Method for monitoring open circuit of high-voltage MSD

Abstract

The invention relates to the technical field of high-voltage circuit break switch monitoring, in particular to a high-voltage MSD circuit break monitoring method which comprises the following steps: setting a sampling window of a plug terminal assembly, recording a displacement track in a plugging process, reading a voltage sequence and eliminating interference, matching a trigger signal with an action section, collecting vibration data and extracting interference characteristics. And generating a high-voltage MSD monitoring result. According to the method, the time sequence tag sequence of the plugging action is established, each stage in the plugging process is accurately identified, the accuracy of action identification is improved by combining signal calibration and time sequence binding, dynamic monitoring of the plugging action is ensured by eliminating noise and interference and enhancing identification and response of abnormal conditions, and the accuracy of action identification is improved in a complex environment. According to the invention, the plugging process can be analyzed with high precision, the monitoring precision, reliability and anti-interference capability of the system on the high-voltage plugging device are improved, and continuous monitoring and accurate judgment of the operation state are ensured.

Description

Technical Field

[0001] This invention relates to the field of high-voltage circuit breaker monitoring technology, and in particular to a monitoring method for high-voltage MSD circuit breaker monitoring. Background Technology

[0002] The field of high-voltage circuit breaker monitoring technology involves identifying, recording, and analyzing the operating status of circuit breakers or disconnectors in high-voltage electrical systems. Its core aspects include detecting the opening and closing status of high-voltage switches, assessing mechanical life, determining arc generation, and identifying the continuity of conductive circuits. This technology is widely used in electric vehicles, rail transit, high-voltage power distribution equipment, and power equipment operation and maintenance management. By monitoring key parameters such as the opening and closing behavior, contact status, and number of operations of circuit breakers, it provides technical support for the safe operation of the system. Its technical implementation typically encompasses switch structure feature analysis, electrical signal status acquisition, auxiliary contact configuration, mechanical structure posture recognition, data acquisition, and judgment strategies, and has become an important component of the intelligentization and safety enhancement of high-voltage electrical systems. The traditional high-voltage MSD circuit breaker monitoring method refers to a monitoring method used to identify the high-voltage on / off status of the manual service disconnection device when it is plugged in or not. It mainly involves setting up plug terminals and plug assemblies, with insertable prongs in the plug and female terminal assemblies in the socket. The insertion of the prongs forms a closed circuit, and the on / off status is determined by the physical contact relationship. In some schemes, auxiliary contacts or mechanical linkage devices are used to determine the circuit breaker status. External mechanical detection mechanisms, contact auxiliary switches, or visual sensing structures are often used to realize the on / off status perception, and then to transmit signals or provide alarm prompts.

[0003] Existing technologies rely on contact structures to determine on / off states, and action state recognition is limited to outcome-based judgments. It is difficult to achieve dynamic analysis of the entire insertion and removal process, resulting in a lack of effective identification of incomplete execution, mid-process pauses, or repeated operations. On / off judgments are subject to delays and deviations. The signal acquisition strategy is simplistic and cannot effectively eliminate non-structural interference. Factors such as vibration and shock can easily lead to signal misjudgment. Auxiliary switches or external detection structures are significantly affected by installation accuracy and structural stability, lacking a direct mapping relationship with the actual action chain. Monitoring data cannot form a high-time-resolution response system, affecting the continuity and reliability of on / off recognition, and failing to meet the high-precision monitoring requirements for the operating status of high-voltage insertion and removal devices. Summary of the Invention

[0004] To achieve the above objectives, the present invention adopts the following technical solution: a monitoring method for high-voltage MSD circuit breaker monitoring, comprising the following steps: S1: Set the axial sampling window of the plug-in terminal assembly in the high-voltage MSD socket housing to include an image acquisition device, record the continuous displacement trajectory of the plug-in terminal assembly during the plugging and unplugging process, divide the action segments according to the direction change, speed trend and stop point, and generate a contact action label sequence; S2: Based on the contact action tag sequence, read the voltage sequence from the high-voltage MSD low-voltage synchronization circuit, eliminate jitter and unstructured fluctuations, screen the descent start point and locate the timing position, and generate a calibrated trigger signal group; S3: Match the calibrated trigger signal group with the action segment of the contact action tag sequence, identify whether it covers the insertion and removal behavior, fill in the missing binding points, extract the time boundary and trigger type, and generate a set of insertion and removal linkage segments. S4: Extract the set of insertion and removal linkage segments, collect vibration data of the high-voltage MSD housing, perform noise reduction and standardization processing on the triaxial signal, extract slope changes and amplitude transitions, filter time periods that meet the interference characteristics, and generate a set of signal interference intervals. S5: Compare the signal interference interval set with the trigger structure of the pull-out action in the insertion and removal linkage segment set, identify the non-triggered and misaligned trigger segments, verify the monitoring results according to the preset interference impact judgment threshold, and generate high voltage MSD monitoring results.

[0005] As a further embodiment of the present invention, the contact action tag sequence includes action type tag, displacement change characteristics, velocity distribution pattern, and action stagnation mark; the calibrated trigger signal group includes signal start point, boundary reference value, stable change point, and time series identifier; the plug-in linkage segment set includes behavior interval mapping, trigger point binding relationship, response missing mark, and start and end corresponding tag; the signal interference interval set includes slope change segment, amplitude change segment, covariance feature region, and duration threshold segment; and the high voltage MSD monitoring result includes untriggered segment, triggered misaligned segment, action signal matching rate, and circuit breaker monitoring judgment value.

[0006] As a further aspect of the present invention, the removal of jitter and unstructured fluctuations refers to removing invalid signal disturbances in the voltage sequence caused by random noise and irregular changes.

