Adaptive control method of inertia synchronous pmsg based on online impedance estimation

Abstract

The application relates to an adaptive control method of inertia synchronization PMSG based on online impedance estimation, and belongs to the field of new energy grid connection. The method comprises the following steps: S1, constructing an RLS-GA hybrid strategy algorithm to quickly and accurately identify the grid impedance, and determining the strong and weak types of the grid; S2, establishing a direct-drive wind turbine grid connection system control strategy, adopting a power control loop, a current control loop and an inertia transmission control loop on the generator side; adopting a direct-current voltage synchronization loop, a reactive voltage control loop and a stabilizing control loop on the grid side; S3, linearizing the control equation, and obtaining the state variables of the direct-current voltage synchronization control grid connection system of the grid-connected direct-drive wind turbine generator set based on the linearized small signal equation; S4, calculating the control parameters of the reactive voltage control loop and the direct-current voltage synchronization loop; and S5, adaptively setting the parameters of the existing permanent magnet wind turbine inertia synchronization strategy, so that the change of the strong and weak grid is adapted, and the stability performance of the PMSG under different grid short-circuit ratios is improved.

Description

Technical Field

[0001] This invention belongs to the field of new energy grid connection and relates to an adaptive control method for inertial synchronization PMSG based on online impedance estimation. Background Technology

[0002] With the increasing proportion of renewable energy in the global energy structure, new energy technologies, represented by wind power, are experiencing rapid development. Currently, most wind turbine converters adopt a grid-following control strategy based on phase-locked loops (PLLs) when connected to the grid. This strategy relies on precise tracking of the grid voltage phase to achieve synchronization and does not possess independent frequency and voltage support capabilities. When large-scale wind farms replace traditional synchronous generator units, the total rotational inertia of the power system decreases significantly, resulting in a "loss of inertia" problem. This weakens the system's ability to resist frequency disturbances, making it susceptible to frequency instability when grid disturbances occur, posing a threat to the safe and stable operation of the power system.

[0003] To address these challenges, grid-forming control strategies have emerged. By simulating the operating mechanism and dynamic characteristics of synchronous generators, grid-forming control enables converters to autonomously synchronize with the grid, presenting themselves as controllable voltage source characteristics. This control method not only provides inertia and damping support to the grid but also avoids oscillation and instability issues when interacting with weak grids. Therefore, it is considered a key technology supporting new power systems dominated by renewable energy sources.

[0004] In the application of Permanent Magnet Synchronous Generators (PMSGs), existing grid-based control strategies are mainly divided into two categories based on their core control objectives: active power control mode and DC voltage control mode. Among them, grid-based PMSGs using active power control mode experience drastic changes in output power when facing high load fluctuations. This can easily lead to abnormal interactions between the generator-side converter and the grid-side converter, resulting in system oscillations or even instability. Therefore, its adaptability under complex operating conditions is poor.

[0005] In contrast, the grid-type PMSG employing DC voltage control mode exhibits superior performance. This mode indirectly controls power balance by stabilizing the DC bus voltage, eliminating the risk of DC-side interactive instability. It not only maintains stable operation under weak grid conditions but also possesses the ability to autonomously establish a grid in islanded mode (i.e., "networking") and withstand high loads and wind power fluctuations, making it more widely applicable.

[0006] However, despite the obvious advantages of DC voltage control mode, further improving the grid connection performance and support capability of PMSG for power grids of different strengths is a technical problem that urgently needs to be solved. Summary of the Invention

[0007] In view of this, the purpose of this invention is to provide an adaptive control method for inertial synchronization PMSG based on online impedance estimation. This method uses an RLS-GA hybrid strategy algorithm to quickly and accurately identify the grid impedance and determine the strength type of the grid. Based on this, adaptive parameter tuning is performed on the existing permanent magnet wind turbine inertial synchronization strategy to adapt to changes in grid strength and improve the stability of the PMSG under different grid short-circuit ratios.

[0008] To achieve the above objectives, the present invention provides the following technical solution:

[0009] An adaptive control method for an inertial synchronization PMSG based on online impedance estimation specifically includes the following steps:

[0010] S1: Construct a power grid impedance identification model based on the RLS-GA hybrid strategy algorithm. Specifically, the initial values ​​of impedance parameters are estimated first by recursive least squares method, which greatly reduces the search range. Then, the crossover rate and mutation rate are dynamically adjusted by adaptive genetic operator to accelerate the convergence speed and estimate the power grid impedance online. Here, RLS represents recursive least squares method and GA represents genetic algorithm.

[0011] S2: Establish a control strategy for the grid-connected system of direct-drive wind turbines. On the turbine side, a power control loop, a current control loop, and an inertia transfer control loop are adopted; on the grid side, a DC voltage synchronization loop, a reactive voltage control loop, and a stabilization control loop are adopted.

[0012] S3: Linearize the equations of the control strategy in step S2. Based on the linearized small-signal equations, obtain the state variables of the grid-connected DC voltage synchronous control system of the grid-connected direct-drive wind turbine.

