Power cable fault monitoring method and system based on big data

Abstract

The invention relates to the technical field of cable fault monitoring, in particular to a power cable fault monitoring method and system based on big data, and the method comprises the following steps: carrying out the optimization and integration of spatial layout and sampling time based on a cable monitoring node, recognizing a data segment with the reverse trend of a current peak and a voltage phase angle, and carrying out the recognition of a current peak; and further analyzing the synchronism of the current-carrying density and the temperature, extracting a periodic intersection distribution characteristic, screening a continuously progressive trend track, and forming a trend directivity factor. According to the method, the linkage accuracy among multi-point signals is guaranteed by combining synchronous acquisition of node data with time sequence calibration, the trend discrimination method performs step-by-step screening for collaborative fluctuations of different characteristic parameters, and a continuous evolution sequence of various types of tiny disturbances under a complex working condition is dynamically extracted; a temperature and current-carrying density correlation mode is adopted to highlight the progressiveness and the positioning capability of a fault precursor, and dynamic risk signals which are difficult to identify at the initial stage of change are effectively captured.

Description

Technical Field

[0001] This invention relates to the field of cable fault monitoring technology, and in particular to a method and system for power cable fault monitoring based on big data. Background Technology

[0002] Cable fault monitoring involves real-time perception of the operating status of power cables, fault diagnosis, and location. It primarily includes key methods for monitoring conductor connection status, insulation aging, grounding faults, and partial discharge behavior. Its development has evolved from low-frequency testing to high-frequency response measurement, and from manual inspection to intelligent identification. By real-time acquisition of electrical parameters such as current, voltage, and impedance, combined with analysis of historical operating data, it enables the identification and monitoring of different types of faults, such as open circuits, short circuits, and joint overheating. Traditional power cable fault monitoring methods refer to fault identification based on cable insulation characteristic testing or impedance characteristic testing. These methods target early, latent fault characteristics such as partial discharge, joint deterioration, and conductor overheating in cross-linked polyethylene insulated power cables and cable accessories during operation.

[0003] Existing technologies are based on segmented data acquisition methods. Asynchronous timing often leads to fragmented information from multiple nodes. Traditional insulation and impedance monitoring, due to its single parameter, is not sensitive enough to weak characteristic signals in continuous development. Under complex operating conditions, small disturbances are easily masked by regular data fluctuations. Data processing lacks cross-parameter trend correlation, making it difficult to make timely judgments on local gradual trends or multi-variable interactions. Latent anomalies are easily overlooked until their impact expands later, causing monitoring and early warning to lose the opportunity for proactive response, resulting in untimely fault handling and increased maintenance cycles and risk control pressure. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies by proposing a power cable fault monitoring method and system based on big data.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a power cable fault monitoring method based on big data, comprising the following steps: S1: Based on cable monitoring nodes, analyze the spatial order of each cable monitoring node and the middle of the structure, adjust the sampling start time, unify the acquisition process of each node, correct the difference in signal response time, and obtain the timing-aligned voltage characteristics. S2: Based on the time-aligned voltage characteristics, determine the occurrence process of current spike signals within the period, analyze the changing trend of voltage phase angle with time, identify data segments with opposite changing directions, and obtain interference separation characteristic sequences; S3: Based on the interference separation feature sequence, compare the changes in conductor current density in each period, analyze the synchronicity between current density and temperature during temperature fluctuations, and statistically analyze the time distribution of synchronization events within the sampling period to obtain the periodic intersection distribution characteristics. S4: Based on the periodic intersection distribution characteristics, filter the periods with continuously progressive time distribution, analyze the starting nodes of the trend trajectory, determine whether the trend evolution requirements are met, optimize the starting node selection method, and obtain the trend starting point trajectory group. S5: Based on the trend starting point trajectory group, adjust and analyze the voltage and conductor flow change characteristics within the target interval, sort out their respective time series changes, determine the order of characteristic changes and calculate the trend directionality to obtain the trend directionality factor.