[0007] As a further aspect of the present invention, the screening of time periods that meet the interference characteristics refers to extracting time periods from the vibration signal that simultaneously possess slope change and amplitude transition characteristics, and identifying behavioral segments affected by interference.

[0008] As a further aspect of the present invention, the specific steps of S1 are as follows: S101: Obtain the insertion and removal image frame sequence of the plug-in terminal assembly along the axial channel of the socket housing, perform pixel intensity projection based on the edge region of the plug in the image, extract the pixel coordinates of the plug edge, call the edge coordinates of the whole frame, and generate a displacement trajectory sequence. S102: Based on the sign change of the displacement direction of adjacent frames in the displacement trajectory sequence, extract the sign change position index, divide the displacement trajectory into intervals, calculate the displacement-to-time ratio of the intervals, call the ratio trend, and generate an action segment division annotation set. S103: Based on the action segment division label set, extract the speed trend and boundary symbol direction, determine the action state within the segment, mark segments with the same direction as insertion and extraction, mark segments with zero speed as termination, mark segments with opposite directions as reversal, and establish a contact action label sequence.

[0009] As a further aspect of the present invention, the specific steps of S2 are as follows: S201: Based on the time range corresponding to the tags in the contact action tag sequence, continuously read the voltage data sequence from the low-voltage synchronization circuit of the high-voltage MSD, and divide each data segment into signal segments according to the time index, corresponding to the time window of each tag, and establish a time segment voltage sequence set. S202: Based on the signal curves in the voltage sequence set of the time segments, detect the baseline level values ​​in the interval between the start boundary and the end boundary, and combine the voltage change amplitude and the signal oscillation period threshold to eliminate the segment intervals with periodic jitter and noise fluctuations, thereby obtaining a stable boundary voltage signal set; S203: Based on the stable boundary voltage signal set, analyze the changing trend of the voltage curve, extract the time point when the voltage first shows a continuous downward trend, and combine the sequential position of the tag in the original time series to perform time sequence calibration on the extracted time point according to the time sequence of the tag, and establish a calibrated trigger signal group.

[0010] As a further aspect of the present invention, the specific steps of S3 are as follows: S301: Based on the trigger point time index in the calibrated trigger signal group, call the corresponding type of tag time period in the contact action tag sequence, determine whether each trigger point is located in the corresponding tag segment, and establish the binding relationship between the trigger point and the tag in chronological order to generate a trigger tag binding reference set; S302: Based on the time period of the tag that has not been bound in the trigger tag binding reference set, identify the insertion position in the original tag order, construct the missing trigger point data frame, fill the missing correspondence, call the original binding relationship for combination correction, and obtain the missing supplementary relationship table; S303: Based on each binding relationship in the missing supplementary relationship table, determine the start and end time coverage range of the trigger point in the tag segment, extract the start and end indexes as identifiers, aggregate the trigger and tag index information, and establish a set of plug-in linkage segments.

[0011] As a further aspect of the present invention, the specific steps of S4 are as follows: S401: Extract the start and end time periods of each group in the set of insertion and removal linkage segments, collect the triaxial vibration signal sequence in the high-voltage MSD structure shell, perform linear trend elimination and mean normalization processing on the channel signal based on the triaxial data channel, and obtain the triaxial normalized signal sequence group corresponding to each linkage segment. S402: Based on each group of linked segment signals in the triaxial normalized signal sequence group, calculate the slope change rate, amplitude transition amplitude and covariance matrix eigenvalues ​​between the three channels in the continuous signal, and aggregate them into a multidimensional feature matrix to obtain the linked segment perturbation feature set. S403: Based on the variation range of duration, slope and amplitude in the disturbance feature set of the linked segment, combined with the preset duration threshold and disturbance joint judgment conditions, the time intervals in which the disturbance index meets the threshold conditions are screened to establish a signal interference interval set.

[0012] As a further aspect of the present invention, the specific steps of S5 are as follows: S501: Based on the start and end time of each interference segment in the signal interference interval set, call the corresponding trigger structure and tag type in the plug-in linkage segment set, perform cross-matching on all segments in the plug-out tag that have not been triggered or whose trigger position has deviated, and mark whether they fall into the interference segment to generate an interference impact matching list. S502: Based on all the marked segments in the interference impact matching list, the matching situation between the corresponding action type and the trigger response is statistically analyzed. The ratio of the number of actions with missing or offset responses to the total number of tags is normalized to obtain the statistical value of matching integrity. S503: Based on the matching integrity statistics and the trigger response status in the interference impact matching list, construct a combined state judgment matrix, and perform assignment operations according to the matching level and interference label mapping relationship, such as the preset encoding rules like the third-order matching priority method and matrix cross-numbering method, to establish high-voltage MSD monitoring results.

[0013] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, by establishing a sequence of timing tags for insertion and removal actions, each stage of the insertion and removal process can be accurately identified. By combining signal calibration and timing binding, the accuracy of action recognition is improved. By eliminating noise and interference, the identification and response to abnormal situations are enhanced, ensuring dynamic monitoring of insertion and removal behavior. In complex environments, the insertion and removal process can be analyzed with high precision, improving the system's monitoring accuracy, reliability, and anti-interference capability for high-voltage insertion and removal devices, and ensuring continuous monitoring and accurate judgment of the operating status. Attached Figure Description

[0014] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0015] Figure 1 This is a schematic diagram of the steps of the present invention; Figure 2 This is a detailed schematic diagram of S1 of the present invention; Figure 3 This is a detailed schematic diagram of S2 of the present invention; Figure 4 This is a detailed schematic diagram of S3 of the present invention; Figure 5 This is a detailed schematic diagram of S4 of the present invention; Figure 6 This is a detailed schematic diagram of S5 of the present invention. Detailed Implementation

