Distribution network voltage transformer metering misalignment online identification method

Abstract

The invention relates to an online identification method for metering misalignment of a voltage transformer of a distribution network. The method comprises the following specific steps: constructing a line dynamic model of the transformer in a normal state according to secondary voltage data, load data and environment temperature data of the voltage transformer among different power distribution rooms; and acquiring real-time secondary side voltage data of the mutual inductors of the two distribution network rooms, regressing predicted voltages of the secondary sides of the two distribution network rooms by using the constructed line dynamic model of the mutual inductors in a normal state, and comparing the predicted voltages with actually acquired secondary voltages to obtain an error evaluation result of the mutual inductors. Through the improved symbolic regression algorithm, the nonlinear symbolic association rules of the voltage transformers in the two adjacent power distribution rooms are automatically learned, and accurate model support is provided for voltage monitoring and fault early warning of the power distribution system.

Description

Technical Field

[0001] This application relates to the field of voltage transformer metering inaccuracy identification technology, and more specifically, to an online identification method for distribution network voltage transformer metering inaccuracy. Background Technology

[0002] In 10kV distribution network systems, voltage transformers are core equipment for energy metering, and their errors and operating status directly affect power supply reliability and metering fairness. Traditional periodic calibration cannot detect sudden errors in a timely manner, which leads to equipment with excessive errors continuing to operate for extended periods. This not only causes energy metering deviations but may also lead to more serious safety accidents due to equipment deterioration. Transformer errors are affected by load and temperature coupling. This multi-factor coupling makes it difficult to describe the error law of voltage transformers using a simple linear model, placing extremely high demands on the complexity and adaptability of error assessment models. Furthermore, establishing an accurate error assessment model requires acquiring a large amount of equipment operating data covering different loads and temperature conditions. However, some older distribution network lines lack comprehensive sensing and monitoring equipment, making data acquisition difficult and lacking in completeness. This hinders the effective training and verification of error assessment models, further increasing the difficulty of implementing error assessment technology in actual distribution network systems. Summary of the Invention

[0003] The purpose of this application is to provide an online identification method for metering inaccuracies of distribution network voltage transformers, which automatically learns the nonlinear symbolic association rules of voltage transformers in two adjacent distribution rooms, providing accurate model support for voltage monitoring and fault early warning in the distribution system.

[0004] To achieve the above objectives, this application provides the following technical solution:

[0005] This application provides an online identification method for metering inaccuracies in distribution network voltage transformers, including the following specific steps:

[0006] A dynamic model of the line under normal conditions of the transformers is constructed based on the secondary voltage data, load data, and ambient temperature data of the voltage transformers in different power distribution rooms.

[0007] Real-time secondary voltage data of instrument transformers in two distribution network rooms are collected. The predicted secondary voltages of the two distribution rooms are regressed using the dynamic line model of the instrument transformers under normal conditions. The error assessment results of the instrument transformers are obtained by comparing the predicted voltages with the actual collected secondary voltages.

[0008] The specific process for constructing a dynamic model of the line under normal transformer conditions is as follows:

[0009] Collect secondary voltage data from the current transformers in power distribution room 1 and power distribution room 2. Line load rate data P and ambient temperature data T were cleaned and statistical measures were calculated.

[0010]

[0011]

[0012]

[0013] The sample variance is:

[0014]

[0015] Where i is the number of sampling points;

[0016] Constructing a symbolic regression model:

[0017] Equation graph initialization:

[0018] Insert terminal symbol Load factor P, ambient temperature T, and the first, second, third, and fourth parameters to be determined in the model. ;

[0019] Initial symbolic expression generation:

[0020] Generate the initial expression:

[0021]

[0022] Solving for the optimal symbolic regression model and parameters:

[0023] Construct the objective function, fix the expression and weights. Solve for the optimal

[0024] Objective: Minimize voltage prediction loss, formula:

[0025] in:

[0026] The symbolic representation of the s-th iteration;

[0027] Global weights:

[0028]

[0029] Input for the target;

[0030] Weighted distance:

[0031] The optimal solution is obtained by using the L-BFGS optimizer.

