diagram of the present invention. Detailed Implementation

[0057] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. These embodiments are based on the technical solution of the present invention and provide detailed implementation methods and specific operating procedures. However, the scope of protection of the present invention is not limited to the following embodiments.

[0058] This invention provides an improved adaptive damping suppression method and apparatus for sub / supersynchronous oscillation suppression using a whale optimization algorithm. The invention utilizes a voltage source converter configured at the grid connection point of a wind power grid-connected system to inject a current with the same frequency as the oscillation component, thereby providing positive damping at the grid connection point corresponding to the oscillation frequency, absorbing the oscillation component power, and achieving oscillation suppression. A multi-channel adaptive damping controller (MC-ADC) modifies the control parameters of the parallel converter, causing it to exhibit positive damping characteristics at the oscillation frequency, absorbing the oscillation component power, and thus disrupting the self-sustaining condition of the oscillation power. The MC-ADC includes mode filtering, proportional phase shifting, and amplitude limiting summation. Based on the multi-strategy adaptive whale optimization algorithm (MSA-WOA), the proportional phase shifting parameters of the adaptive damping controller are tuned to achieve damping control under different oscillation modes. This achieves oscillation component suppression and improves the stability of the grid-connected system.

[0059] In summary, this invention forms a closed-loop synergy from four major stages: detection, identification, control, and optimization. It truly addresses the issue from the energy level and multiple strategy perspectives, achieving efficient and adaptive oscillation suppression of wind power grid-connected systems under different operating conditions in the sub- / super-synchronous phase, significantly enhancing the system's safety, reliability, and economy.

[0060] The beneficial effects of the invention are as follows:

[0061] (1) This invention is not limited to a fixed frequency band and can cover the subsynchronous / supersynchronous spectrum range. Based on multiple oscillation modes, a multi-channel control branch is set up for fusion to solve the system instability caused by multiple oscillation modes.

[0062] (2) The parameters obtained from the algorithm are applied to the model in this invention. Considering the online adjustment, the parameters are fitted by the Lagrange interpolation method to obtain a polynomial rational function which is applied to the model to realize a series of online adjustments such as measurement, identification, parameter adjustment and control.

[0063] (3) This invention differs from the simple control strategy improvement that only focuses on the dual-loop control terminal of the grid-side converter of the unit. Instead, it connects a voltage source converter in parallel at the unit's convergence and grid connection point, and adjusts the control parameters of the parallel voltage source converter in real time through MC-ADC to achieve isolation and complete suppression of oscillation energy between the wind turbine and the power grid.

[0064] (4) This invention uses MSA-WOA to optimize parameter tuning. The proportional phase shifting parameters of the MC-ADC directly affect the damping effect and dynamic speed. In the optimization algorithm, the Sobol sequence is used to initialize the population to improve the quality of the initial solution and strengthen the global search. An optimal neighborhood perturbation strategy is introduced to avoid the algorithm getting trapped in local optima and solve the problem of premature convergence. Adaptive weights are added to dynamically adjust the influence of the optimal position and improve the convergence speed of the algorithm. The exploration and development intensity is controlled by a time-varying function. Multiple improvement strategies balance the optimization intensity and exploration ability, thereby obtaining the optimal solution of the function.

[0065] This invention includes the following steps:

[0066] S1: In the wind power grid connection system model, a voltage source converter is connected in parallel at the grid connection point of the wind turbine, and a multi-channel adaptive damping controller MC-ADC is set in the control loop of the voltage source converter. The controller includes a sub-synchronous channel controller and a super-synchronous channel controller; (an additional converter is added at the grid connection point, and the control loop is set with MC-ADC)

[0067] S2: Determine the input and output signals of the multi-channel adaptive damping controller, and obtain the transfer functions of the subsynchronous channel controller and the supersynchronous channel controller based on the signals; (determine the input and output signals and composition structure of the MC-ADC)

[0068] S3: The equivalent damping coefficient term provided for the wind power grid-connected system model after obtaining the additional adaptive damping controller based on the transfer function; (Based on the Phillips-Heffron model, the electrical damping provided by the MC-ADC is obtained)

[0069] S4: Establish the objective function and constraints based on the equivalent damping coefficient term, optimize the proportional phase shifter parameters based on the multi-strategy adaptive improved whale algorithm, and obtain the optimal proportional phase shifter parameters; (Establish the objective function and constraints, and optimize the MC-ADC parameters based on the metaheuristic algorithm)

[0070] S5: Set the optimal proportional phase shifter parameters to the proportional phase shifter parameters of the subsynchronous channel controller and the supersynchronous channel controller. (Activate the optimized MC-ADC)

[0071] Based on the Phillips-Heffron model, S3 derives the electrical damping provided to the system after adding the MC-ADC.