[0006] The present invention improves upon this invention by including the following: the time-series aligned voltage feature includes a synchronous acquisition identifier, a node response sequence, and aligned baseline parameters; the interference separation feature sequence includes a spike discrimination result, a reverse trend segment, and an interference label index; the periodic intersection distribution characteristic includes intersection time nodes, distribution characteristic indicators, and a synchronous event sequence; the trend starting point trajectory group includes a trend starting point index, a progressive period number, and trajectory grouping information; and the trend directionality factor includes a directionality label, a sequence feature code, and a trend indication parameter.

[0007] The present invention is improved in that the step of obtaining the timing alignment voltage characteristics is specifically as follows: S111: Based on cable monitoring nodes, analyze the layout of cable monitoring nodes and the middle of the structure. Combine the spatial sorting and configuration parameters of the monitoring nodes to optimize the sampling start time of each monitoring node. By adjusting the time order to correspond to the physical distribution, obtain the sampling synchronization parameter set. S112: Based on the sampling synchronization parameter set, analyze the synchronization of the voltage data acquisition start point of each channel, compare the voltage signal response time differences within the same sampling period, integrate and reassemble the response timing of the channels, and obtain the response alignment sequence. S113: Based on the response alignment sequence, compare the synchronization characteristics of each channel, identify the corresponding time index, call the middle part of the structure as a unified reference point, optimize the synchronization process of the voltage fluctuation sequence of each channel, reorganize the response data between channels, and obtain the time-aligned voltage characteristics.

[0008] The present invention is improved in that the step of obtaining the interference separation feature sequence is specifically as follows: S211: Based on the time-aligned voltage characteristics, determine the abrupt change position in each current data frame within the period, identify the time segment of amplitude change in the current signal sequence, and combine the time index when the signal abrupt change occurs to aggregate the identified response intervals to obtain the current peak response interval. S212: Based on the current peak response interval, determine the direction of change of voltage phase angle with time in each response interval, compare the rising and falling process of phase angle sequence under time index, make a continuous judgment on the direction of change, identify segments with trends, and obtain voltage phase trend sequence. S213: Based on the voltage phase trend sequence, filter the time segments within the period where the current spike signal is opposite to the phase angle trend direction, determine the signal data characteristics of the segments, and mark the corresponding time index to obtain the interference separation feature sequence.

[0009] The present invention is improved in that the step of obtaining the periodic intersection distribution characteristics is specifically as follows: S311: Based on the interference separation feature sequence, analyze the current carrying density sequence and temperature sequence collected at each monitoring point along the cable, synchronously compare the changing trends of the two sets of sequences at the same time node, identify continuous data segments with consistent trends, and obtain a set of periodic synchronization events. S312: Based on the set of periodic synchronization events, calculate the change in current density and temperature of all consecutive segments in each period to obtain the average difference factor of the intersection events; S313: Compare the changing trends of the average difference factor of the intersection events in multiple sampling periods, determine the changing pattern over time, identify the periodic segments with changing trends, optimize the data labeling of the periodic segments, and obtain the periodic intersection distribution characteristics.

[0010] The present invention is improved in that the steps for obtaining the trend starting point trajectory group are specifically as follows: S411: Based on the periodic intersection distribution characteristics, filter periodic segments with a continuously progressive time distribution, determine the order of synchronous event nodes on the time axis in each period, optimize the period identification method, and obtain the continuous evolution period index. S412: Based on the continuous evolution cycle index, compare the synchronization event sequence of each cycle, determine the change slope of the current density corresponding to the first time node of each cycle and the change trend of the adjacent cycle, identify the nodes with different change slopes, and obtain the starting node identification sequence. S413: Based on the starting node identifier sequence, calculate the number of synchronization events for each node in the corresponding period, the time interval between adjacent nodes, and the temperature fluctuation amplitude of the node sequence, using the formula: ; Obtain the trend direction preservation parameters for each node to obtain the trend starting point trajectory group, where, Indicates the first The trend direction parameter remains constant at each trend starting node. Indicates the first The absolute difference in temperature fluctuation amplitude within the period in which each trend's starting point is located. Indicates the first The trend starting point is at the... The event quantity factor of each synchronization event location. This indicates the total number of synchronization events within the period. Indicates the first The time interval data between the starting node of a trend and its adjacent nodes.