[0032] Weight update, fixed expression and Solve for the optimal

[0033] According to elastic network regularization, the optimal weight formula is:

[0034]

[0035] in, , To optimize the parameters, they are set based on empirical values;

[0036] According to the optimal Update the symbolic expression and perform a convergence check:

[0037]

[0038] Calculate the MSE of the validation set:

[0039]

[0040] if If the value is less than the empirical threshold, stop the iteration and output the optimal value. and Conversely, it will To proceed to the next iteration, denoted as

[0041] .

[0042] Through the secondary voltage of power distribution room 1 The predicted voltage on the secondary side of distribution room 2 is obtained by regressing the dynamic model of the line under normal transformer conditions. ,use The acquired values ​​of phase A secondary side of transformer 2 in the power distribution room The error assessment result of the transformer in power distribution room 1 is obtained by performing the ratio difference calculation; similarly, the error assessment result of the transformer in power distribution room 2 is obtained.

[0043] Perform current transformer error status judgment:

[0044] when , When both are less than 0.2%, it indicates that both current transformers are in normal condition.

[0045] when , If any one of them exceeds 0.2%, it indicates that there is an abnormality in the current transformer.

[0046] Compared with the prior art, the beneficial effects of the present invention are:

[0047] This invention uses an improved symbolic regression algorithm to automatically learn the nonlinear symbolic association rules of voltage transformers in two adjacent distribution rooms, providing accurate model support for voltage monitoring and fault early warning in power distribution systems. Attached Figure Description

[0048] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0049] Figure 1 This is a flowchart of a method provided in an embodiment of the present invention.

[0050] Figure 2 This is a schematic diagram of a power distribution scenario during the implementation of an embodiment of the present invention.

[0051] Figure 3 A confusion matrix was plotted for the evaluation results of each data point in this application. Detailed Implementation

[0052] The technical solutions of the embodiments of this application will now be described with reference to the accompanying drawings. It should be noted that similar reference numerals and letters in the following drawings indicate similar items; therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings.

[0053] The terms “comprising,” “including,” or any other variations thereof are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase “comprising one…” does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0054] The terms “first,” “second,” etc., are used only to distinguish one entity or operation from another, and should not be construed as indicating or implying relative importance, nor as requiring or implying any such actual relationship or order between these entities or operations.

[0055] This invention proposes an online identification method for metering inaccuracies in distribution network voltage transformers. The method flow is as follows:

[0056] S100: Based on the secondary voltage data, load data, and ambient temperature data of the voltage transformers in different distribution rooms, a dynamic model of the line under normal conditions of the transformers is constructed, specifically as follows:

[0057] like Figure 2 A schematic diagram of a power distribution scenario.

[0058] In the scene:

[0059] Rl12 and Rl23 are the line resistances from power distribution room 1 to power distribution room 2 and from power distribution room 2 to power distribution room 3, respectively.

[0060] Xl12 and Xl23: These are the line reactances from distribution room 1 to distribution room 2 and from distribution room 2 to distribution room 3, respectively.

[0061] I0 represents the bus current of the three power distribution rooms.

[0062] I1, I2, I3: These are the primary currents of power distribution room 1, power distribution room 2, and power distribution room 3, respectively.

[0063] k: Rated transformation ratio of the current transformer (in this scenario, all current transformers have the same specifications and the same transformation ratio).

[0064] f1, f2, f3: These are the ratio errors of the current transformers in distribution room 1, distribution room 2, and distribution room 3, respectively.

[0065] u1a, u2a, and u3a are the measured values ​​of the secondary voltage of phase A in the three power distribution rooms, respectively.

[0066] V1a, V2a, and V3a are the true values ​​of the primary voltage of phase A in the three power distribution rooms, respectively.

