Measurement error evaluation method for voltage transformer

Abstract

The invention discloses a voltage transformer metering error evaluation method, which relates to the technical field of voltage transformer metering and comprises the following steps of: constructing a synchronous acquisition system of multiple groups of voltage transformers in a transformer substation, acquiring original voltage data, constructing a virtual reference voltage generation model and outputting a reference voltage sequence; the method comprises the steps of establishing a virtual reference voltage model, calculating error statistical characteristics, marking the operating state grade of the voltage transformer, evaluating health scores according to the error characteristics, outputting a scoring result, receiving inspection feedback comparison differences, updating the model, generating an analysis chart, and uploading the analysis chart to a client display and linkage operation and maintenance system. On-line error evaluation is realized, multi-source data are fused, error features are dynamically extracted, an operation state is marked, a health scoring and early warning mechanism is established, a feedback closed loop is introduced to optimize model precision, an evaluation result is displayed in a chart form, operation and maintenance scheduling is assisted, and the real-time performance, the accuracy and the self-adaptive capability of evaluation are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of voltage transformer metering, and in particular to a voltage transformer metering error evaluation method. Background Art

[0002] As a critical primary device for voltage monitoring and energy metering in power systems, the accuracy of voltage transformers is directly related to the stability of power system operation and the fairness of user metering. Existing voltage transformer error assessment methods often rely on offline laboratory calibration or periodic manual testing. These methods struggle to reflect the dynamic error changes of the equipment during operation in real time and are unable to adapt to factors such as multi-substation structures, complex load conditions, and environmental fluctuations. Furthermore, some online assessment methods rely on physical standards or single-point benchmark data, resulting in a lack of robustness and universality in the assessment models. These methods suffer from issues such as insufficient error coverage, weak dynamic adjustment capabilities, and a lack of model self-learning capabilities.

[0003] At present, the Chinese patent application number CN201910892188.8 discloses a capacitive voltage transformer metering error situation perception system, including a data acquisition system, a data analysis system and a situation prediction system connected in sequence, wherein the data acquisition system collects the operating data of the CVT to be tested and the power system in which it is located in real time; based on the data collected by the data acquisition system, the data analysis system analyzes the CVT metering error state; based on the CVT metering error state analyzed by the data analysis system, the situation prediction system predicts the CVT metering error situation. The perception system of this application can timely discover the problems existing in the CVT metering error according to the CVT metering error situation, and promptly repair the CVT, thereby avoiding economic losses caused by metering errors. The perception system of this application can also calculate the gradual change process of CVT metering anomalies according to the current metering error situation of the CVT, locate the moment when the CVT fault anomaly occurs, and provide guidance for repairing the CVT and other treatments.

[0004] It is difficult for related technologies to achieve online dynamic evaluation and accurate classification judgment of voltage transformer measurement errors without the need for physical standards. Summary of the Invention

[0005] The technical problem solved by the present invention is that it is difficult for related technologies to realize online dynamic evaluation and accurate classification judgment of voltage transformer measurement errors without the need for physical standards.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: A method for evaluating voltage transformer measurement error comprises the following steps: Step S1, constructing a synchronous data acquisition system for multiple groups of voltage transformers in a substation to obtain raw voltage data; Step S2, constructing a virtual reference voltage generation model and outputting a reference voltage sequence; Step S3, calculating the error statistical characteristics and marking the voltage transformer operating status level; Step S4, evaluating the health score based on the error characteristics and outputting the score result; Step S5: receiving inspection feedback, comparing differences and updating the model; Step S6: Generate an analysis chart and upload it to the client for display and linkage operation and maintenance system.

[0007] Preferably, step S1 includes the following sub-steps: Step S101: Select multiple voltage transformers installed at specific busbars, incoming lines, and outgoing lines in the substation as monitoring objects, covering the main voltage node areas, and record the installation point, operation number, and electrical node information corresponding to each voltage transformer; Step S102: synchronously collecting the voltage signal outputted by the secondary side of the voltage transformer using a high-precision optical fiber communication module with a unified time reference, wherein the voltage signal is derived from the transformed output of the busbar, incoming line, or outgoing line voltage measured by the voltage transformer during actual operation, to form structured raw voltage data; Step S103: Upload the structured raw voltage data to the time series database to complete format standardization conversion and field verification.

