GIS circuit breaker mechanical characteristic curve similarity detection method and system
By matching the standard curve with the action data to be tested, the output curve similarity coefficient is solved, combined with the curve slope method and the minimum difference method, the problem of low mechanical characteristics detection accuracy of GIS circuit breakers in the prior art is solved, and more accurate mechanical characteristics state is achieved Evaluation and fault warning.
Patent Information
- Application Number
- CN202410960652.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-17
- Publication Date
- 2024-11-22
AI Technical Summary
The existing GIS circuit breaker mechanical characteristic curve similarity detection method has low detection accuracy and cannot promptly warn when there is a tendency to deteriorate mechanical characteristics, resulting in low fault detection accuracy.
The standard curve matches the operation data to be tested to output the curve similarity coefficient. If the curve similarity coefficient is greater than the detection threshold, it is considered normal mechanical characteristics. If it is equal to the threshold, it is considered to have a deterioration trend. If it is less than the threshold, it is considered to be a fault. The detection threshold is based on the similarity of multiple curves. The distribution of coefficients is determined, and the standard curve is obtained by using the curve slope method and the minimum difference method. The boundary between the action segment and the non-action segment is clear, which improves the detection accuracy.
Effectively improve the accuracy of mechanical characteristics detection of GIS circuit breakers, can promptly warn when mechanical characteristics tend to deteriorate, improve the accuracy of mechanical characteristics such as stroke current, and help power equipment operation and maintenance personnel to evaluate and judge GIS more accurately Mechanical status of the circuit breaker.
Smart Images

Figure CN119000033A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to a method and a system for detecting similarity of mechanical characteristic curves of a GIS circuit breaker, and belongs to the technical field of mechanical characteristic monitoring of a GIS circuit breaker. Background Art
[0002] As an important switchgear in the power grid, GIS circuit breaker plays an important role in protection and control. Its safe and stable operation is directly related to the stability of the entire power grid. At present, a large number of statistical analyses of GIS circuit breaker failures have shown that 80% of GIS circuit breaker failures are caused by mechanical reasons. Therefore, mechanical characteristic status detection of GIS circuit breakers plays an important early warning role in maintaining the safe and stable operation of GIS circuit breakers.
[0003] In recent years, the number of high-voltage circuit breakers in the power grid has gradually increased. Accurately understanding the mechanical action status of GIS circuit breakers and timely repairing GIS circuit breakers to avoid serious accidents caused by faults are of great significance to the safe and stable operation of the entire power system.
[0004] The existing mechanical characteristic curve similarity detection method for detecting the mechanical characteristics of GIS circuit breakers uses the entire curve data as the detection object and sets a unified detection threshold for detection. The detection method has a simple design idea, but the detection accuracy is low and it cannot provide early warning of equipment failure. The equipment status alarm can only be completed when a more serious mechanical characteristic failure occurs. Generally speaking, equipment failure is a degradation process that changes slowly over time. When the degree of degradation reaches a certain threshold, the circuit breaker will fail. Therefore, the existing detection method has low detection accuracy for the mechanical characteristics of the circuit breaker, resulting in the inability to provide timely warnings when the mechanical characteristics show a degradation trend. Summary of the invention
[0005] The purpose of the present invention is to provide a method and system for detecting similarity of mechanical characteristic curves of a GIS circuit breaker, so as to solve the problem of low accuracy of existing detection methods.
[0006] To achieve the above object, the solution of the present invention includes:
[0007] The present invention provides a GIS circuit breaker mechanical characteristic curve similarity detection method, the method comprising: outputting a curve similarity coefficient by matching a standard curve with action data to be tested; if the curve similarity coefficient is greater than a detection threshold, it is considered that the circuit breaker mechanical characteristic is normal; if the curve similarity coefficient is equal to the detection threshold, it is considered that the circuit breaker mechanical characteristic has a degradation trend; if the curve similarity coefficient is less than the detection threshold, it is considered that the circuit breaker mechanical characteristic is faulty; the detection threshold is determined according to the distribution of multiple curve similarity coefficients, and the multiple curve similarity coefficients are obtained by matching the action segments of multiple mechanical characteristic test data with the action segments of the corresponding standard curves.