[0013] S4: Calculate the control parameter K of the reactive voltage control loop. s Control parameters K of the DC voltage synchronization loop pq and K iq ;

[0014] S5: Online Adaptive Coordinated Control: Under grid-connected operation, the RLS-GA hybrid strategy algorithm in step S1 is immediately activated, and the impedance component R is initialized through RLS. g and X g The estimated value is used to calculate the power term of the small-signal linearization; then the control parameters K of the grid-side converter (MSC) are calculated respectively. s K pq and K iqBy continuously reading and monitoring the DC bus voltage, when the voltage changes and exceeds a certain threshold, the system automatically generates a signal that triggers the RLS-GA hybrid strategy algorithm to re-estimate the grid impedance parameters. Finally, K is recalculated using the updated impedance value. s K pq and K iq This allows it to be compatible with different operating conditions of both strong and weak power grids.

[0015] Furthermore, in step S1, the grid impedance characteristics are represented by the voltage-current relationship at the point of common coupling (PCC), as shown in the following equation:

[0016]

[0017] Among them, v PCC (t) represents the measured voltage value at point PCC at time t; e g (t) represents the grid voltage; i g (t) represents the instantaneous value of the grid-connected current; R g L g The equivalent resistance and inductance of the power grid to be estimated;

[0018] The RLS parameter update formula is shown below:

[0019]

[0020] Where y = v PCC -e g Regression vector Parameter estimation vector Forgetting factor λ∈(0,1]; P is the error covariance matrix; K(k) is the gain vector, e(k) is the prediction error, and k is the number of update iterations;

[0021] Based on the RLS estimation results, the search range of the genetic algorithm is compressed:

[0022]

[0023] in, The equivalent resistance of the power grid is estimated based on RLS. The equivalent inductance of the power grid is estimated based on RLS.

[0024] The initial population generated based on the RLS estimation results is as follows:

[0025] P0={θ i |θ i =[R g,i ,L g,i ] T ,i=1,...,N} (4)

[0026] Where P0 is the initial population, θ i Let N be the i-th individual in the population, and N be the total number of individuals in the initial population.

[0027] The fitness function is:

[0028]

[0029] in, To estimate the voltage at the grid connection point, M is the number of sampling points in the current data window;

[0030] Selection operators are used to perform a survival-of-the-fittest operation on individuals in a population:

[0031]

[0032] Among them, P select (x i Let x be the probability that an individual is selected. i The fitness of an individual, i.e., x i =F(θ) i );

[0033] An adaptive crossover probability mechanism is adopted, which dynamically adjusts the crossover probability P based on the relative relationship between the maximum fitness f′ and the average fitness f of the current population. c Then, the crossover operation is performed as follows:

[0034] L child =βX1+(1-β)X2, β∈[0,1] (7)

[0035] Among them, L child X1 and X2 are the offspring individuals, β is the weighting coefficient, and X1 and X2 are the parent individuals.

[0036] A time-decreasing adaptive mutation strategy is adopted, where the mutation probability changes linearly with the number of algebras to balance global exploration and local convergence; the mutation amplitude is decayed according to the number of algebras.

[0037]

[0038] Where, δ i (g) represents the variable asynchronous length of the i-th individual in the g-th generation, η i Here, g is the variation intensity coefficient, g is the current iteration number, and G is the maximum iteration number. These are the upper and lower boundaries of the i-th individual, respectively.

[0039] Furthermore, in step S2, the machine side employs a power control loop, a current control loop, and an inertia transfer control loop, specifically including: the q-axis current reference value i mqref The calculation formula is:

[0040]

[0041] Where, k p1 k is the proportional coefficient of the power control loop. i1 P is the integral coefficient of the power control loop. mref P is the power reference value. m P is the output power of the direct drive motor. m * y1 is the power transferred by inertia, y1 is the output of the integral regulator of the active power control loop of the machine-side converter (MSC), and s is the complex frequency.

[0042] Inertia transfer power P m * The calculation formula is:

[0043]

[0044] Among them, H m (s) is the inertia controller, K m T is the inertia transfer factor. m u is the filter time constant. dc This is the per-unit value of DC voltage;

[0045] The current control loop is based on a PI current controller to control the output current i of the direct-drive fan in the current control loop. m The tracking is performed as shown in equation (11):

[0046]

[0047] Where, k p2 k p3 k is the proportional coefficient of the current control loop. i2 k i3 i represents the integral coefficient of the current control loop. mqref i is the reference value for the q-axis current. md i mq For the direct-drive fan, output dq axis current, L d L q For the dq axis inductance of the permanent magnet motor, ω m y1 represents the electric angular velocity of the permanent magnet synchronous motor, ψ represents the flux linkage of the permanent magnet synchronous motor, y2 represents the output of the q-axis integral regulator of the current control loop of the machine-side converter, and y3 represents the output of the d-axis integral regulator of the current control loop of the machine-side converter.