[0008] Preferably, step S2 includes the following sub-steps: Step S201: extract the physical voltage nodes corresponding to the voltage transformers in the substation and construct a logical connection relationship diagram; Step S202 , analyzing the voltage conduction paths between the busbar, incoming lines, and outgoing lines based on the logical relationship of the electrical nodes connected to the voltage transformer, and constructing a reference voltage function model; The reference voltage function model includes the voltage relationship, voltage transformation ratio and coupling influence between each physical voltage node; Step S203: collecting historical operation data, wherein the historical operation data includes original voltage measurement data of the voltage transformer in multiple time periods, corresponding substation operation condition information, and actual load characteristics; Based on the regular deviations of historical operating data, coordinated change trends, and static and dynamic relationships between node voltages, combined with a preset rule set that includes node voltage stability criteria, load regulation modes, and grid operating status classification rules, a virtual reference voltage generation model for fitting the reference voltage is constructed. The virtual reference voltage generation model outputs an estimated value based on the comparison of the current node structure with the historical pattern; Step S204 , fitting the current original voltage data through a virtual reference voltage generation model, and outputting a virtual reference voltage sequence matching the actual node.

[0009] Preferably, the reference voltage function model supports a dynamic parameter adaptive adjustment mechanism, and the dynamic parameter adaptive adjustment mechanism includes: Determine the type of topology change based on the real-time collected operating status; When a topology change event occurs, the node function parameters are recalculated and updated to the reference voltage function model; Parameters are adjusted through incremental learning.

[0010] Preferably, the indicators of the synergistic change trend include: Calculate the average error change rate in the sliding window to identify short-term fluctuation trends; Extract the short-term volatility change rate through the second-order difference of the mean; The drift slope is extracted as a long-term trend indicator using the least squares method to fit the error mean trajectory; Short-term fluctuation trend, short-term fluctuation change rate and long-term trend index are used as characteristic variables for status classification and scoring.

[0011] Preferably, step S3 includes the following sub-steps: Step S301 , comparing the original voltage measurement data with the virtual reference voltage sequence to generate a measurement error sequence; Step S302, performing sliding window segmented statistics on the measurement error sequence, extracting the error mean, extreme value range, variance and change rate, and outputting them as voltage transformer error statistical feature data; Step S303: Perform a comprehensive analysis on the statistical characteristic data of the voltage transformer errors in each sliding time window according to the classification standard. The logic of the comprehensive analysis is: The statistical characteristic data of the voltage transformer error is compared with the preset state classification standard. According to the performance of the statistical characteristic data of the voltage transformer error on the preset state classification standard, the operating state level of the corresponding voltage transformer in the corresponding time window is marked and output. The operating state level includes normal, slightly abnormal, moderately abnormal and severely abnormal.

[0012] Preferably, step S4 includes the following sub-steps: Step S401, identifying periodic changes, continuous offsets, and sudden abnormal features of voltage transformer error statistical feature data, and constructing an error feature mapping table; Step S402: Referring to the historical voltage error data and a preset scoring model, the scoring model is constructed based on the aforementioned voltage transformer error statistical feature data, sets independent scoring factor weights for different feature dimensions, and calculates the corresponding health factor score values ​​after normalization. Obtain the health factor score value, perform weighted summation on the health factor score value according to the weighted calculation rules set by the scoring model, and output the comprehensive health score result; Step S403: When the comprehensive health score result is lower than the preset warning threshold, the warning mechanism is triggered and a maintenance task signal is generated.

[0013] Preferably, the logic of step S403 is: Continuously compare transformer health scores with thresholds; When the score is lower than the threshold, an inspection recommendation of the corresponding level is generated according to the scheduling rules; Optimize dispatching plans based on transformer location, voltage level, and maintenance records; Output maintenance scheduling suggestions and upload them to the operation and maintenance platform for generating work dispatch instructions.