[0011] The present invention is improved in that the step of obtaining the trend directionality factor is specifically as follows: S511: Based on the trend starting point trajectory group, determine the correspondence between the target interval voltage change sequence and the conductor flow change sequence within the sampling boundary, optimize the time axis sorting of voltage change nodes and conductor flow state change nodes, identify change segments that occur synchronously or sequentially, and obtain the current-voltage alignment feature set. S512: Based on the current-voltage alignment feature set, analyze the time sequence of voltage change sequence and conductor flow sequence, determine the order of each change segment, filter the sequence segment that changes first, and mark the start and end index information to obtain the feature change sorting sequence. S513: Based on the sorted sequence of the characteristic changes, determine the correlation between the voltage change direction and the conductor flow change direction in a continuous segment, compare the directional labels of each time segment, identify sequence segments with a consistent trend, and obtain the trend directionality factor.

[0012] The present invention is improved in that the monitoring node refers to a data acquisition point set along the power cable to collect electrical parameters in real time; the signal response time difference refers to the time sequence in which different nodes detect the same electrical event; and the current spike signal refers to the instantaneous current signal with a sudden change in amplitude that appears in the monitoring data.

[0013] A power cable fault monitoring system based on big data, the system comprising: The data synchronization module is based on cable monitoring nodes. It analyzes the spatial order of each cable monitoring node and the middle of the structure, adjusts the sampling start time, unifies the acquisition process of each node, corrects the difference in signal response time, and obtains the timing-aligned voltage characteristics. Based on the time-aligned voltage characteristics, the trend separation module determines the occurrence process of current spike signals within the period, analyzes the changing trend of voltage phase angle over time, identifies data segments with opposite changing directions, and obtains the interference separation feature sequence. Based on the interference separation feature sequence, the hot load extraction module compares the changes in conductor current density in each period, analyzes the synchronicity between current density and temperature during temperature fluctuations, and statistically analyzes the time distribution of synchronization events within the sampling period to obtain the periodic intersection distribution characteristics. Based on the periodic intersection distribution characteristics, the trajectory filtering module filters periods with continuously progressive time distribution, analyzes the starting nodes of trend trajectories, determines whether they meet the trend evolution requirements, optimizes the starting node selection method, and obtains the trend starting point trajectory group. Based on the trend starting point trajectory group, the directional factor module adjusts and analyzes the voltage and conductor flow change characteristics within the target interval, sorts out their respective time series changes, determines the order of characteristic changes and calculates the trend directionality to obtain the trend direction factor.

[0014] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, the accuracy of linkage between multiple signals is ensured by synchronously collecting node data and combining it with time-series calibration. The trend discrimination method progressively filters the coordinated fluctuations of different characteristic parameters, dynamically extracts the continuous evolution sequence of various types of minor disturbances under complex operating conditions, and uses the temperature and current density correlation method to highlight the gradual nature and location capability of fault precursors. Dynamic grouping of data segments and trend progression trajectory determination improve the correlation depth of monitoring information. The labeling of trend directionality factors provides direct parameter support for subsequent decision-making and anomaly identification, effectively capturing dynamic risk signals that are difficult to identify in the early stages of change, and realizing step-by-step early warning and fine identification of multi-source weak anomalies in power cables. Attached Figure Description

