Three-phase transformer parameter estimation method and device, computer equipment and medium

Abstract

The invention provides a three-phase transformer parameter estimation method and device, computer equipment and a medium, and belongs to the field of electrical control, and the method comprises the steps: obtaining synchronous phasor data of a target three-phase transformer; an improved LASSO algorithm is constructed; preprocessing the data through an improved LASSO algorithm; constructing a positive sequence component equivalent model of the target three-phase transformer; based on the positive sequence component equivalent model, determining a correlation among a plurality of circuit parameters of the target three-phase transformer, and determining the plurality of circuit parameters; and estimating structure parameters and magnetic circuit parameters of the target three-phase transformer based on the circuit parameters to obtain corresponding estimation results. Thus, complex specific experiments do not need to be carried out on the target three-phase transformer, on one hand, loss is low, the application range is wide, the parameters of the target three-phase transformer can be comprehensively estimated, on the other hand, better adaptability is achieved for complex conditions such as nonlinear characteristics of the transformer, and parameter evaluation is more accurate.

Description

Technical Field

[0001] The present invention belongs to the field of electrical control, and in particular relates to a three-phase transformer parameter estimation method, device, computer equipment and medium. Background Art

[0002] Transformer parameter estimation refers to determining various parameters of the transformer by certain methods, such as electrical parameters (resistance, reactance, inductance, etc.) and physical parameters (number of turns, column cross section, etc.). Traditional parameter estimation methods are mainly based on the physical model and electrical theory of the transformer. For example, the short-circuit test is used to determine the short-circuit impedance of the transformer, and the no-load test is used to determine the no-load loss and excitation reactance of the transformer.

[0003] These traditional methods are based on specific test conditions and theoretical formulas for calculation. Take the short-circuit test as an example. By short-circuiting one side of the transformer and applying the rated current on the other side, the voltage and power loss at this time are measured, and the short-circuit impedance is calculated according to the formula.

[0004] However, traditional parameter estimation methods usually rely on specific test conditions and theoretical formulas, which may not fully reflect the actual operating status of the transformer. Therefore, the accuracy of the results may be limited; or due to the influence of the actual transformer structure and materials, as well as the accuracy limitations of the measuring equipment, the calculated parameters such as short-circuit impedance may contain errors. On the other hand, the test process of traditional methods may be relatively complicated and time-consuming, resulting in relatively high equipment and labor costs. In addition, traditional methods mainly focus on certain specific parameters of the transformer, such as short-circuit impedance, no-load loss, etc., but cannot fully evaluate the overall performance of the transformer.

[0005] Therefore, in the prior art, the parameter estimation of three-phase transformers has high loss cost and limited scope of application, and cannot comprehensively evaluate the transformer performance. On the other hand, the parameter estimation accuracy is low and the error is large. Summary of the invention

[0006] In order to solve the problem of limited application scope and low accuracy of three-phase transformer parameter estimation in the prior art, the present invention provides a three-phase transformer parameter estimation method, device, computer equipment and medium.

[0007] In order to achieve the above object, the present invention provides the following technical solutions:

[0008] Firstly, a three-phase transformer parameter estimation method is provided, the method comprising:

[0009] Acquire the synchronized phasor data of voltage and current on each side of the target three-phase transformer;

[0010] The weight coefficient lateral quantity α of the LASSO algorithm is changed into a parameter column vector w, which is used to scale the linear model; the parameter column vector w is offset so that each element in the parameter column vector w is reduced by 1, and then the L1 norm is calculated to obtain the improved LASSO algorithm;

[0011] The synchronized phasor data are preprocessed by improving the LASSO algorithm to screen the correlation, and the screened synchronized phasor data are obtained.

[0012] Constructing a positive-sequence component equivalent model of a target three-phase transformer according to the screened synchronized phasor data;

[0013] Determine the correlation between multiple circuit parameters of the target three-phase transformer based on the positive-sequence component equivalent model, and substitute multiple groups of synchronous phasor data into the correlation to determine the multiple circuit parameters;

[0014] Based on the circuit parameters, the structural parameters and magnetic circuit parameters of the target three-phase transformer are estimated according to the transformer type, size and winding arrangement to obtain corresponding estimation results.

[0015] Optionally, before constructing the positive-sequence component equivalent model of the target three-phase transformer according to the synchronized phasor data, the synchronized phasor data is also preprocessed, including:

[0016] The time tags of each data sequence of the synchronized phasor data are mapped to a common time axis, and the points to be interpolated are calculated using a linear interpolation method:

[0017]

[0018] Among them, t′ is the time point interpolated at the same time interval, q′ is the interpolated electrical value corresponding to t′, and t1, t2, q1, and q2 are the actual collected time and electrical data;

[0019] Estimate missing time points and corresponding electrical values ​​using existing time series data points.

[0020] Optionally, the preprocessing of correlation screening of the synchronized phasor data by using the improved LASSO algorithm to obtain the screened synchronized phasor data includes:

[0021] Introduce a new variable x i , the synchronized phasor data is converted from nonlinear relationship variables to linear relationship, and the regression model is defined as:

[0022] Y=bw☉X;

[0023] Where b is the variable coefficient, w=(w1,w2,…,w n ) T,X=(x1,x2,…,x n );

[0024] The objective function of the improved LASSO algorithm based on the regression model is:

[0025]

[0026] The nonlinear relationship data in the synchronized phasor data is screened according to the target algorithm to obtain the screened synchronized phasor data.