[0072] In S4, the objective function and constraints are established, and the parameters of the proportional phase shifting stage of the MC-ADC are optimized based on the metaheuristic algorithm.

[0073] Figure 3 This is a system structure diagram of the MC-ADC control of the present invention;

[0074] Figure 4 This is a structural diagram of the MC-ADC of the present invention; Figure 4 middle, The output current of this synchronization channel, This is the output current of the supersynchronous channel; the sum of the two currents equals... Figure 3 The controller output current The current output by the converter is .

[0075] Figure 5 This invention provides a comparison of the active power at the grid connection point of the wind power grid-connected system with and without MC-ADC and with optimization, wherein... Figure 5 (a) represents no oscillation suppression. Figure 5 (b) represents the MC-ADC without algorithm optimization. Figure 5 (c) is a case with MC-ADC and algorithm optimization;

[0076] Figure 6 This is the controller hardware structure of the present invention;

[0077] Figure 7 This is the grid connection point oscillation current spectrum diagram of the present invention.

[0078] The optimized MC-ADC can be used for simulation verification in the model.

[0079] The steps for establishing the objective function of MC-ADC are as follows:

[0080] In step S4, based on the damping method, an additional parallel converter is introduced at the grid connection point, and the damping relationship expression is derived according to the definition of the damping coefficient in the Heffron Phillips model.

[0081] Electromechanical transient processes of the synchronous machine rotor:

[0082] (1)

[0083] In the formula, J For rotational inertia, For the angle of attack, T m , T e These are mechanical torque and electromagnetic torque, respectively.

[0084] Moment of inertia corresponds to the inertial time constant:

[0085] (2)

[0086] In the formula, M The inertial time constant, 0 represents the synchronous angular velocity.

[0087] Considering damping torque (For example, the effect of rotor damping windings is proportional to angular velocity), for small disturbances , ,in

[0088] (3)

[0089] After linearization, we have

[0090] (4)

[0091] After small-signal linearization, the electromagnetic torque T e This can be represented as the angle of action. and angular velocity Linear combination:

[0092] (5)

[0093] In the formula, K S is the synchronization coefficient, which is related to the power angle and reflects the synchronization torque (determining the oscillation frequency). D is the damping coefficient, which is related to the angular velocity and reflects the damping torque (determining the rate of oscillation decay).

[0094] Taking the Laplace transform of the linearized equation yields the characteristic equation.

[0095] (6)

[0096] Compared to the standard equation

[0097] (7)

[0098] achievable

[0099] (8)

[0100] By drawing an analogy between the dynamics of the DC capacitor and the rotor model of a synchronous machine, the additional damping coefficient is derived. The dynamics of the DC side power are as follows:

[0101] (9)

[0102] In the formula, C dc For DC bus capacitors, V dc DC side voltage Pdc This refers to the DC-side power.

[0103] (10)

[0104] In the formula, P in For DC-side external active power input, P out This refers to the active power on the AC side.

[0105] (11)

[0106] In the formula, s It is a differential operator.

[0107] For fluctuations in DC-side output power P out Its dynamic model is

[0108] (12)

[0109] In the formula, G ( s () is the transfer function between the output power increment and the control variable. x To control the input of variables.

[0110] Transforming this transfer function to the frequency domain and constructing a second-order form, it can be expressed as follows:

[0111] (13)

[0112] During the oscillation suppression process, the parallel converter does not "pull" active power from the grid or other power sources. Based on power conservation, the dynamic equation is rewritten as a second-order equation of motion, thus...

[0113] (14)

[0114] Comparing the above equation with the dynamic relationship of the synchronous machine rotor model, we can obtain the equivalent damping coefficient term without additional damping path as follows:

[0115] (15)

[0116] In the formula, D eq This is the equivalent damping coefficient without additional channels.

[0117] With the addition of a damping controller, since the controller is a voltage input and its output current interferes with power, from a power perspective, a dynamic mapping between power and current needs to be introduced, resulting in an equivalent power disturbance injected into the system. P = S i Therefore, the equivalent damping control coefficient on the power path caused by the additional damping controller is:

[0118] (16)

[0119] In the formula, D eq The equivalent damping coefficient with additional channels.

[0120] The objective function and constraints are as follows:

[0121] (17)

[0122] in The whale-based optimization algorithm employs a metaheuristic approach. At the outset, it randomly generates an initial solution set. Different initial distributions influence subsequent search paths, ultimately leading to varying results. Furthermore, in key steps mimicking humpback whale foraging, the algorithm introduces random numbers to determine the search direction and step size. For example, it randomly selects individuals for position updates and randomly adjusts parameters to switch search modes. This randomness results in different search trajectories each time, leading to different optimization parameters. It can be seen that traditional whale optimization algorithms are prone to getting trapped in local optima for this model, neglecting globally optimal parameters. The algorithm's unstable output degrades the performance of the MC-ADC, resulting in poor oscillation suppression.

