Autonomous identification method and system for error degradation form of direct current measurement equipment

Abstract

The invention discloses an autonomous identification method and system for an error degradation form of direct current measurement equipment, and the method comprises the steps: building an evaluation standard quantity based on the normal operation state after operation and historical data measured by the direct current measurement equipment in different error degradation forms; based on the measured historical data, establishing a mathematical model of a metering error degradation form of the direct current measurement equipment; extracting feature sample sets of different error forms based on the evaluation standard quantity; based on the feature sample set, establishing a direct-current measurement equipment error form identification model and training the direct-current measurement equipment error form identification model; and when the evaluation standard quantity detects that the metering error state of the direct current measurement equipment is abnormal, performing online detection on the metering error of the direct current measurement equipment by utilizing the established mathematical model of the metering error degradation form of the direct current measurement equipment, and performing online identification by applying the trained error form identification model of the direct current measurement equipment. According to the invention, workers can be helped to detect the abnormal state of the direct current measurement equipment in time.

Description

Technical Field

[0001] The present invention relates to the field of electric power technology, and in particular to an autonomous identification method and system for error degradation forms of direct current measurement equipment. Background Art

[0002] DC measuring equipment is an important part of the power system. Its accuracy is an important basis for judging the operating status of the power system and adjusting the operating mode of the power system. It has a very important impact on the control, measurement, metering, protection and other functions of the DC system. However, in the actual operation process, affected by the complex environment of varying degrees such as temperature, humidity, frequency, etc., DC measuring equipment is prone to different forms of degradation and error out-of-tolerance phenomenon, which affects the accuracy of electric energy metering and the reliability of the relay protection device.

[0003] In addition, the metering error of DC measuring equipment presents different characteristics under different degradation forms. For example, the cumulative effect of periodic error changes on electric energy metering may be zero. This is because the periodic errors cancel each other out in the positive and negative half-cycles and will not have a substantial impact on electric energy metering. Errors in other degradation forms, such as random errors and drift errors, will have different degrees of cumulative effects on electric energy metering. Therefore, it is necessary to autonomously detect and identify the metering errors of DC measuring equipment to achieve rapid detection and processing of abnormal conditions.

[0004] At present, the detection method for the metering error of DC measuring equipment generally adopts the method of comparison and calibration with standard DC measuring equipment. However, due to the high requirements of standard DC measuring equipment on the operating environment, this method can only be carried out periodically to reflect the short-term error state. In the power system, the change of the metering error of DC measuring equipment is a long-term and continuous accumulation process. This method cannot accurately reflect the long-term operating state of DC measuring equipment. Therefore, in order to improve the safe operation level of the power system, it is necessary to conduct online detection and autonomous identification methods of the metering error of DC measuring equipment, aiming to accurately perceive the change of the metering error state of DC measuring equipment in real time and accurately identify the degradation form of the error state.

[0005] In the prior art, a Chinese invention patent application with publication number CN109444791A discloses a method and system for online monitoring of the error state of a capacitor voltage transformer, which collects measurement data of a calibrated and put into operation three-phase capacitor voltage transformer and uses the principal component analysis method to achieve real-time online evaluation of the error state of the capacitor voltage transformer. However, its online detection method only evaluates the error state and does not further identify the form of error degradation. Summary of the invention

[0006] The technical problem to be solved by the present invention is to provide a method for further identifying the error degradation form based on online detection of the measurement error of a DC measurement device in a data-driven manner.

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

[0008] A method for autonomously identifying the error degradation form of a DC measurement device, comprising:

[0009] S10, establishing an evaluation standard quantity based on the historical data measured by the DC measurement equipment under normal operation status and different error degradation forms after commissioning;

[0010] S20, establishing a mathematical model of the degradation form of the measurement error of the DC measurement equipment based on the measured historical data;

[0011] S30, extracting feature sample sets of different error forms based on the evaluation standard quantity;

[0012] S40, based on the feature sample set, establishing a DC measurement equipment error form identification model and performing training;

[0013] S50, when the evaluation standard detects that the metering error state of the DC measuring device is abnormal, the metering error of the DC measuring device is detected online by using the established mathematical model of the degradation form of the metering error of the DC measuring device, and the trained DC measuring device error form identification model is used for online identification.