[0027] Optionally, determining the correlation between multiple circuit parameters of the target three-phase transformer based on the positive-sequence component equivalent model includes:

[0028] The measurement equations for the three windings of high, medium and low are:

[0029]

[0030] in, Z1=R1+jX1(i=1,2,3) represents the voltage, current phasor and impedance of the high, middle and low sides of the transformer respectively; is the neutral point voltage; Z m is the excitation impedance; k 12 and k 13 are the transformation ratios of the high pressure side, the medium pressure side and the low pressure side respectively;

[0031] The measurement equations of the high, medium and low windings are transformed by term shifting and the neutral point voltage is eliminated to obtain the correlation between multiple circuit parameters:

[0032] A k x=B k ;

[0033] Among them, A k is a 4×6 matrix consisting of the current phasor and transformation ratio at time k, B k is a 4×1 matrix consisting of voltage phasors and transformation ratios at time k, and x is a parameter matrix consisting of multiple circuit parameters.

[0034] Optionally, the structural parameters include the number of turns, main column section, column and yoke length, and yoke and side column section, and the magnetic circuit parameters include magnetic flux density and magnetic field strength and no-load excitation current;

[0035] Based on the conventional setting of a three-phase transformer, the voltage per turn on the high-voltage side is determined as:

[0036]

[0037] Among them, V t is the voltage per turn on the high voltage side; S is the apparent power; C is the correction factor, whose value depends on the transformer type;

[0038] The number of turns of the high voltage side winding is:

[0039]

[0040] Among them, N1 is the number of turns on the high-voltage side; U1 is the rated voltage on the high-voltage side;

[0041] The calculation formula for the main leg column section is:

[0042]

[0043] Among them, A ml is the cross section of the main leg column, ω is the angular velocity, B max is the maximum magnetic flux density;

[0044] Assuming the main leg flux is symmetrical, the magnetic flux is:

[0045]

[0046] When ωt=π / 2, the magnetic flux of phase A reaches its maximum value. Substituting ωt=π / 2 into the magnetic flux density formula, we get:

[0047]

[0048] According to Φ = Β·Α, the formula for magnetic field strength is:

[0049]

[0050] Among them, Φ r For a , Φ b and Φ c are the magnetic flux of phase A, B and C respectively, Φ a1 , Φ b1 and Φ c1 are the maximum magnetic flux of phases A, B and C on the high voltage side, B a1 , B b1 and B c1 are the magnetic field strengths of phases A, B and C on the high voltage side respectively;

[0051] when When the instantaneous excitation current of phase A is a for:

[0052]

[0053] Among them, I no-load is the no-load current value, is the initial phase;

[0054] The formula for calculating the column and yoke length is:

[0055]

[0056] l1 is the length of the main column; l2 is the length of the yoke; H a1,b1,c1 are the magnetic field strength of phases A, B, and C on the high-voltage side respectively; N is the number of winding turns, i a1 is the A phase current on the high voltage side;

[0057] The calculation formula for the yoke cross section is:

[0058] A yo =A ml .

[0059] Secondly, a three-phase transformer parameter estimation device is provided, the device comprising:

[0060] An acquisition module, used for acquiring synchronized phasor data of voltage and current on each side of a target three-phase transformer;

[0061] A construction module is used to construct the LASSO algorithm, and the weight coefficient lateral quantity α is changed into a parameter column vector w for scaling the linear model; the parameter column vector w is offset so that each element in the parameter column vector w is reduced by 1, and then the L1 norm is calculated to obtain an improved LASSO algorithm;

[0062] A preprocessing module is used to preprocess the synchronization phasor data by performing correlation screening through an improved LASSO algorithm to obtain screened synchronization phasor data;

[0063] A parameter estimation module is used to construct a positive-sequence component equivalent model of a target three-phase transformer based on the screened synchronous phasor data; determine the correlation between multiple circuit parameters of the target three-phase transformer based on the positive-sequence component equivalent model, and substitute multiple groups of synchronous phasor data into the correlation to determine multiple circuit parameters; based on the circuit parameters, estimate the structural parameters and magnetic circuit parameters of the target three-phase transformer according to the transformer type, size and winding arrangement to obtain corresponding estimation results.

[0064] In addition, a computer-readable storage medium is provided, wherein the storage medium stores a computer program, and when the computer program is executed by a processor, the above-mentioned three-phase transformer parameter estimation method is implemented.

[0065] Finally, a computer device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned three-phase transformer parameter estimation method when executing the program.