[0123] The following improvements were made to the algorithm:

[0124] 1) Sobol sequences are used to initialize the population to improve the quality of the initial solution, making the initial distribution of elites more uniform and enhancing the global search. Sobol sequences are low-discrepancy sequences used to generate uniformly distributed sample points. Unlike traditional pseudo-random numbers, Sobol sequences exhibit better uniformity and coverage in high-dimensional spaces.

[0125] 2) Introduce an optimal neighborhood perturbation strategy to avoid the algorithm getting trapped in local optima and solve the problem of premature convergence. This strategy involves periodically performing random perturbation searches around the current optimal solution during WOA iterations, thereby expanding the search space and increasing diversity.

[0126] 3) Adaptive weights are added to dynamically adjust the influence of the optimal position and improve the algorithm's convergence speed. The exploration / exploration intensity is controlled by a time-varying function.

[0127] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0128] (1) This invention is not limited to a fixed frequency band and can cover the subsynchronous / supersynchronous spectrum range. Based on multiple oscillation modes, a multi-channel control branch is set up for fusion to solve the system instability caused by multiple oscillation modes.

[0129] (2) The parameters obtained from the algorithm are applied to the model in this invention. Considering the online adjustment, the parameters are fitted by the Lagrange interpolation method to obtain a polynomial rational function which is applied to the model to realize a series of online adjustments such as measurement, identification, parameter adjustment and control.

[0130] (3) This invention differs from the simple control strategy improvement that only focuses on the dual-loop control terminal of the grid-side converter of the unit. Instead, it connects a voltage source converter in parallel at the unit's convergence and grid connection point, and adjusts the control parameters of the parallel voltage source converter in real time through MC-ADC to achieve isolation and complete suppression of oscillation energy between the wind turbine and the power grid.

[0131] (4) This invention uses MSA-WOA to optimize parameter tuning. The proportional phase shifting parameters of the MC-ADC directly affect the damping effect and dynamic speed. In the optimization algorithm, the Sobol sequence is used to initialize the population to improve the quality of the initial solution and strengthen the global search. An optimal neighborhood perturbation strategy is introduced to avoid the algorithm getting trapped in local optima and solve the problem of premature convergence. Adaptive weights are added to dynamically adjust the influence of the optimal position and improve the convergence speed of the algorithm. The exploration and development intensity is controlled by a time-varying function. Multiple improvement strategies balance the optimization intensity and exploration ability, thereby obtaining the optimal solution of the function.

[0132] The key implementation methods and apparatus for adaptive damping suppression of sub / supersynchronous oscillations of the present invention are as follows:

[0133] I. Model Introduction and Control Modification

[0134] The method provided by this invention takes a direct-drive wind turbine grid-connected system as an example to explore suppression measures for different oscillation frequencies caused under different operating conditions, with subsynchronous / supersynchronous oscillations as the main research target.

[0135] The electromagnetic power output from the direct-drive wind turbine is filtered by a filter circuit to remove high-frequency components before being input to the rotor-side converter (RSC). Energy is then transferred via a DC capacitor to the grid-side converter (GSC). The output three-phase AC power is stepped up by a transformer and connected to the grid. During the suppression process, a voltage source converter is connected in parallel at the grid connection point. The control loop of this additional converter injects a sub- / super-synchronous current with the frequency of the oscillation component into the grid connection point, thereby increasing damping within the oscillation frequency range. An MC-ADC is added to the control loop of the additional parallel converter, and its structure is determined.

[0136] An MC-ADC is added to the inner current loop, which controls the converter output current and offers the fastest dynamic response and least delay. Placing damping control in this loop allows for a faster response to subsynchronous / supersynchronous oscillations and weakly coupled grid oscillations, thereby improving the system's electrical damping capability.

[0137] II. Design the MC-ADC structure, the steps are as follows:

[0138] The MC-ADC introduces secondary and superdamped channels respectively, with multiple channels jointly controlling the current output of the damping controller. The input signal is the three-phase voltage signal at the grid connection point. The damping controller consists of a filtering stage, a phase shifting stage, an amplification stage, and a limiting stage. The filter uses a second-order bandpass / bandstop filter, and the phase shifting stage consists of two first-order lead / lag correction stages connected in series. After gain adjustment, the output control current is obtained. The secondary and superdamped channels are connected in parallel. The following formula is the transfer function of a single channel, and the overall control function of the controller is the sum of the transfer functions of the two channels.

