Wind turbine generator control parameter identification method and system based on subsynchronous oscillation characteristics

Abstract

The invention provides a wind turbine generator control parameter identification method and system based on subsynchronous oscillation characteristics, and the method comprises the steps: carrying out the phase margin sensitivity analysis based on the main oscillation mode of a wind turbine generator grid-connected system, and determining a to-be-identified control parameter; constructing a simulation model of the wind turbine generator, determining an initial value of a to-be-identified control parameter, and determining a to-be-identified control parameter-active output parameter group; performing analog simulation on the simulation model of the wind turbine generator, and determining oscillation characteristics; performing model training on a preset neural network model according to the to-be-identified control parameter-active output parameter group and the oscillation characteristics, and determining the neural network model after model training is completed; and inputting the actual oscillation characteristics and the actual active power output into the neural network model which completes model training, and determining an actual to-be-identified control parameter identification result. According to the method and the device, the key control parameter identification of the wind turbine generator is realized by adopting the finite oscillation data, the calculation amount is small, and the identification efficiency is high.

Description

Technical Field

[0001] The present disclosure relates to the field of renewable energy power generation technology, and in particular to a method and system for identifying control parameters of a wind turbine generator system based on subsynchronous oscillation characteristics. Background Art

[0002] Renewable energy sources such as wind power have developed rapidly in recent years and have become my country's second-largest power source. However, as the proportion of wind power connected to the grid continues to increase, wind power has a significant impact on the dynamic stability of the grid, with subsynchronous oscillations being a prominent issue.

[0003] Accurate modeling is crucial for studying wind power grid-connected stability. However, due to commercial confidentiality, actual wind turbines often exhibit "black-box" and "gray-box" problems. Therefore, identifying wind turbine control parameters is of great practical significance. Traditional control parameter identification methods include time-domain methods, which apply continuous excitation to obtain time-domain responses, and frequency-domain methods, which measure frequency-domain characteristics. After obtaining time-domain and frequency-domain models, both methods use optimization algorithms to fit the measured data to achieve parameter identification.

[0004] However, the above method is difficult to implement in actual manufacturer equipment and has low practicality. The oscillation characteristics are closely related to the control parameters. How to mine the control parameter information of wind turbines based on oscillation data and then realize control parameter identification has become the focus of wind power grid stability research. Summary of the Invention

[0005] In view of the defects in the prior art, the present invention aims to provide a method and system for identifying control parameters of a wind turbine generator system based on subsynchronous oscillation characteristics.

[0006] To achieve the above objectives, according to one aspect of the present disclosure, a method for identifying control parameters of a wind turbine generator system based on subsynchronous oscillation characteristics is provided, comprising:

[0007] Conduct phase margin sensitivity analysis based on the main oscillation mode of the wind turbine grid-connected system to determine the control parameters to be identified for the wind turbine;

[0008] According to the actual control structure and parameter information of the wind turbine generator set, a simulation model of the wind turbine generator set is constructed, an initial value of the control parameter to be identified is determined, and a control parameter to be identified-active power output parameter group is determined;

[0009] Simulating the simulation model of the wind turbine generator system according to the control parameter-active power output parameter group to be identified, and determining the oscillation characteristics corresponding to each control parameter-active power output parameter group to be identified;

[0010] Performing model training on a preset neural network model according to the control parameter to be identified-active output parameter group and the oscillation characteristics, and determining a neural network model that has completed model training;

[0011] The actual oscillation characteristics and the actual active power output are input into the neural network model that has completed the model training to determine the identification result of the actual control parameter to be identified.

[0012] Optionally, the method further includes:

[0013] The value range of the actual control parameter to be identified is determined according to the identification result of the actual control parameter to be identified and a preset confidence level.

[0014] Optionally, performing phase margin sensitivity analysis based on a main oscillation mode of a wind turbine grid-connected system to determine a control parameter to be identified for the wind turbine includes:

[0015] Dividing the wind turbine grid-connected system into a load subsystem and a source subsystem;

[0016] Establishing a positive-sequence impedance model of the load subsystem and a positive-sequence impedance model of the source subsystem;

[0017] Drawing a Bode diagram of the positive sequence impedance of the load subsystem and a Bode diagram of the positive sequence impedance of the source subsystem according to the positive sequence impedance model of the load subsystem and the positive sequence impedance model of the source subsystem;

[0018] determining a phase margin of the wind power grid-connected system under each control parameter according to a Bode diagram of the positive sequence impedance of the load subsystem and a Bode diagram of the positive sequence impedance of the source subsystem;

[0019] Determining a phase margin sensitivity ratio of each control parameter according to the phase margin of the wind power grid-connected system under each control parameter;

[0020] A preset number of control parameters to be identified of the wind turbine generator system are determined according to the descending order of the phase margin sensitivity ratio of each control parameter.

[0021] Optionally, determining the control parameter to be identified-active power output parameter group includes:

[0022] Within a preset control parameter range, taking the initial value of the control parameter to be identified as a reference and taking a preset first step length as an interval, determining multiple values ​​of the control parameter to be identified;

[0023] Determining multiple values ​​of the active power output within a preset active power output interval at intervals of a preset second step length;

[0024] The multiple values ​​of the control parameter to be identified and the multiple values ​​of the active output are combined to determine the control parameter to be identified-active output parameter group.

[0025] Optionally, the oscillation characteristics include oscillation frequency and damping ratio.