[0015] Figure 1 This is a flowchart of the main steps of the present invention; Figure 2 This is a flowchart illustrating the acquisition of timing-aligned voltage characteristics in this invention. Figure 3 This is a flowchart illustrating the process of obtaining the interference separation feature sequence in this invention. Figure 4 This is a flowchart illustrating the process of obtaining the periodic intersection distribution characteristics in this invention. Figure 5 This is a flowchart illustrating the process of obtaining the trend starting point trajectory group in this invention. Figure 6 This is a flowchart illustrating the process of obtaining the trend directionality factor in this invention. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0017] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0018] Example Please see Figure 1 This invention provides a technical solution: a method for monitoring power cable faults based on big data, comprising the following steps: S1: Based on cable monitoring nodes, analyze the installation sequence between monitoring nodes and the middle of the structure, adjust the sampling start time of each monitoring node so that the sampling sequence corresponds to the spatial distribution characteristics, optimize the data acquisition process of the acquisition channel, compare the signal response time differences collected by each node, and integrate and align the voltage fluctuations according to a unified baseline to obtain the time-aligned voltage characteristics. S2: Based on the time-aligned voltage characteristics, determine the occurrence process of the current spike signal in the response interval within the period, analyze the change trend of the voltage phase angle with time during the occurrence process, compare the change trend of the current spike signal with the change direction of the voltage phase angle, identify the data segment with the opposite direction as valid data, and obtain the interference separation feature sequence. S3: Based on the interference separation feature sequence, compare the change process of conductor current density in each period, analyze the synchronicity of the change of current density and temperature data during temperature fluctuation, determine the time distribution characteristics of synchronicity events within the sampling period, calculate the distribution law of intersection events in each period, and obtain the period intersection distribution characteristics. S4: Based on the characteristics of periodic intersection distribution, filter the periods in which the time distribution shows a continuous progressive state, analyze the starting node of the trend trajectory in the progressive period, determine whether the starting node meets the judgment condition of trend evolution, optimize the selection method of the starting node, and obtain the trend starting point trajectory group. S5: Based on the trend starting point trajectory group, adjust the voltage change characteristics and conductor flow change characteristics of the target interval, analyze the time series performance of each type of change characteristic, determine the order of characteristic changes, calculate the trend directionality of characteristic changes, and obtain the trend directionality factor.

[0019] The timing alignment voltage characteristics include synchronous acquisition identifier, node response sequence, and alignment baseline parameters; the interference separation characteristic sequence includes peak discrimination results, reverse trend segments, and interference label index; the periodic intersection distribution characteristics include intersection time nodes, distribution characteristic indicators, and synchronous event sequences; the trend starting point trajectory group includes trend starting point index, progressive period number, and trajectory grouping information; and the trend directionality factor includes directional label, sequence feature code, and trend indication parameters.

[0020] In S1, a monitoring node refers to a data acquisition point set along the power cable for real-time acquisition of electrical parameters (such as voltage, current, etc.), typically a sensor or terminal acquisition unit; the center of the structure refers to the geometric center point along a section of the cable's physical length, often used as a time alignment or electrical characteristic reference point; the sampling start time refers to the initial time when each monitoring node begins data acquisition, adjusting this item allows for synchronized acquisition across the entire line; spatial distribution characteristics refer to the distribution relationship of each monitoring node in physical space (along the cable), used to analyze signal transmission paths and timing relationships; the data acquisition process refers to the data processing steps, such as signal acquisition, synchronization, transmission, and storage, completed sequentially by each node; the signal response time difference refers to the time sequence at which different nodes detect the same electrical event (such as a pulse or voltage change), reflecting signal propagation delay; the unified baseline refers to correcting the data acquired by different nodes to the same time reference, achieving time synchronization; and the aligned voltage fluctuation refers to the voltage change process at each point after all acquisition nodes are synchronized, often used for subsequent feature analysis.

[0021] In S2, the current spike signal refers to the instantaneous current signal with a sudden change in amplitude that appears in the monitoring data, usually reflecting abnormal or disturbance events such as partial discharge, short circuit, and impact; the response interval refers to the time range during which the current spike signal appears and lasts; the voltage phase angle refers to the phase position of the AC voltage relative to the reference signal at a certain moment, and its change can reveal the spatial propagation characteristics of electrical events; the trend of change refers to the directionality (rising, falling, or oscillating) of the current spike signal or voltage phase angle as it changes over time; the direction of change refers to the direction of increase or decrease of a certain characteristic parameter within the observation interval (such as phase advance or lag, current increase or decrease); the data segment with opposite direction refers to the time segment in which the current spike signal and voltage phase angle change in opposite directions, and such segments often correspond to physical anomalies (such as the energy release at the moment of discharge).