[0139] (18)

[0140] In the formula, c2 The angular frequency of the power frequency component. c1 This is another oscillating component that is complementary to the oscillating component. c1 =(100- s1 )2 , s1 This is an oscillating component.

[0141] Convert the transfer function to the frequency domain, let s =j s1 , can be obtained

[0142] (19)

[0143] II. Establishing the objective function of MC-ADC, the steps are as follows:

[0144] Based on the damping method, an additional parallel converter is introduced at the grid connection point, and the damping relationship expression is derived according to the definition of the damping coefficient in the Heffron Phillips model.

[0145] Electromechanical transient processes of the synchronous machine rotor:

[0146] (20)

[0147] In the formula, JFor rotational inertia, For the angle of attack, T m , T e These are mechanical torque and electromagnetic torque, respectively.

[0148] Moment of inertia corresponds to the inertial time constant:

[0149] (twenty one)

[0150] In the formula, M The inertial time constant, 0 represents the synchronous angular velocity.

[0151] Considering damping torque (For example, the effect of rotor damping windings is proportional to angular velocity), for small disturbances , ,in

[0152] (twenty two)

[0153] After linearization, we have

[0154] (twenty three)

[0155] After small-signal linearization, the electromagnetic torque T e This can be represented as the angle of action. and angular velocity Linear combination:

[0156] (twenty four)

[0157] In the formula, K S is the synchronization coefficient, which is related to the power angle and reflects the synchronization torque (determining the oscillation frequency). D is the damping coefficient, which is related to the angular velocity and reflects the damping torque (determining the rate of oscillation decay).

[0158] Taking the Laplace transform of the linearized equation yields the characteristic equation.

[0159] (25)

[0160] Compared to the standard equation

[0161] (26)

[0162] achievable

[0163] (27)

[0164] By drawing an analogy between the dynamics of the DC capacitor and the rotor model of a synchronous machine, the additional damping coefficient is derived. The dynamics of the DC side power are as follows:

[0165] (28)

[0166] In the formula, C dc For DC bus capacitors, V dc DC side voltage P dc This refers to the DC-side power.

[0167] (29)

[0168] In the formula, P in For DC-side external active power input, P out This refers to the active power on the AC side.

[0169] (30)

[0170] In the formula, s It is a differential operator.

[0171] For fluctuations in DC-side output power P out Its dynamic model is

[0172] (31)

[0173] In the formula, G ( s () is the transfer function between the output power increment and the control variable. x To control the input of variables.

[0174] Transforming this transfer function to the frequency domain and constructing a second-order form, it can be expressed as follows:

[0175] (32)

[0176] During the oscillation suppression process, the parallel converter does not "pull" active power from the grid or other power sources. Based on power conservation, the dynamic equation is rewritten as a second-order equation of motion, thus...

[0177] (33)

[0178] Comparing the above equation with the dynamic relationship of the synchronous machine rotor model, we can obtain the equivalent damping coefficient term without additional damping path as follows:

[0179] (34)

[0180] In the formula, D eq This is the equivalent damping coefficient without additional channels.

[0181] With the addition of a damping controller, since the controller is a voltage input and its output current interferes with power, from a power perspective, a dynamic mapping between power and current needs to be introduced, resulting in an equivalent power disturbance injected into the system. P = S i Therefore, the equivalent damping control coefficient on the power path caused by the additional damping controller is:

[0182] (35)

[0183] In the formula, D eq The equivalent damping coefficient with additional channels.

[0184] The objective function and constraints are as follows:

[0185] (36)

[0186] IV. MSA-WOA optimization parameter tuning, the steps are as follows:

[0187] The MSA-WOA optimization algorithm is a multi-strategy improvement based on the WOA algorithm, which evolved from the hunting methods of humpback whales. It mainly consists of three phases: the prey encirclement phase (converging towards the current optimal solution), the development phase, and the exploration phase. In the prey encirclement phase, the mathematical modeling process assumes that the current best candidate solution is the target prey or close to the optimum. After defining the optimal search agent, other search agents will attempt to update their positions to move closer to the optimal search agent. This behavior can be represented by the following equation:

[0188] (37)

[0189] (38)

[0190] in t Represents the current iteration number. A and C For the coefficient vector, A Location weighting factor C This is the distance scaling factor.X * This is the position vector of the currently obtained optimal solution. X Given the current position vector, if a better solution is found, it should be updated in each iteration. X * .

[0191] vector A and C The calculation process is as follows:

[0192] (39)

[0193] (40)

[0194] in Parameters that change linearly during the iteration process. r It is a random value between [0, 1].