[0026] Optionally, simulating the simulation model of the wind turbine generator set according to the control parameter-active power output parameter group to be identified to determine the oscillation characteristics corresponding to each control parameter-active power output parameter group to be identified includes:

[0027] Using the simulation model of the wind turbine generator set to simulate each of the control parameter-active power output parameter groups to be identified, and determining the time domain data corresponding to each of the control parameter-active power output parameter groups to be identified;

[0028] Performing empirical mode decomposition on the time domain data corresponding to each of the control parameter-active power output parameter groups to be identified, and determining the intrinsic mode function corresponding to each of the control parameter-active power output parameter groups to be identified;

[0029] Performing a Hilbert transform on the intrinsic mode function corresponding to each of the control parameter-active power parameter groups to be identified to determine the instantaneous frequency, instantaneous amplitude, and phase information corresponding to each of the control parameter-active power parameter groups to be identified;

[0030] The instantaneous frequency, instantaneous amplitude and phase information corresponding to each of the control parameter to be identified-active power output parameter groups are fitted to determine the oscillation characteristics corresponding to each of the control parameter to be identified-active power output parameter groups.

[0031] Optionally, the performing model training on a preset neural network model according to the control parameter to be identified-active output parameter group and the oscillation characteristics, and determining the neural network model that has completed model training includes:

[0032] Constructing a data set of active output-oscillation characteristics-control parameters according to the control parameter-active power output parameter group to be identified and the oscillation characteristics;

[0033] Normalizing samples in the active output-oscillation characteristics-control parameter data set to determine a normalized data set;

[0034] Dividing the normalized data set according to a preset ratio to determine a training set and a test set;

[0035] Using the active output and oscillation characteristics in the training set as inputs of the preset neural network, and the control parameters in the training set as outputs of the preset neural network, performing model training on the preset neural network model, and determining a neural network model that has undergone model training;

[0036] The test set is used to test the neural network model that has undergone model training to determine the neural network model that has completed model training.

[0037] According to a second aspect of the present disclosure, a wind turbine control parameter identification system based on subsynchronous oscillation characteristics is provided, comprising:

[0038] Phase margin sensitivity analysis module, used to perform phase margin sensitivity analysis based on the main oscillation mode of the wind turbine grid-connected system and determine the control parameters to be identified for the wind turbine;

[0039] A simulation model and simulation parameter construction module, configured to construct a simulation model of the wind turbine generator set according to the actual control structure and parameter information of the wind turbine generator set, determine the initial values ​​of the control parameters to be identified, and determine the control parameter to be identified - active power output parameter group;

[0040] a simulation module, configured to simulate a simulation model of the wind turbine generator system according to the control parameter-active power output parameter group to be identified, and determine an oscillation characteristic corresponding to each of the control parameter-active power output parameter groups to be identified;

[0041] A model training module is used to perform model training on a preset neural network model according to the control parameter to be identified-active power output parameter group and the oscillation characteristics, and determine a neural network model that has completed model training;

[0042] The control parameter identification module is used to input the actual oscillation characteristics and the actual active power output into the neural network model that has completed model training, and determine the identification results of the actual control parameters to be identified.

[0043] According to a third aspect of the present disclosure, a non-transitory computer-readable storage medium is provided, on which a computer program is stored, and when the program is executed by a processor, the steps of the method provided in the first aspect of the present disclosure are implemented.

[0044] According to a fourth aspect of the present disclosure, there is provided an electronic device, including:

[0045] a memory having a computer program stored thereon;

[0046] A processor is used to execute the computer program in the memory to implement the steps of the method provided in the first aspect of the present disclosure.

[0047] Compared with the prior art, the embodiments of the present disclosure have at least one of the following beneficial effects:

[0048] Through the above technical solution, a phase margin sensitivity analysis is performed under the main oscillation mode of the wind turbine grid-connected system to determine the control parameters that have a greater impact on the main oscillation mode as the control parameters to be identified, which is helpful to analyze the control parameters that affect the stability of the wind power grid-connected system. Through the simulation model of the wind turbine, simulation is performed to determine the oscillation characteristics corresponding to the parameters to be identified and the active output. The preset neural network model is trained using the control parameter to be identified-active output parameter group and the oscillation characteristics to learn the relationship between the oscillation characteristics and the control parameters, and the ability of the neural network model to identify the control parameters to be identified based on limited oscillation characteristics is realized. The calculation amount is small and the identification efficiency is high. The control strategy of the wind turbine can be adjusted and optimized in a targeted manner through oscillation data to adapt to the dynamic changes of the power grid and improve the operating efficiency and reliability of the unit. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Other features, objects and advantages of the present disclosure will become more apparent upon reading the detailed description of non-limiting embodiments with reference to the following drawings:

[0050] Figure 1 The present invention is a flow chart showing a method for identifying control parameters of a wind turbine generator system based on subsynchronous oscillation characteristics according to an exemplary embodiment.

[0051] Figure 2 The figure is a schematic diagram of an equivalent circuit of a doubly-fed wind farm series-compensated grid-connected system according to an exemplary embodiment.

[0052] Figure 3 The figure is a schematic diagram of the topology and control structure of a single doubly-fed wind turbine generator set according to an exemplary embodiment.

[0053] Figure 4 The figure is a schematic diagram showing the phase margin sensitivity of a control parameter of a doubly-fed wind turbine generator system according to an exemplary embodiment.

[0054] Figure 5 The figure is a schematic diagram showing verification of the generalization ability of a preset neural network model according to an exemplary embodiment.

[0055] Figure 6 The present invention is a block diagram showing a wind turbine control parameter identification system based on subsynchronous oscillation characteristics according to an exemplary embodiment. DETAILED DESCRIPTION

[0056] The present disclosure is described in detail below with reference to specific embodiments. The following embodiments will help those skilled in the art further understand the present disclosure, but are not intended to limit the present disclosure in any way. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the scope of the present disclosure. These modifications and improvements are all within the scope of protection of the present disclosure.