[0008] Furthermore, the standard curve for matching the motion data to be measured is selected according to the motion type of the motion data to be measured.
[0009] Furthermore, multiple groups of test data of normal operating state of each action type are taken as candidate data from the pre-acquired mechanical characteristic test data, and a standard curve corresponding to each action type is selected from the candidate data.
[0010] Further, the action data to be tested includes multiple action test data of each action type of the circuit breaker serving as the test prototype.
[0011] Furthermore, the action segment of the standard curve is obtained using a curve slope method.
[0012] Furthermore, the action segment of the mechanical property test data is obtained using the curve slope method.
[0013] Furthermore, the standard curve was selected according to the minimum difference method.
[0014] Furthermore, the number of groups of test data ranges from 15 to 35.
[0015] Furthermore, the action types include closing and opening.
[0016] A GIS circuit breaker mechanical characteristic curve similarity detection system of the present invention comprises a processor, and the processor is used to execute instructions to implement the above-mentioned GIS circuit breaker mechanical characteristic curve similarity detection method.
[0017] Beneficial effects of the present invention:
[0018] The present invention is a pioneering invention, which provides a GIS circuit breaker mechanical characteristic curve similarity detection method and system, performs curve similarity detection on the action data of the circuit breaker, and rationally sets the detection threshold through the distribution of multiple curve similarity coefficients, effectively improves the detection accuracy of the circuit breaker mechanical characteristics, further improves the accuracy of the mechanical characteristic state evaluation of the GIS circuit breaker stroke current, and helps the power equipment operation and maintenance personnel to more accurately evaluate and judge the mechanical state of the GIS circuit breaker. Specifically, the standard curve is matched with the action data to be tested to output the curve similarity coefficient. If the curve similarity coefficient is greater than the detection threshold, it is considered that the mechanical characteristics of the circuit breaker are normal. If the curve similarity coefficient is equal to the detection threshold, it is considered that the mechanical characteristics of the circuit breaker have a deterioration trend. If the curve similarity coefficient is less than the detection threshold, it is considered that the mechanical characteristics of the circuit breaker are faulty; the detection threshold is determined according to the distribution of multiple curve similarity coefficients, and multiple curve similarity coefficients are obtained by matching the action segments of multiple mechanical characteristic test data with the action segments of the corresponding standard curve. Among them, multiple mechanical characteristic test data are obtained using big data to meet the richness of multiple mechanical characteristic test data as sample data, and can cover the diversified mechanical characteristic information of multiple action types of circuit breakers. BRIEF DESCRIPTION OF THE DRAWINGS
[0019] Figure 1 It is a detection flow chart of a single action of a GIS circuit breaker according to an embodiment of the present invention;
[0020] Figure 2 It is a curve diagram of the opening action section of the opening standard data curve in the mechanical characteristic curve of the GIS circuit breaker in the embodiment of the present invention;
[0021] Figure 3 It is a curve diagram of the closing action section of the closing standard data curve in the mechanical characteristic curve of the GIS circuit breaker in the embodiment of the present invention;
[0022] Figure 4 is a summary scatter plot of similarity coefficients of tripping curves of an embodiment of the present invention;
[0023] Figure 5 It is a summary scatter plot of the similarity coefficients of the closing curves of the embodiment of the present invention. DETAILED DESCRIPTION
[0024] In order to solve the problems in the background technology, the present invention provides a GIS circuit breaker mechanical characteristic curve similarity detection method and system, which performs curve similarity detection on the action data of the circuit breaker, and rationally sets the detection threshold through the distribution of multiple curve similarity coefficients, effectively improves the detection accuracy of the circuit breaker mechanical characteristics, further improves the accuracy of the mechanical characteristic state evaluation of the GIS circuit breaker stroke current and other mechanical characteristics, and helps the power equipment operation and maintenance personnel to more accurately evaluate and judge the mechanical state of the GIS circuit breaker. Specifically, the standard curve is matched with the action data to be tested to output the curve similarity coefficient. If the curve similarity coefficient is greater than the detection threshold, it is considered that the mechanical characteristics of the circuit breaker are normal. If the curve similarity coefficient is equal to the detection threshold, it is considered that the mechanical characteristics of the circuit breaker have a deterioration trend. If the curve similarity coefficient is less than the detection threshold, it is considered that the mechanical characteristics of the circuit breaker are faulty; the detection threshold is determined according to the distribution of multiple curve similarity coefficients, and the multiple curve similarity coefficients are obtained by matching the action segments of multiple mechanical characteristic test data with the action segments of the corresponding standard curve. Among them, multiple mechanical characteristic test data are obtained using big data to meet the richness of multiple mechanical characteristic test data as sample data, and can cover the diversified mechanical characteristic information of multiple action types of circuit breakers.