[0195] (41)

[0196] in, a 1. Used to control the trade-off between exploration and development. a 2. Used for exponential updates in bubble web attack strategies.

[0197] During the prey-hunting phase, humpback whales employ two main strategies based on their behavior: a shrinking encirclement mechanism (optimal localized encirclement) and a spiral-updating position (bubble net attack). The shrinking encirclement mechanism involves reducing the... a Value is achieved by reducing a To reduce A This allows for a narrowed update of the search agent's location. Theoretically, the new location of the search agent can be defined anywhere between the agent's original location and the current best agent's location.

[0198] (42)

[0199] The spiral update position mechanism first calculates the position located at ( X , Y ) whales and ( X * , Y * The distance between the whale and its prey is determined, and then a spiral equation is created between the whale and its prey to mimic the spiral motion of a humpback whale (bubble net attack), as shown below:

[0200] (43)

[0201] in, D ’ = X* ( t )- X ( t ) indicates the first i The distance between the whale and its prey (the optimal solution obtained so far). b It is the bubble mesh shape factor, used to control the tightness of the spiral trajectory. l It is a random number between [-1, 1], representing the spiral constant.

[0202] It is worth noting that humpback whales simultaneously swim around their prey using both a contraction-circling mechanism and a spiral path. To simulate this simultaneous behavior, we assume that during the optimization process, there is a 50% probability that either the contraction-circling mechanism or the spiral model can be chosen to update the whale's position. The mathematical model is as follows:

[0203] (44)

[0204] In summary, the WOA algorithm also starts with initializing the population. In each iteration, the search agent updates its position based on either a randomly selected search agent or the best solution obtained so far. The parameter 'a' is reduced from 2 to 0 to provide exploration and exploitation respectively. When | A When | > 1, a random search agent is selected, while when | A When | < 1, the optimal solution is selected to update the search agent's position. Based on the value of the random number p (the probability of policy switching), WOA can switch between spiral and circular motions. Finally, the WOA algorithm terminates when the termination criterion is met.

[0205] The specific steps for optimizing the proportional phase shifter parameters based on the multi-strategy adaptive improved whale algorithm to obtain the optimal proportional phase shifter parameters are as follows:

[0206] Define supersynchronous decision variables and subsynchronous decision variables. The subsynchronous decision variable is determined by the lead time constant of the subsynchronous channel. The lag time constant of the subsynchronous channel Composition, the lead time constant of the supersynchronous channel and the lag time constant of the supersynchronous channel Composition; Perform the following steps for both hypersynchronous and subsynchronous decision variables:

[0207] Population data were initialized using Sobol low-difference sequences to generate the initial population individuals;

[0208] The equivalent damping control coefficient term corresponding to the decision variable is used as the fitness. The fitness of each initial individual is calculated, and the individual with the largest fitness is used as the initial global optimum. The initial global optimum is used as the current global optimum. For hypersynchronous decision variables, the hypersynchronous equivalent damping control coefficient is used as the fitness, and the same applies to subsynchronous variables.

[0209] Calculate control parameters, update variables, and generate a random number p for strategy switching. If p < 0.5, perform a shrinking encirclement operation based on adaptive weights, calculate the fitness of the new current individual, and select the new current individual with the highest fitness as the current global optimum. Then, update the current global optimum again using the optimal neighborhood perturbation strategy. Otherwise, perform a spiral update, calculate the fitness of the new current individual, and select the new current individual with the highest fitness as the current global optimum. Control parameters include position weight factors. A and distance scaling factor C ;

[0210] Determine if the maximum number of iterations has been reached (whether the termination condition has been met). If so, output the current global optimal solution; otherwise, update the control parameters and return to the step of calculating the control parameters.

[0211] On the other hand, the determination of whether the termination condition has been met can be set after the initial global optimal solution has been calculated. The steps in this case are as follows: Figure 2 As shown.

[0212] The improvements of MSA-WOA over the WOA algorithm are as follows:

[0213] 1) Sobol sequences are used to initialize the population to improve the quality of the initial solution, making the initial distribution of elites more uniform and enhancing the global search. Sobol sequences are low-discrepancy sequences used to generate uniformly distributed sample points. Unlike traditional pseudo-random numbers, Sobol sequences exhibit better uniformity and coverage in high-dimensional spaces.

[0214] 2) Introducing an optimal neighborhood perturbation strategy to avoid the algorithm getting trapped in local optima and address premature convergence. This strategy involves periodically performing random perturbation searches around the current optimal solution during WOA iterations, thereby expanding the search space and increasing diversity. The basic idea is to add random perturbations to the optimal individual position with a small probability or periodically, in addition to updating the original whale position, generating candidate new solutions and comparing them with the original optimal solution. If a new solution is better, the global optimum is updated. The perturbation strategy can break the stable structure that the algorithm may get stuck in, allowing it to escape local optima and helping to overcome premature convergence.