[0022] In S3, conductor current density refers to the actual current density carried by the conductor per unit cross-sectional area, which is a parameter for measuring the conductor's load condition; temperature fluctuation process refers to the change of conductor temperature over time within the monitoring period, which is affected by multiple factors such as current carrying capacity and environment; change synchronicity guides whether the change in current carrying capacity and temperature change are synchronized in time, and abnormal synchronicity indicates potential faults; synchronous event refers to the event node where current carrying capacity and temperature data change together within the same time period; sampling period refers to the set data acquisition interval period, and each period can be analyzed independently; time distribution characteristics refer to the time distribution pattern of synchronous events throughout the entire sampling period; the distribution pattern of intersection events guides the statistical characteristics of the intersection points of the current carrying capacity and temperature change curves within the period, which are used to analyze evolution trends.

[0023] In S4, time distribution refers to the time distribution of synchronous events or feature nodes in the previous cycle or multiple cycles; the cycle of continuous progressive state refers to the trend of synchronous events or feature nodes gradually moving and changing in several consecutive cycles; the starting node of the trend trajectory refers to the key data point corresponding to the beginning of the trend change in the progressive cycle; the judgment condition of trend evolution refers to the standard set according to the change law of monitoring parameters, used to judge whether the trend trajectory meets the fault evolution or development trend; the selection method refers to the method of final inclusion in the analysis by using operations such as screening, judgment, and priority sorting relative to the starting node of the trend trajectory.

[0024] In S5, the target interval refers to the time period or data range that needs to be focused on and analyzed in trend trajectory analysis; voltage change characteristics refer to the key characteristics of voltage change over time within the target interval, such as rising edges, falling edges, and jump points; conductor flow change characteristics refer to the change pattern of conductor current-carrying state (such as current, current density, etc.) within the target interval; time series performance refers to the sequence-like change process of characteristic parameters over time; the sequence of characteristic changes refers to clarifying the relative order of changes of each parameter on the time axis when analyzing the changes of multiple parameters, which helps to identify causal relationships; trend directionality refers to the trend development direction summarized from the changes of the above parameters, which is used for subsequent status warnings or hidden danger location.

[0025] Please see Figure 2 The specific steps for obtaining the timing-aligned voltage characteristics are as follows: S111: Based on cable monitoring nodes, analyze the layout of cable monitoring nodes and the middle of the structure. Combine the spatial sorting and configuration parameters of the monitoring nodes to optimize the sampling start time of each monitoring node. By adjusting the time order to correspond to the physical distribution, obtain the sampling synchronization parameter set. The installation position of each node along the physical length of the cable is obtained, and its coordinates are converted into relative distances from the geometric center of the cable. These nodes are then numbered sequentially from the cable's start to the end point. Combining the node configuration parameters such as sampling frequency, data processing cycle, and start phase, the sampling start time configuration of each node is read sequentially. The signal propagation delay corresponding to the actual physical location of each node is compared, and the propagation time difference required for each node to travel from the signal source to its location is calculated. Following the central symmetry reference principle, the sampling start time of the corresponding node in the middle of the cable is extracted as the baseline sampling time. Based on this, the propagation delay is added or subtracted node by node. The original sampling time is corrected to form a sampling time sequence that presents an ordered progressive relationship from the start to the end. By combining each node number with the corrected sampling time pair, a complete sampling synchronization parameter set is constructed. In specific implementation, for example, if the total length of the cable is 500 meters and the middle node is located at the 250-meter mark, its sampling start time is set to 10 milliseconds. If a node is located at 300 meters, it is offset from the middle by 50 meters. If the signal propagation speed is 200 million meters per second, the node should start sampling 0.25 microseconds in advance, and the corresponding sampling start time is adjusted to 9.9 milliseconds. The synchronization time configuration of all nodes is completed in this way to form a sampling synchronization parameter set covering the entire cable line.

