Broadband impedance identification model, method and system of flexible low-frequency power transmission system

Abstract

The invention relates to the technical field of flexible power transmission, in particular to a broadband impedance identification model, method and system of a flexible low-frequency power transmission system. According to the training method of the broadband impedance identification model of the flexible low-frequency power transmission system, the power conservation loss is introduced in the model training process, and the broadband impedance identification model of the flexible low-frequency power transmission system considering the industrial and low frequency coupling effect is established; and the broadband oscillation generation mechanism of the flexible low-frequency power transmission system considering the M3C working low-frequency side coupling effect can be accurately analyzed. The problems that due to the fact that an M3C topological structure is complex, has a working and low frequency side coupling effect and is limited by commercial technology secrecy, relevant parameters cannot be accurately obtained, and a broadband impedance identification model is difficult to establish through theoretical derivation are solved.

Description

Technical Field

[0001] The present invention relates to the field of flexible power transmission technology, and in particular to a broadband impedance identification model, method and system for a flexible low-frequency power transmission system. Background Art

[0002] Flexible low-frequency transmission systems (FLTSs) are a new type of transmission system with a high proportion of power electronics. They contain numerous power electronic devices, such as AC converters, and nonlinear control elements. Compared to DC transmission systems, FLTSs exhibit more complex dynamic characteristics due to their frequency-interaction characteristics. These power electronic devices interact dynamically with each other, and with the AC network, leading to the risk of broadband oscillations in the system, resulting in degraded power quality, equipment damage, and even system shutdown. Impedance analysis, due to its clear physical meaning and simple application, is currently the primary analytical method for studying broadband oscillations in power systems with power electronics. Impedance modeling is the fundamental prerequisite for using impedance analysis to study broadband oscillations in FLTSs.

[0003] Currently, impedance modeling for addressing the broadband oscillation stability of M3C-based flexible low-frequency transmission systems uses multiple reference DC and AC / DC impedance models. This model only considers the impact of the M3C (modular multilevel matrix converter) low-frequency side or power frequency side on the AC network, without considering the power frequency coupling relationship between the M3C itself and the influence of inter-frequency energy exchange between M3C bridge arms. This modeling results in low impedance modeling accuracy, which in turn affects the accuracy of research on broadband oscillations in flexible low-frequency transmission systems. Furthermore, using conventional neural networks (such as BP neural networks) is a purely data-driven approach, relying solely on training data samples to obtain a specific distribution pattern. This requires a large number of training samples, and because M3C converters involve coupling between the power frequency and low frequency sides, existing training methods lack interpretability, making it difficult to obtain an accurate broadband impedance identification model. Summary of the Invention

[0004] In order to overcome the shortcomings of the above-mentioned prior art in the broadband impedance identification of M3C converters, which directly adopts the existing machine learning methods with low accuracy and poor interpretability, the present invention proposes a training method for a broadband impedance identification model of a flexible low-frequency power transmission system. This method can establish an accurate and efficient broadband impedance identification model of a flexible low-frequency power transmission system that takes into account the mutual coupling between the industrial and low-frequency sides of M3C.

[0005] The present invention proposes a training method for a broadband impedance identification model of a flexible low-frequency power transmission system. The broadband impedance identification model identifies the broadband impedance of a modular multi-level matrix converter M3C based on its operating current, voltage, and frequency f.

[0006] The loss function during the training of the broadband impedance identification model is:

[0007] Loss=Loss a +Loss phy +Loss bc_ld +Loss bc_lq +Loss bc_sd +Loss bc_sq

[0008] Among them, Loss a Loss is the model prediction loss. phy is the power conservation loss; Loss bc_ld is the d-axis loss on the low-frequency side, Loss bc_lq is the q-axis loss on the low-frequency side; Loss bc_sd is the d-axis loss on the power frequency side, Loss bc_sq is the q-axis loss on the power frequency side;

[0009] Power conservation loss, Loss phy The submodule capacitor voltage U dc On the calculation, U dc The power conservation constraint is satisfied:

[0010]

[0011] Among them, C dc Indicates the equivalent capacitance value of the M3C submodule; U sd and I sd Indicates the voltage and current of the d-axis on the power frequency side of M3C, U sq and I sq Indicates the voltage and current of the q-axis on the power frequency side of M3C, U ld and I ld represents the voltage and current of the d-axis on the low-frequency side of M3C, U lq and I lq Indicates the voltage and current on the d-axis of the M3C low-frequency side.

[0012] Preferably, Loss bc_ld and Loss bc_lq The d-axis common mode voltage U on the low-frequency side of M3C is m3cd and q-axis common mode voltage U m3cq U m3cd and U m3cq Satisfy the low-frequency side constraints:

[0013]

[0014] Where L represents the equivalent inductance of the bridge arm of the flexible low-frequency power transmission system, w l Indicates the low-frequency side corner frequency of M3C; U ld and I ldrepresents the voltage and current of the d-axis on the low-frequency side of M3C, U lq and I lq Indicates the voltage and current on the d-axis of the M3C low-frequency side.

[0015] Preferably, Loss bc_sd and Loss bc_sq The d-axis common mode voltage U comd and q-axis common mode voltage U comq On the calculation, U comd and U comq

[0016] The following power frequency side constraints are met:

[0017]

[0018] Among them, w s Indicates the angular frequency of the M3C power frequency side, U sd and I sd Indicates the voltage and current of the d-axis on the power frequency side of M3C, U sq and I sq Indicates the voltage and current of the q-axis on the power frequency side of M3C.