[0057] Figure 1 The present invention is a flow chart showing a method for identifying control parameters of a wind turbine generator system based on subsynchronous oscillation characteristics according to an exemplary embodiment.

[0058] like Figure 1 As shown, the present disclosure provides a method for identifying control parameters of a wind turbine generator system based on subsynchronous oscillation characteristics, including S11 to S15.

[0059] S11, performing phase margin sensitivity analysis based on the main oscillation mode of the wind turbine grid-connected system to determine the control parameters to be identified of the wind turbine.

[0060] There are m control parameters of the wind turbine generator set to be identified.

[0061] The wind turbines in the wind turbine grid-connected system disclosed in this disclosure are "gray box" models, meaning the control structure is known but the control parameters are unknown. Specific examples of wind turbines include permanent magnet direct-drive wind turbines and doubly-fed wind turbines. This disclosure uses doubly-fed wind turbines as an example for illustration.

[0062] S12, constructing a simulation model of the wind turbine generator set according to the actual control structure and parameter information of the wind turbine generator set, determining the initial values ​​of the control parameters to be identified, and determining the control parameter-active power output parameter group to be identified.

[0063] The actual control structure and parameter information of the wind turbine generator set used to construct the simulation model of the wind turbine generator set are all known information.

[0064] Determine the control parameter to be identified-active power output parameter group as follows: take the initial value of the control parameter to be identified as a reference, adjust the value of the control parameter to be identified within a preset control parameter range, and adjust the value of the active output within a preset active power output range to determine the control parameter to be identified-active power output parameter group.

[0065] Each control parameter to be identified-active power output parameter group includes a value of the control parameter to be identified and a value of the active power output.

[0066] S13 , simulating the simulation model of the wind turbine generator system according to the control parameter-active power output parameter group to be identified, and determining the oscillation characteristics corresponding to each control parameter-active power output parameter group to be identified.

[0067] The oscillation characteristics include oscillation frequency and damping ratio.

[0068] S14, performing model training on a preset neural network model according to the control parameter to be identified - the active output parameter group and the oscillation characteristics, and determining a neural network model that has completed model training.

[0069] Among them, a data set of active output-oscillation characteristics-control parameters is constructed according to the control parameter-active output parameter group to be identified and the oscillation characteristics, so as to determine the mapping relationship between active output and oscillation characteristics and control parameters.

[0070] Each sample data in the active output-oscillation characteristics-control parameter data set includes active output, oscillation characteristics and corresponding control parameters. The active output-oscillation characteristics-control parameter data set is used to train a preset neural network.

[0071] S15, inputting the actual oscillation characteristics and the actual active power output into the neural network model that has completed model training, and determining the identification result of the actual control parameter to be identified.

[0072] Through the above technical solution, a phase margin sensitivity analysis is performed under the main oscillation mode of the wind turbine grid-connected system to determine the control parameters that have a greater impact on the main oscillation mode as the control parameters to be identified, which is helpful to analyze the control parameters that affect the stability of the wind power grid-connected system. Through the simulation model of the wind turbine, simulation is performed to determine the oscillation characteristics corresponding to the parameters to be identified and the active output. The preset neural network model is trained using the control parameter to be identified-active output parameter group and the oscillation characteristics to learn the relationship between the oscillation characteristics and the control parameters, and the ability of the neural network model to identify the control parameters to be identified based on limited oscillation characteristics is realized. The calculation amount is small and the identification efficiency is high. The control strategy of the wind turbine can be adjusted and optimized in a targeted manner through oscillation data to adapt to the dynamic changes of the power grid and improve the operating efficiency and reliability of the unit.

[0073] In a possible embodiment, S11 , performing phase margin sensitivity analysis based on the main oscillation mode of the wind turbine grid-connected system to determine the control parameters to be identified of the wind turbine, includes: S111 to S116 .

[0074] S111, divide the wind turbine grid-connected system into a load subsystem and a source subsystem.

[0075] The load subsystem represents the wind turbine side system, such as a permanent magnet direct-drive wind turbine and a doubly-fed wind turbine.

[0076] The source subsystem impedance module represents the power supply side system, such as the grid side and grid equivalent impedance, and grid side equipment.

[0077] S112, establishing a positive-sequence impedance model of the load subsystem and a positive-sequence impedance model of the source subsystem.

[0078] S113 , drawing a Bode diagram of the positive sequence impedance of the load subsystem and a Bode diagram of the positive sequence impedance of the source subsystem according to the positive sequence impedance model of the load subsystem and the positive sequence impedance model of the source subsystem.

[0079] The Bode diagram of the positive sequence impedance of the load subsystem and the Bode diagram of the positive sequence impedance of the atomic system are plotted in the same Bode diagram.

[0080] S114 , determining a phase margin of the wind power grid-connected system under each control parameter according to the Bode diagram of the positive sequence impedance of the load subsystem and the Bode diagram of the positive sequence impedance of the source subsystem.

[0081] Among them, the intersection frequency f of the wind power grid-connected system in the Bode diagram under each control parameter is os The phase margin corresponds to a phase margin, and the Bode diagrams of the positive sequence impedance of the load subsystem and the positive sequence impedance of the source subsystem are observed at the intersection frequency f. os The phase margin at the intersection frequency f is determined by the wind power grid-connected system under each control parameter. os The phase margin at .