[0025] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments.
[0026] An embodiment of a method for detecting similarity of mechanical characteristic curves of a GIS circuit breaker:
[0027] A method for detecting similarity of mechanical characteristic curves of a GIS circuit breaker comprises: outputting a curve similarity coefficient by matching a standard curve with action data to be tested; if the curve similarity coefficient is greater than a detection threshold, it is considered that the mechanical characteristic of the circuit breaker is normal; if the curve similarity coefficient is equal to the detection threshold, it is considered that the mechanical characteristic of the circuit breaker has a deterioration trend; if the curve similarity coefficient is less than the detection threshold, it is considered that the mechanical characteristic of the circuit breaker is faulty; the detection threshold is determined according to the distribution of a plurality of curve similarity coefficients, and the plurality of curve similarity coefficients are obtained by matching action segments of a plurality of mechanical characteristic test data with action segments of corresponding standard curves.
[0028] Of course, the action data to be tested here can also be adjusted according to actual needs. For example: take multiple action test data of each action type of a circuit breaker as a test prototype as the action data to be tested, that is, the action data to be tested includes multiple mechanical characteristic test data of opening action and multiple mechanical characteristic test data of closing action. Or take multiple action test data of the same action type of a circuit breaker as a test prototype as the action data to be tested, that is, the action data to be tested includes multiple mechanical characteristic test data of opening action or multiple mechanical characteristic test data of closing action.
[0029] According to the action type matching detection of these test data, the curve similarity coefficients corresponding to different action types are output (that is, each action type has a curve similarity coefficient), that is, the opening data curve fitted by a group of mechanical characteristics test data of switching action is matched with the standard curve of opening action to output the curve similarity coefficient of opening, and the closing data curve fitted by a combination of mechanical characteristics test data of switching action is matched with the standard curve of closing action to output the curve similarity coefficient of closing, and multiple mechanical characteristics test data of opening action are taken as a group of mechanical characteristics test data of switching action, and multiple mechanical characteristics test data of closing action are taken as a combination of mechanical characteristics test data of switching action.
[0030] Specifically, the standard curve for matching the motion data to be measured is selected according to the motion type of the motion data to be measured, that is, the motion type of the motion data to be measured is the same as the motion type of the standard curve for matching the motion data to be measured.
[0031] Specifically, multiple groups of test data of normal operating state of each action type are taken as candidate data from the pre-acquired mechanical characteristic test data, and a standard curve corresponding to each action type is selected from the candidate data.
[0032] Specifically, the action segment of the standard curve is obtained using the curve slope method.
[0033] Specifically, the action segment of the mechanical property test data is obtained using the curve slope method.
[0034] Specifically, the standard curve was selected according to the minimum difference method.
[0035] Specifically, the number of groups of test data used as candidate data ranges from 15 to 35.
[0036] Specifically, the action types include closing and opening.
[0037] like Figure 1 As shown in the figure, the main implementation steps of the GIS circuit breaker mechanical characteristic curve similarity detection method are as follows:
[0038] ① From the collected historical test data of mechanical characteristics of GIS circuit breakers, select 20 groups of test data with normal operation status according to the action type of the circuit breaker and store them in the standard curve candidate folder. The 20 groups here can also be adjusted according to actual needs, for example, adjusted to 35 groups, 15 groups, 30 groups, or 50 groups. The action type of the circuit breaker includes opening or closing.