[0019] Preferably, Loss bc_ld Using M3C low-frequency side d-axis common mode voltage U m3cd The low-frequency side d-axis common mode voltage U represented by the model hidden layer output m3cd The cross entropy loss of the process quantity; Loss bc_lq Using M3C low-frequency side q-axis common mode voltage U m3cq The low-frequency side q-axis common mode voltage U represented by the model hidden layer output m3cq The cross entropy loss of the process quantity; Loss bc_sd Using M3C power frequency side d-axis common mode voltage U comd The d-axis common mode voltage U on the power frequency side is represented by the output of the hidden layer of the model comd The cross entropy loss of the process quantity; Loss bc_sq Using M3C power frequency side q-axis common mode voltage U comq The q-axis common mode voltage U on the power frequency side is represented by the output of the hidden layer of the model comq The cross entropy loss of the process quantity.

[0020] Preferably, Loss phy Submodule capacitor voltage U using M3C dc The submodule capacitor voltage U representing the output of the hidden layer of the model dc The cross entropy loss of the process quantity.

[0021] Preferably, the broadband impedance identification model adopts a KAN neural network model.

[0022] Preferably, the learning data set of the broadband impedance identification model is denoted as {U' sd 、I' sd 、U' sq 、I' sq 、U' ld 、I' ld 、U' lq 、I' lq 、f;Z' dq (freq)},U' sd 、I' sd 、U' sq 、I' sq 、U' ld 、I' ld 、U' lq 、I' lq and Z' dq (freq) are U sd , I sd 、U sq , I sq 、U ld , I ld 、U lq , I lq and Z dq Normalized value of (freq);

[0023] U sd and I sd Indicates the voltage and current of the d-axis on the power frequency side of M3C, U sq and I sq Indicates the voltage and current of the q-axis on the power frequency side of M3C, U ld and I ld represents the voltage and current of the d-axis on the low-frequency side of M3C, U lq and I lq represents the voltage and current of the d-axis on the low-frequency side of M3C; Z dq (freq) is the broadband impedance of M3C; Loss a The predicted impedance normalized value and Z' output by the model dq The cross entropy loss of (freq).

[0024] The present invention proposes a broadband impedance identification method for a flexible low-frequency power transmission system. First, a broadband impedance identification model is trained. Then, the broadband impedance identification model is combined with the data set {U sd , I sd 、U sq , I sq 、U ld , I ld 、U lq , I lq 、f;Zdq (freq)} to normalize the working condition samples and construct the working condition input samples {U' sd 、I' sd 、U' sq 、I' sq 、U' ld 、I' ld 、U' lq 、I' lq The working condition input samples are combined with different scanning frequencies f0 and input into the broadband impedance identification model. The obtained impedance normalization value is inversely normalized to obtain the predicted impedance at each scanning frequency.

[0025] The present invention proposes a broadband impedance control method, which obtains the target impedance obtained by scanning different frequencies f under a specified steady-state operating point; designs the current and voltage data of the M3C power frequency side and low frequency side ports in a synchronous rotating coordinate system, and constructs a design scheme {U sd , I sd 、U sq , I sq 、U ld , I ld 、U lq , I lq , f}; normalize the design scheme based on the training data set of the broadband impedance identification model, then input the broadband impedance identification model to obtain the predicted impedance corresponding to the design scheme; adjust the current and voltage data in the design scheme until the predicted impedance is consistent with the target impedance.

[0026] The present invention proposes a broadband impedance identification system for a flexible low-frequency power transmission system, comprising:

[0027] The data acquisition module is used to collect the current and voltage data of the M3C power frequency side and low frequency side ports in the synchronous rotating coordinate system at each steady-state operating point;

[0028] The data processing module pre-processes the collected current and voltage data, filters the data samples, and obtains the corresponding broadband impedance through simulation calculation for the filtered data samples to construct the initial sample {U sd , I sd 、U sq , I sq 、U ld , I ld 、U lq , I lq 、f;Z dq (freq)},U sd and I sd Indicates the voltage and current of the d-axis on the power frequency side of M3C, U sq and I sqIndicates the voltage and current of the q-axis on the power frequency side of M3C, U ld and I ld represents the voltage and current of the d-axis on the low-frequency side of M3C, U lq and I lq represents the voltage and current of the d-axis on the low-frequency side of M3C; Z dq (freq) is the broadband impedance of M3C; f is the scanning frequency;

[0029] The model training module normalizes the initial samples and obtains the data set {U' sd 、I' sd 、U' sq 、I' sq 、U' ld 、I' ld 、U' lq 、I' lq 、f;Z' dq (freq)},U' sd 、I' sd 、U' sq 、I' sq 、U' ld 、I' ld 、U' lq 、I' lq and Z' dq (freq) are U sd , I sd 、U sq , I sq 、U ld , I ld 、U lq , I lq and Z dq The normalized value of (freq); a neural network model is trained on the data set as a broadband impedance identification model;

[0030] The impedance identification module obtains the current and voltage data {U sd , I sd 、U sq , I sq 、U ld , I ld 、U lq , I lq} and normalize it, then use the normalized value {U' sd 、I' sd 、U' sq 、I' sq 、U' ld 、I' ld 、U' lq 、I'lq} and frequency f are input into the broadband impedance identification model to obtain the corresponding broadband impedance Z dq (freq).