[0014] In one embodiment of the present invention, establishing an evaluation standard includes:

[0015] S11, based on the measured historical data, construct the data matrix X of the output signal of the DC measurement device 0 ;

[0016] S12, for the data matrix X 0 Perform standardization processing to obtain the standardized matrix X;

[0017] S13, performing dimensionality reduction processing on the standardized matrix X to separate the true value information of the collected DC measurement device data and error information ΔX;

[0018] S14, calculating the evaluation standard quantity Q representing the error state according to the error information ΔX, and calculating the limit value Q of the evaluation standard quantity Q c ;

[0019] in, In the formula, ‖·‖2 represents the 2-norm, and the limit value Q c is the statistical value when the confidence level is c.

[0020] In one embodiment of the present invention, a mathematical model of the degradation form of the measurement error of the DC measurement device is established, through the following formula:

[0021] y t =kx t +v x +f t ;

[0022] In the formula, y t is the real value of the secondary output of the DC measuring device, k is the transformation ratio of the DC measuring device, x t It is expressed as the actual value of the primary side signal amplitude of the DC measuring device, v x and f t They represent random error and systematic error respectively; t represents time;

[0023] Among them, the system error f t include:

[0024] Transformation ratio error f1(t): f1(t) = l(t)Δk; where l(t) represents the true value and Δk represents the difference in transformation ratio before and after the transformation ratio error occurs in the DC measuring device;

[0025] Amplitude fixed error f2(t): Where n ft It is expressed as a step function, t1 represents the start time of the fixed error, and t2 represents the end time of the fixed error;

[0026] Drift error f3: f3 = p(t-t3); where r3 represents the start time of the drift error, and p represents a constant.

[0027] In one embodiment of the present invention, the feature sample set is obtained in the following manner:

[0028] S31, based on the time series of the estimated standard quantity {Q i}, first specify an embedding dimension s0, solve the time delay parameter τ, and reconstruct the phase space using the delay coordinates;

[0029] S32, based on the reconstructed phase space, calculate the time series {Q i}'s correlation integral C i (r);

[0030] S33, according to the correlation score C i (r), obtain the correlation dimension D, perform least squares fitting on the logarithmic curve of the correlation dimension D, and obtain the initial correlation dimension estimate D(s0);

[0031] S34, increase the value of the embedding dimension s0, so that the increased embedding dimension s1>s0. As the value of the dimension increases, the initial correlation dimension estimate D(s0) corresponding to the embedding dimension s0 gradually stabilizes within a certain error range. At this time, the correlation dimension is obtained as the time series {Q i}'s correlation dimension;

[0032] S35, calculating the correlation dimension D of the time series as the characteristic quantity of the time series for estimating the standard quantity, and the correlation dimensions under various error degradation forms constitute the characteristic sample sets E of different error forms.

[0033] In one embodiment of the present invention, the time delay parameter τ is solved in the following manner:

[0034]

[0035] Wherein, R represents the autocorrelation function, n-τ represents the delay of the nth sampling point corresponding to each time point; the variation curve of the autocorrelation function R with the time delay parameter τ is obtained by calculation, and the time delay parameter τ corresponding to when R is 0 for the first time or drops to 1-1 / e of the initial value is taken as the optimal time delay parameter; wherein, e represents the natural constant, i represents the i-th sampling point, and - represents the mean value of Q;

[0036] The time series {Q i}Perform coordinate delay under the time delay parameter τ to obtain the reconstructed phase space.

[0037] In one embodiment of the present invention, the correlation score C i (r) is obtained by the following formula:

[0038]

[0039] Where θ is represented by the Heaviside unit function, where x=r-|y i -y j ∣,y i ,y j To reconstruct the time series vector in the phase space, r represents the specified distance threshold. When the distance between two vectors is less than r, it can be judged that there is a correlation between the two vectors;

[0040] The correlation dimension D is obtained by the following formula:

[0041]

[0042] Where r represents the specified distance threshold.

[0043] In one embodiment of the present invention, a DC measurement device error form identification model is established, including:

[0044] S41, the feature sample set E is represented as E = {a v ,b w}, b w Indicates a v Category, v = 1, 2, ..., n, n is the number of samples collected; a v is the vth input sample; w=1,2,…,m, where m is the number of categories

[0045] S42, establish a mapping function φ(x) to map the points in the feature sample set E to a high-dimensional space: φ(x) = α i j i K(x i ,x); where K(x i ,x) is the kernel function, α i is the Lagrange multiplier, j i Represented as kernel function parameters;

[0046] S43, establishing a DC measurement equipment error form identification model according to the mapping function φ(x).