[0039] Data arrangement in the standard curve candidate folder: All collected historical test data of mechanical characteristics of GIS circuit breakers are stored in the folder to be tested, and 20 groups of test data with normal operating status are selected according to the opening / closing action type and stored in the standard curve candidate folder.
[0040] ② Based on the minimum difference method, select the data closest to the theoretical curve from the standard curve candidate folder as the standard curve.
[0041] Standard curve selection: Based on curve fitting, minimum difference and other methods, the two sets of data (open and closed) closest to the theoretical curve are selected from the standard curve candidate folder as the standard curve.
[0042] The standard curve selection specifically selects mechanical characteristic data of multiple action types under normal operating conditions at different acquisition times from the pre-collected test data, and takes multiple groups of mechanical characteristic data of each action type as sample data; the sample data of each action type in the preprocessed sample data are weighted averaged in phase to obtain the standardized fitting data corresponding to each action type; the sample data are phased and subtracted from the standardized fitting data one by one according to the action type to obtain the difference data; the difference between the maximum value and the minimum value of each phase in each group of the difference data is used as the difference element to form a difference group, and the difference data to which the smallest difference element in the difference group belongs is used as the standard data, and the standard data is used as the standard curve. The number of groups of mechanical characteristic data of each action type ranges from 15 to 50. The action types include opening, closing and reclosing. The preprocessing includes classifying, sorting, statistics and analytical alignment of the sample data according to the action type.
[0043] ③ Based on the curve slope method, the dividing points of the stroke and coil current action segment and non-action segment are extracted from the standard curve, and the curve is divided into the action segment and the non-action segment.
[0044] The curve slope method used for division is to detect point by point from front to back in chronological order in the standard curve, and take the point where the slope suddenly and continuously increases as the starting point of the action segment; to detect point by point from back to front in chronological order in the standard curve, and take the point where the slope suddenly and continuously increases as the end point of the action segment, so as to realize the division of the action segment and non-action segment of the standard curve.
[0045] Divide the action segment: Based on the curve slope method, extract the starting and ending points of the stroke and coil current action segments from the standard curve, and divide the curve into action segments and non-action segments. The action segment is Figure 2 and Figure 3 The area framed in the middle.
[0046] The curve diagram of the opening action section of the selected opening standard data curve is as follows: Figure 2As shown, the opening standard data curve CB_2022-2-16-17-24-11-278-768.dat includes the stroke, opening coil current and closing coil current parts, among which the respective action segments of the stroke and opening coil current are framed in the figure.
[0047] The curve diagram of the closing action section of the selected closing standard data curve is as follows: Figure 3 As shown, the closing standard data curve CB_2024-2-16-16-46-42-88-337.dat includes the stroke, the open coil current and the closing coil current parts, wherein the respective action segments of the stroke and the closing coil current are framed in the figure.
[0048] ④ Match the action segment of the curve to be tested with the action segment of the standard curve, and output the curve similarity coefficient.
[0049] Action segment detection: Match the action segment of the curve to be tested with the action segment of the standard curve, that is, calculate the correlation coefficient between the curve to be tested and the standard curve. The built-in function of the programming software can be used to directly output the curve similarity coefficient.
[0050] ⑤ Match and detect the historical test data one by one according to step ④, draw a curve similarity scatter plot of all the data in the folder, and determine the detection threshold based on the overall similarity coefficient distribution.
[0051] Considering that the similarity coefficient of most normal data is within a stable data range, such as Figure 4 and Figure 5 As shown in FIG, except for some fault data with low similarity coefficients, the similarity coefficients of most normal data are within a stable data range, so the minimum value of the normal data similarity coefficient is selected as the detection threshold.