[0048] Furthermore, in step S2, the grid side employs a DC voltage synchronization loop, a reactive voltage control loop, and a stabilization control loop, specifically including: considering that the grid voltage operates at its rated angular frequency ω g Rotation, therefore the grid voltage angle θ gWith the grid-side converter (GSC) output electromotive force angle θ gsc The phase difference between them can be defined as:

[0049]

[0050] Where δ is the phase angle difference between the converter output voltage and the grid voltage, ω g The rated angular frequency, ω gsc The net-side angular velocity;

[0051] The DC voltage synchronization loop is shown in equation (13):

[0052]

[0053] Among them, U dc U is the instantaneous value of the DC voltage. dcref This is the steady-state value of the DC voltage;

[0054] The reactive voltage control loop consists of a PI controller, as shown in equation (14):

[0055]

[0056] Among them, E ref For the rated voltage, k pq and k iq For the proportional and integral coefficients of the reactive voltage control loop; Q ref The reactive power reference value is 0; Q g U represents the actual value of reactive power. s For stabilizing voltage; K s T is the stability control coefficient; g U is the time constant of the network-side filter. dc is the per-unit value of DC voltage, and s is the complex frequency.

[0057] Furthermore, in step S2, the Thevenin equivalent model of the main circuit of the direct-drive wind turbine grid-connected system is as follows:

[0058]

[0059] Among them, E g =[E gd E gq ] T The dq-axis voltage of the power grid; u pcc =[u pccd u pccq ] T The dq-axis voltage at point pcc; i s =[i sd i sq ] T For the grid-side dq-axis current; and R g For the equivalent inductance and resistance on the grid side; ω g The rated angular frequency;

[0060] The dynamic equation for the capacitor voltage is shown in equation (16):

[0061]

[0062] Among them, C dc For capacitor, U dcref For steady-state DC voltage, U dc P is the instantaneous value of the DC voltage. g To output active power to the grid-side converter;

[0063] The active power on the machine side of the grid-connected system is the output power P of the direct drive motor. m for:

[0064]

[0065] Among them, u md u mq For the machine-side output dq axis voltage, i md i mq The machine-side output dq axis current;

[0066] The active and reactive power outputs at the grid connection point are as follows:

[0067]

[0068] Among them, P pcc and Q pcc For the grid connection point to output active and reactive power, V i and V j These are the effective voltage values ​​at the system grid connection point and on the grid side, respectively. X g δ represents the total impedance between the grid-side converter and the power grid, and δ is the phase angle difference between the converter output voltage and the power grid voltage.

[0069] Furthermore, in step S3, the state variable Δx of the grid-connected system for DC voltage synchronous control of the grid-connected direct-drive wind turbine generator is obtained as follows:

[0070] Δx=[ΔU dc ΔδΔi md Δi mq Δy1Δy2Δy3Δx m Δx g Δi sd Δi sq Δy4] T (19)

[0071] Wherein, ΔUdc Δδ represents the change in DC-side (grid-side) voltage; Δδ represents the phase change in the output voltage of the grid-side converter (GSC); Δi md The change in the d-axis current output by the machine-side converter (MSC); Δi mq Δy1 represents the change in the q-axis current output of the machine-side converter (MSC); Δy2 represents the change in the output of the q-axis integral regulator of the active power control loop of the machine-side converter (MSC); Δy3 represents the change in the output of the d-axis integral regulator of the current control loop of the machine-side converter (MSC); Δx represents the change in the output of the q-axis integral regulator of the current control loop of the machine-side converter (MSC); m The change in the output of the first-order low-pass filter of the inertia transfer control loop of the machine-side converter (MSC); Δx g Δi represents the change in the output of the first-order low-pass filter in the stabilization control loop of the grid-side converter (GSC). sd Δi represents the change in the d-axis component of the output current of the grid-side converter (GSC). sq Δy4 represents the change in the q-axis component of the output current of the grid-side converter (GSC); Δy4 represents the change in the output of the reactive power loop integral regulator of the grid-side converter (GSC).

[0072] Furthermore, in step S4, the control parameters K of the reactive voltage control loop are calculated. s Specifically, this includes: the response performance of DC voltage to grid frequency dynamic processes is mainly determined by the damping ratio of the second-order function. The optimal stability coefficient K is then calculated so that the damping ratio of the second-order response function of DC voltage change to grid frequency disturbance is equal to 0.707. s As in equation (20):

[0073]

[0074] Where ζ is the damping ratio, C dc For capacitor, U dc0 k is the steady-state value of the DC voltage. pdc For inertial gain; Among them, A 11 V is the correlation coefficient between the active power at the grid connection point and the phase angle difference under small disturbances. i0 V is the effective value of the system grid connection point voltage during steady-state operation. j0 The effective value of the grid voltage during steady-state operation is given by δ0, where δ0 is the steady-state value of the phase angle difference between the converter output voltage and the grid voltage. 12 The correlation coefficient between active power and voltage at the grid connection point under small disturbances;

[0075] The maximum value of the stability coefficient K smax As in equation (21):

[0076]

[0077] Among them, U i This represents the output voltage amplitude on the AC side of the receiving-end converter station.

[0078] Furthermore, in step S4, the control parameter K of the DC voltage synchronization loop is calculated. pq and K iq The expression is:

[0079]

[0080] Where τ is the time constant.

[0081] The beneficial effects of this invention are as follows: Based on PMSG inertial synchronous control, this invention adaptively adjusts control parameters under different short-circuit ratio conditions, thereby achieving dynamic stability of the permanent magnet wind turbine generator under varying short-circuit ratios. The main advantages are reflected in the following aspects:

[0082] (1) Significantly improved the stability and adaptability of the system under different power grid intensities.