[0067] Calculate the line voltage drop between distribution room 1 and distribution room 2, and between distribution room 2 and distribution room 3. The line voltage drop consists of the voltage drop caused by resistance and reactance, and can be expressed in complex form:

[0068] (1)

[0069] (2)

[0070] so:

[0071] (3)

[0072] (4)

[0073] According to the formula for calculating the ratio difference:

[0074] (5)

[0075] (6)

[0076] Calculate the ratio of the measured secondary output voltage values ​​of the current transformer:

[0077] (7)

[0078]

[0079]

[0080]

[0081]

[0082] because The value of can be calculated when the line parameters are known. Therefore, the calculated result can be regarded as a coefficient. The above equation can then be simplified to:

[0083] (8)

[0084] (9)

[0085] Therefore, it can be It can be viewed as a dynamic parameter that depends solely on changes in error, ambient temperature, and load. and The balance is achieved through this dynamic parameter.

[0086] Based on this, the secondary voltages of the two transformers under normal conditions can be collected. Based on the data, load, and ambient temperature data, an improved symbolic regression algorithm was used to construct... This represents the dynamic model of the line under normal operating conditions of the current transformer. The specific process is as follows:

[0087] Collect secondary voltage data from the current transformers in power distribution room 1 and power distribution room 2. Line load rate data P and ambient temperature data T were cleaned and statistical measures were calculated.

[0088] (10)

[0089] (11)

[0090] (12)

[0091] The sample variance is:

[0092] (13)

[0093] Where i is the number of sampling points.

[0094] Constructing a symbolic regression model:

[0095] Equation graph initialization:

[0096] Insert terminal symbol Load factor P, ambient temperature T, and the first, second, third, and fourth parameters to be determined in the model. .

[0097] Initial symbolic expression generation:

[0098] Generate the initial expression according to formula (9):

[0099] (13)

[0100] Solving for the optimal symbolic regression model and parameters:

[0101] Construct the objective function, fix the expression and weights. Solve for the optimal

[0102] Objective: Minimize voltage prediction loss, formula:

[0103] (14)

[0104] in:

[0105] The symbolic representation of the s-th iteration;

[0106] Global weights:

[0107] (15)

[0108] Input for the target;

[0109] Weighted distance:

[0110] (16)

[0111] The optimal solution is obtained by using the L-BFGS optimizer.

[0112] Weight update, fixed expression and Solve for the optimal

[0113] According to elastic network regularization, the optimal weight formula is:

[0114] (17)

[0116] in, , To optimize the parameters, you can set them based on empirical values.

[0117] According to the optimal Update the symbolic expression and perform a convergence check:

[0118] (18)

[0119] Calculate the MSE of the validation set:

[0120] (19)

[0121] if If the value is less than the empirical threshold, stop the iteration and output the optimal value. and Conversely, it will Substitute equation (14) into the next iteration. This can be expressed as...

[0122] (20)

[0123] S200: Collects real-time secondary voltage data of the instrument transformers in two distribution network rooms, and uses the constructed line dynamic model under normal conditions of the instrument transformers to regress the predicted secondary voltages of the two distribution rooms. By comparing the predicted voltages with the actual collected secondary voltages, the error assessment results of the instrument transformers are obtained, specifically:

[0124] Through the secondary voltage of power distribution room 1 The predicted voltage on the secondary side of distribution room 2 is obtained by regressing the dynamic model of the line under normal transformer conditions. .use The acquired values ​​of phase A secondary side of transformer 2 in the power distribution room The error assessment results of the current transformer in distribution room 1 are obtained by performing the ratio difference calculation; similarly, the error assessment results of the current transformer in distribution room 2 can be obtained. The results of the model can be summarized as shown in Table 1:

[0125] Table 1 Evaluation Results

[0126]

[0127] The error status of the current transformer can be determined from the table:

[0128] when , When both are less than 0.2%, it indicates that both current transformers are in normal condition.

[0129] when , If any one of them exceeds 0.2%, it indicates that there is an abnormality in the current transformer. To further determine which specific current transformer is abnormal, the transformer turns ratio model can be used.

[0130] I. Validation of the evaluation algorithm based on field data

[0131] By installing data acquisition devices at the distribution network user sites, metering data was collected from three locations: User 1, User 2, and User 3, as well as low-voltage side data of the transformer. The collected field data was then used to experimentally verify the instrument transformer evaluation algorithm proposed in this invention.

