A capacitive voltage transformer error online detection method

Abstract

The present invention discloses an online detection method for capacitor voltage transformer error in the field of electric power technology. The method collects the secondary output signal of a three-phase capacitor voltage transformer in a normal operating state after commissioning to construct a standard amplitude data matrix X. The standard amplitude data matrix X is processed based on an adaptive blind source separation algorithm to calculate and obtain a statistical threshold value Q of the standard amplitude data matrix of the three-phase capacitor voltage transformer. cα ; Collect the real-time measurement data of the three-phase capacitor voltage transformer under the running state, construct the corresponding original amplitude data matrix, and calculate the statistic Q of the original amplitude data matrix of the three-phase capacitor voltage transformer under the running state based on the adaptive blind source separation algorithm; judge the statistic Q and the statistic threshold Q cα The present invention can continuously update the model data and perform long-term effective and accurate capacitor voltage transformer state evaluation.

Description

Technical Field

[0001] The invention relates to the technical field of electric power, and in particular to an online error detection method for a capacitive voltage transformer. Background Art

[0002] Capacitor voltage transformers have been widely used in high-voltage power grids of 110KV and above. As a kind of meter and relay protection equipment, their errors and reliability are related to the safe operation of the power system. Compared with traditional electromagnetic voltage transformers, capacitor voltage transformers have the advantages of small size, light weight, and the ability to avoid grid resonance caused by core saturation. But at the same time, due to the addition of components such as capacitor voltage dividers and compensating reactors, the measurement accuracy will be affected by more factors. For the error monitoring of capacitor voltage transformers, the existing method generally adopts regular offline detection. Since the calibration cycle is relatively long, generally 4 years, the phenomenon of error exceeding the tolerance will occur, which will endanger the operation of the power system. In addition, the difficulty of power outage operation in high-voltage power grids will have a greater impact on the continuous operation of substations. In order to timely and accurately detect the errors of capacitor voltage transformers in operation, it is necessary to propose an online error monitoring method that is separated from the standard. Summary of the invention

[0003] The purpose of the present invention is to propose an online monitoring method for capacitor voltage transformer error, which can realize online monitoring of capacitor voltage transformer error without a standard device. The present invention only performs correlation analysis on the collected data matrix based on the output amplitude data of the capacitor voltage transformer, and performs online evaluation of the capacitor voltage transformer state by comparing the statistical value with the size of the statistical threshold.

[0004] To achieve the above object, the present invention provides the following technical solutions:

[0005] A capacitive voltage transformer error online detection method, the method is as follows:

[0006] Collect the secondary output signal of the three-phase capacitor voltage transformer in normal operation after commissioning, and construct the standard amplitude data matrix X;

[0007] Based on the adaptive blind source separation algorithm, the standard amplitude data matrix X is processed to calculate the statistical threshold Q of the standard amplitude data matrix of the three-phase capacitor voltage transformer. cα ;

[0008] Collect real-time measurement data of the three-phase capacitor voltage transformer under the running state, construct the corresponding original amplitude data matrix, and calculate the statistics Q of the original amplitude data matrix of the three-phase capacitor voltage transformer under the running state based on the adaptive blind source separation algorithm;

[0009] Judging statistic Q and statistic threshold Qcα Size:

[0010] If Q≤Q cα , it is judged that the three-phase capacitive voltage transformer is operating normally at this time;

[0011] If Q>Q cα , it is judged that the measurement error of the three-phase capacitive voltage transformer is in an abnormal state at this time.

[0012] In some embodiments, the standard amplitude data matrix X is represented as:

[0013]

[0014] Where n represents the number of measurements of the three-phase capacitor voltage transformer parameters, x i1 Represents the amplitude data of the A-phase capacitor voltage transformer, x i2 Represents the amplitude data of the B-phase capacitor voltage transformer, x i3 Represents the amplitude data of the C-phase capacitive voltage transformer, 1≤i≤n.