[0031] The advantages of the present invention are:

[0032] (1) The present invention proposes a training method for a broadband impedance identification model for a flexible low-frequency power transmission system. This method introduces power conservation loss during model training and establishes a broadband impedance identification model for the flexible low-frequency power transmission system that takes into account the industrial-low-frequency coupling effect. This method can accurately analyze the mechanism of broadband oscillations in the flexible low-frequency power transmission system that takes into account the industrial-low-frequency coupling effect of the M3C. This invention addresses the difficulty of establishing a broadband impedance identification model through theoretical deduction due to the complex topology of the M3C, the presence of industrial-low-frequency coupling, and the limitations of commercial technology confidentiality, which prevent accurate acquisition of relevant parameters.

[0033] (2) The present invention adds the energy conservation formula of the M3C power frequency side / low frequency side and DC side as a regularization of physical constraints into the loss function of the neural network to construct a physical information neural network model. The present invention adds physical information constraints, considers the coupling effect of the M3C power frequency side and low frequency side and the influence of the heterogeneous frequency energy interaction on the system impedance frequency characteristics, can obtain more accurate broadband impedance data of the flexible low-frequency power transmission system, and improves the interpretability of the broadband impedance identification model, which is used to analyze the broadband oscillation characteristics of the flexible low-frequency power transmission system, is conducive to the study of the generation mechanism of broadband oscillation of the flexible low-frequency power transmission system, and has important guiding significance for the control strategy design and broadband oscillation suppression of actual flexible low-frequency power transmission projects.

[0034] (3) The present invention establishes a broadband impedance identification model for a flexible low-frequency power transmission system by using a method based on a KAN neural network. This eliminates the need to master the topological structure and controller-related detailed parameters of the M3C in the flexible low-frequency power transmission system, thereby solving the problem that the system topology is complex and is subject to restrictions on commercial technology confidentiality, making it impossible to accurately obtain relevant parameters for theoretical deduction.

[0035] (4) The present invention proposes a broadband impedance identification method and control method for a flexible low-frequency power transmission system, which can quickly and accurately predict the broadband impedance under different steady-state operating points through a broadband impedance identification model and adjust it to ensure the working reliability of the flexible low-frequency power transmission system. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 A flow chart of the method for constructing a data set used in the present invention;

[0037] Figure 2 This is a flow chart of the model training method proposed in the present invention;

[0038] Figure 3This is a flow chart of a broadband impedance identification method for a flexible low-frequency power transmission system proposed by the present invention;

[0039] Figure 4 Schematic diagram of a flexible low-frequency power transmission system module in an embodiment;

[0040] Figure 5 Schematic diagram of the operation flow of the embodiment;

[0041] Figure 6 This is a diagram showing some impedance effects in the embodiments;

[0042] Figure 7 The diagram shows the impedance effects of the remaining parts in the embodiment. DETAILED DESCRIPTION

[0043] The following will provide a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0044] The symbols in this application are defined as follows:

[0045] d and q represent the d-axis and q-axis respectively; s and l represent the power frequency side and low frequency side of the modular multilevel matrix converter (M3C), respectively;

[0046] The d-axis of the M3C power frequency side output port is referred to as the M3C power frequency side d-axis, denoted by the subscript sd;

[0047] The d-axis of the M3C low-frequency side output port is referred to as the M3C low-frequency side d-axis, denoted by the subscript ld;

[0048] The q-axis of the M3C power frequency side output port is referred to as the M3C power frequency side q-axis, denoted by the subscript sq;

[0049] The q-axis of the M3C low-frequency side output port is referred to as the M3C low-frequency side q-axis, denoted by the subscript lq;

[0050] Zkj represents the impedance generated by the influence of the k-axis on the j-axis, referred to as the mutual impedance of the k-axis and the j-axis; Zjj represents the impedance generated by the influence of the j-axis on itself, referred to as the j-axis self-impedance; k∈{sd,ld,sq,lq}, j∈{sd,ld,sq,lq}; sd, ld, sq and lq represent the d-axis on the power frequency side, the d-axis on the low frequency side, the q-axis on the power frequency side and the q-axis on the low frequency side, respectively.

[0051] This embodiment proposes a broadband impedance identification model for a flexible low-frequency power transmission system, which is used to identify the broadband impedance of a modular multilevel matrix converter (M3C) based on the operating current, voltage and frequency f of the modular multilevel matrix converter (M3C).

[0052] The working voltage of M3C includes the d-axis voltage U sd , power frequency side q-axis voltage U sq , low-frequency side d-axis voltage U ld and the low-frequency side q-axis voltage U lq ; s represents the power frequency side, l represents the low frequency side;

[0053] The working current of M3C includes the d-axis current I sd , power frequency side q-axis current I sq , low-frequency side d-axis current I ld and the low-frequency side q-axis current I lq ;

[0054] M3C's broadband impedance Z dq (freq) is written as:

[0055]

[0056] Among them, Zsdsd and Zsqsq represent the d-axis self-impedance and q-axis self-impedance of M3C on the power frequency side, respectively; Zldld and Zlqlq represent the d-axis self-impedance and q-axis self-impedance of M3C on the low frequency side, respectively; Zsdsq and Zsqsd represent the mutual impedance of the d-axis and q-axis on the power frequency side, respectively; Zldlq and Zlqld represent the mutual impedance of the d-axis and q-axis on the low frequency side, respectively; Zsdld is the mutual impedance of the d-axis on the power frequency side to the d-axis on the low frequency side. Zsdlq is the mutual impedance of the power frequency side d-axis to the low frequency side q-axis; Zsqld is the mutual impedance of the power frequency side q-axis to the low frequency side d-axis, and Zsqlq is the mutual impedance of the power frequency side q-axis to the low frequency side q-axis; Zldsd is the mutual impedance of the low frequency side d-axis to the power frequency side d-axis, and Zldsq is the mutual impedance of the low frequency side d-axis to the power frequency side q-axis; Zlqsd is the mutual impedance of the low frequency side q-axis to the power frequency side d-axis, and Zlqsq is the mutual impedance of the low frequency side q-axis to the power frequency side q-axis.