[0083] Traditional control strategies rely on fixed parameters that are difficult to adapt to changes in grid conditions. This invention, by estimating grid impedance online in real time, accurately determines the grid strength (i.e., different short-circuit ratios) and adaptively adjusts key control parameters accordingly. Simulation results show that when the grid is weak or undergoes drastic changes, traditional strategies may experience oscillations or even instability, while the method of this invention enables permanent magnet synchronous wind turbines (PMSGs) to maintain stable operation, effectively adapting to both strong and weak grid conditions and significantly enhancing the unit's grid connection adaptability.

[0084] (2) It enables rapid and accurate online identification of grid impedance.

[0085] This invention innovatively employs a hybrid strategy combining Recursive Least Squares (RLS) and Genetic Algorithm (GA). The RLS algorithm can quickly provide a relatively accurate initial impedance estimate, greatly reducing the search space for subsequent optimization; while the improved Genetic Algorithm, through adaptive crossover and mutation rates, accelerates the convergence speed and achieves precise optimization of impedance parameters. This "coarse-tuning + fine-tuning" approach ensures the speed and accuracy of impedance estimation, which is a key prerequisite for achieving adaptive adjustment of control parameters.

[0086] (3) The dynamic response performance of the system has been optimized.

[0087] This invention, based on accurate impedance estimates and a small-signal model of the system, addresses the stability coefficient (K). s ) and reactive voltage control loop PI parameters (K pq K iqOnline adaptive tuning is performed. This enables the system to not only maintain stability when subjected to disturbances (such as changes in grid frequency) but also exhibit superior dynamic performance. Simulation results show that, compared with the traditional fixed-parameter strategy, the system using the method of this invention has a faster response speed, a smaller DC voltage drop depth, and a shorter recovery time, thereby improving power quality and system transient stability.

[0088] (4) This invention provides a clear theoretical basis for control parameter tuning. By establishing a full-order small-signal model that includes grid impedance, this invention profoundly reveals the mathematical relationship between control parameters (especially Ks) and system damping characteristics, and provides an analytical calculation method for Ks based on the optimal damping ratio (0.707). This provides clear theoretical guidance for controller parameter design, avoids the tedious trial-and-error process in traditional methods, and improves the scientificity and reliability of the design.

[0089] Other advantages, objectives, and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination, or may be learned from practice of the invention. The objectives and other advantages of the invention can be realized and obtained through the following description. Attached Figure Description

[0090] To make the objectives, technical solutions, and advantages of the present invention clearer, the preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings, wherein:

[0091] Figure 1 This is a schematic diagram of the control structure of a direct-drive wind turbine grid-connected system.

[0092] Figure 2 This is the flowchart for adaptive control.

[0093] Figure 3 The simulation results are for the adaptive group and the control group. Detailed Implementation

[0094] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0095] Please see Figure 1This invention provides an adaptive control method for an inertial synchronization PMSG based on online impedance estimation, comprising the following steps:

[0096] S1: Construct a power grid impedance identification model based on the RLS-GA hybrid strategy algorithm. First, the initial values ​​of impedance parameters are estimated using the recursive least squares (RLS) method, significantly reducing the search range. Then, an adaptive genetic operator is used to dynamically adjust the crossover and mutation rates to accelerate the convergence speed, enabling online estimation of power grid impedance.

[0097] The impedance characteristics of the power grid are represented by the voltage-current relationship at the point of common coupling (PCC), as shown in equation (1):

[0098]

[0099] Among them, v PCC (t) represents the measured voltage value at point PCC at time t; e g (t) represents the grid voltage; i g (t) represents the instantaneous value of the grid-connected current; R g L g The equivalent resistance and inductance of the power grid are to be estimated.

[0100] The RLS parameter update formula is shown in equation (2):

[0101]

[0102] Where y = v PCC -e g Regression vector Parameter estimation vector Forgetting factor λ∈(0,1]; P is the error covariance matrix; K(k) is the gain vector; e(k) is the prediction error; and k is the number of update iterations.

[0103] Based on the RLS estimation results, the search range of the genetic algorithm is compressed:

[0104]

[0105] in, The equivalent resistance of the power grid is estimated based on RLS. The equivalent resistance of the power grid is estimated based on RLS.

[0106] The initial population generated based on the RLS estimation results is as follows:

[0107] P0={θ i |θ i =[R g,i ,L g,i ] T ,i=1,...,N} (4)

[0108] Where P0 is the initial population, θ i Let N be the i-th individual in the population, and N be the total number of individuals in the initial population.

[0109] The fitness function is:

[0110]

[0111] in, The voltage at the grid connection point is estimated, where M is the number of sampling points in the current data window.

[0112] Selection operators are used to perform a survival-of-the-fittest operation on individuals in a population:

[0113]

[0114] Among them, P select (x i Let x be the probability that an individual is selected. i The fitness of an individual, i.e., x i =F(θ) i ).

