Switch cabinet online monitoring analysis and fault pre-judgment method and system

Abstract

The invention relates to the technical field of equipment fault prediction, in particular to an on-line monitoring analysis and fault pre-judgment method and system for a switch cabinet of a switch cabinet, and the method comprises the following steps: collecting a voltage sequence, recognizing a main peak, recording the main peak, extracting a rear section of the main peak, analyzing amplitude change, judging discontinuous features, constructing a track trend sequence, recognizing continuous offset, and carrying out fault prediction. According to the method, through extraction and serialization processing of voltage main peak characteristics in the contact conduction period and combination of comparison of change amplitudes of sampling sections after main peaks, accurate identification of transient anomalies is achieved, and through dip angle change analysis of main peak tracks in the continuous periods, accurate identification of the transient anomalies is achieved, so that the accuracy of the transient anomalies is improved, and the accuracy of the transient anomalies is improved. According to the method, the deviation behavior of the running trend is recognized, abnormal sections which are consistent in direction and are distributed in a concentrated mode are positioned through cross comparison of abnormal periods and track deviation, the deviation trend is comprehensively judged based on the period number and direction characteristics, and the timeliness and accuracy of fault pre-judgment are effectively improved.

Description

Technical Field

[0001] This invention relates to the field of equipment fault prediction technology, and in particular to a method and system for online monitoring, analysis and fault prediction of switch cabinets. Background Technology

[0002] The field of equipment fault prediction technology involves technical means for continuous perception of the operating status of industrial equipment, identification of fault symptoms, and monitoring and prediction of trend development. Its core aspects include the collection of equipment operating data, classification and analysis of historical operating status, anomaly detection based on data change patterns, and quantitative prediction and judgment of fault probability. The methodological characteristics of this technical field lie in the real-time collection and analysis of parameter information during equipment operation, combined with existing fault modes and evolution patterns, to complete the identification and prediction of potential equipment faults. Among them, the traditional online monitoring analysis and fault prediction method for switch cabinets used for power distribution and protection in high-voltage power transmission and distribution systems relies on the timed collection of basic operating parameters such as temperature, current, and voltage, and judges whether there is an anomaly in the equipment by setting fixed thresholds for these parameters. The data collection method generally uses power sensors installed at key nodes to periodically upload data to local monitoring terminals, and compares them with preset rules to identify fault risks.

[0003] Existing technologies rely on fixed-point acquisition and fixed-threshold comparison of static parameters such as temperature, current, and voltage. This approach cannot accurately capture subtle fluctuations and early abnormal signals during operation. Short-term voltage changes during contact switching are difficult to respond to completely, making it difficult to detect early trend faults in a timely manner. This is especially true in scenarios with frequent load changes or strong environmental interference, where misjudgments and missed diagnoses are prone to occur. The periodically uploaded data lacks the ability to depict the dynamic evolution process. Comparing single-point data with preset rules is insufficient to identify the fault development process. In practical applications, the problem of abnormal data but inability to locate the fault section often occurs, reducing the reliability of the system and the effectiveness of decision-making. Summary of the Invention

[0004] To address the technical problems existing in the prior art, this invention provides a method for online monitoring, analysis, and fault prediction of switch cabinets, including the following steps: To achieve the above objectives, the present invention adopts the following technical solution: a method for online monitoring, analysis, and fault prediction of a switch cabinet, comprising the following steps: S1: Collect the voltage sampling sequence during the contact conduction period of the switch cabinet, identify the main peak feature points within the period, record the time and amplitude information, and construct the main peak feature sequence in the period order; S2: Based on the main peak feature sequence, extract the sampling segment after the main peak, analyze the amplitude changes between sequences, determine whether there are discontinuous features, filter the abnormal cycle number and offset direction, and generate transient abnormal cycle information. S3: Based on the main peak feature sequence and the transient abnormal period information, using the main peak amplitude and time information in the continuous period, construct the trajectory change trend sequence, and by comparing the direction of the change in the tilt angle of adjacent line segments, determine whether there is a continuous trajectory deviation and generate a trajectory deviation period list. S4: Perform cross-comparison between the transient abnormal period information and the trajectory offset period list, count the offset direction in the matching period, identify the concentrated distribution interval of abnormal behavior, and generate a period matching abnormal identifier group. S5: Based on the number of consecutive cycles, direction consistency and distribution in the cycle matching anomaly identifier group, determine the abnormal cycle segment, output the associated contact number, offset feature and trend label, and generate the fault prediction and diagnosis result.

[0005] As a further embodiment of the present invention, the main peak feature sequence includes the main peak time point, the main peak amplitude, and the main peak sequence sorting position; the transient abnormal cycle information includes the abnormal cycle number, the abnormal offset direction, and the discontinuous characteristic amplitude; the trajectory offset cycle list includes the offset cycle number, the offset trend angle change, and the trend continuity mark; the cycle matching abnormal identification group includes the matching cycle number, the directional statistical result, and the concentrated distribution interval range; and the fault prediction and diagnosis result includes the contact number, the offset behavior characteristics, and the trend label.

[0006] As a further aspect of the present invention, the specific steps of S1 are as follows: S101: Obtain the voltage sampling sequence during the contact conduction period of the switch cabinet, organize the data points in chronological order, construct the voltage waveform curve, identify the sections with significant fluctuation amplitude changes, select the positions where the rate of change exceeds the set threshold, and extract the points with the highest voltage values ​​within the sections to generate a voltage jump point sequence. S102: Call the voltage values ​​of the points in the voltage jump point sequence and the data of adjacent points, filter the main peak points that meet the condition that the voltage value is higher than that of the adjacent points, record the time and amplitude information of the main peak points, and organize them into a structured data table in chronological order to generate the main peak feature dataset. S103: Based on the time field and period number information in the main peak feature dataset, extract the amplitude value and time value of each main peak point in periodic order, and combine them into a continuous data sequence to obtain the main peak feature sequence.