[0215] If the random number is less than 0.5, then add a perturbation to the optimal solution:

[0216] (45)

[0217] Otherwise, the current optimal solution remains unchanged. Of which 0.5 rand 1 indicates the intensity of the disturbance.

[0218] Calculate the fitness of the new position, and use a greedy strategy to compare it with the fitness of the original optimal solution. If the fitness of the new position is greater than the fitness of the original optimal solution, then update the optimal solution.

[0219] The optimal neighborhood perturbation strategy expands the algorithm's local search range by adding random searches around the optimal solution. Unlike the original WOA, which only searches along the current optimal direction, it periodically tries new search directions, thus effectively mitigating local traps and premature convergence problems.

[0220] 3) Adaptive weights are added to dynamically adjust the influence of the optimal position and improve the algorithm's convergence speed. A weight formula based on a cosine function is designed to control the exploration / exploration intensity using a time-varying function.

[0221] (46)

[0222] In the formula, t This represents the current iteration number. maxiter This represents the maximum number of iterations.

[0223] The trend shows that the weight is zero at the start of iteration, reaches its maximum value at half the maximum number of iterations, and returns to zero at the maximum number of iterations. This indicates that the weight curve exhibits a symmetrical upward parabola (shaped like a cosine), reaching its maximum in the middle stage. This suggests that the weight is most attractive to the leader whale in the middle stage, and weaker in the early and later stages. Taking a larger value in the early stage approximates a global search, while decreasing the weight as the number of iterations increases in the later stages prevents the algorithm from getting trapped in local optima and helps with finer searching and improved convergence accuracy in the later stages.

[0224] After introducing weights, the spiral update becomes:

[0225] (47)

[0226] Shrink wrap-around updated to:

[0227] (48)

[0228] It is evident that the weight is directly multiplied by the leader whale's position, causing a change in the relationship between the new position and the leader's position: when... By dynamically adjusting the weights, the algorithm can give the group a stronger global attraction in the early stages and weaken the attraction in the later stages, thereby balancing the optimization effort during position updates.

[0229] Under the model's basic operating conditions, the secondary / super oscillation frequencies are 25 / 75Hz.T b1 =0.001608, T p1 =0.000915, T b2 =0.001932, T p2 =0.000970, simulations were performed under multiple operating conditions to obtain the frequencies of multiple sets of sub- / supersynchronous oscillations, and parameters under multiple different oscillation frequencies were obtained. The Lagrange interpolation method was used to fit multiple points, and a polynomial rational function was constructed and applied to the model.

[0230] V. Hardware Structure Design of MC-ADC

[0231] The hardware design of MC-ADC is as follows Figure 6 As shown, multi-channel parallel filtering and independent proportional phase shifting are achieved through FPGA+ARM layered processing to match the multi-channel control requirements of MC-ADC. The MSA-WOA algorithm is deployed on the ARM, and the optimized proportional phase shifting parameters are sent to the control decision layer through a hardware interface. The digital instructions output by the control decision layer are converted into actual damping current injected into the power grid. In the FPGA, the analog signal input (grid connection point voltage / current), analog-to-digital conversion (ADC), signal preprocessing (preliminary extraction of fundamental / oscillatory components), MC-ADC control, analog-to-digital conversion (ADC), and finally analog signal output are processed. The real-time processor includes the ARM main state machine (communication). The host computer includes a PC human-machine interaction layer (parameter monitoring, algorithm debugging, waveform display).

[0232] VI. Feasibility verification of the method of the present invention.

[0233] Based on the PSCAD wind power grid connection model, a voltage source converter is added to the convergence and grid connection point of the wind turbine. An MC-ADC branch is added to the control loop of the converter to analyze the suppression effect of the controller on subsynchronous / supersynchronous oscillations.

[0234] The system operation is set as follows: after the wind turbines start, within 0-2.5 seconds, the number of wind turbines is increased to trigger multiple / supersynchronous oscillations. The damping controller is activated at the 4th second. Figure 5 This section compares the active power at the grid connection point of the wind power grid-connected system with and without MC-ADC and with optimization. Next, a spectral analysis of the subsynchronous / supersynchronous oscillation current at the grid connection point is performed. Figure 7 .