[0052] Detection threshold setting: Perform matching detection of step ④ on all historical test data in the folder to be tested one by one, output the curve similarity coefficient of all data, and draw a scatter plot of similarity coefficient in the order of data collection time. The summary scatter plot of the similarity coefficient of the tripping curve is as follows: Figure 4 As shown in the figure, the summary scatter plot of the closing curve similarity coefficient is as follows Figure 5 As shown, the detection threshold is set rationally according to the distribution of the overall similarity coefficient, and the detection threshold is updated to the YAML parameter configuration file.
[0053] ⑥ Take the single action data as the input to be tested, match the detection output curve similarity coefficient according to its action type, determine the relationship between the coefficient and the detection threshold, and output the corresponding detection result (normal, warning, fault).
[0054] Single action evaluation detection: input single action data, first determine its action type, then match it with the standard curve of the corresponding action type (opening or closing) to detect the output similarity coefficient α (curve similarity coefficient). If α is greater than the detection threshold, the output is "normal"; if α is equal to the detection threshold, the output is "warning"; if α is less than the detection threshold, the output is "fault". The overall detection process is as follows: Figure 1 shown.
[0055] Among them, the selection steps of step ② are as follows:
[0056] ① Create an alternative data folder (alternative sample data folder): select several groups of data with normal operating status from the test data as alternative sample data. Specifically, select more than 15 groups of opening / closing / reclosing test data with different collection times, good operating status and less interference signals from the collected GIS circuit breaker mechanical characteristics test data as alternative sample data and store them in the alternative sample data folder.
[0057] ② Data preprocessing: complete the preprocessing work such as parsing, classification, and alignment of the data in the candidate sample data folder; preprocessing includes data classification statistics, that is, in the algorithm program, first classify and organize the data in the candidate data folder according to the action type. Preprocessing also includes data parsing and alignment, that is, parsing the test data, intercepting the stroke and coil current data of the first 600ms (3000 data points) for data alignment.
[0058] ③ Fitting standardized curve: weighted average the data curves of all single action types to obtain a standardized fitting curve; specifically, weighted average all data of each action type phase by phase to obtain the standardized fitting curve of opening / closing / reclosing.
[0059] ④ Curve difference matrix, the action type data curve is differentiated with the standardized fitting curve one by one to obtain the curve difference matrix.
[0060] Specifically, the curve difference is first performed: according to the action type, the data in the candidate folder is subtracted from the standard fitting curve one by one and one by one to obtain the difference curve. Then the difference matrix is performed: the maximum and minimum values in the difference curve are subtracted, and the difference of each phase of each group of data is a difference element, and finally the difference matrix is obtained.
[0061] ⑤ Output standard data: Difference the maximum and minimum values in the difference curve matrix again, and use the curve to which the minimum difference belongs as standard data and output it to the parameter configuration file. The curve to which the minimum element in the difference matrix belongs is the standard data, and the standard data file name is updated in the yaml parameter configuration file.
[0062] Through the research of this method and the programming implementation of the algorithm module, the present invention lays a foundation for the comprehensive evaluation of the mechanical state of the GIS circuit breaker. It will help the power equipment operation and maintenance personnel to make a more accurate evaluation of the mechanical state of the GIS circuit breaker based on the characteristic values and comprehensive evaluation results. It also provides important visualization input and GIS comprehensive scoring criteria for operation and maintenance systems such as remote operation and maintenance of high-voltage switchgear and digital twin systems.
[0063] The present invention can be applied to the condition monitoring and fault alarm of the mechanical characteristics of GIS circuit breakers. By finding the dividing point between the action segment and the non-action segment of the stroke and coil current curves, the curves are segmented and processed, and the curve similarity detection is performed using the action segment curves. Through comprehensive statistics of a large amount of test data, the curve similarity detection thresholds of various types of data are determined, providing more accurate criteria for equipment operation and maintenance personnel to mechanically evaluate equipment status.