[0007] As a further aspect of the present invention, the specific steps of S2 are as follows: S201: Based on the main peak feature sequence, extract a fixed-length voltage sampling segment after the time corresponding to the main peak point in each period, organize the data sequence according to the period number, construct a sampling segment set indexed by the period number, call the fluctuation range information of the voltage value in the time dimension, and establish the period sampling change amplitude value. S202: Based on the periodic sampling change amplitude value, call the change amplitude difference sequence between adjacent periods, and determine whether each difference exceeds the threshold range according to the set discontinuity identification threshold. Extract the period number corresponding to the one that exceeds the threshold as an abnormal candidate period and generate a list of mutation period numbers. S203: Call the voltage sampling segment corresponding to each cycle number in the mutation cycle number list, compare the voltage value change direction of the start segment and end segment with the sign of the voltage value fluctuation trend direction change, determine the voltage offset direction of the corresponding cycle, combine the cycle number and offset direction into structured data, and obtain transient abnormal cycle information.

[0008] As a further aspect of the present invention, the specific steps of S3 are as follows: S301: Call the transient abnormal periodic information and the main peak feature sequence, extract the amplitude and time information corresponding to the main peak in the continuous period, combine the corresponding amplitude and time of the main peak in the same period, construct a point set according to the period order, and connect them according to the time order to generate line segments and establish a main peak trajectory line segment group. S302: Based on the point information of adjacent line segments in the main peak trajectory line segment group, calculate the corresponding tilt angle value, construct the angle change sequence, and determine whether the trend of the line segments is consistent based on the difference in the direction of the adjacent angle changes in the sequence, identify the set of line segments with continuous direction characteristics, and generate a trajectory direction continuous segment group. S303: Based on the period index in the continuous segment group of the trajectory direction, check whether there is a continuous offset in the direction of the line segment in the corresponding time series, filter the direction segments that meet the duration requirements, record the corresponding period range and direction, and establish a trajectory offset period list.

[0009] As a further aspect of the present invention, the specific steps of S4 are as follows: S401: Obtain the transient abnormal cycle information and the cycle number in the trajectory offset cycle list, match them in the order of the numbers, filter out data items with the same cycle number as matching samples, and establish a set of matching cycle numbers; S402: Call the set of matching cycle numbers, and for the offset feature directional data corresponding to each cycle, divide it into three categories according to the offset direction: forward, reverse and undirected. Count the frequency of each category in all matching cycles and generate the offset direction distribution value. S403: Based on the offset direction distribution value, extract the concentrated occurrence interval of a single direction in a continuous period, identify the position of the period number corresponding to the direction, mark the corresponding period position number, and generate a period matching anomaly identifier group.

[0010] As a further aspect of the present invention, the specific steps of S5 are as follows: S501: Based on the period number marked in the period matching anomaly identifier group, determine the sequential continuity between adjacent period numbers, extract continuous period number segments, filter number segments that meet the continuity condition, and generate continuous period segment number values. S502: Call the continuous periodic segment number value, obtain the offset direction information within the number segment, identify the distribution of direction types, filter periodic segments with the same direction type, and generate a numbered segment value with the same direction. S503: Based on the consistent numbering segment value, extract the offset trend features and contact number information of the corresponding numbering segment, determine the degree of concentration of the numbering segment distribution, mark it in combination with the trend type, and generate the switch cabinet fault prediction and diagnosis result.

[0011] As a further aspect of the present invention, the switch cabinet contact is a metal contact element in the switch cabinet used to conduct or disconnect the main circuit current, which is a typical conductive interface structure in high-voltage electrical equipment. The voltage sampling sequence is a set of voltage data formed by periodically sampling the voltage signal generated during the conduction process of the switch cabinet contacts, and is collected by the sensor at a fixed sampling frequency; The main peak feature point is the sampling point with the largest amplitude in one period of the voltage sampling sequence; The main peak feature sequence is a data sequence formed by arranging the main peak feature points identified in multiple periods in chronological order.

[0012] As a further aspect of the present invention, the sampling segment is a voltage sampling data region that extends backward from the main peak feature point; The discontinuity feature refers to the voltage waveform abrupt changes, sharp turns, or abnormal curve discontinuities that occur within the sampling segment, characterizing the unstable disturbance phenomenon of the signal in a short period of time. The offset direction refers to the changing trend monitored when analyzing voltage or peak data of adjacent periods, which is manifested as an increasing or decreasing path of the value in the time or amplitude dimension. The trajectory change trend sequence is a line segment path composed of multiple consecutive main peak feature points, recording the change trend of the peak position on the dual axes of time and amplitude. The direction of the tilt angle change refers to the changing trend of the angle formed between adjacent line segments in the trajectory change trend sequence, that is, the angle continuously increases or decreases; The abnormal behavior concentration distribution interval refers to the interval in which data segments with abnormal fluctuations in multiple detection cycles continuously cluster within a single time range. The fault prediction and diagnosis results are state judgment results output after analyzing abnormal periodic data and trajectory change characteristics.

[0013] An online monitoring, analysis, and fault prediction system for switch cabinets includes: The periodic feature extraction module acquires the voltage signal sequence during the contact conduction period in the switch cabinet, segments it according to the conduction period, identifies the voltage maximum point in the period and records the corresponding time position and amplitude, constructs a sequence set including the main peak features of all periods, and generates the main peak feature sequence. The transient screening module extracts the voltage sampling segment after the main peak position based on the main peak feature sequence, judges the amplitude change trend of the sampling points in adjacent periods, identifies the period number with abrupt change characteristics, and classifies them in combination with the change direction to generate transient abnormal period information. The offset recognition module calls the main peak feature sequence and the transient abnormal cycle information, constructs a trajectory change sequence based on the time and amplitude of the main peak point in the continuous cycle, analyzes the directional characteristics of the trajectory slope change between adjacent cycles, and combines the abnormal cycle number to filter and identify the corresponding cycle, generating a trajectory offset cycle list. The anomaly integration module matches records with common period numbers based on the transient anomaly period information and the trajectory offset period list, counts the offset direction of the period and judges the continuity and concentration of the direction distribution, identifies period ranges with consistent characteristics, and generates a period matching anomaly identifier group. The fault diagnosis module calls the period matching anomaly identifier group, extracts the number distribution, direction features and period quantity indicators within the period segment, combines them with the corresponding contact numbers in the switch cabinet to classify the period markings, determines the period sequence with behavioral characteristics, and generates fault prediction and diagnosis results.