[0016] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0017] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0018] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0019] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0020] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0021] Please see Figure 1 This invention provides a monitoring method for high-voltage MSD circuit breaker monitoring, comprising the following steps: S1: In the high-voltage MSD, a sampling window is set for the plug-in terminal assembly along the axial channel of the socket housing. The continuous displacement trajectory of the plug-in terminal assembly during the insertion and removal process is obtained by time-progressing recording. Based on the sign of the displacement direction change, the trend distribution of the speed change, and the position of the action stop point, the physical motion process of the plug-in terminal assembly is divided into time sequences. Combined with the boundary behavior judgment of the motion segment, the insertion, removal, abort, and reverse action identifiers are constructed, and the contact action label sequence is output. S2: Based on the time range of the tags in the contact action tag sequence, continuously read the voltage data sequence from the low-voltage synchronization circuit of the high-voltage MSD, and perform boundary limitation and baseline judgment on the signal within each tag time period, eliminate periodic jitter, noise fluctuation and unstructured changes, select the first time point of stable decline of the curve, calibrate in sequence according to the position in the tag time period, and output the calibrated trigger signal group. S3: For the trigger points in the calibrated trigger signal group, perform corresponding operations with the time periods of the same type of label in the contact action label sequence, identify whether the trigger points are located within the insertion and removal behavior range, and establish binding in chronological order. If there are missing or unresponsive situations, fill in the missing corresponding points. By judging the trigger coverage relationship in the matching segment, construct start and end markers and trigger response labels, and output the set of insertion and removal linkage segments. S4: Extract the start and end time periods of each set of plug-in linkage segments, collect vibration data recorded in the high-pressure MSD structure shell, perform linear denoising and mean normalization on the triaxial signal, extract the slope change, amplitude transition and signal covariance in the continuous signal within each linkage segment time period, and filter them in combination with the duration threshold to output the signal interference interval set. S5: Cross-map each marked segment in the signal interference interval set with the label category and trigger structure in the plug-in linkage segment set, identify segments in the plug-out label where the trigger did not occur or the trigger was misaligned, check whether they fall into the interference segment, statistically analyze the matching integrity of the action and the signal under the influence of the interference segment, construct a discrimination structure through the combination relationship and assign values ​​to determine the result, and output the high voltage MSD monitoring result.

[0022] The contact action label sequence includes action type label, displacement change characteristics, velocity distribution pattern, and action stagnation marker. The calibrated trigger signal group includes signal start point, boundary reference value, stable change point, and time series identifier. The plug-in linkage segment set includes behavior interval mapping, trigger point binding relationship, response missing marker, and start and end corresponding labels. The signal interference interval set includes slope change segment, amplitude change segment, covariance characteristic area, and duration threshold segment. The high voltage MSD monitoring results include untriggered segments, triggered misaligned segments, action signal matching rate, and circuit breaker monitoring judgment value.

[0023] Please see Figure 2 The specific steps of S1 are as follows: S101: Obtain the insertion and removal image frame sequence of the plug-in terminal assembly along the axial channel of the socket housing, perform pixel intensity projection based on the edge region of the plug in the image, extract the pixel coordinates of the plug edge, call the edge coordinates of the whole frame, and generate a displacement trajectory sequence. First, the industrial-grade high-speed vision sensor positioned directly above the high-voltage manual maintenance switch socket is activated. The sampling frequency is set to 200 frames per second, and the image resolution is set to 1024 pixels by 1024 pixels to ensure the capture of minute displacement changes in the connector assembly during high-speed insertion and removal. The process continuously acquires real-time images of the connector assembly relative to the socket housing, storing each frame in the frame buffer queue in chronological order of acquisition. Then, the process calls a preset region of interest parameter, which defines a rectangular area in the image containing only the axial channel of the socket housing. For example, a region with a width of 200 pixels and a height of 600 pixels is selected as the processing object, starting from the image center coordinates. For each frame within this region of interest, the process performs a pixel intensity projection operation. Specifically, the process accumulates the grayscale values ​​of each row of pixels within the region along the horizontal direction, generating a one-dimensional grayscale projection vector in the vertical direction. Each element of this vector represents the sum of the brightness of all pixels in the corresponding row. The process iterates through the one-dimensional grayscale projection vector, calculates the difference between adjacent elements to obtain the grayscale gradient distribution, and identifies the position with the largest gradient value as the edge position of the connector component, thereby extracting the edge pixel ordinates of the connector in the current frame. This edge extraction operation is repeated for each frame in the buffer queue, and the extracted edge ordinates are arranged in timestamp order to construct a displacement trajectory sequence reflecting the change in the connector component's position over time. For example, an edge ordinate of 100 pixels is detected in frame 1, an edge ordinate of 120 pixels is detected in frame 10, and so on, forming continuous trajectory data.

[0024] S102: Based on the sign change of the displacement direction of adjacent frames in the displacement trajectory sequence, extract the position index of the sign change, divide the displacement trajectory into intervals, calculate the ratio of displacement to time in the interval, call the ratio trend, and generate the action segment division annotation set. The process performs a difference operation on the displacement coordinates of two adjacent time points in the sequence. Specifically, the process subtracts the coordinate value of the previous time point from the coordinate value of the current time point to obtain the displacement change. The process detects the sign of this displacement change; a positive sign represents positive movement, a negative sign represents negative movement, and zero sign represents stillness. The process records the time index position where the sign of the displacement direction reverses, such as the moment it changes from positive to zero or from negative to positive. These moments are used as dividing points to segment the continuous displacement trajectory sequence into several independent action intervals. For each action interval, the process calculates the ratio of its total displacement to its duration, i.e., subtracting the starting coordinate from the coordinate at the end of the interval to obtain the displacement difference, and then dividing by the product of the number of sampling points and the sampling period for that interval to obtain the average velocity value for that interval. The process associates the calculated velocity values ​​with the time index to generate a ratio trend sequence containing information on the magnitude and direction of the velocity. To ensure the accuracy of action segmentation, a minimum displacement threshold, such as 5 pixels, is set. If the total displacement of a certain interval is less than this threshold, it is usually considered measurement noise and discarded or merged into an adjacent interval. However, if the interval has complete velocity trends or boundary features in subsequent identification, its label is retained to form an independent action segment. Finally, the process integrates the filtered and calculated intervals and their corresponding velocity attributes to generate an action segmentation annotation set.