[0066] The three-phase transformer parameter estimation method provided by the present invention has the following beneficial effects:

[0067] First, the LASSO algorithm is improved so that it can handle multi-parameter problems more flexibly, and the synchronous phasor data is optimized to improve its accuracy; secondly, a positive-sequence component equivalent model is constructed, which simplifies the complexity of the three-phase system and only considers the positive-sequence component, which helps to more accurately describe the behavior of the transformer under normal conditions; then, the circuit parameter values ​​are calculated by substituting multiple sets of synchronous phasor data, and the structural parameters and magnetic circuit parameters are estimated according to the transformer type, size and winding arrangement combined with the known circuit parameters. In this way, a comprehensive preliminary parameter estimation result of the target three-phase transformer including circuit parameters, structural parameters and magnetic circuit parameters can be obtained, and it can adapt to specific data characteristics, which helps to improve the accuracy and robustness of the algorithm and provide strong support for subsequent error correction; finally, the improved LASSO algorithm is used to correct the error of the estimation result, which can further reduce the estimation error and improve the accuracy of the parameters, which helps to ensure the reliability of transformer parameter estimation. In this way, through the above steps, there is no need to conduct complex specific experiments on the target three-phase transformer, and the normal operation of the transformer will not be interfered with. On the one hand, it has low loss and a wide range of applicability, and can comprehensively estimate the parameters of the target three-phase transformer, which is more extensive for the performance evaluation of the transformer. On the other hand, it has better adaptability to complex situations such as the nonlinear characteristics of the transformer, and the evaluation of parameters is more accurate, which not only helps to improve the design and optimization level of the transformer, but also can provide strong support for the operation and maintenance and troubleshooting of the transformer. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] In order to more clearly illustrate the embodiment of the present invention and its design scheme, the following briefly introduces the drawings required for this embodiment. The drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0069] Figure 1 The present invention is a flowchart of a three-phase transformer parameter estimation method provided according to an exemplary embodiment of the present invention.

[0070] Figure 2 The present invention is a flowchart of another three-phase transformer parameter estimation method provided according to an exemplary embodiment of the present invention.

[0071] Figure 3 It is a schematic diagram of an equivalent model of a positive-sequence component of a three-winding transformer provided according to an exemplary embodiment of the present invention.

[0072] Figure 4 A schematic diagram of a three-phase three-column transformer core and magnetic circuit according to an exemplary embodiment of the present invention; Figure 4 (a) is a schematic diagram of the structure of the iron core. Figure 4 (b) is a schematic diagram of the magnetic circuit.

[0073] Figure 5 The present invention is a block diagram of a three-phase transformer parameter estimation device according to an exemplary embodiment of the present invention. DETAILED DESCRIPTION

[0074] In order to enable those skilled in the art to better understand the technical solution of the present invention and implement it, the present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and cannot be used to limit the scope of protection of the present invention.

[0075] The present invention proposes a data-driven transformer parameter estimation method. First, a transformer parameter identification model is constructed based on a large amount of PMU measurement data (such as voltage, current, power and other data) after preprocessing. The state estimation method such as the least squares method can effectively identify and correct the key parameters of the transformer, such as circuit parameters such as resistance, reactance, inductance, etc. Secondly, based on the above detailed and corrected circuit parameters, and the corresponding measurement data, on the basis of fully considering the operating principle of the transformer and related physical laws, reasonable and practical assumptions can be used to accurately calculate the key parameters related to the transformer structure, such as core size, number of winding turns, winding arrangement, etc., and the magnetic circuit, including flux, magnetic resistance, flux density, etc., which are crucial for analyzing the performance and working state of the transformer. Finally, due to the existence of estimation errors, the accuracy of the relevant parameters calculated after the parameter adjustment still cannot meet the requirements. The improved LASSO algorithm is used to improve the data credibility, improve the stability of the model and the accuracy of variable selection. Based on the above parameter estimation method, a detailed algorithm flow is constructed, such as Figure 1 shown.

[0076] The technical solutions provided by various embodiments of the present invention are described in detail below in conjunction with the accompanying drawings.

[0077] First, the present invention provides a three-phase transformer parameter estimation method, specifically as follows Figure 2 As shown, the following steps are included:

[0078] S201. Acquire synchronized phasor data of voltage and current on each side of a target three-phase transformer.

[0079] In this step, the synchronous phasor data of the voltage and current on each side of the transformer are obtained through the phasor measurement unit (PMU), which relies on an accurate data synchronization mechanism. This may lead to differences in their sampling time intervals, resulting in deviations in the timing of the data. To solve this problem, this step uses a timing unification algorithm to align the time series data of multiple electrical quantities. The main purpose of this algorithm is to unify the time axis of each time series for more accurate and efficient data comparison and analysis. Its core method is to map the time tags of each data series of the synchronous phasor data to a common time axis to achieve temporal consistency.

[0080] Use linear interpolation method to calculate the points to be interpolated:

[0081]

[0082] Among them, t' is the time point interpolated at the same time interval, q' is the interpolated electrical value corresponding to t', and t1, t2, q1, and q2 are the actual collected time and electrical data. The missing time points and corresponding electrical values ​​are estimated through the existing time series data points. Similarly, the power data is also interpolated according to the above method to make the data have a consistent time interval and sampling rate to achieve time series unification.

[0083] S202. Construct an improved LASSO algorithm.

[0084] Specifically, we can first construct the LASSO algorithm, and change the weight coefficient lateral quantity α of the LASSO algorithm into a parameter column vector w for scaling the linear model; offset the parameter vector w, subtract 1 from each element in the vector w, and then calculate the L1 norm to obtain the improved LASSO algorithm.

[0085] The least absolute shrinkage and selection operator (LASSO) is a linear regression method widely used in statistics and machine learning. It is mainly used to solve the overfitting problem that may occur in traditional linear regression, and it can also perform variable selection.