[0057] Reference Figure 1 The method for constructing a data set for training a broadband impedance identification model for a flexible low-frequency power transmission system includes the following steps:

[0058] Step S1: building an electromagnetic transient time-domain simulation model of a flexible low-frequency power transmission system based on electromagnetic transient simulation software or building a hardware-in-the-loop simulation model of a flexible low-frequency power transmission system based on real-time simulation;

[0059] Step S2: Set the output active power of the flexible low-frequency power transmission system simulation model, that is, set different steady-state operating points. The output active power range is 0 to 1 pu, representing the per-unit value. Specifically, when implementing, the steady-state operating points can be set at equal intervals within the output active power range, and the interval is n. In the embodiment, n = 0.005 is set.

[0060] Step S3: Collect the voltages and currents {U sd 、I sd 、U sq 、I sq 、U ld 、I ld 、U lq 、I lq} of the M3C power frequency side and low-frequency side in the synchronous rotating coordinate system (dq axis) at each steady-state operating point.

[0061] Step S4: Use the frequency scanning method to obtain the broadband impedance at different steady-state operating points, so as to construct a data set {U sd 、I sd 、U sq 、I sq 、U ld 、I ld 、U lq 、I lq 、f; Z dq (freq)}; where, {U sd 、I sd 、U sq 、I sq 、U ld 、I ld 、U lq 、I lq} determines the steady-state operating point, f is the scanning frequency, and Z dq (freq) is the broadband impedance obtained by scanning with the frequency f.

[0062] In this step, first determine the broadband impedance scanning frequency f of the M3C, 0 < f < 1000 Hz; obtain the broadband impedance Z dq (freq) obtained by scanning at different frequencies at each steady-state operating point through the frequency sweeping method, so as to construct a data set {U sd 、I sd 、U sq 、I sq 、U ld 、I ld 、U lq 、I lq 、f; Z dq (freq)}. Z dq (freq) includes the impedance amplitude and impedance phase angle.

[0063] In determining the working condition {U sd , I sd 、U sq , I sq 、U ld , I ld 、U lq , I lq} and scanning frequency f, the broadband impedance Z is obtained by frequency sweep dq The method of (freq) is as follows:

[0064] Set the amplitude m of the disturbance source at the steady-state operating point, inject a disturbance voltage of amplitude m into the specified position of M3C, scan the simulation model according to the scanning frequency f, and obtain the voltage response matrix U of M3C dq (freq) and the voltage response matrix I dq (freq); calculate broadband impedance Z dq (freq)=U dq (freq)·I dq (freq) -1 .

[0065] It is worth noting that the steady-state operating point and the working condition {U sd , I sd 、U sq , I sq 、U ld , I ld 、U lq , I lq Correspondingly, at the same steady-state operating point, the impedance Z obtained by scanning at different frequencies f is dq (freq) is different; therefore, multiple frequencies can be used for scanning at the same steady-state operating point, thereby obtaining multiple data set samples at the same steady-state operating point.

[0066] In the embodiment, m is 5% to 10% of the rated voltage of the system port, the scanning frequency f is in the range of (0-1000] Hz, and the scanning frequency interval g is 10 Hz.

[0067]

[0068] Among them, Usj1 and Isj1 respectively represent the voltage disturbance m=aU injected into the d-axis of the power frequency side of M3C when the q-axis injection voltage is 0. sd After that, the voltage response and current response generated by the j-axis of M3C; a is the rated coefficient, 5%≤a≤10%;

[0069] Ulj1 and Ilj1 represent the voltage disturbance m=aU injected into the d-axis of the low-frequency side of M3C when the q-axis injection voltage is 0. ldAfter that, the voltage response and current response generated by the j-axis of M3C;

[0070] Usj2 and Isj2 represent the voltage disturbance m=aU injected into the q-axis of the power frequency side of M3C when the d-axis injection voltage is 0. sq After that, the voltage response and current response generated by the j-axis of M3C;

[0071] Ulj1 and Ilj1 represent the voltage disturbance m=aU injected into the q-axis of the low-frequency side of M3C when the d-axis injection voltage is 0. lq After that, the voltage response and current response generated by the j-axis of M3C;

[0072] j∈{sd,ld,sq,lq}, sd, ld, sq and lq represent the d-axis on the power frequency side, the d-axis on the low frequency side, the q-axis on the power frequency side and the q-axis on the low frequency side, respectively.