[0014] Preferably, step S5 includes the following sub-steps: Step S501: receiving log feedback information, wherein the log feedback information includes inspection time, equipment number, inspection judgment label and on-site remarks information; Compare the log feedback information with the voltage transformer error statistical characteristic data and the operating status level one by one, identify sample pairs that are consistent with the voltage transformer error statistical characteristic data and sample pairs that are inconsistent, output the comparison results, and generate a difference comparison report based on the comparison results, wherein the difference comparison report includes a consistency ratio, a difference reason mark, and an impact range; Step S502 , marking the sample pairs determined to be consistent as credible training samples and inputting them into the model training process of the error classification model and the virtual reference voltage generation model; For sample pairs with inconsistent state labels, they are submitted to the expert system interface for manual verification and labeling, and high-confidence samples are output. The high-confidence samples are input into the model training process of the error classification model and the virtual reference voltage generation model.

[0015] Preferably, step S6 includes the following sub-steps: Step S601: construct an error line graph, a score trend graph, and an equipment level distribution graph based on the measurement error sequence, the comprehensive health score result, and the operating status. The line graph, the score trend graph, and the equipment level distribution graph are converted into a unified format and then visualized and displayed. Step S602: Upload the line graph, the rating trend graph, and the device level distribution graph to the client platform.

[0016] The beneficial effects of the present invention are as follows: the present invention provides a method for constructing a virtual reference voltage model to replace the physical standard, realizes online error evaluation, integrates multi-source data, dynamically extracts error characteristics and marks the operating status, establishes a health scoring and early warning mechanism, and introduces a feedback closed-loop optimization model accuracy. The evaluation results are displayed in a graphical form to assist operation and maintenance scheduling, thereby improving the real-time, accuracy and adaptability of the evaluation. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A flowchart of the steps of a voltage transformer measurement error evaluation method provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0018] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, but not all of the embodiments.

[0019] Example, see Figure 1 , provides a voltage transformer measurement error evaluation method, comprising the following steps: Step S1: construct a synchronous data acquisition system for multiple groups of voltage transformers in a substation to obtain raw voltage data.

[0020] Step S2: construct a virtual reference voltage generation model and output a reference voltage sequence.

[0021] Step S3: Calculate the error statistical characteristics and mark the voltage transformer operating status level.

[0022] Step S4: Evaluate the health score based on the error characteristics and output the score result.

[0023] Step S5: receiving inspection feedback, comparing differences and updating the model.

[0024] Step S6: Generate an analysis chart and upload it to the client for display and linkage operation and maintenance system.

[0025] Step S1 includes the following sub-steps: Step S101: select multiple specific voltage transformers arranged at busbar, incoming line and outgoing line positions in the substation as monitoring objects, covering the main voltage node areas, and record the installation point, operation number and electrical node information corresponding to each voltage transformer.

[0026] Step S101 implements monitoring coverage of key voltage nodes in the substation, ensuring the representativeness of the collected objects and the rationality of spatial distribution. At the same time, it collects necessary transformer attribute information to provide support for data tracing and node modeling.

[0027] In step S102, the voltage signal outputted from the secondary side of the voltage transformer is synchronously collected with a unified time reference based on a high-precision optical fiber communication module. The voltage signal is derived from the transformed output of the busbar, incoming line, or outgoing line voltage measured by the voltage transformer during actual operation, forming structured raw voltage data.

[0028] Step S102 completes the high-precision, unified time-base synchronous acquisition of the output voltage signals of each voltage transformer, ensures the alignment and time consistency of multi-source data in timing analysis, and records the sampling results in a structured manner.

[0029] Step S103: Upload the structured raw voltage data to the time series database to complete format standardization conversion and field verification.

[0030] Step S103 implements standardized storage of the original voltage data and field legitimacy verification to ensure accuracy, consistency, and traceability in subsequent data processing.