[0014] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, by extracting and serializing the voltage peak characteristics during contact conduction, and comparing the change amplitude of the sampling segment after the peak, the transient anomaly can be accurately identified. Furthermore, by analyzing the tilt angle change of the peak trajectory within a continuous cycle, the deviation behavior of the operating trend can be identified. By cross-comparing the abnormal cycle and the trajectory deviation, the abnormal segments with consistent direction and concentrated distribution can be located. Based on the cycle number and directional characteristics, the deviation trend can be comprehensively judged, effectively improving the timeliness and accuracy of fault prediction. Attached Figure Description

[0015] 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.

[0016] 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; Figure 7 This is a system module diagram of the present invention. Detailed Implementation

[0017] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0018] 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.

[0019] 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.

[0020] 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.

[0021] 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.

[0022] Please see Figure 1 This invention provides a method for online monitoring, analysis, and fault prediction of switch cabinets, comprising the following steps: S1: Obtain the voltage sampling sequence during the contact conduction period of the switch cabinet, identify the main peak feature points within the period, record the corresponding time and amplitude information, and generate the main peak feature sequence in the order of the period. Switchgear contacts are metal contact elements in switchgear used to conduct or disconnect the main circuit current. They are typical conductive interface structures in high-voltage electrical equipment and are the direct source of voltage acquisition in online monitoring. The voltage sampling sequence is a set of voltage data formed by periodically sampling the voltage signal generated during the conduction process of the switch cabinet contacts. It is collected by a sensor at a fixed sampling frequency and reflects the waveform trend of voltage change over time. The main peak feature point is the sampling point with the largest amplitude in one period of the voltage sampling sequence. It is used to reflect the maximum fluctuation point of the voltage curve in the conduction state and is a key feature basis for judging the waveform trend and anomalies. The main peak feature sequence is a data sequence formed by arranging the main peak feature points identified in multiple cycles in chronological order. It is used to analyze the voltage trajectory trend and the stability changes between cycles. S2: Based on the main peak feature sequence, extract the sampling segment after the main peak of each cycle for analysis, determine whether there are discontinuous features by the change amplitude between sequences, filter out the cycle number and corresponding offset direction of the abnormality, and generate transient abnormal cycle information. The sampling segment is a voltage sampling data area that extends backward from the main peak feature point. It is used to reflect the behavior pattern of voltage fluctuations in a short period of time after the main peak, and is especially used to identify rebound transient change characteristics. Discontinuity features refer to abrupt changes, sharp turns, or abnormal curve discontinuities in the voltage waveform that occur within the sampling segment, characterizing unstable disturbances in the signal within a short period of time. Offset direction refers to the trend of change monitored when analyzing voltage or peak data of adjacent periods. It is manifested as an increasing or decreasing path of value in the time or amplitude dimension and is used to identify the direction of evolution of abnormal behavior. S3: Call the transient abnormal cycle information and the main peak feature sequence. Based on the amplitude and time information corresponding to the main peak in the continuous cycle, establish the trajectory change trend sequence. By comparing the direction of the tilt angle change between adjacent line segments, determine whether there is a continuous trajectory offset and generate a trajectory offset cycle list. The trajectory change trend sequence is a line segment path composed of multiple consecutive main peak feature points. It records the change trend of the peak position on the dual axes of time and amplitude, and is used to determine whether the overall waveform deviates from the original stable trajectory. The direction of tilt angle change refers to the changing trend of the angle formed between adjacent line segments in the trajectory change trend sequence, that is, whether the angle continuously increases or decreases, which is used to identify whether the trajectory is in a continuous drift state. S4: Cross-compare the transient anomaly period information with the trajectory offset period list by period number, statistically analyze the directionality of offset features in the matching period, identify the concentrated distribution range of abnormal behavior, and generate a period matching anomaly identifier group. An abnormal behavior concentration distribution interval refers to an interval in which data segments with abnormal fluctuations in multiple detection cycles continuously cluster within a single time range, used to determine the location of potential risk concentration areas; S5: Based on the number of consecutive cycles, directional consistency, and concentrated distribution of the cycles marked in the cycle matching anomaly identifier group, determine the cycle segment and output the associated contact number, offset behavior characteristics, and trend label to generate the fault prediction and diagnosis results. Trend labels are classification markers based on the directional, persistent, and amplitude characteristics of cyclical shifts. They are used for pattern recognition and output classification of cyclical behavior, including "continuous drift" and "instantaneous jump". The fault prediction and diagnosis results of the switch cabinet are status judgment results output after analyzing abnormal periodic data and trajectory change characteristics. The content includes contact number, fault type characteristics and trend marking information, which serve as the basis for system operation and maintenance decisions.

[0023] The main peak feature sequence includes the main peak time point, main peak amplitude, and main peak sequence sorting position. The transient abnormal cycle information includes the abnormal cycle number, abnormal offset direction, and discontinuous characteristic amplitude. The trajectory offset cycle list includes the offset cycle number, offset trend angle change, and trend continuity mark. The cycle matching abnormal identification group includes the matching cycle number, directional statistical results, and concentrated distribution range. The switch cabinet fault prediction and diagnosis results include the contact number, offset behavior characteristics, and trend label.

[0024] Please see Figure 2 The specific steps of S1 are as follows: S101: Obtain the voltage sampling sequence during the contact conduction period of the switch cabinet, organize the data points in chronological order, construct the voltage waveform curve, identify the section where the fluctuation amplitude changes beyond the set threshold, select the position where the rate of change exceeds the set threshold, and extract the point with the highest voltage value within the section to generate a voltage jump point sequence. During the contact conduction period of the switchgear, the closing state is set as the sampling trigger signal through a high-speed data acquisition device. The sampling frequency is set to 100kHz, equivalent to sampling once every 10 microseconds, and the sampling duration is set to 20 milliseconds. The voltage values ​​collected during this period are arranged in chronological order into a sequence consisting of multiple data points. For example, the first point's voltage is 120, followed by 125, 130, 185, 160, etc. The voltage difference between each adjacent point is compared. If the voltage at a certain point rises by more than 20 compared to the previous moment, it is considered to have entered the fluctuation range. For example, from 130 to 185, the voltage difference is 55, which exceeds the limit. If the voltage exceeds a set threshold, the segment is determined to be a voltage change segment. Based on this, the voltage change rate is calculated. If the voltage rises by 50 in 10 microseconds, the change rate is 5 per microsecond. If it exceeds the set rate threshold of 2 per microsecond, the point is confirmed as a rate change point. In the marked fluctuation segment, the voltage values ​​of each point are compared, and the maximum value is extracted. For example, if the voltage values ​​of a certain segment are 145, 172, 165, and 178, then 178 is taken as the maximum voltage point in that segment. Finally, the multiple maximum voltage points that meet the conditions are arranged in the order of sampling time to form a point sequence representing the characteristics of rapid voltage change. Each item in this sequence contains time and voltage values.