[0025] S103: Based on the action segment division of the annotation set, extract the speed trend and boundary symbol direction, determine the action state within the segment, mark segments with the same direction as insertion and extraction, mark segments with zero speed as termination, mark segments with opposite directions as reversal, and establish a contact action label sequence. The process iterates through each segment of the action segment label set. For each segment, the process first extracts the average value of its velocity trend and the displacement direction sign at the boundary. The process determines the action state within the segment based on preset logical rules: if the displacement direction sign within a segment remains consistent and the absolute velocity value is greater than zero, the process determines it as a continuous action; when the direction is positive, it is marked as an insertion action, and when the direction is negative, it is marked as a withdrawal action. If the average velocity calculation result within a segment is zero, or the displacement change is within a preset static tolerance range (e.g., within ±1 pixel), the process marks it as an aborted state. If a direction sign within a segment is detected to reverse instantly within a short period, the process marks it as a reverse adjustment action. The process binds the determined state labels (e.g., insertion, withdrawal, abort, reverse) to the corresponding start and end points of the time period, arranges them in chronological order, and establishes a contact action label sequence.

[0026] As shown in Table 1, Table 1 lists the displacement and motion recognition data obtained through the above steps during a complete insertion and removal monitoring process. The table records in detail the edge coordinate changes, the calculated average velocity, and the final motion label within different time windows, verifying that the trajectory analysis method based on visual projection can accurately distinguish between different operation states such as insertion, stationary, and removal.

[0027] Table 1. Data table for generating contact action tag sequences.

[0028] In the above calculation process, the setting of the action judgment threshold adopted a statistical analysis method based on historical data. The process selected 100 sets of trajectory data from standard insertion and removal operations, calculated the standard deviation of background noise displacement in each set of stationary states, and the calculated standard deviation was 0.8 pixels. Based on the principle of three times the standard deviation, the stationary tolerance threshold was set to 0.8 pixels multiplied by 3, i.e., 2.4 pixels (rounded to 3 pixels). The advantage of this threshold setting is that by determining the noise boundary through statistical methods, it can effectively filter out minute coordinate drifts caused by environmental vibrations or sensor noise, preventing the misjudgment of a stationary state as a minor movement, thereby improving the robustness of action state recognition.

[0029] Please see Figure 3 The specific steps of S2 are as follows: S201: Based on the time range corresponding to the tag in the contact action tag sequence, continuously read the voltage data sequence from the low-voltage synchronization circuit of the high-voltage MSD, and divide each data segment into signal segments according to the time index, corresponding to the time window of each tag, and establish a time segment voltage sequence set. First, the voltage analog signal is continuously read at a sampling rate of 10 kHz through the low-voltage synchronous monitoring circuit interface integrated inside the high-voltage manual maintenance switch, and the analog signal is converted into a digital voltage data sequence. The process reads the contact action tag sequence generated in step S103, extracting the start and end timestamps corresponding to each tag. Using these timestamps as indexes, the process performs a segmentation operation on the continuous voltage data sequence, extracting the voltage data belonging to the time range of each action tag to form an independent signal segment. For example, if the time range of tag 1 is 1.0 to 1.5 seconds, the process extracts 5000 voltage sampling points corresponding to that 0.5-second duration. The process repeats this operation for the data segments corresponding to all tags, and associates and stores the extracted voltage data segments with the corresponding action tag IDs to establish a time segment voltage sequence set.

[0030] S202: Based on the signal curves in the voltage sequence set of time segments, detect the baseline level values ​​in the interval between the start boundary and the end boundary, and combine the voltage change amplitude and the signal oscillation period threshold to remove the segment intervals with periodic jitter and noise fluctuations, thereby obtaining a stable boundary voltage signal set; First, the first 10% of data points in a signal segment are selected, and their arithmetic mean is calculated as the baseline level for that segment. Then, the process calculates the difference between the maximum and minimum values ​​of all data points within the signal segment to obtain the voltage change amplitude. Simultaneously, the process uses zero-crossing detection to count the number of oscillations of the signal per unit time, calculating the oscillation period. The process introduces preset period jitter thresholds and noise fluctuation amplitude thresholds, for example, setting the oscillation period threshold to 0.02 seconds and the noise fluctuation amplitude threshold to 0.5 volts. The process compares the calculated actual oscillation period with the period thresholds and the actual voltage change amplitude with the noise fluctuation amplitude thresholds. If the oscillation period of a segment is less than the threshold and the voltage change amplitude is greater than the noise threshold, the process determines that high-frequency period jitter or electromagnetic interference noise exists in that interval, and marks the data in that interval as invalid or performs smoothing filtering. After elimination or correction, the process retains segments with clear signal characteristics and no significant high-frequency noise interference, obtaining a stable boundary voltage signal set.