[0235] Analysis of the above grid-connected model's operational results clearly shows that the addition of the MC-ADC provides positive damping for the wind power grid-connected system. The subsynchronous / supersynchronous oscillation components at the system's grid connection point and the unit's DC side are significantly attenuated within a short period, disrupting the self-sustaining condition of the oscillation components. Simultaneously, after tuning the MC-ADC parameters using the MSA-WOA algorithm, the MC-ADC damping is maximized, allowing the controller to attenuate the energy of the subsynchronous / supersynchronous components even faster after activation, thus improving the suppression effect and demonstrating superior suppression of subsynchronous / supersynchronous oscillations in the wind power grid-connected system.

[0236] The present invention also proposes a sub / supersynchronous oscillation adaptive suppression device based on MSA-WOA. The device includes a multi-channel adaptive damping controller setting module, a transfer function calculation module, an equivalent damping coefficient term calculation module, an optimal proportional phase shifting element parameter calculation module, and a parameter setting module.

[0237] Among them, the multi-channel adaptive damping controller setting module is used to connect a voltage source converter in parallel at the grid connection point of the wind turbine in the wind power grid connection system model, and to set a multi-channel adaptive damping controller MC-ADC in the control loop of the voltage source converter. The controller includes a sub-synchronous channel controller and a super-synchronous channel controller.

[0238] The transfer function calculation module is used to determine the input and output signals of the multi-channel adaptive damping controller, and obtain the transfer functions of the subsynchronous channel controller and the supersynchronous channel controller based on the signals;

[0239] The equivalent damping coefficient calculation module is used to provide the equivalent damping coefficient term for the wind power grid-connected system model after obtaining the additional adaptive damping controller based on the transfer function.

[0240] The optimal proportional phase shifter parameter calculation module is used to establish the objective function and constraints based on the equivalent damping coefficient term, and optimize the proportional phase shifter parameters based on the multi-strategy adaptive improved whale algorithm to obtain the optimal proportional phase shifter parameters.

[0241] The parameter setting module is used to set the optimal proportional phase shifting parameters to the proportional phase shifting parameters of the subsynchronous channel controller and the supersynchronous channel controller.

[0242] Specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art can make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. Unless otherwise specified, the embodiments and features described in this application can be arbitrarily combined with each other.

Claims

1. An adaptive suppression method for sub / supersynchronous oscillations based on MSA-WOA, characterized in that, The method includes the following steps: S1: A voltage source converter is connected in parallel at the grid connection point of the wind turbine in the wind power grid connection system model, and a multi-channel adaptive damping controller MC-ADC is set in the control loop of the voltage source converter. The controller includes a sub-synchronous channel controller and a super-synchronous channel controller. S2: Determine the input and output signals of the multi-channel adaptive damping controller, and obtain the transfer functions of the subsynchronous channel controller and the supersynchronous channel controller based on the signals; S3: Equivalent damping control coefficient term provided for the wind power grid-connected system model after obtaining the additional adaptive damping controller based on the transfer function; S4: Based on the equivalent damping control coefficient, establish the objective function and constraints, and optimize the proportional phase shifter parameters based on the multi-strategy adaptive improved whale algorithm to obtain the optimal proportional phase shifter parameters. S5: Set the optimal proportional phase-shifting element parameters to the proportional phase-shifting element parameters of the sub-synchronous channel controller and the super-synchronous channel controller; the objective function and constraints are: ； in, The lead time constant of this synchronization channel. The time constant of this synchronization channel. It is the lead time constant of the supersynchronous channel. It is the lag time constant of the supersynchronous channel. This represents the equivalent damping control coefficient term; the equivalent damping control coefficient term is: in, s For differential operators, Represents angular frequency. For subsynchronous channel controllers or supersynchronous channel controllers, the transfer function is... When the transfer function is the subsynchronous channel controller, the equivalent damping control coefficient term is the subsynchronous equivalent damping control coefficient. When the transfer function is the supersynchronous channel controller, the equivalent damping control coefficient term is the supersynchronous equivalent damping control coefficient.

2. The adaptive suppression method for sub / supersynchronous oscillations based on MSA-WOA according to claim 1, characterized in that, The controller's transfer function is: ； in, c2 The angular frequency of the power frequency component. c1 This is another oscillating component that is complementary to the oscillating component. Indicates the lead time constant. This represents the time lag constant when the controller is a subsynchronous channel controller. and These are the lead time constant and lag time constant of the subsynchronous channel, respectively. When the controller is a supersynchronous channel controller... and These are the lead time constant and lag time constant of the supersynchronous channel, respectively. Describes the differential operator. Indicates the damping ratio. This represents the proportionality coefficient.