[0047] In one embodiment of the present invention, the DC measurement device error form identification model f(x) is the following formula:

[0048]

[0049] In the formula, d represents the bias term of the function, α i j i K(x i ,x) is represented as a nonlinear mapping function.

[0050] In one embodiment of the present invention, the evaluation standard detects whether the measurement error state of the DC measurement device is abnormal by:

[0051] If Q<Q c , then the DC measuring device is in normal operation;

[0052] If Q>Q c , the DC measuring device is in an abnormal state at this time, and abnormal measurement error occurs.

[0053] Among them, Q represents the evaluation standard quantity, Q c Expressed as the limit value of the evaluation standard quantity Q.

[0054] The present invention also provides an autonomous identification system for the error degradation form of a DC measurement device, comprising:

[0055] An evaluation standard module is used to establish evaluation standard quantities based on historical data measured by DC measurement equipment under normal operating conditions and different error degradation forms after commissioning;

[0056] An error detection module is used to establish a mathematical model of the degradation form of the measurement error of the DC measurement equipment based on the measured historical data;

[0057] A sample extraction module is used to extract feature sample sets of different error forms based on the evaluation standard quantity;

[0058] An identification model training module is used to establish and train a DC measurement equipment error form identification model based on a feature sample set;

[0059] The online detection module is used to perform online detection of the measurement error of the DC measurement equipment when the evaluation standard detects that the measurement error state of the DC measurement equipment is abnormal, using the established mathematical model of the measurement error degradation form of the DC measurement equipment, and applying the trained DC measurement equipment error form identification model for online identification.

[0060] Compared with the prior art, the beneficial effects of the present invention are as follows: on the basis of online detection of the metering error of the DC measuring device, the present invention further identifies the form of error degradation, reconstructs the phase space using the embedding dimension and the time delay parameter, and obtains the correlation dimension as the characteristic parameter of the error identification model, and further constructs the error identification model by establishing a mapping function and a decision function, thereby realizing the online identification of the form of error degradation. The present invention does not need to establish a knowledge base of different types of DC measuring devices, but only based on a data-driven method, by establishing an error identification model, calculating the error characteristic vector using the online sampling data of different devices, and then identifying the specific form of error degradation, the identification result is more accurate and has a wider range of application.

[0061] The present invention analyzes the collected DC measurement data, constructs an evaluation standard quantity Q that characterizes the error state, records it as a statistic, and realizes online detection of the error state of the DC measurement equipment; by establishing a mathematical model of the measurement error degradation form, calculating the correlation dimension of the statistical time series, analyzing the difference of the correlation dimension under different error degradation forms, and forming the correlation dimension into a characteristic quantity, the online identification of the error degradation form of the DC measurement equipment is realized. It can help the staff to detect the abnormal state of the DC measurement equipment in time, and help to formulate more differentiated DC measurement equipment maintenance and repair strategies, thereby improving the reliability and stability of the DC measurement equipment and ensuring the safe, economical and stable operation of the power grid.

[0062] Extracting the correlation dimension of the statistic characterizing the error state of the DC measurement equipment as its identification characteristic parameter can fully reflect the difference in the error state degradation form. Taking the correlation dimension of the statistical time series as the input feature, an error identification model is established to realize the autonomous identification of the error form of the DC measurement equipment.

[0063] The present invention collects operating data of DC measuring equipment in normal operating states and in different error degradation forms after commissioning, performs error detection and identification in an autonomous identification system of error degradation forms, and feeds back the results to on-site staff. It can detect in real time whether the error state of the DC measuring equipment is abnormal and identify the error form, which is helpful to formulate more differentiated maintenance and inspection strategies for DC measuring equipment, thereby improving the reliability and stability of the DC measuring equipment and ensuring the safe, economical and stable operation of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 The present invention is a flowchart of a method for autonomously identifying the error degradation form of a DC measurement device according to an embodiment of the present invention.

[0065] Figure 2 The present invention is a flowchart of the specific steps of autonomously identifying the error degradation form of a DC measurement device according to an embodiment of the present invention.

[0066] Figure 3 The present invention is a block diagram of an autonomous identification system of error degradation forms of a DC measurement device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0067] In order to facilitate those skilled in the art to understand the technical solution of the present invention, the technical solution of the present invention is further described in conjunction with the accompanying drawings of the specification.