[0086] In regression analysis, we usually try to find a set of coefficients to establish the relationship between the independent variable (feature) and the dependent variable. The traditional least squares method is to minimize the residual sum of squares between the predicted value and the true value to determine the coefficient. The LASSO algorithm adds a penalty term to the coefficient on this basis. This penalty term is based on the sum of the absolute values ​​of the coefficients. By adjusting the strength of the penalty term, some unimportant coefficients can be shrunk to zero, thereby achieving the purpose of variable selection.

[0087] The core of the traditional LASSO algorithm is to achieve the purpose of variable screening and model simplification by imposing penalties on coefficients. Suppose there are n observations and m features in the data set (X, Y). Among them, X = (x 1 ,x 2 ,…,x j ,…x m ), x j =(x 1j ,x 2j ,...,x nj ) T represents the independent variable, Y=(y1,y2,...,y i ,...,y n ) T , yi represents the dependent variable, i=1,2,...,n,j=1,2,...,m, and the independent variables are standardized and the dependent variables are centered. When performing linear regression on the dependent variable, the LASSO method limits the weight coefficient α=(α1,α2,…,α m )’s L1 norm must not exceed a given threshold λ. LASSO minimizes the residual sum of squares as:

[0088]

[0089] Wherein, α is an adjustable parameter. Specifically, LASSO aims to minimize the residual sum of squares while controlling the L1 norm of the coefficient through a parameter α. When the parameter α is set to be small, the coefficients of the variables with low correlation will be reduced to 0, thereby achieving feature screening. The present invention defines minimizing the residual sum of squares as the credibility of the data and the model. A small residual sum of squares indicates a high credibility of the model, whereas a small residual sum of squares indicates a low credibility.

[0090] The traditional LASSO algorithm is based on the above formula and is applied to feature selection and linear regression analysis of high-dimensional data. By introducing the L1 regularization term into the objective function to impose sparsity constraints, some elements in the parameter vector α in the above formula are made 0, thereby achieving feature selection and model simplification. The present invention proposes an improved LASSO algorithm to address the problem of processing low-credibility data points of actual three-phase transformers. The LASSO algorithm is used to identify and screen data points with low credibility in the case of one-dimensional data. The specific method is as follows.

[0091] Introduce a new variable x i , the data to be optimized can be converted from nonlinear relationship variables to linear relationships, and the regression model is defined as:

[0092] Y=bw☉X;

[0093] Where b is the variable coefficient, w=(w1,w2,…,w n )T ,X=(x1,x2,…,x n ).

[0094] Based on the regression model, the objective function of LASSO minimizing the residual sum of squares is rewritten as:

[0095]

[0096] It can be seen from the above formula that compared with the traditional LASSO algorithm, the improved LASSO algorithm proposed in the present invention has two characteristics: (1) The improved LASSO algorithm is based on the traditional LASSO algorithm, and the weight coefficient lateral quantity α is changed to a parameter column vector w, which is used to scale the linear model. Some coefficients in the formula are assumed to be b and introduced into the regression model formula to simplify the formula and facilitate the solution of the objective function. (2) The parameter vector w is offset, and each element in the vector w is reduced by 1, and then the L1 norm is calculated. By improving the LASSO algorithm, the deviation of certain elements in the parameter vector w relative to 1 is controlled. w i If it is close to 1, it means that the model minimizes the residual sum of squares, and the corresponding data is retained.

[0097] The improved LASSO algorithm is adaptive and automatically selects the most relevant features for the prediction target. The LASSO algorithm considers the correlation and weight between features, that is, the regularization coefficient λ and the regression coefficient w constrain each other, so as to more accurately identify low-confidence points. It is not just a simple judgment of outliers based on thresholds, but a screening by considering the statistical characteristics of the entire data set, which is more robust to outliers and outliers.

[0098] Use the gradient descent method to solve the parameters w and b in the objective function above. The gradient descent method is an optimization algorithm that uses the first-order information of the objective function and the gradient to perform iterations. By observing the changes in the loss value during the training process, we can determine whether the algorithm has converged. Use the gradient descent method to solve the values ​​of w and b when the loss function is minimized. i When it is equal to 1 or close to 1, the result is retained. If it deviates from 1 by a certain threshold range, it means that the data point is outlier and has low credibility. After identifying and eliminating the data with low credibility, the parameter adjustment method is applied again to improve the accuracy of the model. Through the selection of variables and the estimation of coefficients, the data is reconstructed or the credibility data is cleaned.

[0099] S203, preprocessing the synchronized phasor data by performing correlation screening by means of an improved LASSO algorithm to obtain screened synchronized phasor data.

[0100] First, the synchronized phasor data is used as input, and the appropriate improved LASSO algorithm model is selected according to the characteristics of the synchronized phasor data and the optimization goal. Then, the data is analyzed and processed through the model, for example, by selecting variables and estimating coefficients, reconstructing data or cleaning data. Finally, the processed data is obtained, and its credibility is improved compared to the original data. This data can be used for subsequent data analysis, parameter estimation, and other work.

[0101] Specifically, a new variable x is introduced into the synchronized phasor data. i , the synchronized phasor data is converted from nonlinear relationship variables to linear relationship, and the regression model is defined as:

[0102] Y=bw☉X;

[0103] Where b is the variable coefficient, w=(w1,w2,...,w n ) T ,X=(x1,x2,...,x n ).

[0104] The objective function of the improved LASSO algorithm based on the regression model is:

[0105]

[0106] According to the target algorithm, nonlinear relationship data in the synchronized phasor data is screened to obtain screened synchronized phasor data.