[0073] Right now:

[0074] Ussd1 and Issd1 represent the voltage and current of the d-axis on the power frequency side of M3C after the d-axis voltage disturbance on the power frequency side is injected into M3C when the q-axis voltage is 0;

[0075] Ussq1 and Issq1 represent the voltage and current of the q-axis on the power frequency side of M3C after the d-axis voltage disturbance on the power frequency side is injected into M3C when the q-axis voltage is 0;

[0076] Usld1 and Isld1 represent the voltage and current of the d-axis on the low-frequency side of M3C after the d-axis voltage disturbance on the power frequency side is injected into M3C when the q-axis voltage is 0;

[0077] Uslq1 and Islq1 represent the voltage and current of the q-axis on the low-frequency side of M3C after the d-axis voltage disturbance on the power frequency side is injected into M3C when the q-axis voltage is 0;

[0078] Ulsd1 and Ilsd1 represent the voltage and current of the d-axis on the power frequency side of M3C after the d-axis voltage disturbance on the low frequency side is injected into M3C when the q-axis voltage is 0;

[0079] Ulsq1 and Ilsq1 represent the voltage and current of the q-axis on the power frequency side of M3C after the d-axis voltage disturbance on the low frequency side is injected into M3C when the q-axis voltage is 0;

[0080] Ulld1 and Illd1 represent the voltage and current of the d-axis on the low-frequency side of M3C after the d-axis voltage disturbance on the low-frequency side is injected into M3C when the q-axis voltage is 0;

[0081] Ullq1 and Illq1 represent the voltage and current of the q-axis on the low-frequency side of M3C after the d-axis voltage disturbance on the low-frequency side is injected into M3C when the q-axis voltage is 0;

[0082] Ussd2 and Issd2 represent the voltage and current of the d-axis on the power frequency side of M3C after the q-axis voltage disturbance on the power frequency side is injected into M3C when the d-axis voltage is 0;

[0083] Ussq2 and Issq2 represent the voltage and current on the q-axis of the power frequency side of M3C after the q-axis voltage disturbance on the power frequency side is injected into M3C when the d-axis voltage is 0;

[0084] Usld2 and Isld2 represent the voltage and current of the d-axis on the low-frequency side of M3C after the q-axis voltage disturbance on the power frequency side is injected into M3C when the d-axis voltage is 0;

[0085] Uslq2 and Islq2 represent the voltage and current of the q-axis on the low-frequency side of M3C after the q-axis voltage disturbance on the power frequency side is injected into M3C when the d-axis voltage is 0;

[0086] Ulsd2 and Ilsd2 represent the voltage and current of the d-axis on the power frequency side of M3C after the q-axis voltage disturbance on the low frequency side is injected into M3C when the d-axis voltage is 0;

[0087] Ulsq2 and Ilsq2 represent the voltage and current of the q-axis on the power frequency side of M3C after the q-axis voltage disturbance on the low frequency side is injected into M3C when the d-axis voltage is 0;

[0088] Ulld2 and Illd2 represent the voltage and current of the d-axis on the low-frequency side of M3C after the q-axis voltage disturbance on the low-frequency side is injected into M3C when the d-axis voltage is 0;

[0089] Ullq2 and Illq2 represent the voltage and current on the low-frequency side q-axis of M3C after the low-frequency side q-axis voltage disturbance is injected into M3C when the d-axis voltage is 0.

[0090] Reference Figure 2 The present embodiment proposes a method for training a broadband impedance identification model for a flexible low-frequency power transmission system, comprising the following steps:

[0091] SA1, the dataset {U sd , I sd 、U sq , I sq 、U ld , I ld 、U lq , I lq 、f;Z dq (freq)} normalization processing, the normalized data set is recorded as the normalized data set {U' sd 、I' sd 、U' sq 、I' sq 、U' ld 、I' ld 、U' lq、I' lq 、f;Z' dq (freq)}. The normalization method specifically adopts the minimum-maximum normalization method; U' sd 、I' sd 、U' sq 、I' sq 、U' ld 、I' ld 、U' lq 、I' lq and Z' dq (freq) are U sd , I sd 、U sq , I sq 、U ld , I ld 、U lq , I lq and Z dq Normalized value of (freq).

[0092] SA2, build the basic model and initialize it. The input of the basic model is {U' sd 、I' sd 、U' sq 、I' sq 、U' ld 、I' ld 、U' lq 、I' lq , f}, the output is Z' dq (freq);

[0093] The basic model can specifically adopt the KAN neural network model, which includes an inner layer (hidden layer) and an outer layer (output layer). The input data is forward propagated in the basic model. The specific forward propagation calculation process is as follows:

[0094] Input vector x={U' sd 、I' sd 、U' sq 、I' sq 、U' ld 、I' ld 、U' lq 、I' lq , f} is passed to each neuron in the hidden layer, and the univariate function output of all the input variables it receives is calculated and summed; the univariate activation function is applied to perform nonlinear transformation to obtain the hidden layer output; the weighted summation of the outputs of all hidden layer neurons is used to obtain the final output y = Z' dq (freq), specifically the normalized values ​​of the amplitude and phase angle of the broadband impedance data.

[0095]

[0096] Where x n represents the nth feature of the input vector x, ψ represents the hidden layer neuron, φ represents the univariate activation function; 2N+1 is the total number of neurons, N is the total number of features of x, and m is the neuron number; h' m is the transition parameter, φ m represents the activation function corresponding to neuron m; ψ m,n is the mth neuron pair x n The processing function, ψ m,n (x n ) is the mth neuron pair x n The processing result.