3. The adaptive suppression method for sub / supersynchronous oscillations based on MSA-WOA according to claim 1, characterized in that, The specific steps for optimizing the proportional phase shifter parameters based on the multi-strategy adaptive improved whale algorithm to obtain the optimal proportional phase shifter parameters are as follows: Define supersynchronous decision variables and subsynchronous decision variables, wherein the subsynchronous decision variables are determined by the lead time constant of the subsynchronous channel. The lag time constant of the subsynchronous channel Composition, the lead time constant of the supersynchronous channel and the lag time constant of the supersynchronous channel Composition; Perform the following steps for both hypersynchronous and subsynchronous decision variables: A1. Initial population individuals are generated using Sobol low-difference sequences; A2. Take the equivalent damping control coefficient term corresponding to the decision variable as the fitness, calculate the fitness of each initial individual, take the individual with the largest fitness as the initial global optimal solution, and take the initial global optimal solution as the current global optimal solution. A3. Calculate the control parameters and generate a random number p for strategy switching. If p < 0.5, perform a shrinking encirclement operation based on adaptive weights, calculate the fitness of the new current individual, take the new current individual with the highest fitness as the current global optimal solution, and update the current global optimal solution again using the optimal neighborhood perturbation strategy. Conversely, a spiral update is performed, and the fitness of the new current individual is calculated. The new current individual with the highest fitness is taken as the current global optimum. The control parameters include the position weight factor. A and distance scaling factor C ; A4. Determine if the maximum number of iterations has been reached. If yes, output the current global optimal solution; otherwise, update the control parameters and return to A3.

4. The adaptive suppression method for sub / supersynchronous oscillations based on MSA-WOA according to claim 3, characterized in that, The specific steps for performing the adaptive weighted shrinking and wrapping operation are as follows: When | A When | > 1, a new current individual is calculated by randomly selecting an agent based on adaptive weights; otherwise, a new current individual is calculated based on the current global optimal solution and adaptive weights.

5. The adaptive suppression method for sub / supersynchronous oscillations based on MSA-WOA according to claim 4, characterized in that, The new current individual obtained by randomly selecting an agent based on adaptive weights is: in, Indicates the new current individual, For adaptive weights, To randomly select agents, that is, to randomly select individuals. Let t represent the current individual and t represent the current iteration number. The new current individual, calculated based on the current global optimal solution and adaptive weights, is: in, This is the current globally optimal solution.

6. The adaptive suppression method for sub / supersynchronous oscillations based on MSA-WOA according to claim 5, characterized in that, The new current individual obtained by spiral update is: in, This represents the distance between the current individual and the current global optimal solution. l It is a random number. b It is the bubble mesh shape factor. Update the parameters for the exponent.

7. The adaptive suppression method for sub / supersynchronous oscillations based on MSA-WOA according to claim 6, characterized in that, The index update parameters are: in, This indicates the maximum number of iterations.

8. A subsynchronous / supersynchronous oscillation adaptive suppression device based on MSA-WOA, characterized in that, The device includes a multi-channel adaptive damping controller setting module, a transfer function calculation module, an equivalent damping control coefficient term calculation module, an optimal proportional phase shifting element parameter calculation module, and a parameter setting module. Among them, the multi-channel adaptive damping controller setting module is used to connect a voltage source converter in parallel at the grid connection point of the wind turbine in the wind power grid connection system model, and to set a multi-channel adaptive damping controller MC-ADC in the control loop of the voltage source converter. The controller includes a sub-synchronous channel controller and a super-synchronous channel controller. The transfer function calculation module is used to determine the input and output signals of the multi-channel adaptive damping controller, and obtain the transfer functions of the subsynchronous channel controller and the supersynchronous channel controller based on the signals; The equivalent damping control coefficient calculation module is used to provide the equivalent damping control coefficient term for the wind power grid-connected system model after obtaining the additional adaptive damping controller based on the transfer function. The optimal proportional phase shifter parameter calculation module is used to establish the objective function and constraints based on the equivalent damping control coefficient term, and optimize the proportional phase shifter parameters based on the multi-strategy adaptive improved whale algorithm to obtain the optimal proportional phase shifter parameters. The parameter setting module is used to set the optimal proportional phase shifting parameters to the proportional phase shifting parameters of the subsynchronous channel controller and the supersynchronous channel controller; The objective function and constraints are as follows: ； in, The lead time constant of this synchronization channel. The time constant of this synchronization channel. It is the lead time constant of the supersynchronous channel. It is the lag time constant of the supersynchronous channel. This represents the equivalent damping control coefficient term; The equivalent damping control coefficient term is: in, s For differential operators, Represents angular frequency. For subsynchronous channel controllers or supersynchronous channel controllers, the transfer function is... When the transfer function is the subsynchronous channel controller, the equivalent damping control coefficient term is the subsynchronous equivalent damping control coefficient. When the transfer function is the supersynchronous channel controller, the equivalent damping control coefficient term is the supersynchronous equivalent damping control coefficient.

Images