[0107] S204: construct a positive-sequence component equivalent model of the target three-phase transformer according to the synchronized phasor data.

[0108] In the analysis and research of power systems, the three-winding transformer, an important power equipment, can be characterized by using the positive sequence component equivalent model. Figure 3 shown.

[0109] exist Figure 3 middle, Z1=R1+jX1(i=1,2,3) represents the voltage, current phasor and impedance of the high, middle and low sides of the transformer respectively; is the neutral point voltage; Z m is the excitation impedance; k 12 and k 13 They are the transformation ratios of the high pressure side, medium pressure side and low pressure side respectively.

[0110] S205. Determine the correlation between multiple circuit parameters of the target three-phase transformer based on the positive-sequence component equivalent model, and substitute multiple groups of synchronous phasor data into the correlation to determine multiple circuit parameters.

[0111] In this step, we can Figure 3 The positive-sequence component equivalent model shown determines the correlation between multiple circuit parameters of the target three-phase transformer.

[0112] Among them, the measurement equations of the three windings of high, medium and low can be expressed as:

[0113]

[0114]

[0115]

[0116] in, Z1=R1+jX1(i=1,2,3) represents the voltage, current phasor and impedance of the high, middle and low sides of the transformer respectively; is the neutral point voltage; Z m is the excitation impedance; k 12 and k 13 They are the transformation ratios of the high pressure side, medium pressure side and low pressure side respectively.

[0117] Expand the voltage and current phasors of the three windings according to their real and imaginary parts, that is, And shifting the items we get:

[0118] I1R1cosθ1-I1X1sinθ1=e1-e0

[0119] I1R1sinθ1+I1X1cosθ1=f1-f0

[0120] I2R2cosθ2-I2X2sinθ2=e2-e0 / k 12

[0121] I2R2sinθ2+I2X2cosθ2=f2-f0 / k 12

[0122] I3R3cosθ3-I3X3sinθ3=e3-e0 / k 13

[0123] I3R3sinθ3+I3X3cosθ3=f3-f0 / k 13 ;

[0124] If the neutral point voltage is eliminated, it can be expressed as:

[0125] I1R1cosθ1-I1X1sinθ1-(k 12 I2R2cosθ2-k 12 I2X2sinθ2)=e1-k 12 e2

[0126] I1R1sinθ1+I1X1cosθ1-(k 12 I2R2sinθ2+k 12 I2X2cosθ2)=f1-k 12 f2

[0127] I1R1cosθ1-I1X1sinθ1-(k 13 I3R3cosθ3-k 13 I3X3sinθ3)=e1-k 13 e3

[0128] I1R1sinθ1+I1X1cosθ1-(k 13 I3R3sinθ3+k 13 I3X3cosθ3)=f1-k 13 f3;

[0129] For the above formula, the transformer parameters are x = [R1, X1, R2, X2, R3, X3] T , satisfies the linear equation at time k:

[0130] A k x=B k ;

[0131] Among them, A k is a 4×6 matrix consisting of the current phasor and transformation ratio at time k, B k It is a 4×1 matrix consisting of the voltage phasor and transformation ratio at time k.

[0132] Since PMU has the ability to obtain data at multiple different times, it is possible to write a linear equation as shown above. In this equation, the dimension of matrix A is 4n×6, and the dimension of matrix B is 4n×1. The data information contained in matrix A and matrix B here is closely related to the data collected by PMU at multiple times, and its matrix form can be expressed as:

[0133]

[0134] Using multiple sets of measurement data to obtain the parameters x of each transformer winding, the correlation between multiple circuit parameters can be used as the objective function:

[0135]

[0136] Substituting the measured multiple sets of synchronous phasor data into the above objective function, the circuit parameters of each winding of the transformer are obtained.

[0137] S206. Based on the circuit parameters, estimate the structural parameters and magnetic circuit parameters of the target three-phase transformer according to the transformer type, size and winding arrangement to obtain corresponding estimation results.

[0138] In this step, the structural parameters include the number of turns, main column section, column and yoke length, and yoke and side column section, and the magnetic circuit parameters include magnetic flux density and magnetic field strength as well as no-load excitation current.

[0139] First, the calculation of the number of turns.

[0140] The voltage per turn is closely related to the transformer type, winding arrangement, and insulation requirements. According to the empirical formula:

[0141]

[0142] Where V t is the voltage per turn on the high voltage side; S is the apparent power; C is the correction factor, whose value depends on the transformer structure type.

[0143] The number of turns of the high voltage side winding is:

[0144]

[0145] In the formula, N1 is the number of turns on the high-voltage side; U1 is the rated voltage on the high-voltage side.

[0146] The number of turns of the medium and low voltage side windings can be obtained from the voltage ratio:

[0147]

[0148]

[0149] Among them, N2 and N3 are the number of turns on the medium voltage and low voltage sides respectively; U2 and U3 are the rated voltages on the medium voltage and low voltage sides respectively.

[0150] The second aspect is the calculation of the main column cross section.

[0151] The voltage source applied to the transformer winding is as follows:

[0152]

[0153] Based on the above formula, the induced electromotive force can be defined as:

[0154]

[0155] according to We can get:

[0156]

[0157] In the formula, Φ ais the magnetic flux of phase A; Φ0 is the DC magnetic flux; N1 is the number of turns on the high-voltage side, and ω is the angular velocity.