[0031] S203: Based on the stable boundary voltage signal set, analyze the changing trend of the voltage curve, extract the time point when the voltage first shows a continuous downward trend, and combine the sequential position of the tag in the original time series to perform time-series calibration on the extracted time point according to the time sequence of the tag, and establish a calibrated trigger signal group. Trend analysis is performed on each voltage curve in the stable boundary voltage signal set. The process employs a sliding window slope calculation method, setting the window size to 10 sampling points. Linear regression analysis is performed on the data within each window to calculate the instantaneous slope, ensuring the window width is less than the typical signal change duration to improve the accuracy of capturing short-term dynamic changes. The process monitors the slope value in real time. When a continuous negative slope value is detected, and the duration of the negative value exceeds a preset drop confirmation time (e.g., 5 milliseconds), the process records the moment the slope first becomes negative and extracts it as the voltage drop start point. Similarly, the process extracts the voltage recovery start point. The process aligns and compares the extracted voltage change time points with the time axis in the original contact action tag sequence, sorting the extracted trigger signals according to their chronological order. The process calibrates for time deviations; for example, if the voltage drop point lags behind the visual insertion end point, the lag difference is recorded. Finally, a calibrated trigger signal group containing precise trigger times and sorting information is established.

[0032] As shown in Table 2, this table presents the specific data for baseline analysis and trend detection of the extracted voltage signal segments. By calculating the baseline level and identifying the continuous downward trend, the circuit triggering moment corresponding to the insertion and removal actions was accurately captured, and some interference segments containing high-frequency noise were eliminated.

[0033] Table 2. Analysis and Trigger Calibration of Steady Boundary Voltage Signals

[0034] In the above steps, the setting of the voltage drop confirmation time parameter was experimentally verified. The process selected 50 transient waveforms caused by poor contact and 50 normal plug-in / plug-out waveforms. The process calculations showed that the average voltage drop duration of the transient waveform was approximately 1 millisecond, while the average voltage drop duration caused by normal plug-in / plug-out was greater than 10 milliseconds. To distinguish between the two, the process selected a value near the midpoint of their averages, setting the voltage drop confirmation time threshold to 5 milliseconds. In one test, a duration of continuously negative voltage slope was detected for 1.2 milliseconds, far less than the system's original 5-millisecond confirmation time and sliding window width. Because the duration of the signal change was insufficient to support a valid trend judgment, the process classified it as jitter. In another test, a continuously negative slope lasted for 12 milliseconds, significantly greater than the judgment threshold, and the process judged it as a valid trigger. In yet another test, the duration was 12 milliseconds, greater than 5 milliseconds, and the process judged it as a valid trigger. The advantage of this parameter setting is that, through threshold filtering in the time dimension, it effectively avoids false alarms caused by instantaneous voltage drops due to contact surface oxidation or fretting friction, thus ensuring the reliability of trigger signal extraction.

[0035] Please see Figure 4 The specific steps of S3 are as follows: S301: Based on the trigger point time index in the calibrated trigger signal group, call the corresponding type of tag time period in the contact action tag sequence, determine whether each trigger point is located in the corresponding tag segment, and establish the binding relationship between the trigger point and the tag in time order to generate a trigger tag binding reference set; The process reads the time index of each trigger point in the calibrated trigger signal group and simultaneously retrieves the time interval of the tag marked as "pull out" or "insert" in the contact action tag sequence. The process performs an inclusion check, verifying that the timestamp of each trigger point falls within the start and end time range defined by any tag. If the timestamp of a trigger point is within the time interval of a tag (allowing a preset time tolerance, e.g., ±0.1 seconds), the process determines that the trigger point corresponds to the action tag, binds the two, and records the ID of the binding pair. Following the original timeline, the process lists and stores all successfully bound trigger point and tag pairs, generating a trigger tag binding lookup set. This process ensures that changes in the physical circuit signals can be accurately attributed to specific mechanical actions.

[0036] S302: Based on the time periods of tags that have not been bound in the trigger tag binding reference set, identify the insertion position in the original tag order, construct a missing trigger point data frame, fill in the missing correspondence, call the original binding relationship for combination correction, and obtain the missing supplementary relationship table; The process scans the trigger tag binding reference set, searching for tag time periods that exist in the original contact action tag sequence but lack a corresponding trigger point in the reference set, identifying these as "incomplete binding" tags. The process analyzes the position index of these unbound tags in the original sequence to determine whether they belong to an insertion or removal sequence. For each tag corresponding to a missing trigger, the process constructs a missing trigger point data frame containing a virtual trigger time (the midpoint of the tag time period) and a status bit marked as "missing." The process fills these constructed missing data frames into the existing binding relationship list, maintaining the continuity of the time sequence. Subsequently, the process uses the existing binding relationships as a reference to logically correct the combined sequence; for example, if a removal trigger is missing between two consecutive insertion actions, the process marks that position as requiring special verification, ultimately obtaining a missing completion relationship table containing both measured binding and virtual completion information.

[0037] S303: Based on each binding relationship in the missing and supplementary relationship table, determine the start and end time coverage range of the trigger point in the tag segment, extract the start and end indices as identifiers, aggregate the trigger and tag index information, and establish a set of plug-in linkage segments. The process iterates through each record in the missing and supplementary relationship table. For each binding relationship (whether measured or supplementary), the process extracts the start time of the tag segment as the segment start time and the end time of the tag segment as the segment end time. Alternatively, if a corresponding measured trigger point exists, the trigger point time is included in the range consideration, and the union of the two is taken. The process uses the extracted start time index and end time index as the identifier of the linked segment. The process aggregates all information within the segment, including visual action type, voltage trigger status (normal or missing), time span, etc., to construct a comprehensive data object. The process performs this aggregation operation on all records to establish a set of plug-and-play linked segments, which provides a well-defined and segmented data foundation for subsequent multidimensional signal analysis. In the logical deduction of this step, a matching completeness coefficient is set for the identification logic of the "incomplete binding" tag. The process counts the number of bound tags and divides it by the total number of tags to obtain the matching rate. For example, if there are a total of 10 action tags, and 8 of them successfully match the voltage trigger signal, the matching rate is 0.8. The process extracts the time periods of the two unmatched tags, assuming their time spans are 0.5 seconds and 0.6 seconds respectively. During subsequent data reconstruction, virtual trigger markers are inserted at the midpoints of these two time periods (i.e., 0.25 seconds and 0.3 seconds). The advantage of this processing logic is that it ensures the integrity of the data structure, enabling subsequent steps to perform specialized attribution analysis on abnormal segments of vibration interference that exhibit "action without signal," without directly overlooking these critical potential fault points.