[0026] S112: Based on the sampling synchronization parameter set, analyze the synchronization of the voltage data acquisition start point of each channel, compare the voltage signal response time differences within the same sampling period, integrate and reassemble the response timing of the channels, and obtain the response alignment sequence. After sorting the voltage data of each channel according to the actual start time, the time series is expanded. The time point of the first occurrence of the same electrical event (such as instantaneous voltage jump) in each channel is analyzed. The response time difference of the event between each channel is calculated. By comparing the difference, the channel with the earliest response time is identified as a reference. The starting index of the voltage data of other channels is adjusted, and the voltage sequence of the channel with the later response time is moved forward. The channel with the earlier time can remain unchanged or delay to fill the missing segment, thereby unifying the response process of each channel under the same time reference. Then, the time series of each channel is rearranged to align the voltage response start points of all channels. In actual operation, if multiple channels detect a ground short circuit event, and one node first detects the event at 10.005 milliseconds, while another node detects the same event at 10.001 milliseconds, then the latter channel needs to shift the voltage data backward by 4 sampling points (if the sampling rate is 1000 points per millisecond). After rearrangement and time coordinate mapping, the voltage change process of all channels is aligned in the time dimension, forming a response alignment sequence under a unified reference.

[0027] S113: Based on the response alignment sequence, compare the synchronization characteristics of each channel, identify the corresponding time index, call the middle part of the structure as a unified reference point, optimize the synchronization process of the voltage fluctuation sequence of each channel, reorganize the response data between channels, and obtain the time-aligned voltage characteristics. The variation characteristics of each channel are extracted from the voltage data sequence after response alignment. Within each sampling period, the rate trend of voltage rise or fall is calculated, and the consistency of the change direction of each channel at the same moment is judged. Time periods with similar variation characteristics are identified and marked accordingly. The variation characteristics are then compared one by one with the voltage change trends recorded by the channels in the middle of the structure. Channels with similar change trends to those in the middle of the structure are selected as retention targets, while channels with opposite change directions or fluctuation amplitudes greater than a specified range are excluded. The selected range can be set based on the statistical characteristics of voltage fluctuations in the channels in the middle of the structure. For example, if the voltage of a node in the middle of the structure shows a positive upward trend for a certain period of time, and other nodes also show the same upward trend in the same period of time, and their fluctuation values ​​fluctuate within ±5 volts, then it can be considered that it is consistent with the voltage change in the middle of the structure, and the data of that channel is retained. Otherwise, it is considered an abnormal data channel and is excluded. Subsequently, the voltage data of the retained channels are reordered and resampled to unify them to the same time interval and starting reference point, constructing a multi-channel synchronous data matrix and generating time-aligned voltage characteristics that can be used for further analysis.

[0028] Please see Figure 3 The specific steps for obtaining the interference separation feature sequence are as follows: S211: Based on the time-aligned voltage characteristics, determine the abrupt change position in each current data frame within the period, identify the time segment of amplitude change in the current signal sequence, and combine the time index when the signal abrupt change occurs to aggregate the identified response intervals to obtain the current peak response interval. Extract current signal data frame sequences synchronized with voltage characteristics from the current sampling period. Calculate the current amplitude difference between two adjacent moments for each sampling point, and determine if the difference exceeds a set amplitude change threshold. This threshold is set to three times the current fluctuation range under normal operating conditions. For example, when the background current fluctuation range is ±2 amps, the change threshold can be set to 6 amps. Traverse the current sequence frame by frame for each channel, detecting whether the change between consecutive data points exceeds this threshold. If the current value at a certain point is 5 amps and the next point is 12 amps, the difference is 7 amps, exceeding the threshold of 6 amps. This position can be determined as the start of a change. Continue to determine whether there is a continuous drastic change in the same direction as the change in the subsequent sampling points. If the continuity judgment criteria are met, such as 8 out of 10 consecutive sampling points... If the amplitude of the sampling point change remains above 5 amperes and the direction is consistent, then this time period is identified as a sudden change segment. The start and end time indices are recorded. Then, all sudden change points within the period are traversed sequentially. If the interval between two adjacent sudden change segments is less than the set minimum window threshold, such as within 1 millisecond, the two sudden change segments are merged to form a single response interval. All response intervals are sorted, merged, and numbered to generate a set of current peak response intervals. For example, if four current sudden change response segments are identified within a 50-millisecond sampling period, located at 5-6 milliseconds, 12-13 milliseconds, 13.2-14 milliseconds, and 30-31 milliseconds respectively, the 12-13 and 13.2-14 millisecond segments are merged to form a new 12-14 millisecond response interval. The output consists of three response segments, which serve as the current peak response intervals.