[0158] The proposed estimation method takes the symmetrical flux of the three main branches into consideration. Due to the asymmetry of the flux paths, the flux of the three main branches cannot be completely symmetrical. However, for transformer design, the assumption of symmetry of the main branch flux is reasonable and acceptable. Then the maximum flux Φ max It can be expressed as:

[0159]

[0160] The maximum magnetic flux density under no-load and non-saturated conditions is:

[0161] Φ max =B max ·A ml ;

[0162] From the above two equations, we can know that the cross section of the main leg column can be expressed as follows:

[0163]

[0164] Among them, A ml is the cross section of the main leg column, ω is the angular velocity, B max is the maximum magnetic flux density.

[0165] The third aspect is the calculation of magnetic flux density and magnetic field strength.

[0166] Assuming the main leg flux is symmetrical, the flux density can be expressed as follows:

[0167]

[0168] When ωt=π / 2, the magnetic flux of phase A reaches its maximum value. Substituting ωt=π / 2 into the above formula, we get:

[0169]

[0170] According to Φ = Β·Α, we can get:

[0171]

[0172] Among them, Φ r For a , Φ b and Φ c are the magnetic flux of phase A, B and C respectively, Φ a1 , Φ b1 and Φ c1 The maximum magnetic flux of phase A, B and C respectively, B a1 , B b1 and Bc1 They are the magnetic field strengths of phases A, B and C on the high voltage side respectively.

[0173] Fourthly, calculation of no-load excitation current.

[0174] The no-load excitation current is closely related to the transformer size and is crucial to this method. The no-load excitation current can be obtained through the no-load test of the physical transformer. The calculation formula for the no-load current of phase A is:

[0175]

[0176] when When , the instantaneous excitation current of phase A can be calculated as:

[0177]

[0178] Among them, I no-load For an unloaded circuit, is the initial phase;

[0179] Fifth aspect, calculation of column and yoke length.

[0180] When calculating the length of the support and yoke, the present invention uses a three-leg transformer because there are fewer unknown variables. The core size and its magnetic circuit are as follows Figure 4 a and Figure 4 As shown in b. Figure 4 In a, l1 is the length of the main column; l2 is the length of the yoke; H a,b,c are the magnetic field strengths of phases A, B, and C respectively; Figure 4 b, F a,b,c are the magnetomotive force of phases A, B, and C respectively; R ma,mb,mc Expressed as magnetic resistance.

[0181] according to Figure 4 The magnetic circuit in b can be obtained as follows:

[0182]

[0183] For large power transformers, which are often not grounded, there are:

[0184] i a1 +i b1 +i c1 =0;

[0185] According to Maxwell's equation, we can get:

[0186] ∑ i N i I i =∫Hdl;

[0187] Combining the above two equations, we get:

[0188]

[0189] Assuming l1=l2, we have:

[0190]

[0191] Where N is the number of winding turns, i a1 、i b1 and i c1 They are the three-phase currents on the high voltage side respectively.

[0192] Sixth aspect: calculation of yoke and side column cross-section.

[0193] In the above content, the general calculation process of three-column and five-column transformers has been explained. The main column size and number of winding turns of three-column and five-column transformers are the same, but the yoke and side column sections of three-column and five-column transformers are different. The calculation of the yoke and side column sections will be discussed separately in two cases.

[0194] Three-phase three-winding three-leg transformer: For a three-leg transformer, since the magnetic flux of the leg and the corresponding yoke is the same, the cross-section of the leg and the yoke should have an area. The calculation formula for the cross-section of the yoke is as follows:

[0195] A yo =A ml .

[0196] Three-phase three-winding five-column transformer: For large power transformers, due to transportation restrictions, it is often necessary to reduce the height of the transformer. Using a five-branch iron core in a three-phase transformer can reduce the yoke section area of ​​the three-phase transformer to a 1 / 2 of the yoke section area of ​​the three-branch iron core. About, while keeping the full power of the transformer and the main column size unchanged. According to ABB's data sheet, the transformer design is approximated by the following formula:

[0197]

[0198] Among them, A yo and A sl They are the cross sections of the yoke and the jamb, respectively.

[0199] In addition, the improved LASSO algorithm can be used to correct the error of the estimation result, which can further reduce the estimation error and improve the accuracy of the parameters, which helps to ensure the reliability of transformer parameter estimation.

[0200] Using the above method, firstly, synchronous phasor data is obtained and an equivalent model of the positive-sequence component is constructed, which simplifies the complexity of the three-phase system. Only the positive-sequence component is considered, which helps to more accurately describe the behavior of the transformer under normal conditions; secondly, the circuit parameter values ​​are calculated by substituting multiple sets of synchronous phasor data, and the structural parameters and magnetic circuit parameters are estimated according to the transformer type, size and winding arrangement combined with the known circuit parameters. In this way, a comprehensive preliminary parameter estimation result of the target three-phase transformer including circuit parameters, structural parameters and magnetic circuit parameters can be obtained, and then the LASSO algorithm is improved so that it can handle multi-parameter problems more flexibly and adapt to specific data characteristics, which helps to improve the accuracy and robustness of the algorithm and provide strong support for subsequent error correction; finally, the improved LASSO algorithm is used to correct the error of the estimation result, which can further reduce the estimation error and improve the accuracy of the parameters, which helps to ensure the reliability of the transformer parameter estimation. In this way, through the above steps, there is no need to conduct complex specific experiments on the target three-phase transformer, and the normal operation of the transformer will not be interfered with. On the one hand, it has low loss and a wide range of applicability, and can comprehensively estimate the parameters of the target three-phase transformer, which is more extensive for the performance evaluation of the transformer. On the other hand, it has better adaptability to complex situations such as the nonlinear characteristics of the transformer, and the evaluation of parameters is more accurate, which not only helps to improve the design and optimization level of the transformer, but also can provide strong support for the operation and maintenance and troubleshooting of the transformer.