[0025] S102: Call the voltage values ​​of the points in the voltage jump point sequence and the data of adjacent points, filter the main peak points that meet the condition that the voltage value is higher than the adjacent points, record the time and amplitude information of the main peak points, and organize them into a structured data table in chronological order to generate the main peak feature dataset. Based on the voltage jump point sequence, the voltage value of each point is compared with the two adjacent points. If the voltage of the middle point is higher than the two points on either side, and the difference is greater than 10, it is marked as a main peak point. For example, if the voltages of three points are 160, 190, and 175, and the middle point is 30 higher than the left and 15 higher than the right, both exceeding the set standard, it is confirmed as a main peak point. The time of this point is recorded as a timestamp, such as 0.008 seconds, with a voltage of 190. At the same time, period number information, such as the 5th period, is added. All main peak points are organized and arranged in chronological order to form a structured data set. Each row contains three fields: time, voltage value, and cycle number. During the data collection process, if multiple main peaks appear consecutively, it is still necessary to compare adjacent points and judge the difference to ensure that only points with voltages significantly higher than adjacent points are selected. The comparison process can be performed through logical judgment. For example, if the consecutive voltage values ​​are 172, 198, and 185, then 198 is higher than the values ​​on both sides and is considered a main peak. If the values ​​are 182, 183, and 180, the difference is too small and does not meet the condition, so it is judged as a non-main peak. All the main peak data obtained through filtering are organized into a dataset for subsequent analysis.

[0026] S103: Based on the time field and cycle number information in the main peak feature dataset, extract the amplitude value and time value of each main peak point in cycle order, and combine them into a continuous data sequence to obtain the main peak feature sequence; Based on the main peak feature dataset, the time and voltage amplitude of the main peak point in each period are extracted sequentially according to the period number, forming a main peak sequence that evolves with the period. If the period numbers are 1, 2, and 3, and the main peak points appear at 0.005 seconds, 0.025 seconds, and 0.045 seconds respectively, corresponding to voltages of 192, 185, and 198, then these three sets of data constitute a continuous sequence. If there are multiple main peak points within a certain period, the point with the largest voltage value is selected. For example, if there are three main peak points in period 2, the voltage value is selected. If the pressure values ​​are 176, 185, and 180, then 185 is selected as the representative peak. During the extraction process, the start time and duration of each period are used to determine whether each peak belongs to that period. For example, if the period time is 20 milliseconds, then the first period is 0 to 0.02 seconds, and the second period is 0.02 to 0.04 seconds. If the peak time is 0.021 seconds, then it is assigned to the second period. The processed data sequence will be used to show the time position and amplitude changes of the peak in different periods, forming a complete set of peak feature sequences.

[0027] Please see Figure 3 The specific steps of S2 are as follows: S201: Based on the main peak feature sequence, extract the fixed-length voltage sampling segment after the time corresponding to the main peak point in each period, organize the data sequence according to the period number, construct a sampling segment set indexed by the period number, call the voltage value fluctuation range information in the time dimension, and establish the period sampling change amplitude value. Based on the preceding main peak characteristic sequence, voltage data within a fixed-length time interval following the occurrence of the main peak in each cycle needs to be sampled. The sampling length can be set to 5ms. At a sampling frequency of 100kHz, this is equivalent to extracting 500 consecutive voltage data points per cycle. Each data point includes the sampling time and the corresponding voltage value. During the extraction process, starting from the main peak time, an equal-length time window is extended forward, and the corresponding segment of voltage data is extracted from the original voltage sequence as the sampling segment for that cycle. For example, if the main peak time is 0.015s, the sampling segment range is from 0.015s to 0.020s. Within this segment, data points such as 0.01501s have a voltage of 180, 0.01502s has a voltage of 185, etc. Each sampling segment is indexed by its corresponding cycle number, and the sampling segments are arranged in numerical order to form a set. After all sampling segments are extracted, the voltage data within each segment is analyzed to identify the maximum and minimum voltage values. The difference between these values ​​is then calculated as the voltage fluctuation amplitude for that period. For example, if the maximum value is 210 and the minimum value is 175, the fluctuation amplitude is 35. This process is repeated for all periods to obtain a complete set of voltage fluctuation amplitude values. Each value is associated with its corresponding period number, forming a set of periodic sampling change amplitude values, which serves as the basis for subsequent abnormal period identification.

[0028] S202: Based on the periodic sampling change amplitude value, call the change amplitude difference sequence between adjacent periods, and determine whether each difference exceeds the threshold range according to the set discontinuity identification threshold. Extract the period number corresponding to the one that exceeds the threshold as an abnormal candidate period and generate a list of mutation period numbers. Based on the set of periodic sampling amplitude values, the fluctuation amplitude values ​​between two adjacent periods need to be compared, their differences calculated, and a difference sequence formed. For example, if the fluctuation in period 4 is 38 and in period 3 is 26, the difference is 12. This process is performed on all periods to obtain a complete difference sequence. When determining whether the difference is abnormal, a threshold needs to be set. This threshold can be set based on the historical data fluctuation range of the equipment, for example, set to 15. This means that when the fluctuation between two adjacent periods exceeds 15, the period can be considered potentially abnormal. When the difference between periods exceeds this set value, for example, if period 5 is 50 and period 4 is 30, the difference is 20, which is greater than the threshold of 15, then period 5 is listed as an abnormal candidate period. Here, the cases where the difference is positive or negative need to be handled uniformly, using their absolute values ​​for judgment to avoid missing abrupt changes in periods with a downward trend. The entire judgment process is executed cycle by cycle. All cycle numbers that exceed the set threshold are extracted and compiled into a list. The list only contains cycle numbers with obvious fluctuations and abrupt changes. For example, the 3rd, 6th, and 8th cycles have anomalies with differences of 18, 22, and 19, respectively. The final list is a set of numbers 3, 6, and 8, indicating that these cycles have significant abrupt changes.