[0068] The terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0069] See also Figure 1 and Figure 2 As shown, the present invention discloses an autonomous identification method of the error degradation form of a DC measurement device, comprising:

[0070] S10, establishing an evaluation standard quantity based on the historical data measured by the DC measurement equipment under normal operating conditions and different error degradation forms after commissioning.

[0071] In one embodiment of the present invention, establishing an evaluation standard includes:

[0072] S11, based on the measured historical data, construct the data matrix X of the output signal of the DC measurement device 0 Specifically, construct the data matrix X of the output signal of the DC measurement device 0 ∈R m×n , n is the number of samples collected, m is the number of output signals collected, X 0 for:

[0073]

[0074] Where: X nm Indicates the nth sampling point of the acquired mth output signal.

[0075] S12, for the data matrix X 0 The standardized matrix X is obtained by standardization. The processed data matrix is:

[0076] X=(X 0 -Ib T )∑ -1 ;

[0077] Where: I = [1,1,…,1] T ∈R n×1 , T represents the transposition. b=(X 0 ) T I / n is the mean vector of the data, Σ=diag(σ1 2 ,...,σ m 2 ) is the variance matrix of the data, σ m is the variance of the mth column.

[0078] S13, performing dimensionality reduction processing on the standardized matrix X to separate the main space and the residual space, that is, separating the true value information of the collected DC measurement device data and error information ΔX;

[0079]

[0080] In the formula, It represents the true value information of the collected DC measurement device data, and ΔX represents the error information of the collected DC measurement device data.

[0081] S14, based on the error information ΔX, calculate the evaluation standard quantity Q that characterizes the error state, and calculate the evaluation standard quantity Q and its limit value Q by the following formula c : Q c is the statistical value when the confidence level is c, that is, when the confidence level is 99%, 99% of the verification samples can be found below this statistical value.

[0082] S20, based on the measured historical data, a mathematical model of the degradation form of the measurement error of the DC measurement equipment is established.

[0083] In one embodiment of the present invention, a mathematical model of the degradation form of the measurement error of a DC measurement device is established, including:

[0084] S21, establish a mathematical model of the DC measurement device output:

[0085] y t =kx t +v x +f t ;

[0086] In the formula, y t is the true value of the secondary output of the DC measuring device, k represents the transformation ratio of the DC measuring device, and x t is the actual value of the primary side signal amplitude of the DC measuring device, v x and f t represent random error and systematic error respectively, and t represents time.

[0087] It should be noted that random errors are mainly caused by noise interference and the randomness of the measurement process itself, which are difficult to eliminate through correction. Systematic errors are mainly caused by design defects of the DC measurement equipment itself or measurement accuracy deviations caused by changes in the use environment. Errors can be eliminated or reduced by correction or replacement of components.

[0088] S22, based on the mathematical model of the DC measurement device output, uses different error functions to represent various system errors of the DC measurement device.

[0089] The mathematical model based on the output of the DC measuring device described in step S22 uses different error functions to represent various system errors of the DC measuring device, specifically:

[0090] S221, specifically, when the transformation ratio of the DC measuring device changes due to the influence of temperature or other environments, a transformation ratio error will occur, and the error function is expressed as: f1(t)=l(t)Δk, where Δk is the difference in the transformation ratio of the DC measuring device before and after the transformation ratio error occurs, and l(t) is the true value.

[0091] S222, when the DC measurement device error remains unchanged in phase and a fixed value is added to the amplitude, it is a fixed amplitude error, and the error function is expressed as: Where n ft is a step function, t1 is the start time of the fixed error, t2 is the end time of the fixed error, and it approaches infinity.

[0092] S223, when the DC measuring equipment ages due to the influence of environmental factors such as electromagnetic field, temperature and humidity, its measurement value will drift over time. The mathematical model of the drift error is: f3 = p(t-t3), where t3 is the start time of the drift error and p is a constant.

[0093] S30, extracting feature sample sets of different error forms based on the evaluation standard quantity.

[0094] In one embodiment of the present invention, the feature sample set is obtained in the following manner:

[0095] S31, based on the time series of the estimated standard quantity {Q i}, first specify an embedding dimension s0, solve the time delay parameter τ, and reconstruct the phase space using the delay coordinates.