[0201] Secondly, the present invention also provides a three-phase transformer parameter estimation device, such as Figure 5 As shown, including:

[0202] An acquisition module 501 is used to acquire the synchronous phasor data of the voltage and current on each side of the target three-phase transformer;

[0203] A construction module 502 is used to construct a LASSO algorithm, and the weight coefficient lateral quantity α is changed into a parameter column vector w for scaling the linear model; the parameter column vector w is offset so that each element in the parameter column vector w is reduced by 1, and then the L1 norm is calculated to obtain an improved LASSO algorithm;

[0204] A preprocessing module 503 is used to preprocess the synchronized phasor data by performing correlation screening using an improved LASSO algorithm to obtain screened synchronized phasor data;

[0205] The parameter estimation module 504 is used to construct a positive-sequence component equivalent model of the target three-phase transformer based on the screened synchronous phasor data; determine the correlation between multiple circuit parameters of the target three-phase transformer based on the positive-sequence component equivalent model, and substitute multiple groups of synchronous phasor data into the correlation to determine multiple circuit parameters; based on the circuit parameters, estimate the structural parameters and magnetic circuit parameters of the target three-phase transformer according to the transformer type, size and winding arrangement to obtain corresponding estimation results.

[0206] By using the above-mentioned device, firstly, synchronous phasor data is acquired and an equivalent model of the positive-sequence component is constructed, which simplifies the complexity of the three-phase system and only considers the positive-sequence component, which helps to more accurately describe the behavior of the transformer under normal conditions; secondly, the circuit parameter values ​​are calculated by substituting multiple sets of synchronous phasor data, and the structural parameters and magnetic circuit parameters are estimated according to the transformer type, size and winding arrangement combined with the known circuit parameters, so that a comprehensive preliminary parameter estimation result of the target three-phase transformer including circuit parameters, structural parameters and magnetic circuit parameters can be obtained, and then the LASSO algorithm is improved so that it can handle multi-parameter problems more flexibly and adapt to specific data characteristics, which helps to improve the accuracy and robustness of the algorithm and provide strong support for subsequent error correction; finally, the improved LASSO algorithm is used to perform error correction on the estimation results, which can further reduce the estimation error and improve the accuracy of the parameters, which helps to ensure the reliability of the transformer parameter estimation. In this way, through the above steps, there is no need to conduct complex specific experiments on the target three-phase transformer, and the normal operation of the transformer will not be interfered with. On the one hand, it has low loss and a wide range of applicability, and can comprehensively estimate the parameters of the target three-phase transformer, which is more extensive for the performance evaluation of the transformer. On the other hand, it has better adaptability to complex situations such as the nonlinear characteristics of the transformer, and the evaluation of parameters is more accurate, which not only helps to improve the design and optimization level of the transformer, but also can provide strong support for the operation and maintenance and troubleshooting of the transformer.

[0207] The present invention also provides a computer-readable storage medium, which stores a computer program, which can be used to execute the above Figure 1 The steps of a three-phase transformer parameter estimation method are provided.

[0208] The present invention also provides a computer device. At the hardware level, the computer device includes a processor, an internal bus, a network interface, a memory, and a non-volatile memory. Of course, it may also include hardware required for other services. The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it to achieve the above Figure 1 The steps of a three-phase transformer parameter estimation method are provided.

[0209] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0210] The present invention is described with reference to flowcharts and / or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as a combination of processes and / or blocks 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.

[0211] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0212] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0213] It should be noted that the above specific implementation method can enable those skilled in the art to understand the invention more comprehensively, but does not limit the invention in any way. Therefore, although the present invention has been described in detail in this specification, those skilled in the art should understand that the invention can still be modified or replaced by equivalents; and all technical solutions and improvements that do not deviate from the spirit and scope of the invention are included in the protection scope of the patent for the invention. Any figure mark in the claims should not be regarded as limiting the claims involved.

Claims

1. A three-phase transformer parameter estimation method, characterized in that: The method comprises: Acquire the synchronized phasor data of voltage and current on each side of the target three-phase transformer; The weight coefficient lateral quantity α of the LASSO algorithm is changed into a parameter column vector w, which is used to scale the linear model; the parameter column vector w is offset so that each element in the parameter column vector w is reduced by 1, and then the L1 norm is calculated to obtain the improved LASSO algorithm; The synchronized phasor data are preprocessed by improving the LASSO algorithm to screen the correlation, and the screened synchronized phasor data are obtained. Constructing a positive-sequence component equivalent model of a target three-phase transformer according to the screened synchronized phasor data; Determine the correlation between multiple circuit parameters of the target three-phase transformer based on the positive-sequence component equivalent model, and substitute multiple groups of synchronous phasor data into the correlation to determine the multiple circuit parameters; Based on the circuit parameters, the structural parameters and magnetic circuit parameters of the target three-phase transformer are estimated according to the transformer type, size and winding arrangement to obtain corresponding estimation results.