[0015] In some embodiments, the statistical threshold Q of the three-phase capacitor voltage transformer standard amplitude data matrix cα The calculation steps are as follows:

[0016] Step 1: Matrix X0 is obtained by processing the standard amplitude data matrix X with zero mean;

[0017] Step 2: Whiten the matrix X0 to obtain the whitened matrix W:

[0018] W=D -1 / 2 U T

[0019] Where D represents a diagonal matrix, U represents the eigenvector of each column, and is an orthogonal matrix;

[0020] Step 3: Obtain the observed signal vector X0(t) in the mixed signal, X0(t)=As(t), where s(t) is the source signal vector and A is a mixing matrix composed of various coefficients;

[0021] make

[0022] Then Z=WX0(t)=WAs(t)=U T s(t);

[0023] in represents the estimated value of the source signal, Z represents the matrix obtained after the matrix X0 is whitened, and E{ZZ T}=I;B represents the separation matrix;

[0024] Step 4: Update the separation matrix B in real time through the update formula, which is expressed as:

[0025]

[0026] Where g() is the derivative of the non-quadratic function G, λ t represents a positive adaptive compensation coefficient;

[0027] Step 5: Calculate the statistic Q under healthy operation status c , the calculation formula is:

[0028] Q c =s ec T s ec

[0029] s ec Indicates the source signal matrix columns mapped to the columns with smaller matrix two norms of the mixing matrix A under normal operation;

[0030] Step 6: Based on the amplitude data of the three-phase capacitor voltage transformer under normal operating conditions, the kernel density method is used to calculate the statistic Q when the significance level is α. c The control limits are:

[0031]

[0032] Q cα =F -1 (1-α)

[0033] in, is the probability density distribution function expression, Q cα is the statistic Q when the significance level is α c Control limits.

[0034] In some embodiments, step 2 specifically includes:

[0035] Calculate the covariance matrix R of the matrix X0: R = E{x0(t)x′0(t) T};

[0036] Decompose the eigenvalues ​​of the covariance matrix R to obtain the eigenvector: R = UDU T ;

[0037] According to the whitening formula, the whitening matrix W is obtained: W = D -1 / 2 U T .

[0038] In some embodiments, the method further includes substituting the real-time measurement data of the three-phase capacitor voltage transformer under the operating state, repeating steps 1 to 5, and obtaining the statistic Q of the original amplitude data matrix of the three-phase capacitor voltage transformer under the operating state. In step 5, Q=s e T s e ,s e It represents the source signal matrix column mapped by the column with smaller matrix bi-norm of matrix A in the running state.

[0039] Beneficial effects: The present invention establishes a new comparison standard quantity based on the time-invariant physical relationship during the operation of the power grid. The three-phase output of the capacitor voltage transformer has correlation. The metering error is separated from the primary power grid fluctuation based on adaptive blind source separation, and the metering error information is quantified to establish a new detection standard quantity. The operating state of the capacitor voltage transformer is judged by comparing with the quantized value of the metering error under normal operating conditions. In addition, the present invention can continuously update the model data according to the obtained data, and perform long-term effective and accurate capacitor voltage transformer state evaluation, which solves the problem that other online detection methods in the prior art can only perform short-term capacitor voltage transformer state evaluation. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 It is the online monitoring flow chart of the present invention;

[0041] Figure 2 This is a flow chart of error state evaluation of a three-phase capacitor voltage transformer according to the present invention. DETAILED DESCRIPTION

[0042] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0043] A capacitive voltage transformer error online detection method, such as Figure 1-2 As shown, the method is as follows:

[0044] The secondary output signal of the three-phase capacitor voltage transformer in normal operation after commissioning is collected to construct a standard amplitude data matrix X; the standard amplitude data matrix X is expressed as:

[0045]

[0046] Where n represents the number of measurements of the three-phase capacitor voltage transformer parameters, x i1 Represents the amplitude data of the A-phase capacitor voltage transformer, xi2 Represents the amplitude data of the B-phase capacitor voltage transformer, x i3 Represents the amplitude data of the C-phase capacitive voltage transformer, 1≤i≤n.