[0082] S115 , determining a phase margin sensitivity ratio of each control parameter according to the phase margin of the wind power grid-connected system under each control parameter.

[0083] Among them, firstly, according to each control parameter, the wind power grid-connected system is at the intersection frequency f os The phase margin at , determines the phase margin sensitivity of the wind power grid-connected system under each control parameter; and then, according to the phase margin sensitivity of the wind power grid-connected system under each control parameter, determines the phase margin sensitivity ratio of the wind power grid-connected system under each control parameter.

[0084] Specifically, the phase margin sensitivity is defined as follows:

[0085]

[0086] Among them, a i represents the i-th control parameter, It represents the wind power grid-connected system under the i-th control parameter at the intersection frequency f os Phase margin sensitivity at Δa i represents the increment of the i-th control parameter, P p,n (a i ,f os ) represents the wind power grid-connected system under the i-th control parameter at the intersection frequency f os Phase margin at P p,n (a i +Δa i ,f os ) indicates that the i-th control parameter increases by Δa i Then at the intersection frequency f os The phase margin at .

[0087] Among them, the i-th control parameter increases by Δa iAfter that, the Bode diagram of the positive sequence impedance of the load subsystem and the Bode diagram of the positive sequence impedance of the source subsystem will produce a new intersection frequency f os .

[0088] The Bode diagram of the positive sequence impedance of the load subsystem or the source subsystem corresponding to the control parameter is plotted at the intersection frequency f os Substituting the phase margin at into the above definition formula of phase margin sensitivity, the phase margin with respect to the control parameter a can be calculated. i The partial derivative of the control parameter a is quantitatively evaluated i The degree to which changes near its set value affect the phase margin.

[0089] The phase margin sensitivity ratio is defined as follows:

[0090]

[0091] Among them, P seni % represents the phase margin sensitivity ratio of the wind power grid-connected system under the i-th control parameter, l represents the total number of control parameters, It represents the wind power grid-connected system under the i-th control parameter at the intersection frequency f os Phase margin sensitivity at .

[0092] S116 , determining a preset number of control parameters to be identified for the wind turbines according to the descending order of the phase margin sensitivity ratio of each control parameter.

[0093] The phase margin sensitivity ratio of each control parameter is sorted in descending order, and the first m control parameters whose phase margin sensitivity ratio is far greater than that of other control parameters are taken as the control parameters to be identified.

[0094] The value of m is determined according to the difference between the phase margin sensitivity ratio of a specific control parameter and the phase margin sensitivity ratio of the next control parameter.

[0095] In a possible embodiment, S12, constructing a simulation model of the wind turbine generator set based on the actual control structure and parameter information of the wind turbine generator set and determining the initial values ​​of the control parameters to be identified, may include:

[0096] A simulation model of the wind turbine is constructed based on the control structure of the machine-side converter and the grid-side converter of the known new energy unit, as well as the rated capacity, main circuit parameters, control circuit parameters and grid equivalent impedance parameters of the known new energy unit.

[0097] Among them, the control structure of the wind turbine includes the control structure of the known wind turbine side converter and the control structure of the grid side converter; the parameter information includes the known rated capacity, main circuit parameters, control circuit parameters and grid equivalent impedance parameters of the wind turbine.

[0098] In the present disclosure, a general design method of control parameters and a common control loop bandwidth configuration of a wind turbine generator system may be used to determine the initial value of the control parameter to be identified, such as a general design method of a doubly-fed wind turbine generator system.

[0099] Figure 2 The figure is a schematic diagram of an equivalent circuit of a doubly-fed wind farm series-compensated grid-connected system according to an exemplary embodiment. Figure 3 The figure is a schematic diagram of the topology and control structure of a single doubly-fed wind turbine generator set according to an exemplary embodiment.

[0100] As an example, the simulation model of the wind turbine generator system constructed in the present disclosure may be a simulation model of a doubly-fed wind farm series compensation system, such as Figure 2 As shown in the figure, it is the equivalent circuit of the doubly fed wind farm series compensation system. Figure 3 As shown in the figure, the topology and control structure of a single doubly-fed wind turbine generator set is shown, including the main circuit, the rotor-side converter control structure and the grid-side converter control structure.

[0101] Among them, the wind turbine adopts a 1.5MW doubly-fed wind turbine. The research scenario is a doubly-fed wind farm composed of 25 doubly-fed wind turbines, which are equally aggregated and connected to the grid through transformers, line impedance and series compensation, with a series compensation degree of 40%.

[0102] Specifically, the parameter information for constructing the simulation model includes: the inner loop control parameters of the rotor-side converter, including the current loop proportional coefficient and integral coefficient; the outer loop control parameters of the rotor-side converter, including the power outer loop proportional coefficient and integral coefficient; the inner loop control parameters of the grid-side converter, including the current loop proportional coefficient and integral coefficient, and the outer loop control parameters of the grid-side converter, including the voltage loop proportional coefficient and integral coefficient; the main circuit parameters, including the filter inductance L in the filter. f , filter resistor R f and line impedance r L1 +jx L1 、r L2 +jx L2 .

[0103] like Figure 2 As shown in the figure, the wind farm equivalent model first passes through the 0.69kV / 35kV transformer T1 and the 35kV / 220kV transformer T2, and then passes through the line impedance r L1 +jx L1 After that, it passes through 220kV / 500kV transformer T3 and line impedance r L2 +jxL2 With series capacitor compensator x C Then it is connected to the infinite power grid with a series compensation of 40%.