[0038] Please see Figure 5 The specific steps of S4 are as follows: S401: Extract the start and end time periods of each group in the set of insertion and removal linkage segments, collect the triaxial vibration signal sequence inside the high-voltage MSD structure shell, perform linear trend elimination and mean normalization processing on the channel signal based on the triaxial data channel, and obtain the triaxial normalized signal sequence group corresponding to each linkage segment. For each start and end time period in the set of plug-in / plug-out linkage segments, the corresponding vibration signal sequence is read from a triaxial accelerometer installed inside the housing of the high-voltage manual maintenance switch. The sensor collects acceleration data in three orthogonal directions: X-axis, Y-axis, and Z-axis. The process performs linear trend removal on each segment of raw triaxial data. Specifically, for each axis, a linear regression line is fitted using the least squares method, and the corresponding value on this line is subtracted from the raw signal to eliminate DC deviation caused by sensor zero-point drift or changes in the gravitational component. Subsequently, the process performs mean normalization on the detrended data, calculating the mean and standard deviation of the data segment. The mean is subtracted from each data point, and the result is divided by the standard deviation, transforming the data into a dimensionless metric space with a mean of 0 and a standard deviation of 1. This process is repeated for each of the three axes to obtain the triaxial normalized signal sequence set corresponding to each linkage segment.

[0039] S402: Based on each group of linked segment signals in the triaxial normalized signal sequence group, calculate the slope change rate, amplitude transition amplitude and covariance matrix eigenvalues ​​between the three channels in the continuous signal, and aggregate them into a multidimensional feature matrix to obtain the linked segment perturbation feature set. Feature calculations are performed on the data in the triaxial normalized signal sequence. The process begins by calculating the first-order difference of the continuous signal, taking the average of the absolute values ​​of the differences as the slope change rate to characterize the intensity of the vibration. Next, the difference between the maximum and minimum values ​​in the signal sequence is calculated to obtain the amplitude transition amplitude. Then, the eigenvalues ​​of the covariance matrix among the X, Y, and Z axis signals are calculated. Specifically, a 3x3 covariance matrix is ​​constructed, where the elements represent the correlation between signals along different axes. Eigenvalue decomposition is then performed on this matrix, and the largest eigenvalue is extracted as the principal component characterizing the triaxial coordinated vibration energy. Finally, the calculated slope change rate, amplitude transition amplitude, and largest covariance eigenvalue are vectorized and concatenated into a multidimensional feature matrix, yielding the linkage segment perturbation feature set.

[0040] S403: Based on the variation range of duration, slope and amplitude in the disturbance feature set of the linked segment, combined with the preset duration threshold and disturbance joint judgment conditions, the time intervals in which the disturbance index meets the threshold conditions are screened to establish a set of signal interference intervals. The process analyzes various indicators within the disturbance feature set of the linked segments. It sets preset duration thresholds (e.g., 50 milliseconds) and joint disturbance judgment conditions (e.g., amplitude transition amplitude greater than 3 and slope change rate greater than 2). The process scans the time series corresponding to the feature set, filtering out time intervals where all indicators simultaneously meet the aforementioned threshold conditions. Specifically, if the vibration characteristics continuously exceed the threshold for a period of time and the duration exceeds 50 milliseconds, the process determines that there is significant external mechanical disturbance or impact during that period. The process records the start and end points of these time intervals that meet the conditions, establishing a set of signal interference intervals.

[0041] Table 3 shows the triaxial vibration features extracted for different linkage segments and their determination results. By calculating the covariance eigenvalues ​​and amplitude indices, the potential impact of environmental vibration on the monitoring signal can be effectively quantified, providing a basis for subsequent fault attribution.

[0042] Table 3. Extraction of disturbance features and judgment of interference in linked segments.

[0043] In the above feature extraction calculation logic, the following operational logic is used for calculating the eigenvalues ​​of the covariance matrix: First, the normalized signal sequences for the X, Y, and Z axes are obtained. The sum of the products of the deviations of the X-axis and Y-axis signals is calculated and divided by the sample number minus one to obtain the XY covariance. Similarly, the autovariances of XZ, YZ, and each axis are calculated to construct a 3x3 covariance matrix. Subsequently, the process calls an eigenvalue decomposition algorithm (such as the Jacobi iteration method) to solve the matrix, obtaining three eigenvalues ​​λ1, λ2, and λ3. The process selects the maximum value λmax. For example, in the Frag_02 segment, the calculated principal eigenvalue of the covariance matrix is ​​12.4, which is much higher than the 0.8 of the normal segment Frag_01. The process compares this eigenvalue of 12.4 with a preset interference energy threshold (e.g., 5.0). Since 12.4 is greater than 5.0, the process determines that the segment is subjected to strong multi-axis coupled vibration interference. The advantage of this operational logic is that it can capture the spatial correlation and total energy of triaxial vibrations by using covariance eigenvalues, which reflects structural impact disturbances more accurately than threshold judgments for a single axis.