[0047] Based on the adaptive blind source separation algorithm, the standard amplitude data matrix X is processed to calculate the statistical threshold Q of the three-term capacitor voltage transformer standard amplitude data matrix under normal operating conditions. cα ; Specifically, the following steps are included:

[0048] Step 1: The matrix X0 is obtained after zero mean processing of the standard amplitude data matrix X, that is, each standard amplitude data is subtracted from the homogenization of this batch of data to obtain zero mean data. Adaptive blind source separation is applicable to zero mean signals.

[0049] Step 2: Normal operation The output data of the capacitor voltage transformer is correlated. By whitening the matrix X0, the second-order statistical correlation between the data can be eliminated, the subsequent calculations can be simplified, and the algorithm performance can be improved:

[0050] The whitening process is to remove the correlation through linear transformation. The covariance matrix R of the matrix X0 can be set as:

[0051] R=E{x0(t)x′0(t) T};

[0052] Decompose the eigenvalues ​​of the covariance matrix R to obtain the eigenvector: R = UDU T ;

[0053] Where D represents a diagonal matrix, U represents the eigenvectors of each column of D, and is an orthogonal matrix.

[0054] According to the whitening formula, the whitening matrix W is obtained: W = D -1 / 2 U T .

[0055] Step 3: Separate the metering error information from the primary power grid fluctuation, thereby quantifying the metering error information.

[0056] The observed signal vector X0(t) in the mixed signal is obtained, X0(t) = As(t), where s(t) is the source signal vector, and A is a mixing matrix composed of various coefficients; each source signal in X0(t) is independent.

[0057] make

[0058] Represents the estimated value of the source signal.

[0059] Then Z=WX0(t)=WAs(t)=U Ts(t);

[0060] Where Z represents the matrix obtained after whitening the matrix X0, and E{ZZ T} = I;

[0061] Step 4: Update the separation matrix B in real time through the update formula. In the adaptive blind source separation algorithm, the whitening matrix and the orthogonalization matrix are updated in real time, so the separation matrix B is obtained by combining the two. In the case of continuous operation of the capacitor voltage transformer, the online status of the capacitor voltage transformer can be continuously evaluated and monitored by continuously collecting and updating the measurement error information data.

[0062] The update formula of the whitening matrix is:

[0063] W t+1 =W t -λ t [ZZ T -I]W t

[0064] Among them, λ t is a positive adaptive compensation coefficient.

[0065] The update formula of the orthogonal matrix is:

[0066]

[0067] in, It can be defined as:

[0068]

[0069] Where f is any differentiable function. t+1 =U t+1 W t+1 , we can get the update formula of the separation matrix:

[0070]

[0071] In order to reduce the impact of abnormal data on program stability without changing the basic properties of the program, the separation matrix update formula is normalized:

[0072]

[0073] Here, g() is the derivative of a non-quadratic function G.

[0074] Step 5: Establish a new test standard and calculate the statistic Q under the healthy operation state c , the calculation formula is:

[0075] Q c =s ecT s ec

[0076] s ec Indicates the source signal matrix columns mapped to the columns with smaller matrix two norms of the mixing matrix A under normal operation;

[0077] Step 6: Based on the amplitude data of the three-phase capacitor voltage transformer under normal operating conditions, the kernel density method is used to calculate the statistic Q when the significance level is α. c The control limits are:

[0078]

[0079] Q cα =F -1 (1-α)

[0080] in, is the probability density distribution function expression, Q cα is the statistic Q when the significance level is α c Control limits.

[0081] Collect the real-time measurement data of the three-phase capacitor voltage transformer under the running state, obtain the corresponding original amplitude data matrix, substitute the real-time measurement data of the three-phase capacitor voltage transformer under the running state, repeat steps 1 to 5, and obtain the statistics Q of the original amplitude data matrix of the three-phase capacitor voltage transformer under the running state. In step 5, Q = s e T s e ,s e It represents the source signal matrix column mapped by the column with smaller matrix bi-norm of matrix A in the running state.

[0082] Finally, the statistic Q and the statistic threshold Q are determined cα Size:

[0083] If Q cα , it is judged that the three-phase capacitive voltage transformer is operating normally at this time;

[0084] If Q>Q cα , it is judged that the measurement error of the three-phase capacitive voltage transformer is in an abnormal state at this time.