[0097]

[0098] in, The submodule capacitor voltage U representing M3C output by the model hidden layer dc The process quantity, The output of the hidden layer of the model represents the d-axis common mode voltage U at the low frequency side of M3C m3cd The process quantity, The output of the hidden layer of the model represents the q-axis common mode voltage U at the low frequency side of M3C m3cq The process quantity, The output of the hidden layer of the model represents the d-axis common mode voltage U of M3C on the power frequency side comd The process quantity, The output of the hidden layer of the model represents the q-axis common mode voltage U of M3C on the power frequency side comq process quantity.

[0099] SA3, extract learning samples from the normalized data set and substitute them into the basic model, calculate the loss function Loss, and reversely update the basic model according to the loss function;

[0100] The loss function is:

[0101] Loss=Loss a +Loss phy +Loss bc_ld +Loss bc_lq +Loss bc_sd +Loss bc_sq (2)

[0102] Among them, Loss a The model prediction loss is the combination of the output label y of the base model and the true label Z' dq The calculation loss of (freq), specifically the root mean square loss can be used;

[0103]

[0104] In the formula, NE represents the number of samples in a single batch, y i Represents the output of the basic model for the i-th learning sample, Z' dq (freq) i represents the true label of the i-th learning sample; ||·||2 represents the 2-normal calculation formula. When calculating specifically, first calculate the 2-normal of the amplitude and the 2-normal of the phase angle at each frequency, then add the calculation results at each frequency and calculate the mean as the loss Loss a .

[0105] Loss phy is the power conservation loss;

[0106]

[0107] Among them, U dci Represents the submodule capacitor voltage U of the i-th learning sample dc , the submodule capacitor voltage U in the flexible low-frequency power transmission system dc The following power conservation constraints are met:

[0108]

[0109] C dc Indicates the equivalent capacitance value of the submodule;

[0110] Loss bc_ld is the d-axis loss on the low-frequency side, Loss bc_lq is the q-axis loss on the low-frequency side;

[0111]

[0112] Among them, U m3cdi 、U m3cqi They represent the d-axis common mode voltage U of the M3C low-frequency side of the i-th learning sample respectively. m3cd and q-axis common mode voltage U m3cq ;

[0113] The calculation of the common mode voltage on the low-frequency side of the M3C in the flexible low-frequency power transmission system meets the following low-frequency side constraints:

[0114]

[0115]

[0116] L represents the equivalent inductance of the bridge arm of the flexible low-frequency power transmission system, w l Indicates the low-frequency side corner frequency of M3C.

[0117] Loss bc_sdis the d-axis loss on the power frequency side, Loss bc_sq is the q-axis loss on the power frequency side;

[0118]

[0119] Among them, U comdi 、U comqi They represent the d-axis common mode voltage U of the M3C power frequency side of the i-th learning sample respectively. comd and q-axis common mode voltage U comq ;

[0120] The calculation of the common mode voltage on the power frequency side of the M3C in the flexible low-frequency power transmission system meets the following power frequency side constraints:

[0121]

[0122] Among them, w s Indicates the angular frequency of the M3C power frequency side

[0123] and Both are the outputs of the hidden layer of the basic model and are the process quantities of impedance identification.

[0124] In this step SA3, after the loss function Loss is calculated, for each hidden layer neuron m and input variable x n , use the adaptive optimization algorithm to perform gradient backpropagation and parameter update on the basic model. The specific calculation formula is as follows:

[0125]

[0126] Among them, ψ m,n is the mth neuron pair x n The processing function, h' m is the transition parameter, y is the output of the basic model, h m is the activated value of the mth hidden layer neuron; θ represents the weight parameter of the updated univariate function, that is, the parameter to be updated in the basic model, and μ represents the parameter update learning rate.

[0127] SA4. Determine whether the basic model training has reached the convergence condition; the convergence condition is set as the number of basic model updates reaching the set value, or the difference in loss function over H consecutive adjacent rounds is less than the set threshold;

[0128] If not, return to step SA3;

[0129] If yes, the fixed basic model is used as the broadband impedance identification model for the flexible low-frequency power transmission system.

[0130] Reference Figure 3The present embodiment proposes a broadband impedance identification method for a flexible low-frequency power transmission system, comprising the following steps:

[0131] St1. Obtain the working condition sample of the flexible low-frequency transmission system at the specified steady-state operating point {U sd , I sd 、U sq , I sq 、U ld , I ld 、U lq , I lq};

[0132] St2, combined with the data set {U sd , I sd 、U sq , I sq 、U ld , I ld 、U lq , I lq 、f;Z dq (freq)} to normalize the working condition samples and construct the working condition input samples {U' sd 、I' sd 、U' sq 、I' sq 、U' ld 、I' ld 、U' lq 、I' lq ,f0};

[0133] St3, input the working condition input sample into the trained broadband impedance identification model, and the broadband impedance identification outputs the predicted impedance normalized value corresponding to different scanning frequencies f0;

[0134] Step 4, performing inverse normalization processing on the predicted impedance normalized value to obtain the predicted impedance, where the predicted impedance includes a predicted amplitude and a predicted phase angle;

[0135] St5. Summarize the predicted impedances at different scanning frequencies to obtain the broadband impedance-frequency characteristic curve required for the flexible low-frequency power transmission system.

[0136] The broadband impedance identification model is verified below with reference to specific embodiments.

[0137] In this embodiment, a KAN neural network is used as a basic model to train the impedance recognition model.