[0132] After cleaning the data collected from User 1 and User 2, 2497 valid data points were extracted for experimental verification. A partial screenshot of the data is shown below:

[0133]

[0134] At this point, the model output variable is set to the line voltage of the current transformer under test at user 1's metering point: UABzlbz. The model input variables are set to the line voltage data of the current transformer not under test at user 1 and the line voltage data of the current transformer at the metering point of the adjacent site (user 2): UBCzlbz, UABfgs, UBCfgs.

[0135] Regression modeling was performed using the methods described above, and the fully trained regression model was used to evaluate 200 newly collected data points from June 18th. A confusion matrix was plotted for each data point based on its evaluation results. Figure 3 As shown:

[0136] The results show that out of 200 normal data points tested, 199 of them were accurately identified as normal, with an overall accuracy rate of 99.00%.

[0137] The technical effects brought about by the technical solution of this invention are as follows:

[0138] This invention uses an improved symbolic regression algorithm to automatically learn the nonlinear symbolic association rules of voltage transformers in two adjacent distribution rooms, providing accurate model support for voltage monitoring and fault early warning in power distribution systems.

[0139] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for online identification of metering inaccuracies in distribution network voltage transformers, characterized in that, The specific steps include the following: A dynamic model of the line under normal conditions of the transformers is constructed based on the secondary voltage data, load data, and ambient temperature data of the voltage transformers in different power distribution rooms. Real-time secondary voltage data of instrument transformers in two distribution network rooms are collected. The predicted secondary voltages of the two distribution rooms are regressed using the dynamic line model of the instrument transformers under normal conditions. The error assessment results of the instrument transformers are obtained by comparing the predicted voltages with the actual collected secondary voltages.

2. The online identification method for metering inaccuracies of distribution network voltage transformers according to claim 1, characterized in that, The specific process for constructing a dynamic model of the line under normal transformer conditions is as follows: Collect secondary voltage data from the current transformers in power distribution room 1 and power distribution room 2. The line load rate data P and the ambient temperature data T were cleaned, and statistical measures were calculated. ， ， ， The sample variance is: ， Where i is the number of sampling points; Constructing a symbolic regression model: Equation graph initialization: Insert terminal symbol Load factor P, ambient temperature T, and the first, second, third, and fourth parameters to be determined in the model. ; Initial symbolic expression generation: Generate the initial expression: ， Solving for the optimal symbolic regression model and parameters: Construct the objective function, fix the expression and weights. Solve for the optimal , Objective: Minimize voltage prediction loss, formula: in: The symbolic representation of the s-th iteration; Global weights: ， Input for the target; Weighted distance: The optimal solution is obtained by using the L-BFGS optimizer. , Weight update, fixed expression and Solve for the optimal , According to elastic network regularization, the optimal weight formula is: ， in, , To optimize the parameters, they are set based on empirical values; According to the optimal Update the symbolic expression and perform a convergence check: ， Calculate the MSE of the validation set: ， if If the value is less than the empirical threshold, stop the iteration and output the optimal value. and Conversely, it will To proceed to the next iteration, denoted as 。 3. The online identification method for metering inaccuracies of distribution network voltage transformers according to claim 1, characterized in that, Through the secondary voltage of power distribution room 1 The predicted voltage on the secondary side of distribution room 2 is obtained by regressing the dynamic model of the line under normal transformer conditions. ,use The acquired values ​​of phase A secondary side of transformer 2 in the power distribution room The error assessment result of the transformer in power distribution room 1 is obtained by performing the ratio difference calculation; similarly, the error assessment result of the transformer in power distribution room 2 is obtained.

4. The online identification method for metering inaccuracies of distribution network voltage transformers according to claim 3, characterized in that, Perform current transformer error status judgment: when , When both are less than 0.2%, it indicates that both current transformers are in normal condition. when , If any one of them exceeds 0.2%, it indicates that there is an abnormality in the current transformer.

Images