[0104] like Figure 3 As shown, one end of the wind turbine is connected to the transmission system, and the other end of the transmission system is connected to the doubly fed generator. A capacitor is connected in parallel between the rotor-side converter and the grid-side converter. The rotor-side converter is connected to the rotor-side converter control. The rotor-side converter control includes a power outer loop and a current inner loop. The grid-side converter is connected to the grid-side converter control. The grid-side converter control includes a current outer loop and a voltage outer loop. The output of the grid-side converter is connected to the filter resistor R f One end of the doubly fed generator is connected to the rotor side converter, and the other output is connected to the filter inductor L f The other end is connected as a transmission line and connected to the subsequent transformer; the inductor L f The other end of the filter resistor R f connect.

[0105] In a possible embodiment, S12, determining a control parameter-active power output parameter group to be identified, may include S121 to S123.

[0106] S121 , within a preset control parameter range, taking the initial value of the control parameter to be identified as a reference and taking the preset first step length as an interval, determining multiple values ​​of the control parameter to be identified.

[0107] As an example, for m control parameters to be identified, within the preset control parameter interval corresponding to each control parameter to be identified, the initial value of the control parameter to be identified is used as a reference, and the value of a control parameter to be identified is determined at each preset first step length, thereby determining multiple values ​​of the control parameter to be identified.

[0108] S122 , determining multiple values ​​of active power within a preset active power range at intervals of a preset second step length.

[0109] As an example, within the rated power range of the wind turbine generator, a value of active output is determined at each preset second step length, thereby obtaining multiple values ​​of active output.

[0110] The preset second step size may be 0.05 pu.

[0111] S123 , combining multiple values ​​of the control parameter to be identified with multiple values ​​of the active output to determine a control parameter to be identified-active output parameter group.

[0112] Each value of the control parameter to be identified is combined with each value of the active output in pairs, and each control parameter to be identified-active output parameter group includes a value of the control parameter to be identified and a value of the active output.

[0113] In a possible embodiment, S13, simulating the simulation model of the wind turbine according to the control parameter-active output parameter group to be identified to determine the oscillation characteristics corresponding to each control parameter-active output parameter group to be identified, may include S131 to S134.

[0114] Wherein, in step S13, Hilbert-Huang Transform (HHT) is used to extract the oscillation features.

[0115] S131 , simulating each to-be-identified control parameter-active power output parameter group using a simulation model of the wind turbine generator system to determine time domain data corresponding to each to-be-identified control parameter-active power output parameter group.

[0116] S132 , performing Empirical Mode Decomposition (EMD) on the time domain data corresponding to each control parameter-active power parameter pair to be identified, and determining the Intrinsic Mode Function (IMF) corresponding to each control parameter-active power parameter pair to be identified.

[0117] S133, performing Hilbert transform on the intrinsic mode function (IMF) corresponding to each control parameter to be identified-active power parameter group to determine the instantaneous frequency, instantaneous amplitude and phase information corresponding to each control parameter to be identified-active power parameter group.

[0118] S134 , fitting the instantaneous frequency, instantaneous amplitude, and phase information corresponding to each control parameter-active power parameter group to be identified, and determining the oscillation characteristics corresponding to each control parameter-active power parameter group to be identified.

[0119] The oscillation characteristics may include an oscillation frequency and a damping ratio.

[0120] In a possible embodiment, S14, performing model training on a preset neural network model according to the control parameter to be identified-active output parameter group and the oscillation characteristics, and determining the neural network model that has completed model training, may include S141 to S144.

[0121] S141 : Constructing a data set of active output-oscillation characteristics-control parameters according to the control parameter-active power parameter group to be identified and the oscillation characteristics.

[0122] Each sample data in the active power output-oscillation characteristics-control parameter data set includes active power output, oscillation characteristics and corresponding control parameters.

[0123] S142 , performing normalization processing on samples in the active power-oscillation characteristic-control parameter data set to determine a normalized data set.

[0124] Among them, through normalization processing, the active output, oscillation characteristics and control parameters in the active output-oscillation characteristics-control parameters dataset are scaled to the [-1, 1] interval to eliminate the influence of feature magnitude differences on model training.

[0125] S143: Divide the normalized data set according to a preset ratio to determine a training set and a test set.

[0126] The preset ratio can be 80% for the training set and 20% for the test set. The training set is used to train the preset neural network, and the test set is used to evaluate the ability of the trained neural network model to fit the data and verify its generalization performance.

[0127] S144, using the active output and oscillation characteristics in the training set as the input of the preset neural network, and the control parameters in the training set as the output of the preset neural network, performing model training on the preset neural network model, and determining the neural network model that has undergone model training.

[0128] The preset neural network may adopt a typical back propagation (BP) neural network.

[0129] S145, using the test set to test the neural network model that has undergone model training to determine the neural network model that has completed model training.

[0130] Among them, when using the test set to test the neural network model that has completed model training, the active output and oscillation characteristics in the test set are used as the input of the preset neural network, and the identification results of the predicted control parameters to be identified are output. The generalization performance of the neural network model that has completed model training is evaluated based on the identification results of the predicted control parameters to be identified and the control parameters in the test set.

[0131] Based on the above steps S11 to S14, the following is explained by taking the doubly-fed wind turbine generator system connected to the grid via series compensation as an example:

[0132] Figure 4 The figure is a schematic diagram showing the phase margin sensitivity of a control parameter of a doubly-fed wind turbine generator system according to an exemplary embodiment.

[0133] like Figure 4 As shown, the current inner loop proportional coefficient K in the rotor-side converter control p The phase margin sensitivity of K is about -16.1, and its phase margin sensitivity accounts for 70.18%, which is much higher than other control parameters. Therefore, in this embodiment, K p Set as the control parameter to be identified.