[0029] S212: Based on the current peak response interval, determine the direction of change of voltage phase angle with time in each response interval, compare the rising and falling process of phase angle sequence under time index, make a continuous judgment on the direction of change, identify segments with trends, and obtain voltage phase trend sequence. Read the voltage phase angle data sequence corresponding to each response interval time period, compare the phase angle values ​​between adjacent sampling points point by point, and determine whether the phase angle shows a monotonically increasing or decreasing trend. If the phase angle difference between two consecutive points is positive, it is an increasing trend; if the difference is negative, it is a decreasing trend. For sub-segments with a length of more than 5 sampling points, count the number of consecutively increasing or decreasing points. If the number of consecutively increasing points exceeds 70% of the total, the sub-segment is determined to be an increasing trend segment. Conversely, if the number of consecutively decreasing points exceeds 70%, it is a decreasing trend segment. For example, if the voltage phase angle is 20 degrees, 25 degrees, 28 degrees, 30 degrees, 29 degrees, and 31 degrees in a certain period of time, then the first three points and the last two points can constitute two increasing segments. If the length of the increasing segments after merging reaches 6 sampling points, accounting for 80% of the whole, then the whole can be determined to be an increasing trend segment. Then, repeat the above judgment process for all response intervals, identify and record the phase angle sequence with obvious continuous changing trend, and mark its time index. Collect all segments marked as continuously increasing or decreasing into a voltage phase trend sequence.

[0030] S213: Based on the voltage phase trend sequence, filter the time segments within the period where the current peak signal is opposite to the phase angle trend direction, determine the signal data characteristics of the segments, and mark the corresponding time index to obtain the interference separation feature sequence; All identified current spike response intervals within the cycle are read, and their time indexes are analyzed one by one against the phase trend sequence. For each current spike response interval, the direction of current change is compared with the direction of voltage phase angle change in the same time period. If the current shows an upward trend while the voltage phase angle shows a downward trend, or vice versa, the interval is identified as having opposite directions. The start and end time indices of this interval are marked as candidate interference data intervals. The current change characteristics of the candidate intervals are then analyzed, including the maximum amplitude, fluctuation frequency, and duration. For example, if the duration of a certain interval is 1.5 milliseconds, the maximum current change amplitude is 10 amperes, and the fluctuation frequency is greater than 1 kHz, it indicates that the interval is a transient impulse interference. Based on this, the characteristic index thresholds for interference identification are set. Among them, the fluctuation frequency exceeding 0.8 kHz, the amplitude change exceeding 8 amperes, and the opposite direction of phase change are used as interference judgment conditions. Time intervals that meet all three conditions are identified as interference data intervals. Interference labels are applied to the intervals, and the corresponding time indices are recorded to construct an interference separation feature sequence.