[0138] Reference Figure 5 In this embodiment, construct Figure 4 The flexible low-frequency power transmission system shown in FIG. sets multiple steady-state operating points and corresponding operating condition data {U sd , I sd、U sq , I sq 、U ld , I ld 、U lq , I lq}, simulate the frequency sweep in the PSCAD / EMTP environment to obtain the broadband impedance frequency characteristics of the flexible low-frequency transmission system; the sweep frequency f is swept at intervals of 10 Hz in the interval (0,1000] Hz, that is, at each steady-state operating point, the frequency is swept at 10 Hz, 20 Hz, 30 Hz, ..., 990 Hz, 1000 Hz respectively to obtain the broadband impedance at each frequency, thereby constructing the data set {U sd , I sd 、U sq , I sq 、U ld , I ld 、U lq , I lq 、f;Z dq (freq)}, and the data in the dataset U sd , I sd 、U sq , I sq 、U ld , I ld 、U lq , I lq 、Z dq After normalization of (freq), the normalized data set {U' sd 、I' sd 、U' sq 、I' sq 、U' ld 、I' ld 、U' lq 、I' lq 、f;Z' dq (freq)}.

[0139] In this embodiment, the normalized data set is divided into a training set and a test set. The above steps SA2-SA4 are used to train the KAN neural network as a broadband impedance identification model on the training set. Then, in the test set, the normalized sample {U' sd 、I' sd 、U' sq 、I' sq 、U' ld 、I' ld 、U' lq 、I' lq, f} input the broadband impedance identification model, and inversely normalize the output value to obtain the corresponding predicted impedance; during the test, by continuously changing the frequency f, the broadband impedance frequency characteristic at the steady-state operating point can be obtained.

[0140] In this embodiment, the simulated frequency sweep data on the test set is compared with the predicted data of the broadband impedance identification model. Figure 6 、 Figure 7 As shown in the figure, the red dots are the broadband impedance frequency characteristics obtained by frequency sweep simulation in the PSCAD / EMTP environment, that is, the broadband impedance marked on the test sample; the blue curve is the real-time broadband impedance frequency characteristic result calculated by the broadband impedance identification model. The two have a high degree of agreement, which proves the correctness of the broadband impedance online identification model of the flexible low-frequency transmission system.

[0141] Of course, it will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, but also encompasses the same or similar structures that can be implemented in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims, not the foregoing description, and it is intended that all variations that fall within the meaning and range of equivalents of the claims be encompassed within the present invention. Any reference signs in the claims should not be construed as limiting the claim to which they relate.

[0142] In addition, it should be understood that although this specification is described in terms of implementation methods, not every implementation method contains only one independent technical solution. This narrative method of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment can also be appropriately combined to form other implementation methods that can be understood by those skilled in the art.

[0143] The technology, shape, and structure not described in detail in the present invention are all well-known technologies.

Claims

1. A training method for a broadband impedance identification model of a flexible low-frequency power transmission system, characterized in that: The broadband impedance identification model identifies the broadband impedance of the modular multilevel matrix converter M3C based on its operating current, voltage and frequency f. The loss function during the training of the broadband impedance identification model is: Loss=Loss a +Loss phy +Loss bc_ld +Loss bc_lq +Loss bc_sd +Loss bc_sq Among them, Loss a Loss is the model prediction loss. phy is the power conservation loss; Loss bc_ld is the d-axis loss on the low-frequency side, Loss bc_lq is the q-axis loss on the low-frequency side; Loss bc_sd is the d-axis loss on the power frequency side, Loss bc_sq is the q-axis loss on the power frequency side; Power conservation loss, Loss phy The submodule capacitor voltage U dc On the calculation, U dc The power conservation constraint is satisfied: Among them, C dc Indicates the equivalent capacitance value of the M3C submodule; U sd and I sd Indicates the voltage and current of the d-axis on the power frequency side of M3C, U sq and I sq Indicates the voltage and current of the q-axis on the power frequency side of M3C, U ld and I ld represents the voltage and current of the d-axis on the low-frequency side of M3C, U lq and I lq Indicates the voltage and current on the d-axis of the M3C low-frequency side.

2. The method for training a broadband impedance identification model for a flexible low-frequency power transmission system according to claim 1, wherein: Loss bc_ld and Loss bc_lq The d-axis common mode voltage U on the low-frequency side of M3C is m3cd and q-axis common mode voltage U m3cq U m3cd and U m3cq Satisfy the low-frequency side constraints: Where L represents the equivalent inductance of the bridge arm of the flexible low-frequency power transmission system, w l Indicates the low-frequency side corner frequency of M3C; U ld and I ld represents the voltage and current of the d-axis on the low-frequency side of M3C, U lq and I lq Indicates the voltage and current on the d-axis of the M3C low-frequency side.

3. The method for training a broadband impedance identification model for a flexible low-frequency power transmission system according to claim 2, wherein: Loss bc_sd and Loss bc_sq The d-axis common mode voltage U comd and q-axis common mode voltage U comq On the calculation, U comd and U comq The following power frequency side constraints are met: Among them, w s Indicates the angular frequency of the M3C power frequency side, U sd and I sd Indicates the voltage and current of the d-axis on the power frequency side of M3C, U sq and I sq Indicates the voltage and current of the q-axis on the power frequency side of M3C.