[0134] For the rotor side current inner loop proportional coefficient K p , the typical control parameter 0.035 corresponding to the single doubly fed wind turbine model is used as the initial parameter and as the benchmark, and the control parameter is adjusted within the preset parameter range [0.025, 0.045] with a value of 0.2×10 -3 Adjust the rotor side current inner loop proportional coefficient K for the preset first step length p The value of the active output is adjusted within the preset active output range [0.3pu, 0.4pu] with a preset second step of 0.01pu, thereby determining the control parameter to be identified - active output parameter group.

[0135] The simulation model of the doubly-fed wind farm series compensation system constructed above and the control parameter-active power parameter group to be identified were used for operation simulation to extract the oscillation characteristics under the corresponding active power output and control parameters. The oscillation characteristics include oscillation frequency and damping ratio. Finally, an 1111×4 data sample set was obtained, namely the active power output-oscillation characteristic-control parameter data set. The first three columns are the active power output, damping ratio and oscillation frequency, respectively, and the last column is the control parameter to be identified.

[0136] The BP neural network model is trained using the dataset of active output, oscillation characteristics and control parameters, so that the BP neural network model can be used to realize the parameter identification task with oscillation characteristics and active output as input and control parameters to be identified as output.

[0137] Specifically, the input layer dimension of the BP neural network model is set to 3 layers and the output layer dimension is set to 1 layer.

[0138] The loss function was set to mean squared error (MSE), the learning rate to 0.001, and the maximum number of training rounds to 2000. After debugging, 80% of the data samples were used as the training set, and the remaining 20% ​​as the test set. Four hidden layers were set, with the number of neurons being 10, 12, 8, and 10, respectively. After 1000 training rounds, the mean squared error (MSE) was reduced to 0.0002164. At this point, the BP neural network model established a mapping relationship between active power output, oscillation characteristics, and the dominant control parameters.

[0139] The formula for calculating the mean square error (MSE) is as follows:

[0140]

[0141] Among them, m represents the number of test set samples, y i represents the true value of the control parameter to be identified, Represents the predicted value of the control parameter to be identified.

[0142] Figure 5 The figure is a schematic diagram showing verification of the generalization ability of a preset neural network model according to an exemplary embodiment.

[0143] like Figure 5 As shown in the figure, the test set is input into the BP neural network model after model training. The obtained output effect is compared. It can be determined that the output result of the test set is basically consistent with the predicted output result, and the mean square error (MSE) is 1.1691×10 -10 , which verifies the generalization ability of the BP neural network model.

[0144] In a possible embodiment, S15, inputting the actual oscillation characteristics and the actual active power output into the trained neural network model to determine the identification result of the actual control parameter to be identified includes:

[0145] The actual oscillation characteristics and the actual active power output are normalized to determine the normalized actual oscillation characteristics and the normalized actual active power output.

[0146] The normalized actual oscillation characteristics and the normalized actual active output are input into the neural network for completing model training, and the identification results of the actual control parameters to be identified are output.

[0147] As an example, oscillation characteristics of n groups of actual oscillation data are extracted and normalized with the actual active power output, and input into a trained BP neural network model to obtain n groups of control parameter identification results.

[0148] In the present disclosure, it is assumed that the wind turbines in the same wind farm have the same active output, and three groups of related oscillation data with active output sizes of 0.3pu, 0.35pu, and 0.4pu are obtained to determine the three groups of oscillation characteristics corresponding to the oscillation data. The oscillation characteristics corresponding to the active output are input into the trained BP neural network model, and the identification results of the corresponding control parameters to be identified are determined to be 0.0359, 0.0348, and 0.0342, respectively.

[0149] A confidence interval is a statistical method used to estimate the range of a population parameter. It determines the range of values ​​for the actual parameter to be identified based on a preset confidence level. The confidence level represents the probability that the interval covers the true value.

[0150] In a possible embodiment, a method for identifying control parameters of a wind turbine generator system based on subsynchronous oscillation characteristics further includes S16.

[0151] S16, determining a value interval of the actual control parameter to be identified according to the identification result of the actual control parameter to be identified and a preset confidence interval.

[0152] Specifically, S16 includes:

[0153]

[0154] in, represents the sample mean, z represents the preset confidence level, that is, the critical value of the corresponding confidence level under the normal distribution, s represents the standard deviation of the sample, n represents the number of samples, and CI represents the value interval of the actual control parameter to be identified;

[0155] The actual value range of the control parameter to be identified is:

[0156]

[0157] When the preset confidence level is 95%, z=1.96; when the preset confidence level is 99%, z=2.576.

[0158] Continuing with the above example, in this embodiment, the preset confidence level is 95%, and the rotor side current inner loop proportional coefficient K is calculated as p The confidence interval using a 95% confidence level is [0.033991, 0.035942], that is, the actual value interval of the control parameter to be identified is [0.033991, 0.035942].

[0159] The present disclosure provides a wind turbine control parameter identification method based on subsynchronous oscillation characteristics, which uses limited oscillation data and artificial intelligence algorithms to realize the identification of key control parameters of wind turbines, with the characteristics of small computational complexity and high identification efficiency.

[0160] Figure 6 The present invention is a block diagram showing a wind turbine control parameter identification system based on subsynchronous oscillation characteristics according to an exemplary embodiment.

[0161] Based on the same concept, the present disclosure also provides a wind turbine control parameter identification system 100 based on subsynchronous oscillation characteristics, such as Figure 6 As shown, it includes: a phase margin sensitivity analysis module 110 , a simulation model and simulation parameter construction module 120 , a simulation module 130 , a model training module 140 , and a control parameter identification module 150 .