[0115] An adaptive crossover probability mechanism is adopted, which dynamically adjusts the crossover probability P based on the relative relationship between the maximum fitness f′ and the average fitness f of the current population. c Then, the crossover operation is performed as follows:

[0116] L child =βX1+(1-β)X2, β∈[0,1] (7)

[0117] Among them, L child X1 and X2 are the offspring individuals, β is the weighting coefficient, and X1 and X2 are the parent individuals.

[0118] A time-decreasing adaptive mutation strategy is adopted, where the mutation probability changes linearly with the number of algebras to balance global exploration and local convergence; the mutation amplitude is decayed according to the number of algebras.

[0119]

[0120] Where, δ i (g) represents the variable asynchronous length of the i-th individual in the g-th generation, η i Here, g is the variation intensity coefficient, g is the current iteration number, and G is the maximum iteration number. These are the upper and lower boundaries of the i-th individual, respectively.

[0121] S2: Establish a control strategy for the direct-drive wind turbine grid-connected system, such as... Figure 1 As shown, where P mref P is the power reference value. mP is the output power of the direct drive motor. m * For power transferred by inertia, u dc H is the per-unit value of DC voltage. m (s) is the inertia controller, K m Let be the inertia transfer coefficient. md i mq For the direct-drive fan, output dq axis current, L d L q ω is the dq-axis inductance of the permanent magnet motor. gsc and θ gsc These are the net-side angular velocity and angle, U ref The virtual electromotive force, U, provided for reactive voltage control dc U is the instantaneous value of the DC voltage. dcref E is the steady-state value of the DC voltage. ref Rated voltage; Q ref The reactive power reference value is 0; Q g This represents the actual value of reactive power; H g (s) is a stabilizing controller.

[0122] The machine side employs a power control loop, a current control loop, and an inertia transfer control loop. The q-axis current reference value is i. mqref As shown in equation (9):

[0123]

[0124] Where, k p1 k is the proportional coefficient of the power control loop. i1 P is the integral coefficient of the power control loop. mref P is the power reference value. m P is the output power of the direct drive motor. m * y1 is the power transferred by inertia, y1 is the output of the integral regulator of the active power control loop of the machine-side converter (MSC), and s is the complex frequency.

[0125] Inertia transfer power P m * As shown in equation (10):

[0126]

[0127] Among them, H m (s) is the inertia controller, K m T is the inertia transfer factor. m is the filter time constant.

[0128] The current control loop is based on a PI current controller to control the output current i of the direct-drive fan in the current control loop. mThe tracking is performed as shown in equation (11):

[0129]

[0130] Where, k p2 k p3 k is the proportional coefficient of the current control loop. i2 k i3 i represents the integral coefficient of the current control loop. mqref i is the reference value for the q-axis current. md i mq For the direct-drive fan, output dq axis current, L d L q For the dq axis inductance of the permanent magnet motor, ω m y1 represents the electric angular velocity of the permanent magnet synchronous motor, ψ represents the flux linkage of the permanent magnet synchronous motor, y2 represents the output of the q-axis integral regulator of the current control loop of the machine-side converter, and y3 represents the output of the d-axis integral regulator of the current control loop of the machine-side converter.

[0131] The grid side employs a DC voltage synchronization loop, a reactive voltage control loop, and a stabilization control loop. This is considering that the grid voltage operates at its rated angular frequency ω. g Rotation, therefore the grid voltage angle θ g With the grid-side converter (GSC) output electromotive force angle θ gsc The phase difference between them can be defined as:

[0132]

[0133] The DC voltage synchronization loop is shown in equation (13):

[0134]

[0135] Among them, U dc U is the instantaneous value of the DC voltage. dcref This is the steady-state value of the DC voltage.

[0136] The reactive voltage control loop consists of a PI controller, as shown in equation (14):

[0137]

[0138] Among them, E ref For the rated voltage, k pq and k iq For the proportional and integral coefficients of the reactive voltage control loop; Q ref The reactive power reference value is 0; Q g U represents the actual value of reactive power. s For stabilizing voltage; K s T is the stability control coefficient; g U is the time constant of the network-side filter.dc is the per-unit value of DC voltage, and s is the complex frequency.

[0139] The equivalent model of the main circuit Thevenin is:

[0140]

[0141] Among them, E g =[E gd E gq ] T The dq-axis voltage of the power grid; u pcc =[u pccd u pccq ] T The dq-axis voltage at point pcc; i s =[i sd i sq ] T For the grid-side dq-axis current; and R g For the equivalent inductance and resistance on the grid side; ω g This is the rated angular frequency.

[0142] The dynamic equation for the capacitor voltage is shown in equation (16):

[0143]

[0144] Among them, C dc For capacitor, U dcref For steady-state DC voltage, U dc P is the instantaneous value of the DC voltage. g It outputs active power to the grid-side converter.

[0145] The active power on the machine side of the grid-connected system is the output power P of the direct drive motor. m for:

[0146]

[0147] Among them, u md u mq For the machine-side output dq axis voltage, i md i mq This is the output dq-axis current on the machine side.