4. The method for training a broadband impedance identification model for a flexible low-frequency power transmission system according to claim 3, wherein: Loss bc_ld Using M3C low-frequency side d-axis common mode voltage U m3cd The low-frequency side d-axis common mode voltage U represented by the model hidden layer output m3cd The cross entropy loss of the process quantity; Loss bc_lq Using M3C low-frequency side q-axis common mode voltage U m3cq The low-frequency side q-axis common mode voltage U represented by the model hidden layer output m3cq The cross entropy loss of the process quantity; Loss bc_sd Using M3C power frequency side d-axis common mode voltage U comd The d-axis common mode voltage U on the power frequency side is represented by the output of the hidden layer of the model comd The cross entropy loss of the process quantity; Loss bc_sq Using M3C power frequency side q-axis common mode voltage U comq The q-axis common mode voltage U on the power frequency side is represented by the output of the hidden layer of the model comq The cross entropy loss of the process quantity.

5. The method for training a broadband impedance identification model for a flexible low-frequency power transmission system according to claim 1, wherein: Loss phy Submodule capacitor voltage U using M3C dc The submodule capacitor voltage U representing the output of the hidden layer of the model dc The cross entropy loss of the process quantity.

6. The method for training a broadband impedance identification model for a flexible low-frequency power transmission system according to claim 1, wherein: The broadband impedance identification model adopts the KAN neural network model.

7. The method for training a broadband impedance identification model for a flexible low-frequency power transmission system according to any one of claims 1 to 7, characterized in that: The learning data set of the broadband impedance identification model is denoted as {U' sd 、I' sd 、U' sq 、I' sq 、U' ld 、I' ld 、U' lq 、I' lq 、f;Z' dq (freq)},U' sd 、I' sd 、U' sq 、I' sq 、U' ld 、I' ld 、U' lq 、I' lq and Z' dq (freq) are U sd , I sd 、U sq , I sq 、U ld , I ld 、U lq , I lq and Z dq Normalized value of (freq); U sd and I sd Indicates the voltage and current of the d-axis on the power frequency side of M3C, U sq and I sq Indicates the voltage and current of the q-axis on the power frequency side of M3C, U ld and I ld represents the voltage and current of the d-axis on the low-frequency side of M3C, U lq and I lq represents the voltage and current of the d-axis on the low-frequency side of M3C; Z dq (freq) is the broadband impedance of M3C; Loss a The predicted impedance normalized value and Z' output by the model dq The cross entropy loss of (freq).

8. A method for identifying the broadband impedance of a flexible low-frequency power transmission system using the training method for the broadband impedance identification model of the flexible low-frequency power transmission system according to claim 7, characterized in that: First, the broadband impedance identification model is trained, and then the dataset {U sd , I sd 、U sq , I sq 、U ld , I ld 、U lq , I lq 、f;Z dq (freq)} to normalize the working condition samples and construct the working condition input samples {U' sd 、I' sd 、U' sq 、I' sq 、U' ld 、I' ld 、U' lq 、I' lq The working condition input samples are combined with different scanning frequencies f0 and input into the broadband impedance identification model. The obtained impedance normalization value is inversely normalized to obtain the predicted impedance at each scanning frequency.

9. A broadband impedance control method using the broadband impedance identification method of the flexible low-frequency power transmission system according to claim 8, characterized in that: Obtain the target impedance obtained by scanning at different frequencies f under a specified steady-state operating point; Design the current and voltage data of the M3C power frequency side and low frequency side ports in the synchronous rotating coordinate system, and build a design scheme {U sd , I sd 、U sq , I sq 、U ld , I ld 、U lq , I lq , f}; normalize the design scheme based on the training data set of the broadband impedance identification model, and then input the broadband impedance identification model to obtain the predicted impedance corresponding to the design scheme; Adjust the current and voltage data in the design until the predicted impedance is consistent with the target impedance.

10. A broadband impedance identification system for a flexible low-frequency power transmission system, characterized in that: include: The data acquisition module is used to collect the current and voltage data of the M3C power frequency side and low frequency side ports in the synchronous rotating coordinate system at each steady-state operating point; The data processing module pre-processes the collected current and voltage data, filters the data samples, and obtains the corresponding broadband impedance through simulation calculation for the filtered data samples to construct the initial sample {U sd , I sd 、U sq , I sq 、U ld , I ld 、U lq , I lq 、f;Z dq (freq)},U sd and I sd Indicates the voltage and current of the d-axis on the power frequency side of M3C, U sq and I sq Indicates the voltage and current of the q-axis on the power frequency side of M3C, U ld and I ld represents the voltage and current of the d-axis on the low-frequency side of M3C, U lq and I lq represents the voltage and current of the d-axis on the low-frequency side of M3C; Z dq (freq) is the broadband impedance of M3C; f is the scanning frequency; The model training module normalizes the initial samples and obtains the data set {U' sd 、I' sd 、U' sq 、I' sq 、U' ld 、I' ld 、U' lq 、I' lq 、f;Z' dq (freq)},U' sd 、I' sd 、U' sq 、I' sq 、U' ld 、I' ld 、U' lq 、I' lq and Z' dq (freq) are U sd , I sd 、U sq , I sq 、U ld , I ld 、U lq , I lq and Z dq The normalized value of (freq); a neural network model is trained on the data set as a broadband impedance identification model; The impedance identification module obtains the current and voltage data {U sd , I sd 、U sq , I sq 、U ld , I ld 、U lq , I lq } and normalize it, then use the normalized value {U' sd 、I' sd 、U' sq 、I' sq 、U' ld 、I' ld 、U' lq 、I' lq } and frequency f are input into the broadband impedance identification model to obtain the corresponding broadband impedance Z dq (freq).

Images