[0096] In one embodiment of the present invention, based on the time series {Q i}, first specify a smaller embedding dimension s0, solve the time delay parameter τ, and reconstruct the phase space using the delay coordinates. Specifically, the calculation method for solving the time delay parameter τ is:

[0097] For the time series {Q i}, its autocorrelation function R can be expressed as:

[0098]

[0099] In the formula, R represents the autocorrelation function, n-τ represents the delay of the nth sampling point corresponding to each time point. The variation curve of the autocorrelation function R with the time delay τ is obtained by calculation, and the τ corresponding to when R is 0 for the first time or drops to 1-1 / e of the initial value is taken as the optimal time delay parameter. Among them, e represents a natural constant. i represents the i-th sampling point, - represents the mean value of Q. After obtaining the time delay parameter τ, the time series {Q i}Perform coordinate delay under the time delay parameter τ to obtain the reconstructed phase space.

[0100] S32, based on the reconstructed phase space, calculate the time series {Q i}'s correlation integral C i (r).

[0101] In one embodiment of the present invention, the correlation score C i (r) is obtained by the following formula:

[0102]

[0103] Where θ is represented by the Heaviside unit function, where x=r-|y i -y j ∣,y i ,y j To reconstruct the time series vector in the phase space, r is represented as the specified distance threshold. When the distance between two vectors is less than r, it can be determined that there is a correlation between the two vectors.

[0104] S33, according to the correlation score C i (r), obtain the correlation dimension D, perform least squares fitting on the logarithmic curve of the correlation dimension D, and obtain the initial correlation dimension estimate D(s0).

[0105] In one embodiment of the present invention, the correlation dimension D is obtained by the following formula:

[0106]

[0107] Where r represents the specified distance threshold.

[0108] S34, increase the value of the embedding dimension s0, so that the increased embedding dimension s1>s0. As the value of the dimension increases, the initial correlation dimension estimate D(s0) corresponding to the embedding dimension s0 gradually stabilizes within a certain error range. At this time, the correlation dimension is obtained as the time series {Q i}'s correlation dimension;

[0109] S35, calculating the correlation dimension D of the time series as the characteristic quantity of the time series for estimating the standard quantity, and the correlation dimensions under various error degradation forms constitute the characteristic sample sets E of different error forms.

[0110] S40, based on the feature sample set, establishing a DC measurement equipment error form identification model and performing training.

[0111] In one embodiment of the present invention, a DC measurement device error form identification model is established, including:

[0112] S41, the feature sample set E is represented as E = {a v ,b w}, v = 1, 2, ..., n, n is the number of collected samples; a v is the vth input sample; w=1,2,…,m, m is the number of categories, b w Indicates a v Category.

[0113] S42, establish a mapping function φ(x) to map the points in the feature sample set E to a high-dimensional space: φ(x) = α i j i K(x i ,x); where K(x i,x) is the kernel function, and you can choose linear kernel, polynomial kernel, radial basis kernel and other kernel functions. i is the Lagrange multiplier, j i Represented as kernel function parameters.

[0114] S43, establishing a DC measurement equipment error form identification model according to the mapping function φ(x).

[0115] In this embodiment, the DC measurement device error form identification model f(x) is expressed as:

[0116]

[0117] Among them, α i j i K(x i ,x) is the nonlinear mapping function, and d is the bias term of the function.

[0118] S50, when the evaluation standard quantity detects that the DC measurement device metering error state is abnormal, the DC measurement device metering error is detected online using the established mathematical model of the DC measurement device metering error degradation form, and the DC measurement device error form identification model that has been trained is used to perform online identification.

[0119] In one embodiment of the present invention, the evaluation standard detects whether the measurement error state of the DC measurement device is abnormal by:

[0120] If Q<Q c , then the DC measuring device is in normal operation;

[0121] If Q>Q c , the DC measuring device is in an abnormal state at this time, and abnormal measurement error occurs.

[0122] In one embodiment of the present invention, during online detection:

[0123] Collect online operation data of DC measurement equipment and construct a data matrix of output signal amplitude;

[0124] Based on the data matrix, calculate the Q, Q c , and make comparison to determine whether the error state of the online operation data of the DC measuring equipment is abnormal;

[0125] If an exception occurs:

[0126] Inputting the online operation data into the mathematical model of the degradation form of the measurement error of the DC measurement equipment, and performing online detection on the measurement error of the DC measurement equipment;

[0127] At the same time, the correlation dimension of the abnormal time series is obtained to extract the characteristics of the error form;

[0128] The characteristics are identified using a DC measurement equipment error form identification model to obtain an autonomous identification result of the error degradation form.