[0148] The active and reactive power outputs at the grid connection point are as follows:

[0149]

[0150] Among them, P pcc and Q pcc For the grid connection point to output active and reactive power, V i and V jThese are the effective voltage values ​​at the system grid connection point and on the grid side, respectively. X g δ represents the total impedance between the grid-side converter and the power grid, and δ is the phase angle difference between the converter output voltage and the power grid voltage.

[0151] S3: Linearize the control equations in step S2. Based on the linearized small-signal equations, obtain the state variable Δx of the grid-connected system for DC voltage synchronous control of grid-connected direct-drive wind turbine generators.

[0152] Δx=[ΔU dc ΔδΔi md Δi mq Δy1Δy2Δy3Δx m Δx g Δi sd Δi sq Δy4] T (19)

[0153] Where, ΔU dc Δδ represents the change in DC-side (grid-side) voltage; Δδ represents the phase change in the output voltage of the grid-side converter (GSC); Δi md The change in the d-axis current output by the machine-side converter (MSC); Δi mq Δy1 represents the change in the q-axis current output of the machine-side converter (MSC); Δy2 represents the change in the output of the q-axis integral regulator of the active power control loop of the machine-side converter (MSC); Δy3 represents the change in the output of the d-axis integral regulator of the current control loop of the machine-side converter (MSC); Δx represents the change in the output of the q-axis integral regulator of the current control loop of the machine-side converter (MSC); m The change in the output of the first-order low-pass filter of the inertia transfer control loop of the machine-side converter (MSC); Δx g Δi represents the change in the output of the first-order low-pass filter in the stabilization control loop of the grid-side converter (GSC). sd Δi represents the change in the d-axis component of the output current of the grid-side converter (GSC). sq Δy4 represents the change in the q-axis component of the output current of the grid-side converter (GSC); Δy4 represents the change in the output of the reactive power loop integral regulator of the grid-side converter (GSC).

[0154] S4: The small-signal expression of active power in equation (18) is shown in equation (20); the small-signal expression of equation (16) is shown in equation (21).

[0155]

[0156] in, Subscript 0 represents the steady-state value; A 11 V is the correlation coefficient between the active power at the grid connection point and the phase angle difference under small disturbances. i0V is the effective value of the system grid connection point voltage during steady-state operation. j0 The effective value of the grid voltage during steady-state operation is given by δ0, where δ0 is the steady-state value of the phase angle difference between the converter output voltage and the grid voltage. 12 ΔP is the correlation coefficient between active power and voltage at the grid connection point under small disturbances. pcc ΔU represents the active power at the grid connection point under small disturbances. dc This is the DC voltage under small disturbances.

[0157] Combining equations (20) and (21), with the generator-side input power remaining constant, the transfer function between the dynamic change of DC voltage and the change of grid frequency is shown in equation (22). The response performance of the dynamic process of DC voltage and grid frequency is mainly determined by the damping ratio of the second-order function. To make the damping ratio of the second-order response function of DC voltage change to grid frequency disturbance equal to 0.707, the optimal stability coefficient K is calculated. s As shown in equation (23). The maximum value of the stabilization coefficient K smax As shown in equation (24).

[0158]

[0159] S5: The relationship between the reactive power after small-signal linearization of equation (18) and the DC voltage control loop and reactive power reference value is shown in equation (25). If the influence of the DC voltage loop is ignored, it can be written as equation (26).

[0160] ΔQ pcc =H 21 Δu dc +H 22 ΔQ ref (25)

[0161] ΔQ pcc =H 22 ΔQ ref (26)

[0162] Among them, H 21 This represents the correlation coefficient between reactive power and DC voltage at the grid connection point under small disturbances. A 22 This represents the correlation coefficient between reactive power and voltage at the grid connection point under small disturbances. To eliminate steady-state errors and ensure the reactive power loop gain is 1, K is chosen. pq for To achieve a faster reactive power response and avoid the influence of high-frequency noise, K iq for: Where τ is the time constant.

[0163] S6: Online adaptive coordinated control, such as Figure 2As shown. Under grid-connected operation, the RLS-GA hybrid strategy algorithm in step S1 is immediately activated, initializing the impedance component R. g and X g The estimated values ​​are then used to calculate the power term for small-signal linearization. The control parameters K of the grid-side converter (MSC) are then calculated. s K pq and K iq By continuously reading and monitoring the DC bus voltage, when the signal changes and exceeds a certain threshold, the system automatically generates a signal that triggers the RLS-GA hybrid strategy algorithm to re-estimate the grid impedance parameters. Finally, K is recalculated using the updated impedance values. s K pq and K iq This allows it to be compatible with different operating conditions of both strong and weak power grids.

[0164] Verification experiment:

[0165] In the simulation model, the system outputs 4MW of active power and a reactive power reference value of 0. Online adaptive coordinated control is initiated at t=0.5s, and the grid frequency drops by 0.1Hz at t=1s. Figure 3 As shown, adaptive K-axis was observed under different SCR values ​​(1.5, 3, 8, 15). s and K pq and K iq The DC voltage waveform is shown. Using the adaptive group and the traditional strategy as a control group, the adaptive group reaches stability faster and has a smaller voltage drop depth. When the SCR exceeds 8, the traditional strategy exhibits instability and oscillation, while the adaptive group maintains stability. Therefore, the impedance estimation method of this invention demonstrates better applicability and more accurate parameter tuning compared to traditional methods in maintaining overall system stability.