[0085] The adaptive blind source separation and adaptive updating model data adopted by the present invention improve the accuracy of the amplitude change that can be evaluated, and can effectively perform online evaluation of the state of the capacitor voltage transformer for a long time, and facilitate the operation and maintenance of the capacitor voltage transformer in a timely manner. The present invention can effectively avoid the loss of manpower, material resources, and financial resources, and is of great significance for ensuring the safe and economical operation of the power system.

[0086] ​Although this specification is described according to implementation modes, not every implementation mode includes only one independent technical solution. This description 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 may also be appropriately combined to form other implementation modes that can be understood by those skilled in the art.

[0087] Therefore, the above description is only a preferred embodiment of the present application and is not intended to limit the scope of implementation of the present application; that is, all equivalent changes made according to the scope of the claims of the present application are within the protection scope of the claims of the present application.

Claims

1. A capacitive voltage transformer error online detection method, characterized in that: Here’s how: Collect the secondary output signal of the three-phase capacitor voltage transformer in normal operation after commissioning and construct the standard amplitude data matrix ; Based on the adaptive blind source separation algorithm, the standard amplitude data matrix Processing, calculation to obtain the statistical threshold of the standard amplitude data matrix of the three-phase capacitor voltage transformer ; The real-time measurement data of the three-phase capacitor voltage transformer under the running state is collected, the corresponding original amplitude data matrix is ​​constructed, and the statistics of the original amplitude data matrix of the three-phase capacitor voltage transformer under the running state are calculated based on the adaptive blind source separation algorithm. ; Judgment Statistics With statistical threshold Size: like , it is judged that the three-phase capacitive voltage transformer is operating normally at this time; like , it is judged that the metering error of the three-phase capacitive voltage transformer is in an abnormal state at this time; Among them, the statistical threshold of the standard amplitude data matrix of the three-phase capacitor voltage transformer is The calculation steps are as follows: Step 1: Standard amplitude data matrix The matrix obtained after zero mean processing ; Step 2: Matrix Perform whitening to obtain the whitening matrix : Where D represents a diagonal matrix, U represents the eigenvector of each column, and is an orthogonal matrix; Step 3: Obtain the observed signal vector in the mixed signal , , where s(t) is the source signal vector and A is the mixing matrix composed of the coefficients; make , but ; in represents the estimated value of the source signal, Z represents the matrix The matrix obtained after whitening is ; B represents the separation matrix; Step 4: Update the separation matrix B in real time through the update formula, which is expressed as: in, Is a non-quadratic function The derivative of represents a positive adaptive compensation coefficient; Step 5: Calculate statistics under healthy operating conditions , the calculation formula is: Indicates the source signal matrix columns mapped to the columns with smaller matrix two norms of the mixing matrix A under normal operation; Step 6: According to the amplitude data of the three-phase capacitor voltage transformer under normal operating conditions, at a significance level of When , the kernel density method is used to calculate the statistic The control limits are: ； in, is the probability density distribution function expression, The significance level is Time Statistics Control limits; Substitute the real-time measurement data of the three-phase capacitor voltage transformer under the operating state, repeat steps 1 to 5, and obtain the statistics of the original amplitude data matrix of the three-phase capacitor voltage transformer under the operating state. , in step 5, , It represents the source signal matrix column mapped by the column with smaller matrix bi-norm of matrix A in the running state.

2. A capacitive voltage transformer error online detection method according to claim 1, characterized in that: Standard Amplitude Data Matrix It is expressed as: Where n represents the number of measurements of the three-phase capacitor voltage transformer parameters. Represents the amplitude data of the A-phase capacitor voltage transformer, Represents the amplitude data of the B-phase capacitor voltage transformer, Represents the amplitude data of the C-phase capacitive voltage transformer, .

3. A capacitive voltage transformer error online detection method according to claim 1, characterized in that: Step 2 specifically includes: Calculating the Matrix The covariance matrix R is: ; Decompose the eigenvalues ​​of the covariance matrix R to obtain the eigenvectors: ; According to the whitening formula, the whitening matrix W is obtained: .

Images