[0044] Please see Figure 6 The specific steps of S5 are as follows: S501: Based on the start and end times of each interference segment in the signal interference interval set, call the corresponding trigger structure and tag type in the plug-in linkage segment set, perform cross-matching on all segments in the plug-out tags that have not been triggered or whose trigger positions have deviated, and mark whether they fall into the interference segment to generate an interference impact matching list. The process reads the start and end data of all time periods identified as having interference from the signal interference interval set. Simultaneously, it calls the plug-in / plug-out linkage segment set and filters out all segments containing the "plug-out" tag. The process further subdivides these plug-out segments, focusing on extracting abnormal segments marked with "trigger not occurred" (i.e., action without voltage change) or "trigger position deviation" (i.e., the voltage change time deviates from the action time by more than a threshold). The process performs a cross-matching operation on the time axis to determine if the time range of each abnormal segment overlaps with any interference segment in the signal interference interval set. If overlap exists, and the overlap time percentage exceeds a preset proportion (e.g., 30%), the process marks the abnormal segment as an "affected segment"; otherwise, it marks it as a "non-interference anomaly." The process summarizes all cross-matched segments and their interference marking statuses to generate an interference impact matching list.

[0045] S502: Based on all the marked segments in the interference-affected matching list, the matching situation between the corresponding action type and the trigger response is statistically analyzed. The ratio of the number of actions with missing or offset responses to the total number of tags is normalized to obtain the statistical value of matching integrity. The process counts the number of anomalous segments marked as "affected by interference" in the list, as well as the total number of anomalous segments. It also counts the original total number of tags. The process calculates the number of segments with missing responses (triggers not occurring) confirmed as affected by interference, and the number of segments with offset responses confirmed as affected by interference. These statistics are then normalized to calculate a match integrity statistic. Specifically, the calculation logic is as follows: the process subtracts the number of segments that "both experienced anomalies but were not covered by interference ranges" from the total number of tags, and then divides this subtracted by the total number of tags to obtain a ratio reflecting the actual reliability after excluding external interference. A higher value indicates that the anomalies are mostly caused by vibration interference rather than a hard fault in the circuit itself; conversely, a lower value suggests a substantial failure in the circuit.

[0046] S503: Based on the statistical value of matching integrity and the trigger response status in the interference impact matching list, construct a combined state judgment matrix, and perform assignment operations according to the mapping relationship between matching level and interference label, such as the preset coding rules such as the third-order matching priority method and matrix cross numbering method, to establish the high voltage MSD monitoring results. Based on the calculated matching integrity statistics and the detailed states in the interference impact matching list, a multi-dimensional combined state judgment matrix is ​​constructed. The rows of this matrix represent the matching integrity level (e.g., high, medium, low), and the columns represent the interference coverage (e.g., full coverage, partial coverage, no coverage). The process searches for the corresponding cell in the matrix based on the actual calculation results and assigns a value to the cell according to preset encoding rules. For example, if the matching integrity is high and all anomalies fall within the interference zone, the value is assigned as "mechanical disturbance warning"; if the matching integrity is low and the anomalies do not fall within the interference zone, the value is assigned as "interlock circuit fault". The process outputs the final encoded value to establish the monitoring results for the high-voltage manual maintenance switch.

[0047] In the above evaluation logic, the calculation of the matching integrity statistics is performed using the following specific example: Assume there are a total of 20 plugged-in tags during the monitoring process (total tag count = 20). Four abnormal segments are detected (including two missing responses and two offset responses). After cross-matching by S501, it is found that the time periods of three of the abnormal segments overlap with the vibration signal interference interval (i.e., number affected by interference = 3), leaving only one abnormal segment that cannot be explained by vibration (i.e., number of substantial anomalies = 1). The process calculates the matching integrity statistics, first calculating the substantial anomaly rate as 1 divided by 20, which equals 0.05. The process subtracts 0.05 from 1 to obtain 0.95 (i.e., 95%). The process compares this result 0.95 with a preset baseline (e.g., 0.90). Since 0.95 is greater than 0.90, the process determines the monitoring result as "functionally normal, but with environmental interference." The advantage of this calculation logic is that by quantitatively eliminating environmental vibration factors, the true circuit failure rate can be accurately separated, avoiding false alarms caused by external factors such as vehicle bumps, and improving the accuracy of fault diagnosis.

[0048] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of protection of the technical solution.

Claims

1. A monitoring method for high-voltage MSD circuit breaker monitoring, characterized in that, Includes the following steps: S1: Set the axial sampling window of the plug-in terminal assembly in the high-voltage MSD socket housing to include an image acquisition device, record the continuous displacement trajectory of the plug-in terminal assembly during the plugging and unplugging process, divide the action segments according to the direction change, speed trend and stop point, and generate a contact action label sequence; S2: Based on the contact action tag sequence, read the voltage sequence from the high-voltage MSD low-voltage synchronization circuit, eliminate jitter and unstructured fluctuations, screen the descent start point and locate the timing position, and generate a calibrated trigger signal group; S3: Match the calibrated trigger signal group with the action segment of the contact action tag sequence, identify whether it covers the insertion and removal behavior, fill in the missing binding points, extract the time boundary and trigger type, and generate a set of insertion and removal linkage segments. S4: Extract the set of insertion and removal linkage segments, collect vibration data of the high-voltage MSD housing, perform noise reduction and standardization processing on the triaxial signal, extract slope changes and amplitude transitions, filter time periods that meet the interference characteristics, and generate a set of signal interference intervals. S5: Compare the signal interference interval set with the trigger structure of the pull-out action in the insertion and removal linkage segment set, identify the non-triggered and misaligned trigger segments, verify the monitoring results according to the preset interference impact judgment threshold, and generate high voltage MSD monitoring results.

2. The monitoring method for high-voltage MSD circuit breaker monitoring according to claim 1, characterized in that, The contact action label sequence includes action type label, displacement change characteristics, velocity distribution pattern, and action stagnation marker. The calibrated trigger signal group includes signal start point, boundary reference value, stable change point, and time series identifier. The plug-in linkage segment set includes behavior interval mapping, trigger point binding relationship, response missing marker, and start and end corresponding labels. The signal interference interval set includes slope change segment, amplitude change segment, covariance feature region, and duration threshold segment. The high-voltage MSD monitoring result includes untriggered segment, triggered misaligned segment, action signal matching rate, and circuit breaker monitoring judgment value.