[0129] See also Figures 1 to 3 As shown, the present invention also provides an autonomous identification system of the error degradation form of a DC measurement device, which uses the above-mentioned autonomous identification method of the error degradation form of a DC measurement device, including:

[0130] An evaluation standard module is used to establish evaluation standard quantities based on historical data measured by DC measurement equipment under normal operating conditions and different error degradation forms after commissioning;

[0131] The error detection module is used to establish a mathematical model of the degradation form of the measurement error of the DC measurement equipment based on the measured historical data.

[0132] The sample extraction module is used to extract feature sample sets of different error forms based on the evaluation standard quantity.

[0133] The identification model training module is used to establish and train the error form identification model of the DC measurement equipment based on the feature sample set.

[0134] The online detection module is used to perform online detection of the measurement error of the DC measurement equipment when the evaluation standard detects that the measurement error state of the DC measurement equipment is abnormal, using the established mathematical model of the measurement error degradation form of the DC measurement equipment, and applying the trained DC measurement equipment error form identification model for online identification.

[0135] It is obvious to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential features of the present invention. Therefore, the embodiments should be regarded as exemplary and non-limiting from any point of view, and the scope of the present invention is defined by the appended claims rather than the above description, and it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present invention, and any reference numerals in the claims should not be regarded as limiting the claims involved.

[0136] The above-described embodiments merely represent implementation methods of the invention. The protection scope of the present invention is not limited to the above-described embodiments. For those skilled in the art, several modifications and improvements may be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention.

Claims

1. A method for autonomously identifying the error degradation form of a DC measurement device, characterized in that: include: S10, establishing an evaluation standard quantity based on the historical data measured by the DC measurement equipment under normal operation status and different error degradation forms after commissioning; S20, establishing a mathematical model of the degradation form of the measurement error of the DC measurement equipment based on the measured historical data; S30, extracting feature sample sets of different error forms based on the evaluation standard quantity; S40, based on the feature sample set, establishing a DC measurement equipment error form identification model and performing training; S50, when the evaluation standard detects that the metering error state of the DC measuring device is abnormal, the metering error of the DC measuring device is detected online by using the established mathematical model of the degradation form of the metering error of the DC measuring device, and the trained DC measuring device error form identification model is used for online identification.

2. The autonomous identification method of the error degradation form of the DC measurement device according to claim 1, characterized in that: Establish evaluation criteria, including: S11, based on the measured historical data, construct the data matrix X of the output signal of the DC measurement device 0 ; S12, for the data matrix X 0 Perform standardization processing to obtain the standardized matrix X; S13, performing dimensionality reduction processing on the standardized matrix X to separate the true value information of the collected DC measurement device data and error information ΔX; S14, calculating the evaluation standard quantity Q representing the error state according to the error information ΔX, and calculating the limit value Q of the evaluation standard quantity Q c ; in, In the formula, ‖·‖ 2 Expressed as 2-norm, limit Q c is the statistical value when the confidence level is c.

3. The autonomous identification method of the error degradation form of the DC measurement equipment according to claim 1, characterized in that: The mathematical model of the measurement error degradation form of DC measurement equipment is established through the following formula: yes t =kx t +v x +f t ; In the formula, y t is the real value of the secondary output of the DC measuring device, k is the transformation ratio of the DC measuring device, x t It is expressed as the actual value of the primary side signal amplitude of the DC measuring device, v x and f t They represent random error and systematic error respectively; t represents time; Among them, the system error f t include: Transformation ratio error f1(t): f1(t) = l(t)Δk; where l(t) represents the true value and Δk represents the difference in transformation ratio before and after the transformation ratio error occurs in the DC measuring device; Amplitude fixed error f2(t): Where n ft It is expressed as a step function, t1 represents the start time of the fixed error, and t2 represents the end time of the fixed error; Drift error f3: f3 = p(t-t3); where t3 represents the start time of the drift error, and p represents a constant.