[0029] S203: Call the voltage sampling segment corresponding to each cycle number in the mutation cycle number list, compare the voltage value change direction of the start segment and end segment with the sign of the voltage value fluctuation trend direction change, determine the voltage offset direction of the corresponding cycle, combine the cycle number and offset direction into structured data, and obtain transient abnormal cycle information. For each cycle in the list of mutation cycle numbers, the corresponding voltage sampling segment is called, and the voltage values ​​at the start and end positions within the segment are compared to determine the direction of voltage change within that cycle. The comparison method uses the average voltage of several points at the very beginning of the sampling segment as the starting value and the average voltage of several points at the very end as the ending value to reduce the influence of single-point mutation interference. For example, if the voltages at the first three points are 172, 175, and 178, with an average of 175, and the voltages at the last three points are 193, 195, and 191, with an average of 193, then the ending value is higher than the starting value, and the voltage offset direction for that cycle is considered positive. If the starting value is 190 and the ending value is 175, then it is a negative offset. During the judgment process, if the difference between the start and end values ​​is within 5, the trend is considered insignificant, and the direction is not recorded. This process is performed on all mutation cycles, forming structured data for each cycle number and its offset direction, such as cycle number 5 having a positive offset direction, cycle number 6 having a negative offset direction, and cycle number 8 having no significant offset, etc. After all the results are compiled, a complete set of transient abnormal cycle information is formed, which can be used for further fault identification or trend analysis.

[0030] Please see Figure 4 The specific steps of S3 are as follows: S301: Call the transient abnormal periodic information and the main peak feature sequence, extract the amplitude and time information corresponding to the main peak in the continuous period, combine the corresponding amplitude and time of the main peak in the same period, construct the point set according to the period order, and connect them according to the time order to generate line segments and establish the main peak trajectory line segment group. The transient anomaly cycle information and the main peak feature sequence are retrieved. For each anomaly cycle number, the amplitude and occurrence time corresponding to its main peak point are extracted. If multiple main peak points exist in a cycle, the main peak with the largest amplitude is selected first, and the time information of that point is also extracted. The main peak amplitude and its corresponding time are paired one-to-one to form data points. For example, for cycle number 7, its main peak voltage is 215 and its time is 0.135s, so the data points are 215 and 0.135s. After performing the same operation on all cycles, these points are arranged in the order of cycle number to ensure... By gradually increasing the time interval, a set of data points is formed. In the sorted set of data points, adjacent data points are connected by straight line segments. For example, points 215 and 0.135s are connected to 218 and 0.145s to form the first line segment, and then connected to 213 and 0.155s to form the second line segment. This process is repeated to form a continuous broken line structure. In a two-dimensional coordinate system with the time axis as the horizontal axis and the voltage amplitude as the vertical axis, a group of main peak trajectory segments is established. Finally, a complete main peak trajectory structure is obtained for subsequent analysis of trajectory changes.

[0031] S302: Based on the point information of adjacent line segments in the main peak trajectory line segment group, calculate the corresponding tilt angle value, construct the angle change sequence, and determine whether the trend of the line segments is consistent based on the difference in the direction of the adjacent angle changes in the sequence. Identify the set of line segments with continuous directional features and generate a trajectory direction continuous segment group. Based on the established main peak trajectory segment group, the values ​​of the two endpoints of any pair of adjacent segments are extracted segment by segment to obtain the voltage amplitude change and time span between adjacent segments. Combining the changing trends of these two values, the segment direction is estimated. The segment direction is set with the rate of voltage change over time as a reference: an increase in voltage amplitude represents an upward direction, and a decrease represents a downward direction. For example, an increase from 215 to 218 is upward, and a decrease from 218 to 213 is downward. The extraction and direction estimation operations are performed sequentially on all segments to form a set of angle sequences representing the trajectory change trend. To avoid errors in direction change judgment... Judgment conditions can be set between angles. For example, if the change direction of two consecutive line segments is consistent, the trajectory direction is considered unchanged. If there is a sudden change in direction, it is a direction switching point, such as from continuous rise to fall, which is marked as a change in direction. In addition, a tolerance value for direction change can be set to evaluate the degree of continuity. For example, setting it to 30 means that the change of two segments does not exceed this range and the direction is consistent. When the voltage value change trend of adjacent line segments is consistent and the time difference does not exceed the set upper limit, it is considered to be continuous in direction and merged into the same segment. All line segments are processed repeatedly to finally form a group of continuous trajectory direction segments with consistent direction for further behavior judgment.

[0032] S303: Based on the period index in the continuous segment group of the trajectory direction, check whether there is a continuous offset in the direction of the line segment in the corresponding time series, filter the direction segments that meet the duration requirements, record the corresponding period range and direction, and establish a trajectory offset period list. Based on the cycle number information contained in the continuous segment group of the trajectory direction, the time segment corresponding to each cycle is returned and extracted. The direction of the trajectory line segment within each segment is checked to see if it remains consistent. The direction is determined by the fact that the voltage change trend has not reversed in multiple consecutive cycles. If the trend remains upward or downward within the entire segment, it is judged as a stable direction segment. Then, the time span condition is combined for filtering. The continuous offset time threshold is set to 30ms. If the start and end time interval of a continuous segment in a certain direction reaches or exceeds this value, the segment is identified as an offset segment that meets the continuous condition. For example, if the segment extends from 0.135s to 0.175s, a total of 40ms, the time meets the requirement. The offset direction and cycle number range of the segment are recorded. If the direction is positive, it is recorded as a positive offset; if the direction is negative, it is recorded as a negative offset. All segment numbers and directions that meet the continuous direction condition are organized, and finally, a trajectory offset cycle list is constructed. The list includes the start and end numbers of the cycle corresponding to each segment that meets the condition and the corresponding offset direction. For example, the direction of cycle numbers 7 to 11 is positive, and the direction of numbers 12 to 15 is negative, etc.