[0031] Step S1 is used to establish a synchronous data acquisition system for multiple groups of voltage transformers in the substation, forming a data foundation with a unified time reference and structured standards, and providing original voltage data support for subsequent error analysis and model construction.

[0032] Step S2 includes the following sub-steps: Step S201: extract the physical voltage nodes corresponding to the voltage transformers in the substation and construct a logical connection relationship diagram.

[0033] Step S201 clarifies the identity of the physical voltage nodes to which each voltage transformer is connected, constructs an electrical connection relationship diagram between the nodes, and provides basic topological information for subsequent voltage function modeling.

[0034] Step S202 : Analyze the voltage conduction paths between the busbar, the incoming line, and the outgoing line according to the logical relationship of the electrical nodes to which the voltage transformer is connected, and construct a reference voltage function model.

[0035] The reference voltage function model includes the voltage relationship, transformation ratio and coupling effect between each physical voltage node.

[0036] Step S202 extracts the conduction structure in each voltage path according to the node logical relationship, constructs a function model reflecting the voltage distribution and transformation relationship between the busbar, incoming line and outgoing line, and establishes a mathematical expression of the voltage dependency and coupling effect of each node.

[0037] Step S203 : collecting historical operation data, which includes original voltage measurement data of the voltage transformer in multiple time periods, corresponding substation operation condition information, and actual load characteristics.

[0038] Based on the regular deviations of historical operating data, coordinated change trends and static and dynamic relationships between node voltages, combined with a preset rule set, which includes node voltage stability criteria, load regulation modes and grid operation status classification rules, a virtual reference voltage generation model for fitting the reference voltage is constructed. The virtual reference voltage generation model outputs an estimated value based on the comparison between the current node structure and the historical pattern.

[0039] Step S203 systematically collects historical operating data for multiple time periods, including original voltage data, load characteristics, and operating conditions. Combined with the coordinated change trends and voltage response patterns between nodes, a virtual reference voltage generation model is trained according to preset rules to achieve structured error learning capabilities.

[0040] Step S204 , fitting the current original voltage data through a virtual reference voltage generation model, and outputting a virtual reference voltage sequence matching the actual node.

[0041] The reference voltage function model supports a dynamic parameter adaptive adjustment mechanism, which includes: The type of topology change is determined based on the real-time collected operating status.

[0042] When a topology change event occurs, the node function parameters are recalculated and updated to the reference voltage function model.

[0043] Parameters are adjusted through incremental learning.

[0044] Indicators of collaborative change trends include: The average rate of change of error is calculated in the sliding window to identify short-term fluctuation trends.

[0045] The short-term volatility change rate is extracted through the second-order difference of the mean.

[0046] The drift slope is extracted as a long-term trend indicator using the least squares method to fit the error mean trajectory.

[0047] Short-term fluctuation trend, short-term fluctuation change rate and long-term trend index are used as characteristic variables for status classification and scoring.

[0048] In step S204, the original voltage measurement data at the current moment is input into the trained virtual reference voltage generation model, and a virtual reference voltage sequence that matches the actual electrical node structure is fitted and output as a reference value source for subsequent error calculations. The reference voltage function model has dynamic parameter adaptation capabilities, and can automatically identify the type of structural change based on real-time monitored topological changes, reconstruct function parameters, and update the model through incremental learning to achieve adaptive response to changes in the operating environment and maintain estimation accuracy.

[0049] Step S2 is used to construct a virtual reference voltage generation mechanism that can dynamically reflect the electrical structure characteristics and historical operating laws of the substation. By integrating node voltage relationships, historical measurement characteristics and expert rules, it outputs virtual reference voltage data with timing accuracy and structural adaptability, providing a benchmark sequence with high comparability and strong dynamic response capability for voltage transformer error assessment.

[0050] Step S3 includes the following sub-steps: Step S301 : comparing original voltage measurement data with a virtual reference voltage sequence to generate a measurement error sequence.