[0162] A phase margin sensitivity analysis module 110 is configured to perform a phase margin sensitivity analysis based on the main oscillation mode of the wind turbine grid-connected system to determine the control parameters to be identified for the wind turbine;

[0163] A simulation model and simulation parameter construction module 120 is used to construct a simulation model of the wind turbine according to the actual control structure and parameter information of the wind turbine, determine the initial value of the control parameter to be identified, and determine the control parameter to be identified - active power output parameter group;

[0164] The simulation module 130 is configured to simulate the simulation model of the wind turbine generator system according to the control parameter-active power output parameter group to be identified, and determine the oscillation characteristics corresponding to each of the control parameter-active power output parameter groups to be identified;

[0165] A model training module 140 is configured to perform model training on a preset neural network model based on the control parameter to be identified-active power output parameter group and the oscillation characteristics, and determine a neural network model that has completed model training;

[0166] The control parameter identification module 150 is used to input the actual oscillation characteristics and the actual active power output into the neural network model that has completed model training, and determine the identification results of the actual control parameters to be identified.

[0167] Through the above technical solution, a phase margin sensitivity analysis is performed under the main oscillation mode of the wind turbine grid-connected system to determine the control parameters that have a greater impact on the main oscillation mode as the control parameters to be identified, which is helpful to analyze the control parameters that affect the stability of the wind power grid-connected system. Through the simulation model of the wind turbine, simulation is performed to determine the oscillation characteristics corresponding to the parameters to be identified and the active output. The preset neural network model is trained using the control parameter to be identified-active output parameter group and the oscillation characteristics to learn the relationship between the oscillation characteristics and the control parameters, and the ability of the neural network model to identify the control parameters to be identified based on limited oscillation characteristics is realized. The calculation amount is small and the identification efficiency is high. The control strategy of the wind turbine can be adjusted and optimized in a targeted manner through oscillation data to adapt to the dynamic changes of the power grid and improve the operating efficiency and reliability of the unit.

[0168] Regarding the embodiment of the above system, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.

[0169] Based on the same concept as above, in another embodiment of the present disclosure, an electronic device is also provided, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor, and the processor is used to execute a method for identifying wind turbine control parameters based on subsynchronous oscillation characteristics when executing the program.

[0170] Optionally, the memory is used to store programs; the memory may include volatile memory (English: volatile memory), such as random-access memory (English: random-access memory, abbreviated: RAM), such as static random-access memory (English: static random-access memory, abbreviated: SRAM), double data rate synchronous dynamic random access memory (English: Double Data Rate Synchronous Dynamic Random Access Memory, abbreviated: DDR SDRAM), etc.; the memory may also include non-volatile memory (English: non-volatile memory), such as flash memory (English: flash memory). The memory is used to store computer programs (such as applications, functional modules, etc. that implement the above-mentioned methods), computer instructions, etc., and the above-mentioned computer programs, computer instructions, etc. can be partitioned and stored in one or more memories. In addition, the above-mentioned computer programs, computer instructions, data, etc. can be called by the processor.

[0171] The aforementioned computer programs, computer instructions, etc. may be partitioned and stored in one or more memories, and the aforementioned computer programs, computer instructions, data, etc. may be called by a processor.

[0172] The processor is configured to execute the computer program stored in the memory to implement the various steps of the method involved in the above embodiment. For details, please refer to the relevant description in the above method embodiment.

[0173] The processor and memory can be independent structures or integrated structures. When the processor and memory are independent structures, the memory and processor can be coupled via a bus.

[0174] In an embodiment of the present disclosure, a non-transitory computer-readable storage medium is further provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of a wind turbine control parameter identification method based on subsynchronous oscillation characteristics in any of the above embodiments are implemented.

[0175] Those skilled in the art will appreciate that the embodiments of the present disclosure may be provided as methods, systems, or computer program products. Therefore, the present disclosure may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present disclosure may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0176] The present disclosure is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present disclosure. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0177] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0178] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0179] Although the preferred embodiments of the present disclosure have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concepts. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present disclosure.

[0180] Obviously, those skilled in the art may make various changes and modifications to the present disclosure without departing from the spirit and scope of the present disclosure. Thus, if these modifications and variations of the present disclosure fall within the scope of the claims of the present disclosure and their equivalents, the present disclosure is intended to include these modifications and variations.

Claims

1. A method for identifying control parameters of a wind turbine generator system based on subsynchronous oscillation characteristics, characterized in that: include: Conduct phase margin sensitivity analysis based on the main oscillation mode of the wind turbine grid-connected system to determine the control parameters to be identified for the wind turbine; According to the actual control structure and parameter information of the wind turbine generator set, a simulation model of the wind turbine generator set is constructed, an initial value of the control parameter to be identified is determined, and a control parameter to be identified-active power output parameter group is determined; Simulating the simulation model of the wind turbine generator system according to the control parameter-active power output parameter group to be identified, and determining the oscillation characteristics corresponding to each control parameter-active power output parameter group to be identified; Performing model training on a preset neural network model according to the control parameter to be identified-active output parameter group and the oscillation characteristics, and determining a neural network model that has completed model training; The actual oscillation characteristics and the actual active power output are input into the neural network model that has completed the model training to determine the identification result of the actual control parameter to be identified.

2. The method according to claim 1, characterized in that The method further comprises: The value range of the actual control parameter to be identified is determined according to the identification result of the actual control parameter to be identified and a preset confidence level.