[0033] Please see Figure 5 The specific steps of S4 are as follows: S401: Obtain the transient anomaly cycle information and the cycle number in the trajectory offset cycle list, match them in the order of the numbers, filter the data items with the same cycle number as the matching samples, and establish a set of matching cycle numbers; After obtaining the transient abnormal cycle information and the cycle numbers in the trajectory offset cycle list, both are extracted and sorted in ascending order. Then, they are compared using a sequential matching method. First, the abnormal cycle set numbers are extracted, such as numbers 7, 8, 9, 10, 12, and 13. Next, the set of numbers in the trajectory offset cycle list is extracted, such as offset segments 7 to 9 and 12 to 13, which expand to 7, 8, 9, 12, and 13. Then, each abnormal cycle in the set is checked to see if it completely matches the offset cycle number, including the value and the order of arrangement. If an item in the abnormal cycle, such as 10, is not in the offset cycle list, the entire group is excluded. A matching standard must be set during the screening process, meaning all numbers must be completely identical and without omissions. A comparative search logic is used to verify whether there are any missing or redundant items. For completely matching cycle numbers, such as numbers 7, 8, 9, 12, and 13, they are added to the matching sample set after confirming that they meet the conditions. The cycle numbers in this set constitute the basic data for subsequent analysis. Cycle number pairs that do not form a complete match are removed, thus constructing a cycle set that meets the precise matching standard for numbers.

[0034] S402: Call the set of matching period numbers, and for the offset feature directional data corresponding to each period, divide it into three categories according to the offset direction: forward, reverse and undirected. Count the frequency of each category in all matching periods and generate the offset direction distribution value. Based on the established set of matching cycle numbers, the offset directional features corresponding to each cycle are extracted one by one. The direction categories are divided into three types: positive, negative, and undirected, represented by the symbols +, −, and 0, respectively. For example, if direction 7 is positive, direction 8 is positive, direction 9 is negative, direction 12 is undirected, and direction 13 is positive, then the direction sequence is +, +, −, 0, +. Next, the frequency is counted according to the direction type. Positive appears 3 times, negative once, and undirected once, for a total of 5 times. This translates to a percentage of 60%, 20%, and 20%, respectively. The entire process uses a group counting method for classification and statistics, which can be implemented manually or through simple programming. The number of occurrences of each type of direction is recorded separately and then divided by the total number of matching samples to obtain the proportion. For example, the proportion of positive is 3 divided by 5, which equals 60%, and the proportions of negative and undirected are each 1 divided by 5, which equals 20%. Finally, a complete set of offset direction distribution values ​​is formed, which will serve as an important basis for judging the degree of concentration of directional trends.

[0035] S403: Based on the offset direction distribution value, extract the concentrated occurrence interval of a single direction in a continuous period, identify the position of the period number corresponding to the direction, mark the corresponding period position number, and generate a period matching anomaly identification group. Based on the directions that appear most frequently in the offset direction distribution values, we retrieve whether there is a continuous concentrated interval for that direction in the directional sequence corresponding to the matching period number set. For example, if a positive direction appears twice consecutively in period numbers 7 and 8, and although there are reverse and undirected occurrences in between, it appears again in number 13. Since this occurrence does not form a continuous interval, only the positive continuous interval formed by numbers 7 and 8 is retained. After extracting these period numbers, their positions in the original matching set are marked. For example, number 7 is the first position in the set, and number 8 is the second position. These numbers are then assigned positions. The numbers are identified and organized into a periodic matching anomaly identification group. The data structure of this group includes the period number range and the corresponding direction. For example, the identification group content is "number 7 to 8, positive direction". When judging the concentration interval, a continuity judgment standard needs to be set. For example, at least two consecutive periods have the same direction. If it only appears once, it does not constitute continuity. The directional sequence is scanned from beginning to end using a position sliding method. Each time, the current position and several subsequent positions are checked to see if the direction is consistent. If they are, the number interval and its direction type are recorded. After completion, all period number ranges that meet the central tendency are output and summarized into anomaly identification groups.

[0036] Please see Figure 6 The specific steps of S5 are as follows: S501: Based on the period number marked in the period matching anomaly identifier group, determine the sequential continuity between adjacent period numbers, extract continuous period number segments, filter number segments that meet the continuity condition, and generate continuous period segment number values. Based on the cycle numbers marked in the cycle matching anomaly identifier group, all numbers are first collected and organized into a sequence, and then sorted in ascending order of numerical value. Next, the numerical differences between adjacent numbers are compared sequentially. If the difference between two adjacent numbers is 1, it is considered a continuous cycle; otherwise, it is considered a discontinuous node. Multiple number segments are constructed through this sequential comparison method. Each number segment consists of continuous cycles. For example, if the number sequence is 7, 8, 9, 11, 12, 14, then 7 to 9 are continuous, 11 and 12 are also continuous, while numbers 10 and 13 are absent, resulting in segmentation. The final number segments are 7 to 9 and 11 to 12, which are recorded as continuous cycle segment numbers. To enhance the accuracy of the judgment, filtering conditions need to be set, such as retaining only number segments with a length of at least 2. If a number segment contains only one number, such as 14, it is discarded because it does not meet the minimum segment length condition. The segment length threshold can be set to 2, meaning that each number segment should contain at least 2 consecutive numbers. All number segments that meet this condition are organized into the final result, forming the data foundation for subsequent directional analysis.

[0037] S502: Call the continuous periodic segment number value, obtain the offset direction information within the number segment, identify the distribution of direction types, filter periodic segments with the same direction type, and generate numbered segment values ​​with the same direction. The system retrieves the continuous periodic segment number values, reads each period number contained within it segment by segment, and searches for the corresponding offset direction information. Direction values ​​are generally recorded in three categories: + for positive, - for negative, and 0 for no direction. By statistically analyzing all direction types contained in each number segment, it determines whether the segment has consistent directions. For example, number segments 7 to 9 contain +, +, and - directions, representing two types of directions, indicating inconsistent directions. Number segments 11 to 12, however, contain only - directions, making them consistent directions. The consistency of directions is determined by checking if all direction values ​​within a number segment are identical. If the number of direction types is only one, it is considered consistent. Direction information can be obtained by indexing each row in an established data table. A threshold condition can be set for consistency, such as requiring over 80% of the period directions in a number segment to be the same. When dealing with number segments like 7 to 9 containing multiple directions, those that do not meet the consistency condition are discarded. Finally, all number segment values ​​with consistent directions are collected and archived to form consistent direction number segment values, laying the foundation for subsequent offset trend and equipment correlation analysis.