[0064] In the current similarity detection of mechanical characteristic curves of GIS circuit breakers, the detection object is the entire action curve, and different types of curves use a unified detection threshold. This detection method references a large number of interference signals, which overall reduces the detection accuracy of curve similarity; due to the obvious characteristic differences between the stroke curve and the coil current curve, the similarity coefficients of the stroke curve and the coil current curve are also different. Using a unified detection threshold, it is impossible to accurately complete the operating state detection of the mechanical characteristics of the operating mechanism. The present invention adopts a segmented comparison method, only extracting the action segment of the curve for similarity detection, which greatly improves the signal-to-noise ratio, and then improves the reliability of the detection threshold under normal conditions. Different detection thresholds are set for different types of curves, which improves the detection accuracy.
[0065] The present invention divides the mechanical characteristic curve into an action section and a non-action section, and only performs curve similarity detection on the action section curve; adopts the curve slope method to accurately calibrate the dividing point between the action section and the non-action section; and sets a reasonable detection threshold through the statistical results of a large amount of test data, thereby improving the accuracy of mechanical characteristic status evaluation such as GIS circuit breaker stroke current, which helps power equipment operation and maintenance personnel to more accurately evaluate and judge the GIS mechanical status from the mechanical characteristic sensor data.
[0066] An embodiment of a GIS circuit breaker mechanical characteristic curve similarity detection system:
[0067] A GIS circuit breaker mechanical characteristic curve similarity detection system includes a processor, and the processor is used to execute instructions to implement a GIS circuit breaker mechanical characteristic curve similarity detection method. Among them, a GIS circuit breaker mechanical characteristic curve similarity detection method has been described in detail in an embodiment of a GIS circuit breaker mechanical characteristic curve similarity detection method, which will not be repeated here.
Claims
1. A method for detecting similarity of mechanical characteristic curves of GIS circuit breakers, characterized in that: The method comprises: outputting a curve similarity coefficient by matching a standard curve with the action data to be tested; if the curve similarity coefficient is greater than a detection threshold, it is considered that the mechanical characteristics of the circuit breaker are normal; if the curve similarity coefficient is equal to the detection threshold, it is considered that the mechanical characteristics of the circuit breaker have a degradation trend; if the curve similarity coefficient is less than the detection threshold, it is considered that the mechanical characteristics of the circuit breaker are faulty; the detection threshold is determined according to the distribution of a plurality of curve similarity coefficients, and the plurality of curve similarity coefficients are obtained by matching the action segments of a plurality of mechanical characteristic test data with the action segments of the corresponding standard curve.
2. The GIS circuit breaker mechanical characteristic curve similarity detection method according to claim 1 is characterized in that: The standard curve for matching the motion data to be measured is selected according to the motion type of the motion data to be measured.
3. The GIS circuit breaker mechanical characteristic curve similarity detection method according to claim 1 is characterized in that: A plurality of groups of test data of normal operating state of each action type are taken as candidate data from the pre-acquired mechanical characteristic test data, and a standard curve corresponding to each action type is selected from the candidate data.
4. The GIS circuit breaker mechanical characteristic curve similarity detection method according to claim 1 is characterized in that: The action data to be tested includes multiple action test data of each action type of the circuit breaker serving as a test prototype.
5. The GIS circuit breaker mechanical characteristic curve similarity detection method according to claim 1 is characterized in that: The action segment of the standard curve is obtained using the curve slope method.
6. The GIS circuit breaker mechanical characteristic curve similarity detection method according to claim 1 is characterized in that: The action segment of the mechanical property test data is obtained using the curve slope method.
7. The method for detecting similarity of mechanical characteristic curves of GIS circuit breakers according to claim 3 is characterized in that: The standard curve was selected according to the minimum difference method.
8. The method for detecting similarity of mechanical characteristic curves of GIS circuit breakers according to claim 3 is characterized in that: The number of groups of the test data ranges from 15 to 35.
9. The GIS circuit breaker mechanical characteristic curve similarity detection method according to claim 2 or 3, characterized in that: The action types include closing and opening.
10. A GIS circuit breaker mechanical characteristic curve similarity detection system, comprising a processor, characterized in that: The processor is used to execute instructions to implement the GIS circuit breaker mechanical characteristic curve similarity detection method as described in any one of claims 1 to 9.