[0051] Step S301 compares the original voltage measurement data of each voltage transformer at the current moment with the corresponding virtual reference voltage value, outputs a measurement error sequence representing the change in error amplitude at each time point, and establishes a basic data source for error feature extraction.

[0052] Step S302 , performing sliding window segmented statistics on the measurement error sequence, extracting the error mean, extreme value range, variance and change rate, and outputting them as voltage transformer error statistical feature data.

[0053] Step S302 performs sliding window segmentation processing on the measurement error sequence, extracts key statistical indicators in each window segment, including error mean, extreme value range, variance and change rate, and forms structured voltage transformer error statistical feature data for subsequent state evaluation.

[0054] Step S303: Perform a comprehensive analysis on the statistical characteristic data of the voltage transformer errors in each sliding time window according to the classification criteria. The logic of the comprehensive analysis is: The statistical characteristic data of the voltage transformer error is compared with the preset state classification standard. According to the performance of the statistical characteristic data of the voltage transformer error on the preset state classification standard, the operating state level of the corresponding voltage transformer in the corresponding time window is marked and output. The operating state levels include normal, slightly abnormal, moderately abnormal and seriously abnormal.

[0055] Step S303 compares and analyzes the extracted error statistical feature data with the system's preset status grading standards. Based on the placement of each indicator in the grading standards, the operating status level of the voltage transformer in the current time window is comprehensively determined and marked as normal, slightly abnormal, moderately abnormal or severely abnormal, and the status level result is output for subsequent scoring and early warning module calls.

[0056] Step S3 is used to dynamically extract the error change characteristics of the voltage transformer based on the comparative analysis between the original voltage measurement data and the virtual reference voltage, and identify its operating status level in combination with the status classification standard, providing data support and classification basis for subsequent health scoring and early warning strategies.

[0057] Step S4 includes the following sub-steps: Step S401 : identifying periodic variation, continuous offset, and sudden abnormality features of voltage transformer error statistical feature data, and constructing an error feature mapping table.

[0058] Step S401 extracts the time series change pattern from the error statistical feature data, identifies representative periodic changes, persistent offsets and sudden abnormal features, constructs an error feature mapping table, and provides an input data source for the scoring factor.

[0059] Step S402 refers to the historical voltage error data and the preset scoring model. The scoring model is constructed based on the aforementioned voltage transformer error statistical feature data, and independent scoring factor weights are set for different feature dimensions. After normalization processing, the corresponding health factor score values ​​are calculated respectively.

[0060] Obtain the health factor score value, perform weighted summation on the health factor score value according to the weighted calculation rules set by the scoring model, and output the comprehensive health score result.

[0061] In step S402, the extracted error features are input into a preset scoring model. The scoring model sets weights based on the error mean, extreme value range, variance, rate of change, and other dimensions, and calculates the corresponding health factor score after normalizing each indicator. The scores of all factors are then combined using weighted rules to output a comprehensive health score result reflecting the operating status of the voltage transformer.

[0062] Step S403: When the comprehensive health score result is lower than the preset warning threshold, the warning mechanism is triggered and a maintenance task signal is generated.

[0063] The logic of step S403 is: Continuously compare the transformer health score with the threshold.

[0064] When the score is lower than the threshold, an inspection recommendation of the corresponding level is generated according to the scheduling rules.

[0065] Optimize the scheduling plan by combining transformer location, voltage level and maintenance records.

[0066] Output maintenance scheduling suggestions and upload them to the operation and maintenance platform for generating work dispatch instructions.

[0067] Step S403 continuously compares the output comprehensive health score result with the warning threshold set by the system; if the score is lower than the warning threshold, the warning mechanism is triggered according to the built-in scheduling rules, and preliminary inspection recommendations including inspection priority and recommended response period are generated; combined with the actual installation location, voltage level and historical maintenance records of the transformer, the maintenance scheduling path and resource allocation plan are dynamically optimized, and finally a maintenance scheduling recommendation is generated and uploaded to the operation and maintenance platform for automatic generation and execution of work dispatch instructions.