3. The monitoring method for high-voltage MSD circuit breaker monitoring according to claim 1, characterized in that, The removal of jitter and unstructured fluctuations refers to removing invalid signal disturbances in the voltage sequence caused by random noise and irregular changes.

4. The monitoring method for high-voltage MSD circuit breaker monitoring according to claim 1, characterized in that, The selection of time periods that meet the interference characteristics refers to extracting time periods from vibration signals that simultaneously possess slope changes and amplitude transitions, thereby identifying behavioral segments affected by interference.

5. The monitoring method for high-voltage MSD circuit breaker monitoring according to claim 1, characterized in that, The specific steps of S1 are as follows: S101: Obtain the insertion and removal image frame sequence of the plug-in terminal assembly along the axial channel of the socket housing, perform pixel intensity projection based on the edge region of the plug in the image, extract the pixel coordinates of the plug edge, call the edge coordinates of the whole frame, and generate a displacement trajectory sequence. S102: Based on the sign change of the displacement direction of adjacent frames in the displacement trajectory sequence, extract the sign change position index, divide the displacement trajectory into intervals, calculate the displacement-to-time ratio of the intervals, call the ratio trend, and generate an action segment division annotation set. S103: Based on the action segment division label set, extract the speed trend and boundary symbol direction, determine the action state within the segment, mark segments with the same direction as insertion and extraction, mark segments with zero speed as termination, mark segments with opposite directions as reversal, and establish a contact action label sequence.

6. The monitoring method for high-voltage MSD circuit breaker monitoring according to claim 1, characterized in that, The specific steps of S2 are as follows: S201: Based on the time range corresponding to the tags in the contact action tag sequence, continuously read the voltage data sequence from the low-voltage synchronization circuit of the high-voltage MSD, and divide each data segment into signal segments according to the time index, corresponding to the time window of each tag, and establish a time segment voltage sequence set. S202: Based on the signal curves in the voltage sequence set of the time segments, detect the baseline level values ​​in the interval between the start boundary and the end boundary, and combine the voltage change amplitude and the signal oscillation period threshold to eliminate the segment intervals with periodic jitter and noise fluctuations, thereby obtaining a stable boundary voltage signal set; S203: Based on the stable boundary voltage signal set, analyze the changing trend of the voltage curve, extract the time point when the voltage first shows a continuous downward trend, and combine the sequential position of the tag in the original time series to perform time sequence calibration on the extracted time point according to the time sequence of the tag, and establish a calibrated trigger signal group.

7. The monitoring method for high-voltage MSD circuit breaker monitoring according to claim 1, characterized in that, The specific steps for S3 are as follows: S301: Based on the trigger point time index in the calibrated trigger signal group, call the corresponding type of tag time period in the contact action tag sequence, determine whether each trigger point is located in the corresponding tag segment, and establish the binding relationship between the trigger point and the tag in chronological order to generate a trigger tag binding reference set; S302: Based on the time period of the tag that has not been bound in the trigger tag binding reference set, identify the insertion position in the original tag order, construct the missing trigger point data frame, fill the missing correspondence, call the original binding relationship for combination correction, and obtain the missing supplementary relationship table; S303: Based on each binding relationship in the missing supplementary relationship table, determine the start and end time coverage range of the trigger point in the tag segment, extract the start and end indexes as identifiers, aggregate the trigger and tag index information, and establish a set of plug-in linkage segments.

8. The monitoring method for high-voltage MSD circuit breaker monitoring according to claim 1, characterized in that, The specific steps of S4 are as follows: S401: Extract the start and end time periods of each group in the set of insertion and removal linkage segments, collect the triaxial vibration signal sequence in the high-voltage MSD structure shell, perform linear trend elimination and mean normalization processing on the channel signal based on the triaxial data channel, and obtain the triaxial normalized signal sequence group corresponding to each linkage segment. S402: Based on each group of linked segment signals in the triaxial normalized signal sequence group, calculate the slope change rate, amplitude transition amplitude and covariance matrix eigenvalues ​​between the three channels in the continuous signal, and aggregate them into a multidimensional feature matrix to obtain the linked segment perturbation feature set. S403: Based on the variation range of duration, slope and amplitude in the disturbance feature set of the linked segment, combined with the preset duration threshold and disturbance joint judgment conditions, the time intervals in which the disturbance index meets the threshold conditions are screened to establish a signal interference interval set. The joint disturbance determination condition refers to the slope change rate exceeding a specific threshold and the amplitude transition amplitude being greater than a set value.

9. The monitoring method for high-voltage MSD circuit breaker monitoring according to claim 1, characterized in that, The specific steps of S5 are as follows: S501: Based on the start and end time of each interference segment in the signal interference interval set, call the corresponding trigger structure and tag type in the plug-in linkage segment set, perform cross-matching on all segments in the plug-out tag that have not been triggered or whose trigger position has deviated, and mark whether they fall into the interference segment to generate an interference impact matching list. S502: Based on all the marked segments in the interference impact matching list, the matching situation between the corresponding action type and the trigger response is statistically analyzed. The ratio of the number of actions with missing or offset responses to the total number of tags is normalized to obtain the statistical value of matching integrity. S503: Based on the matching integrity statistics and the trigger response status in the interference impact matching list, construct a combined state judgment matrix, and perform assignment operations according to the matching level and interference label mapping relationship, such as the preset encoding rules like the third-order matching priority method and matrix cross-numbering method, to establish high-voltage MSD monitoring results.

10. The monitoring method for high-voltage MSD circuit breaker monitoring according to claim 9, characterized in that, Whether the marker falls within the interference zone refers to determining and recording whether the time range of the triggered abnormal segment overlaps with any signal interference zone.

Images