[0038] S503: Based on the consistent numbering segment value, extract the offset trend characteristics and contact number information of the corresponding numbering segment, determine the degree of concentration of the numbering segment distribution, mark it in combination with the trend type, and generate fault prediction and diagnosis results. Based on the consistent numbering segment values, the offset trend characteristics and contact number data corresponding to each numbering segment are extracted. The offset trend is usually recorded as three types: rising, falling, or stable. The contact number is used to identify the switchgear component to which each cycle belongs. By analyzing the trend data of each cycle in the numbering segment, it is determined whether the trend type is consistent. For example, if all trends in a certain numbering segment are rising, then the trend of that segment is identified as rising. If there is a mixed trend, that segment is not included in further judgment. At the same time, the contact number information of the corresponding cycle is extracted. For example, if the contact number of cycles 11 and 12 is CT01, it means that the numbering segment corresponds to a single contact, which is used for judgment. The degree of concentration can be determined by setting conditions. For example, if more than 80% of the contact numbers in a numbering segment are the same, the distribution can be considered concentrated. If there are 5 cycles in a numbering segment and 4 cycles have the same contact number, accounting for 80%, the concentration condition is met. Otherwise, the distribution is considered scattered. After completing the trend consistency and concentration judgment, the numbering segments that meet the conditions are marked. For example, if the trend of numbering segments 11 to 12 is upward and the contact numbers are the same, it is marked as an upward trend and concentrated contact. Finally, a set of switchgear fault prediction and diagnosis results is constructed. The content structure is usually a combination of three elements: numbering segment range, trend type, and contact concentration status.

[0039] Please see Figure 7 An online monitoring, analysis, and fault prediction system for switch cabinets includes: The periodic feature extraction module acquires the voltage signal sequence during the contact conduction period in the switch cabinet, segments it according to the conduction period, identifies the voltage maximum point in the period and records the corresponding time position and amplitude, constructs a sequence set including the main peak features of all periods, and generates the main peak feature sequence. The transient screening module extracts the voltage sampling segment after the main peak position based on the main peak feature sequence, judges the amplitude change trend of the sampling points in adjacent periods, identifies the period number with abrupt change characteristics, and classifies them in combination with the change direction to generate transient abnormal period information. The offset recognition module calls the main peak feature sequence and transient abnormal cycle information, constructs the trajectory change sequence based on the time and amplitude of the main peak point in the continuous cycle, analyzes the directional characteristics of the trajectory slope change between adjacent cycles, and combines the abnormal cycle number to filter and identify the corresponding cycle, generating a list of trajectory offset cycles. The anomaly integration module matches records with common period numbers based on transient anomaly period information and trajectory offset period list, counts the offset direction of the period and judges the continuity and concentration of the direction distribution, identifies period ranges with consistent characteristics, and generates period matching anomaly identifier groups. The fault diagnosis module calls the period matching anomaly identifier group to extract the number distribution, direction features and period quantity indicators within the period segment. It then combines these with the corresponding contact numbers in the switch cabinet to classify the period markings, determine the period sequences with behavioral characteristics, and generate fault prediction and diagnosis results.

[0040] 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 the claims.

Claims

1. A method for online monitoring analysis and fault prediction of a switchgear cabinet of a switch, characterized in that, The method comprises the following steps: S1: collecting a voltage sampling sequence during the conduction of a switch cabinet contact, identifying a main peak characteristic point in a period, recording time and amplitude information, and constructing a main peak characteristic sequence in order of period; S2: based on the main peak characteristic sequence, extracting a sampling segment after the main peak, analyzing the amplitude change between sequences, judging whether there is a discontinuous feature, screening abnormal period numbers and offset directions, and generating transient abnormal period information; S3: according to the main peak characteristic sequence and the transient abnormal period information, using the main peak amplitude and time information in the continuous period, constructing a trajectory change trend sequence, comparing the change direction of the inclination angle of adjacent line segments, judging whether the trajectory has a continuous offset, and generating a trajectory offset period list; S4: cross-comparing the transient abnormal period information and the trajectory offset period list, counting the offset direction in the matching period, identifying the distribution interval of the abnormal behavior, and generating a period matching abnormal identification group; S5: determining the abnormal period section according to the number of continuous periods, the consistency of the direction and the distribution in the period matching abnormal identification group, outputting the associated contact number, the offset feature and the trend label, and generating a fault prediction diagnosis result.

2. The method for online monitoring analysis and fault predication of the switchgear cabinet of the switch according to claim 1, characterized in that, The main peak characteristic sequence includes the main peak time point, the main peak amplitude, and the main peak sequence sorting position. The transient abnormal period information includes the abnormal period number, the abnormal offset direction, and the discontinuous feature amplitude. The trajectory offset period list includes the offset period number, the offset trend angle change, and the trend continuity mark. The period matching abnormal identification group includes the matching period number, the directionality statistical result, and the centralized distribution interval range. The fault prediction diagnosis result includes the contact number, the offset behavior feature, and the trend label.

3. The method of online monitoring analysis and fault pre-judging of the switchgear cabinet of the switch according to claim 1, characterized in that, The specific steps of S1 are as follows: S101: obtaining a voltage sampling sequence during the conduction of a switch cabinet contact, arranging data points in time order, constructing a voltage waveform curve, identifying a section with obvious fluctuation amplitude change, selecting a position with a change rate exceeding a set threshold, and extracting the point with the highest voltage value in the section to generate a voltage jump point sequence; S102: calling the voltage value of the point in the voltage jump point sequence and the data of adjacent points, selecting the main peak point meeting the condition that the voltage value is higher than that of adjacent points, recording the time and amplitude information of the main peak point, and arranging it into a structured data table in time order to generate a main peak feature data set; S103: according to the time field and period number information in the main peak feature data set, extracting the amplitude value and time value of each main peak point in order of period, and combining them into a continuous data sequence to obtain the main peak characteristic sequence.

4. The method of online monitoring analysis and fault predication of the switchgear cabinet of the switch according to claim 3, characterized in that, The specific steps of S2 are as follows: S201: based on the main peak characteristic sequence, extracting a fixed length voltage sampling segment after the main peak point corresponding to each period, arranging the data sequence in order of period number, constructing a sampling segment set indexed by period number, calling the fluctuation range information of the voltage value in the time dimension, and establishing the period sampling change amplitude value; S202: According to the periodic sampling amplitude value, the change amplitude difference value sequence between adjacent periods is called, and whether each difference value exceeds the threshold range is judged according to the set discontinuous identification threshold, and the period number corresponding to the exceeding one is extracted as the abnormal candidate period, and a mutation period number list is generated; S203: The voltage sampling segment corresponding to each period number in the mutation period number list is called, and the voltage value change direction of the starting segment and the ending segment is compared, the voltage value fluctuation trend direction change sign is compared, the voltage offset direction of the corresponding period is judged, the period number and the offset direction are combined as structured data, and the transient abnormal period information is obtained.