[0068] Step S4 is used to identify the evolution pattern of the voltage transformer operating status based on the error statistical characteristics, build a health scoring model and dynamically output the comprehensive scoring results. When the scoring results are abnormal, the early warning and maintenance scheduling mechanisms are linked to complete the closed-loop transformation from error identification to active maintenance.

[0069] Step S5 includes the following sub-steps: Step S501: receiving log feedback information, which includes inspection time, equipment number, inspection judgment label and on-site remarks.

[0070] The log feedback information is compared one by one with the voltage transformer error statistical characteristic data and the operating status level, and the sample pairs that are consistent with the voltage transformer error statistical characteristic data and the sample pairs that are inconsistent are identified. The comparison results are output and a difference comparison report is generated based on the comparison results. The difference comparison report includes the consistency ratio, difference reason labeling and impact range.

[0071] Step S501 receives log feedback information from the operation and maintenance end or the monitoring system, including the equipment number, inspection time, manual judgment label and on-site records, and compares the feedback information with the operating status level previously output by the system based on the error statistical characteristics, identifies consistent and inconsistent sample pairs, and outputs a difference comparison report based on the comparison results. The report indicates the consistency ratio, possible sources of deviation and their affected areas, which are used to evaluate the system accuracy and potential correction needs.

[0072] In step S502 , the sample pairs determined to be consistent are marked as credible training samples, and are input into the model training process of the error classification model and the virtual reference voltage generation model.

[0073] For sample pairs with inconsistent state labels, they are submitted to the expert system interface for manual verification and labeling, and high-confidence samples are output. The high-confidence samples are input into the model training process of the error classification model and the virtual reference voltage generation model.

[0074] In step S502, the sample pairs marked as consistent in the comparison are identified as credible samples, and are input into the training process of the error classification model and the virtual reference voltage generation model as high-quality training samples to improve the model accuracy; sample pairs with inconsistent labels are submitted to the expert system interface for manual verification and correction labeling. The confirmed samples are re-input into the model training process as high-confidence samples to further enhance the robustness and generalization ability of the model in abnormal sample identification and complex scenarios.

[0075] Step S5 is used to build a closed-loop self-optimization mechanism for the voltage transformer error evaluation system. By performing consistency comparison and difference analysis on the operation and maintenance inspection feedback and the system identification results, high-quality samples are screened for model iterative training, thereby continuously improving the accuracy and adaptability of the error identification and estimation model.

[0076] Step S6 includes the following sub-steps: Step S601: construct an error line graph, a score trend graph, and an equipment level distribution graph based on the measurement error sequence, the comprehensive health score result, and the operating status. The line graph, the score trend graph, and the equipment level distribution graph are converted into a unified format and then visualized and displayed.

[0077] Step S601 uses the measurement error sequence, comprehensive health score results and operating status level obtained in the previous step as input data to generate an error line chart, a score trend chart and an equipment level distribution chart respectively. The chart content reflects the dynamic changes in the errors of each voltage transformer, the operating stability assessment results and the overall status distribution. After generation, various charts are uniformly converted into a data format that meets the access requirements of the system visualization module to complete the format standardization processing before graphic rendering.

[0078] Step S602: Upload the line graph, the rating trend graph, and the device level distribution graph to the client platform.

[0079] Step S602 uploads the formatted chart data to the client interactive platform for real-time viewing and comparison by operation and maintenance personnel. The platform supports functions such as chart switching, node filtering, and historical trend backtracking, enabling synchronous sharing of analysis results between the local terminal and the remote operation and maintenance system, thereby improving the efficiency of early warning response and scheduling deployment.

[0080] Step S6 is used to uniformly display the voltage transformer error assessment results in a graphical manner, thereby improving the operation and maintenance personnel's intuitive perception of error change trends and equipment status, and uploading the chart data to the client platform to support the visual linkage between remote operation and maintenance decision-making and equipment status monitoring.