[0166] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. An adaptive control method for an inertial synchronization PMSG based on online impedance estimation, characterized in that, The method specifically includes the following steps: S1: Construct a power grid impedance identification model based on the RLS-GA hybrid strategy algorithm. Specifically, the initial values ​​of impedance parameters are estimated first by recursive least squares method; then, the crossover rate and mutation rate are dynamically adjusted by adaptive genetic operator to estimate the power grid impedance online. Here, RLS represents recursive least squares method and GA represents genetic algorithm. S2: Establish a control strategy for the grid-connected system of direct-drive wind turbines. On the turbine side, a power control loop, a current control loop, and an inertia transfer control loop are adopted; on the grid side, a DC voltage synchronization loop, a reactive voltage control loop, and a stabilization control loop are adopted. S3: Linearize the equations of the control strategy in step S2. Based on the linearized small-signal equations, obtain the state variables of the grid-connected DC voltage synchronous control system of the grid-connected direct-drive wind turbine. S4: Calculate the control parameters of the reactive power control loop. K s Control parameters of DC voltage synchronization loop K pq and K iq ; S5: Online Adaptive Coordination Control: Under grid-connected operation, the RLS-GA hybrid strategy algorithm in step S1 is immediately activated, and the impedance component is initialized through RLS. R g and X g The estimated values ​​are used to calculate the power term of the small-signal linearization; then the control parameters of the grid-side converter are calculated respectively. K s , K pq and K iq By continuously reading and monitoring the DC bus voltage, when the voltage changes and exceeds a threshold, the system automatically generates a signal that triggers the RLS-GA hybrid strategy algorithm to re-estimate the grid impedance parameters. Finally, the updated impedance values ​​are used to recalculate the impedance parameters. K s , K pq and K iq This ensures compatibility with different operating conditions of both strong and weak power grids; In step S1, the grid impedance characteristics are expressed by the voltage-current relationship at the point of common coupling (PCC), as shown in the following equation: （1） in, v PCC (t) is t The measured voltage value at point PCC at time 10:00; e g ( t () represents the grid voltage; i g ( t () represents the instantaneous value of the grid-connected current; R g , L g The equivalent resistance and inductance of the power grid to be estimated; The RLS parameter update formula is shown below: （2） in, y = v PCC e g Regression vector ; Parameter estimation vector Forgetting factor λ ∈(0,1]; P Here is the error covariance matrix; For the gain vector, For prediction error, k To update the number of iterations; Based on the RLS estimation results, the search range of the genetic algorithm is compressed: （3） in, The equivalent resistance of the power grid is estimated based on RLS. The equivalent inductance of the power grid is estimated based on RLS. The initial population generated based on the RLS estimation results is as follows: （4） in, For the initial population, The first in the population i Individual, N This represents the total number of individuals in the initial population. The fitness function is: （5） in, Estimate the voltage at the grid connection point. M This represents the number of sampling points in the current data window. Selection operators are used to perform a survival-of-the-fittest operation on individuals in a population: （6） in, The probability of an individual being selected. x i For the fitness of an individual, i.e. x i = F ( ); An adaptive crossover probability mechanism is adopted, that is, based on the maximum fitness of the current population. ′ and average fitness The relative relationship dynamically adjusts the crossover probability P c Then, the crossover operation is performed as follows: （7） in, For offspring individuals, These are the weighting coefficients. X 1 and X 2 represents the parent generation individual; A time-decreasing adaptive mutation strategy is adopted, where the mutation probability changes linearly with the number of generations; the mutation amplitude is decayed according to the number of generations. （8） in, For the first g The middle generation i The variable and asynchronous growth of each individual, The coefficient of variation intensity g This represents the current iteration number. G The maximum number of iterations, , The first i The upper and lower boundaries of each individual.

2. The adaptive control method for inertial synchronization PMSG based on online impedance estimation according to claim 1, characterized in that, In step S2, the machine side employs a power control loop, a current control loop, and an inertia transfer control loop, specifically including: q-axis current reference value. i mqref The calculation formula is: （9） in, k p1 This is the proportional coefficient of the power control loop. k i1 The integral coefficient of the power control loop. P mref This is a power reference value. P m This refers to the output power of the direct drive motor. P m * Power is transferred by inertia. y 1 represents the output of the integral regulator in the active power control loop of the generator-side converter. s It is a complex frequency; Inertia transfer power P m * The calculation formula is: （10） in, H m (s) is an inertia controller. K m The inertia transfer factor, T m The filter time constant is This is the per-unit value of DC voltage; The current control loop uses a PI current controller to control the output current of the direct-drive fan in the current control loop. i m The tracking is performed as shown in equation (11): （11） in, k p2 , k p3 This is the proportional coefficient of the current control loop. k i2 , k i3 The integral coefficient of the current control loop. This is the reference value for the q-axis current. i md , i mq Output dq axis current for direct drive fan. L d , L q For the dq axis inductance of a permanent magnet motor, ω m The electric angular velocity of the permanent magnet synchronous motor. ψ For permanent magnet synchronous motor flux linkage; y 2 represents the output of the q-axis integral regulator of the current control loop of the machine-side converter; y 3 represents the output of the d-axis integral regulator of the current control loop of the machine-side converter.