3. The method according to claim 1, characterized in that The phase margin sensitivity analysis is performed based on the main oscillation mode of the wind turbine grid-connected system to determine the control parameters to be identified of the wind turbine, including: Dividing the wind power grid-connected system into a load subsystem and a source subsystem; Establishing a positive-sequence impedance model of the load subsystem and a positive-sequence impedance model of the source subsystem; Drawing a Bode diagram of the positive sequence impedance of the load subsystem and a Bode diagram of the positive sequence impedance of the source subsystem according to the positive sequence impedance model of the load subsystem and the positive sequence impedance model of the source subsystem; determining a phase margin of the wind power grid-connected system under each control parameter according to a Bode diagram of the positive sequence impedance of the load subsystem and a Bode diagram of the positive sequence impedance of the source subsystem; Determining a phase margin sensitivity ratio of each control parameter according to the phase margin of the wind power grid-connected system under each control parameter; A preset number of control parameters to be identified of the wind turbine generator system are determined according to the descending order of the phase margin sensitivity ratio of each control parameter.

4. The method according to claim 3, characterized in that Determining the phase margin sensitivity ratio of each control parameter according to the phase margin of the wind power grid-connected system under each control parameter includes: Among them, a i represents the i-th control parameter, It represents the wind power grid-connected system under the i-th control parameter at the intersection frequency f os Phase margin sensitivity at Δa i represents the increment of the i-th control parameter, P p,n (a i ,f os ) represents the wind power grid-connected system at the intersection frequency f under the i-th control parameter os Phase margin at P p,n (a i +Δa i ,f os ) indicates that the i-th control parameter increases by Δa i Then at the intersection frequency f os Phase margin at ; Among them, P seni % represents the phase margin sensitivity ratio of the i-th control parameter, l represents the total number of the control parameters, It represents the wind power grid-connected system under the i-th control parameter at the intersection frequency f os Phase margin sensitivity at .

5. The method according to claim 1, wherein The step of determining the control parameter to be identified-active power output parameter group includes: Within a preset control parameter range, taking the initial value of the control parameter to be identified as a reference and taking a preset first step length as an interval, determining multiple values ​​of the control parameter to be identified; Determining multiple values ​​of the active power output within a preset active power output interval at intervals of a preset second step length; The multiple values ​​of the control parameter to be identified and the multiple values ​​of the active output are combined to determine the control parameter to be identified-active output parameter group.

6. The method according to claim 1, wherein The oscillation characteristics include oscillation frequency and damping ratio; The simulating the simulation model of the wind turbine generator system according to the control parameter-active power output parameter group to be identified to determine the oscillation characteristics corresponding to each control parameter-active power output parameter group to be identified includes: Using the simulation model of the wind turbine generator set to simulate each of the control parameter-active power output parameter groups to be identified, and determining the time domain data corresponding to each of the control parameter-active power output parameter groups to be identified; Performing empirical mode decomposition on the time domain data corresponding to each of the control parameter-active power output parameter groups to be identified, and determining the intrinsic mode function corresponding to each of the control parameter-active power output parameter groups to be identified; Performing a Hilbert transform on the intrinsic mode function corresponding to each of the control parameter-active power parameter groups to be identified to determine the instantaneous frequency, instantaneous amplitude, and phase information corresponding to each of the control parameter-active power parameter groups to be identified; The instantaneous frequency, instantaneous amplitude and phase information corresponding to each of the control parameter to be identified-active power output parameter groups are fitted to determine the oscillation characteristics corresponding to each of the control parameter to be identified-active power output parameter groups.

7. The method according to claim 1, characterized in that The performing model training on a preset neural network model according to the control parameter to be identified-active output parameter group and the oscillation characteristics, and determining a neural network model that has completed model training, includes: Constructing a data set of active output-oscillation characteristics-control parameters according to the control parameter-active power output parameter group to be identified and the oscillation characteristics; Normalizing samples in the active output-oscillation characteristics-control parameter data set to determine a normalized data set; Dividing the normalized data set according to a preset ratio to determine a training set and a test set; Using the active output and oscillation characteristics in the training set as inputs of the preset neural network, and the control parameters in the training set as outputs of the preset neural network, performing model training on the preset neural network model, and determining a neural network model that has undergone model training; The test set is used to test the neural network model that has undergone model training to determine the neural network model that has completed model training.

8. A wind turbine control parameter identification system based on subsynchronous oscillation characteristics, characterized in that: include: Phase margin sensitivity analysis module, used to perform phase margin sensitivity analysis based on the main oscillation mode of the wind turbine grid-connected system and determine the control parameters to be identified for the wind turbine; A simulation model and simulation parameter construction module, configured to construct a simulation model of the wind turbine generator set according to the actual control structure and parameter information of the wind turbine generator set, determine the initial values ​​of the control parameters to be identified, and determine the control parameter to be identified-active power output parameter group; a simulation module, configured to simulate a simulation model of the wind turbine generator system according to the control parameter-active power output parameter group to be identified, and determine an oscillation characteristic corresponding to each of the control parameter-active power output parameter groups to be identified; A model training module is used to perform model training on a preset neural network model according to the control parameter to be identified-active power output parameter group and the oscillation characteristics, and determine a neural network model that has completed model training; The control parameter identification module is used to input the actual oscillation characteristics and the actual active power output into the neural network model that has completed model training, and determine the identification results of the actual control parameters to be identified.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

10. An electronic device, characterized in that: include: a memory having a computer program stored thereon; A processor, configured to execute the computer program in the memory to implement the steps of the method according to any one of claims 1 to 7.

Images