[0081] The present invention does not require an external physical standard device. By constructing a virtual reference voltage generation model and using operating data to dynamically estimate a high-precision benchmark, it effectively solves the problem of traditional methods relying on manual comparison and standard equipment. By introducing node voltage logic, historical collaborative characteristics and operating rules, the model's ability to adapt to complex substation topologies is improved. A sliding window is used to extract error feature data, and the equipment status level is output according to preset grading standards, so as to achieve the distinction and location of errors of different severity. A health scoring model is constructed based on error statistical indicators, and a comprehensive scoring value is output. The scoring is linked to the warning threshold to dynamically generate inspection tasks and maintenance suggestions, integrate inspection feedback and expert verification samples, continuously iterate and train error identification and reference estimation models, improve the system's self-learning ability and evaluation accuracy, and convert error evolution trends, scoring results and status levels into charts and upload them to the client, so that operation and maintenance personnel can make intuitive judgments and assist in scheduling decisions.

[0082] Those skilled in the art will appreciate 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 an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media containing computer-usable program code. The storage medium may be implemented by any type of volatile or non-volatile storage device, or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0083] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A voltage transformer measurement error evaluation method, characterized in that: The steps include: Step S1, constructing a synchronous data acquisition system for multiple groups of voltage transformers in a substation to obtain raw voltage data; Step S2, constructing a virtual reference voltage generation model and outputting a reference voltage sequence; Step S3, calculating the error statistical characteristics and marking the voltage transformer operating status level; Step S4, evaluating the health score based on the error characteristics and outputting the score result; Step S5: receiving inspection feedback, comparing differences and updating the model; Step S6: Generate an analysis chart and upload it to the client for display and linkage operation and maintenance system.

2. A voltage transformer measurement error evaluation method according to claim 1, characterized in that: The step S1 includes the following sub-steps: Step S101: Select multiple voltage transformers installed at specific busbars, incoming lines, and outgoing lines in the substation as monitoring objects, covering the main voltage node areas, and record the installation point, operation number, and electrical node information corresponding to each voltage transformer; Step S102: synchronously collecting the voltage signal outputted by the secondary side of the voltage transformer using a high-precision optical fiber communication module with a unified time reference, wherein the voltage signal is derived from the transformed output of the busbar, incoming line, or outgoing line voltage measured by the voltage transformer during actual operation, to form structured raw voltage data; Step S103: Upload the structured raw voltage data to the time series database to complete format standardization conversion and field verification.

3. A voltage transformer measurement error evaluation method according to claim 2, characterized in that: The step S2 includes the following sub-steps: Step S201: extract the physical voltage nodes corresponding to the voltage transformers in the substation and construct a logical connection relationship diagram; Step S202 , analyzing the voltage conduction paths between the busbar, incoming lines, and outgoing lines based on the logical relationship of the electrical nodes connected to the voltage transformer, and constructing a reference voltage function model; The reference voltage function model includes the voltage relationship, voltage transformation ratio and coupling influence between each physical voltage node; Step S203: collecting historical operation data, wherein the historical operation data includes original voltage measurement data of the voltage transformer in multiple time periods, corresponding substation operation condition information, and actual load characteristics; Based on the regular deviations of historical operating data, coordinated change trends, and static and dynamic relationships between node voltages, combined with a preset rule set that includes node voltage stability criteria, load regulation modes, and grid operating status classification rules, a virtual reference voltage generation model for fitting the reference voltage is constructed. The virtual reference voltage generation model outputs an estimated value based on the comparison of the current node structure with the historical pattern; Step S204 , fitting the current original voltage data through a virtual reference voltage generation model, and outputting a virtual reference voltage sequence matching the actual node.

4. A voltage transformer measurement error evaluation method according to claim 3, characterized in that: The reference voltage function model supports a dynamic parameter adaptive adjustment mechanism, which includes: Determine the type of topology change based on the real-time collected operating status; When a topology change event occurs, the node function parameters are recalculated and updated to the reference voltage function model; Parameters are adjusted through incremental learning.