3. The adaptive control method for inertial synchronization PMSG based on online impedance estimation according to claim 2, characterized in that, In step S2, the grid side employs a DC voltage synchronization loop, a reactive voltage control loop, and a stabilization control loop, specifically including: considering that the grid voltage operates at its rated angular frequency... ω g Rotation, therefore the grid voltage angle θ g Angle with the output electromotive force of the grid-side converter θ gsc The phase difference between them can be defined as: （12） in, The phase angle difference between the converter output voltage and the grid voltage. ω g The rated angular frequency, ω gsc The net-side angular velocity; The DC voltage synchronization loop is shown in equation (13): （13） in, U dc This is the instantaneous value of the DC voltage. U dcref This is the steady-state value of the DC voltage; The reactive voltage control loop consists of a PI controller, as shown in equation (14): （14） in, E ref Rated voltage, k pq and k iq These are the proportional and integral coefficients of the reactive voltage control loop; Q ref The reactive power reference value is 0. Q g This represents the actual value of reactive power. U s For stabilizing voltage; K s This is the stability control coefficient; T g The time constant of the network-side filter. u dc This is the per-unit value of DC voltage. s It is a complex frequency.

4. The adaptive control method for inertial synchronization PMSG based on online impedance estimation according to claim 3, characterized in that, In step S2, the Thevenin equivalent model of the main circuit of the direct-drive wind turbine grid-connected system is as follows: （15） in, The dq-axis voltage of the power grid; The voltage across the dq axis at point pcc; For the grid-side dq-axis current; and R g For the equivalent inductance and resistance on the grid side; ω g The rated angular frequency; The dynamic equation for the capacitor voltage is shown in equation (16): （16） in, C dc For capacitors, U dcref The steady-state DC voltage, U dc This is the instantaneous value of the DC voltage. To output active power to the grid-side converter; The active power on the machine side of the grid-connected system is the output power of the direct drive motor. P m for: （17） in, u md , u mq This is the output dq axis voltage from the machine side. i md , i mq The machine-side output dq-axis current; The active and reactive power outputs at the grid connection point are as follows: （18） in, and Output active and reactive power to the grid connection point. V i and V j These are the effective voltage values ​​at the system grid connection point and on the grid side, respectively. X g δ represents the total impedance between the grid-side converter and the power grid, and δ is the phase angle difference between the converter output voltage and the power grid voltage.

5. The adaptive control method for inertial synchronization PMSG based on online impedance estimation according to claim 4, characterized in that, In step S3, the state variables of the grid-connected system for DC voltage synchronous control of grid-connected direct-drive wind turbine generators are obtained. for: （19） Where, Δ U dc Δδ represents the change in DC-side voltage; Δδ represents the phase change in the grid-side converter output voltage; Δ i md The change in the d-axis current output by the machine-side converter; Δ i mq The change in the q-axis current output by the machine-side converter; Δ y 1 represents the change in the output of the integral regulator in the active power control loop of the generator-side converter; Δ y 2 represents the change in the output of the q-axis integral regulator of the current control loop of the machine-side converter; Δ y 3 represents the change in the output of the d-axis integral regulator of the current control loop of the machine-side converter; Δ x m The change in the output of the first-order low-pass filter of the inertia transfer control loop of the machine-side converter; Δ x g The change in the output of the first-order low-pass filter in the stabilization control loop of the grid-side converter; Δ i sd Δ represents the change in the d-axis component of the grid-side converter output current. i sq Δ represents the change in the q-axis component of the grid-side converter output current. y 4 represents the change in the output of the reactive power loop integral regulator of the grid-side converter.

6. The adaptive control method for inertial synchronization PMSG based on online impedance estimation according to claim 5, characterized in that, In step S4, the control parameters of the reactive voltage control loop are calculated. K s Specifically, this includes: the response performance of the DC voltage to the grid frequency dynamic process is determined by the damping ratio of the second-order function, so that the damping ratio of the second-order response function of the DC voltage change to the grid frequency disturbance is equal to 0.707, and then determining the optimal stability coefficient at this point. K s As in equation (20): （20） in, For the damping ratio, C dc For capacitors, This is the steady-state value of the DC voltage. For inertial gain; , The subscript 0 represents the steady-state value, where, This is the correlation coefficient between the active power at the grid connection point and the phase angle difference under small disturbances. This is the effective value of the system grid connection point voltage during steady-state operation. This represents the effective value of the grid-side voltage during steady-state operation. This represents the steady-state value of the phase angle difference between the converter output voltage and the grid voltage. The correlation coefficient between active power and voltage at the grid connection point under small disturbances; Maximum value of the stability coefficient K smax As shown in equation (21): （21） in, This represents the output voltage amplitude on the AC side of the receiving-end converter station.

7. The adaptive control method for inertial synchronization PMSG based on online impedance estimation according to claim 6, characterized in that, In step S4, the control parameters of the DC voltage synchronization loop are calculated. K pq and K iq The expression is: （22） （23） in, τ is the time constant.

Images