5. The method of online monitoring analysis and fault pre-judging of the switchgear cabinet of the switch according to claim 4, characterized in that, The specific steps of S3 are: S301: The transient abnormal period information and the main peak feature sequence are called, the amplitude and time information corresponding to the main peak in the continuous period are extracted, the main peak amplitude and time in the same period are combined, the point set is constructed in the order of period, and the line segment is generated by connecting in time sequence, and the main peak trajectory line segment group is established; S302: According to the point information of adjacent line segments in the main peak trajectory line segment group, the corresponding inclination angle value is calculated, the angle change sequence is constructed, whether the line segment trend is consistent is judged according to the direction difference of adjacent angles in the sequence, the line segment set with the direction continuous characteristic is identified, and the trajectory direction continuous section group is generated; S303: According to the period index in the trajectory direction continuous section group, it is checked whether there is a continuous offset of line segment direction in the corresponding time sequence, the direction section meeting the continuous time requirement is screened, and the corresponding period range and direction are recorded to establish a trajectory offset period list.

6. The method of online monitoring analysis and fault predication of the switchgear cabinet of the switch according to claim 5, characterized in that, The specific steps of S4 are: S401: The period number in the transient abnormal period information and the trajectory offset period list is obtained, matched in the order of number, and the data items with the same period number are selected as matching samples to establish a matching period number set; S402: The matching period number set is called, and the offset characteristic directionality data corresponding to each period is divided into three categories of positive, negative and no direction according to the offset direction category, the frequency of each category appearing in all matching periods is counted, and the offset direction distribution value is generated; S403: According to the offset direction distribution value, the centralized occurrence interval of single direction in continuous period is extracted, the position of the period number corresponding to the direction is identified, the corresponding period position number is marked, and a period matching abnormal identification group is generated.

7. The method of online monitoring analysis and fault pre-determination of the switchgear cabinet of the switchboard according to claim 6, characterized in that, The specific steps of S5 are: S501: According to the period number marked in the period matching abnormal identification group, the order continuity between adjacent period numbers is judged, the continuous period number segment is extracted, the number segment meeting the continuity condition is screened, and the continuous period section number value is generated; S502: The offset direction information in the number segment is obtained by calling the continuous period section number value, the distribution of the direction type is identified, the period segment with consistent direction type is screened, and the direction consistent number section value is generated; S503: According to the direction consistent number section value, the offset trend characteristic and contact number information of the corresponding number segment are extracted, the concentration degree of the number segment distribution is judged, and the trend type is marked to generate the fault prediction diagnosis result.

8. The method for online monitoring analysis and fault predication of the switchgear cabinet of the switch according to claim 1, characterized in that, The switch cabinet contact is a metal contact element for conducting or disconnecting main loop current in a switch cabinet, and belongs to a typical conductive interface structure in high-voltage electrical equipment. The voltage sampling sequence is a voltage data set formed by timing sampling of a voltage signal generated in the conduction process of the switch cabinet contact, and is collected by a sensor at a fixed sampling frequency. The main peak feature point is a sampling point with the maximum amplitude in one cycle in the voltage sampling sequence. The main peak feature sequence is a data sequence formed by arranging the main peak feature points identified in multiple cycles in time sequence.

9. The method for online monitoring analysis and fault predication of the switchgear cabinet of the switch according to claim 1, characterized in that, The sampling segment is a section of voltage sampling data region extending backward from the main peak feature point; The discontinuous feature refers to a voltage waveform mutation, sharp turn or abnormal change mode of a curve fault in the sampling segment, indicating a non-stable disturbance phenomenon of the signal in a short time; The offset direction refers to the change trend monitored when analyzing the voltage or wave peak data of adjacent cycles, which is manifested as an increasing or decreasing path in the time or amplitude dimension; The trajectory change trend sequence is a line segment path composed of multiple main peak feature points in succession, recording the change trend of the wave peak position in the time and amplitude dual axes; The inclination angle change direction refers to the change trend of the included angle between adjacent line segments in the trajectory change trend sequence, that is, the angle continuously increases or decreases; The abnormal behavior concentrated distribution interval refers to an interval in which data segments with abnormal fluctuations in multiple detection cycles are continuously concentrated in a single time range; The fault pre-diagnosis result is a state judgment result output based on the analysis of the abnormal cycle data and the trajectory change feature.

10. A switchgear online monitoring analysis and fault prediction system, characterized in that, The system is used to implement the switch cabinet online monitoring analysis and fault pre-diagnosis method of any one of claims 1-9, and the system comprises: The cycle feature extraction module acquires a voltage signal sequence during the conduction of the switch cabinet contact, processes in segments according to the conduction cycle, identifies the voltage maximum point in the cycle and records the corresponding time position and amplitude, constructs a sequence set including the main peak features of all cycles, and generates the main peak feature sequence; The transient screening module extracts the voltage sampling segment after the main peak position based on the main peak feature sequence, judges the amplitude change trend of the sampling points in adjacent cycles, identifies the cycle number with the mutation feature, and classifies the change direction to generate the transient abnormal cycle information; The offset identification module calls the main peak feature sequence and the transient abnormal cycle information, constructs a trajectory change sequence according to the time and amplitude of the main peak points of consecutive cycles, analyzes the direction feature of the trajectory slope change between adjacent cycles, and screens and discriminates the corresponding cycle in combination with the abnormal cycle number to generate a list of trajectory offset cycles; The abnormal integration module matches the records with the same cycle number according to the transient abnormal cycle information and the trajectory offset cycle list, counts the offset direction of the cycle and judges the continuity and concentration of the direction distribution, identifies the cycle range with consistent features, and generates a cycle matching abnormal identification group. The fault judgment module calls the cycle matching exception identification group, extracts the number distribution, direction characteristics and cycle number indicators in the cycle section, combines the cycle marking classification of the corresponding contact numbers in the switch cabinet of the switch, judges the cycle sequence with behavior characteristics, and generates a fault pre-judgment diagnosis result.

Images