5. A voltage transformer measurement error evaluation method according to claim 4, characterized in that: The indicators of the coordinated change trend include: Calculate the average error change rate in the sliding window to identify short-term fluctuation trends; Extract the short-term volatility change rate through the second-order difference of the mean; The drift slope is extracted as a long-term trend indicator using the least squares method to fit the error mean trajectory; Short-term fluctuation trend, short-term fluctuation change rate and long-term trend index are used as characteristic variables for status classification and scoring.

6. A voltage transformer measurement error evaluation method according to claim 5, characterized in that: The step S3 includes the following sub-steps: Step S301 , comparing the original voltage measurement data with the virtual reference voltage sequence to generate a measurement error sequence; Step S302, performing sliding window segmented statistics on the measurement error sequence, extracting the error mean, extreme value range, variance and change rate, and outputting them as voltage transformer error statistical feature data; Step S303: Perform a comprehensive analysis on the statistical characteristic data of the voltage transformer errors in each sliding time window according to the classification standard. The logic of the comprehensive analysis is: The statistical characteristic data of the voltage transformer error is compared with the preset state classification standard. According to the performance of the statistical characteristic data of the voltage transformer error on the preset state classification standard, the operating state level of the corresponding voltage transformer in the corresponding time window is marked and output. The operating state level includes normal, slightly abnormal, moderately abnormal and severely abnormal.

7. A voltage transformer measurement error evaluation method according to claim 6, characterized in that: The step S4 includes the following sub-steps: Step S401, identifying periodic changes, continuous offsets, and sudden abnormal features of voltage transformer error statistical feature data, and constructing an error feature mapping table; Step S402: Referring to the historical voltage error data and a preset scoring model, the scoring model is constructed based on the aforementioned voltage transformer error statistical feature data, sets independent scoring factor weights for different feature dimensions, and calculates the corresponding health factor score values ​​after normalization. Obtain the health factor score value, perform weighted summation on the health factor score value according to the weighted calculation rules set by the scoring model, and output the comprehensive health score result; Step S403: When the comprehensive health score result is lower than the preset warning threshold, the warning mechanism is triggered and a maintenance task signal is generated.

8. A voltage transformer measurement error evaluation method according to claim 7, characterized in that: The logic of step S403 is: Continuously compare transformer health scores with thresholds; When the score is lower than the threshold, an inspection recommendation of the corresponding level is generated according to the scheduling rules; Optimize dispatching plans based on transformer location, voltage level, and maintenance records; Output maintenance scheduling suggestions and upload them to the operation and maintenance platform for generating work dispatch instructions.

9. A voltage transformer measurement error evaluation method according to claim 8, characterized in that: The step S5 includes the following sub-steps: Step S501: receiving log feedback information, wherein the log feedback information includes inspection time, equipment number, inspection judgment label and on-site remarks information; Compare the log feedback information with the voltage transformer error statistical characteristic data and the operating status level one by one, identify sample pairs that are consistent with the voltage transformer error statistical characteristic data and sample pairs that are inconsistent, output the comparison results, and generate a difference comparison report based on the comparison results, wherein the difference comparison report includes a consistency ratio, a difference reason mark, and an impact range; Step S502 , marking the sample pairs determined to be consistent as credible training samples and inputting them into the model training process of the error classification model and the virtual reference voltage generation model; For sample pairs with inconsistent state labels, they are submitted to the expert system interface for manual verification and labeling, and high-confidence samples are output. The high-confidence samples are input into the model training process of the error classification model and the virtual reference voltage generation model.

10. A voltage transformer measurement error evaluation method according to claim 9, characterized in that: The step S6 includes the following sub-steps: Step S601: construct an error line graph, a score trend graph, and an equipment level distribution graph based on the measurement error sequence, the comprehensive health score result, and the operating status. The line graph, the score trend graph, and the equipment level distribution graph are converted into a unified format and then visualized and displayed. Step S602: Upload the line graph, the rating trend graph, and the device level distribution graph to the client platform.

Images