2. A three-phase transformer parameter estimation method according to claim 1, characterized in that: Before constructing the positive sequence component equivalent model of the target three-phase transformer according to the synchronized phasor data, the synchronized phasor data is preprocessed, including: The time tags of each data sequence of the synchronized phasor data are mapped to a common time axis, and the points to be interpolated are calculated using a linear interpolation method: Among them, t′ is the time point interpolated at the same time interval, q′ is the interpolated electrical value corresponding to t′, and t1, t2, q1, and q2 are the actual collected time and electrical data; Estimate missing time points and corresponding electrical values ​​using existing time series data points.

3. A three-phase transformer parameter estimation method according to claim 1, characterized in that: The preprocessing of correlation screening of synchronized phasor data by improving the LASSO algorithm to obtain screened synchronized phasor data includes: Introduce a new variable x i , the synchronized phasor data is converted from nonlinear relationship variables to linear relationship, and the regression model is defined as: Y=bw☉X; Where b is the variable coefficient, w=(w1,w2,...,w n ) T ,X=(x1,x2,…,x n ); The objective function of the improved LASSO algorithm based on the regression model is: The nonlinear relationship data in the synchronized phasor data is screened according to the target algorithm to obtain the screened synchronized phasor data.

4. A three-phase transformer parameter estimation method according to claim 1, characterized in that: Determining the correlation between multiple circuit parameters of the target three-phase transformer based on the positive sequence component equivalent model includes: The measurement equations for the three windings of high, medium and low are: in, Z1=R1+jX1(i=1,2,3) represents the voltage, current phasor and impedance of the high, middle and low sides of the transformer respectively; is the neutral point voltage; Z m is the excitation impedance; k 12 and k 13 are the transformation ratios of the high pressure side, the medium pressure side and the low pressure side respectively; The measurement equations of the high, medium and low windings are transformed by shifting terms and eliminating the neutral point voltage to obtain the correlation between multiple circuit parameters: A k x=B k ; Among them, A k is a 4×6 matrix consisting of the current phasor and transformation ratio at time k, B k is a 4×1 matrix consisting of voltage phasors and transformation ratios at time k, and x is a parameter matrix consisting of multiple circuit parameters.

5. A three-phase transformer parameter estimation method according to claim 1, characterized in that: The structural parameters include the number of turns, main column section, column and yoke length, and yoke and side column section, and the magnetic circuit parameters include magnetic flux density and magnetic field strength and no-load excitation current; Based on the conventional setting of a three-phase transformer, the voltage per turn on the high-voltage side is determined as: Among them, V t is the voltage per turn on the high voltage side; S is the apparent power; C is the correction factor, whose value depends on the transformer type; The number of turns of the high voltage side winding is: Among them, N1 is the number of turns on the high-voltage side; U1 is the rated voltage on the high-voltage side; The calculation formula for the main leg column section is: Among them, A ml is the cross section of the main leg column, ω is the angular velocity, B max is the maximum magnetic flux density; Assuming the main leg flux is symmetrical, the magnetic flux is: When ωt=π / 2, the magnetic flux of phase A reaches its maximum value. Substituting ωt=π / 2 into the magnetic flux density formula, we get: According to Φ = Β·Α, the formula for magnetic field strength is: Among them, Φ r For a , Φ b and Φ c are the magnetic flux of phase A, B and C respectively, Φ a1 , Φ b1 and Φ c1 are the maximum magnetic flux of phases A, B and C on the high voltage side, B a1 , B b1 and B c1 are the magnetic field strengths of phases A, B and C on the high voltage side respectively; when When the instantaneous excitation current of phase A is a for: Among them, I no-load is the no-load current value, is the initial phase; The formula for calculating the column and yoke length is: l1 is the length of the main column; l2 is the length of the yoke; H a1,b1,c1 are the magnetic field strength of phases A, B, and C on the high-voltage side respectively; N is the number of winding turns, i a1 is the A phase current on the high voltage side; The calculation formula for the yoke cross section is: A yo =A ml 。 6. A three-phase transformer parameter estimation device, characterized in that: The device comprises: An acquisition module, used for acquiring synchronized phasor data of voltage and current on each side of a target three-phase transformer; A construction module is used to construct the LASSO algorithm, and the weight coefficient lateral quantity α is changed into a parameter column vector w for scaling the linear model; the parameter column vector w is offset so that each element in the parameter column vector w is reduced by 1, and then the L1 norm is calculated to obtain an improved LASSO algorithm; A preprocessing module is used to preprocess the synchronization phasor data by performing correlation screening through an improved LASSO algorithm to obtain screened synchronization phasor data; A parameter estimation module is used to construct a positive-sequence component equivalent model of a target three-phase transformer based on the screened synchronous phasor data; determine the correlation between multiple circuit parameters of the target three-phase transformer based on the positive-sequence component equivalent model, and substitute multiple groups of synchronous phasor data into the correlation to determine multiple circuit parameters; based on the circuit parameters, estimate the structural parameters and magnetic circuit parameters of the target three-phase transformer according to the transformer type, size and winding arrangement to obtain corresponding estimation results.

7. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

8. A computer device, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the method according to any one of claims 1 to 5 when executing the program.

Images