4. The autonomous identification method of the error degradation form of the DC measurement device according to claim 1, characterized in that: Get the feature sample set in the following ways: S31, based on the time series of the estimated standard quantity {Q i }, first specify an embedding dimension s0, solve the time delay parameter τ, and reconstruct the phase space using the delay coordinates; S32, based on the reconstructed phase space, calculate the time series {Q i }'s correlation integral C i (r); S33, according to the correlation score C i (r), obtain the correlation dimension D, perform least squares fitting on the logarithmic curve of the correlation dimension D, and obtain the initial correlation dimension estimate D(s0); S34, increase the value of the embedding dimension s0, so that the increased embedding dimension s1>s0. As the value of the dimension increases, the initial correlation dimension estimate D(s0) corresponding to the embedding dimension s0 gradually stabilizes within a certain error range. At this time, the correlation dimension is obtained as the time series {Q i }'s correlation dimension; S35, calculating the correlation dimension D of the time series as the characteristic quantity of the time series for estimating the standard quantity, and the correlation dimensions under various error degradation forms constitute the characteristic sample sets E of different error forms.

5. The autonomous identification method of the error degradation form of the DC measurement device according to claim 4, characterized in that: Solve for the time delay parameter τ by: Wherein, R represents the autocorrelation function, n-τ represents the delay of the nth sampling point corresponding to each time point; the variation curve of the autocorrelation function R with the time delay parameter τ is obtained by calculation, and the time delay parameter τ corresponding to when R is 0 for the first time or drops to 1-1 / e of the initial value is taken as the optimal time delay parameter; wherein, e represents the natural constant, i represents the i-th sampling point, and - represents the mean value of Q; The time series {Q i }Perform coordinate delay under the time delay parameter τ to obtain the reconstructed phase space.

6. The autonomous identification method of the error degradation form of the DC measurement device according to claim 4, characterized in that: Correlation score C i (r) is obtained by the following formula: Where θ is represented by the Heaviside unit function, where x=r-|y i -y j ∣,y i ,y j To reconstruct the time series vector in the phase space, r represents the specified distance threshold. When the distance between two vectors is less than r, it can be judged that there is a correlation between the two vectors; The correlation dimension D is obtained by the following formula: Where r represents the specified distance threshold.

7. The autonomous identification method of the error degradation form of a DC measurement device according to claim 1, characterized in that: Establish a DC measurement equipment error form identification model, including: S41, the feature sample set E is represented as E = {a v ,b w }, b w Indicates a v Category, v = 1, 2, ..., n, n is the number of samples collected; a v is the vth input sample; w=1,2,…,m, where m is the number of categories S42, establish a mapping function φ(x) to map the points in the feature sample set E to a high-dimensional space: φ(x) = α i j i K(x i ,x); where K(x i ,x) is the kernel function, α i is the Lagrange multiplier, j i Represented as function parameters; S43, establishing a DC measurement equipment error form identification model according to the mapping function φ(x).

8. The autonomous identification method of the error degradation form of the DC measurement device according to claim 7, characterized in that: The DC measurement equipment error form identification model f(x) is as follows: In the formula, d represents the bias term of the function.

9. The autonomous identification method of the error degradation form of a DC measurement device according to claim 1, characterized in that: The evaluation standard detects whether the measurement error state of the DC measurement equipment is abnormal by the following methods: If Q<Q c , then the DC measuring device is in normal operation; If Q>Q c , the DC measuring device is in an abnormal state at this time, and abnormal measurement error occurs. Among them, Q represents the evaluation standard quantity, Q c Expressed as the limit value of the evaluation standard quantity Q.

10. An autonomous identification system for the error degradation form of a DC measurement device, characterized in that: The method for autonomously identifying the error degradation form of a DC measurement device according to any one of claims 1 to 9 comprises: An evaluation standard module is used to establish evaluation standard quantities based on historical data measured by DC measurement equipment under normal operating conditions and different error degradation forms after commissioning; An error detection module is used to establish a mathematical model of the degradation form of the measurement error of the DC measurement equipment based on the measured historical data; A sample extraction module is used to extract feature sample sets of different error forms based on the evaluation standard quantity; An identification model training module is used to establish and train a DC measurement equipment error form identification model based on a feature sample set; The online detection module is used to perform online detection of the measurement error of the DC measurement equipment when the evaluation standard detects that the measurement error state of the DC measurement equipment is abnormal, using the established mathematical model of the measurement error degradation form of the DC measurement equipment, and applying the trained DC measurement equipment error form identification model for online identification.

Images