[0031] Please see Figure 4 The specific steps for obtaining the periodic intersection distribution characteristics are as follows: S311: Based on the interference separation feature sequence, analyze the current carrying density sequence and temperature sequence collected at each monitoring point along the cable, synchronously compare the changing trends of the two sets of sequences at the same time node, identify continuous data segments with consistent trends, and obtain a set of periodic synchronization events. Extract the conductor current-carrying density time series and corresponding temperature time series recorded by each monitoring node along the cable path. For each monitoring node, read the current-carrying density and temperature values ​​arranged in ascending order within the same sampling period. For the same time index point, calculate the change difference between adjacent data points in the two series and mark the direction of change as increasing or decreasing. Compare the change direction at each moment in the two series. If the two directions are consistent, that moment is determined as a point of consistent trend. This judgment operation is continuously performed on each node within one sampling period, and the time period of continuous consistent trend is counted. Set the minimum continuous length threshold to 5 sampling points. If a node has consecutive values ​​between 2 milliseconds and 5 milliseconds within a 10-millisecond period, it is considered a point of consistent trend. If the trend is consistent across four consecutive sampling points, the threshold requirement is not met, and the segment is not marked as a periodic synchronization event. If the number of consecutive points with consistent trends reaches or exceeds five, the segment is marked as a candidate synchronization event segment, and its start and end time indices are recorded. To enhance accuracy, the maximum fluctuation value is determined for each synchronization event segment. The maximum amplitude change in the current density and temperature sequence of that segment is calculated. If the amplitude of any sequence change is less than the minimum fluctuation threshold of two units (e.g., 2A / mm² or 2℃), the segment is deemed unrepresentative and is removed. Segments with consistent trend direction, sufficient duration, and amplitude fluctuation greater than the threshold are retained. The valid segments of all monitoring nodes are sorted by time to form a set of periodic synchronization events.

[0032] S312: Based on the set of periodic synchronization events, calculate the changes in current density and temperature for all consecutive segments within each set of periodic synchronization events, using the following formula: ; The average difference factor of the intersection events is obtained, where, This is the average difference factor for intersection events, representing the average degree of difference between changes in current density and temperature within the sampling period. This represents the number of synchronization event segments, specifically the total number of consecutive segments that meet the conditions identified within a sampling period. Indicates the first The change in current density within a single synchronous event segment represents the magnitude of the change in current density at the beginning and end of that segment. Indicates the first The temperature change within a single synchronous event segment represents the magnitude of temperature change at the beginning and end of that segment. It represents the equivalent proportionality coefficient, which is used to unify the change in current density and the change in temperature into the same physical dimension to achieve comparability; The average difference factor of the intersection event is a calculated technical indicator used to quantitatively measure the degree of difference between the change in current density and the change in temperature within a synchronous event segment during the monitoring of power cables. It reflects the change in current density within each sampling period in all segments identified as "synchronous event segments" (i.e., intervals where the trends of current density and temperature change are consistent). ) and temperature change ( The average absolute difference after unification of dimensions. The smaller the value, the more synchronized and highly coupled the changes in current density and temperature are within the sampling period, and the intersection event exhibits a "high degree of matching"; The larger the value, the more significant the deviation between the two changes, the weaker the synchronicity, and the existence of abnormal energy transfer, local faults, or disturbances. In essence, the average difference factor of the intersection event is a numerical and traceable descriptive index of the dynamic coupling state of local thermoelectric characteristics during the operation of power cables, providing basic data support for subsequent technical aspects such as periodic distribution characteristic analysis and fault trend judgment.

[0033] For example, analysis period number The continuous segment distribution of the synchronous event set is used to extract the current density and temperature data of the first and last sampling nodes for each segment. Excerpt The current carrying capacity is from 5.1 A / mm 2 Change to 5.9A / mm 2 The temperature changed from 40℃ to 42.8℃; Excerpt The current carrying capacity changed from 4.8 to 5.6, and the temperature changed from 39.5 to 42.1. Excerpt The current carrying capacity changed from 5.3 to 6, and the temperature changed from 40.7 to 43.2. Excerpt The current carrying density changed from 5 to 6.1, and the temperature changed from 40.5 to 43.6.

[0034] By calculating the changes in each segment using the data from the beginning and end, we can obtain: , , , ; , , , .

[0035] The data is mapped to a linear normalization method for all changes. Intervals, after normalization: , , , ; , , , The data corresponds one-to-one with the variation characteristics of each segment. To coordinate the dimensions of the influence of normalized current density and temperature, a scaling factor is set. This coefficient is determined based on the thermoelectric properties of the conductor material, and is selected as follows: , representing the equivalent adjustment ratio of the normalized change in current carrier density to the change in temperature, is given by substituting the above parameters into the formula: ; ; ; The summation symbol in the formula represents summation over all The absolute differences of each synchronized segment are summed and then averaged. The denominator is... Number of synchronized segments in the current period The multiplication sign is used to calculate the proportionality coefficient. The product of the current carrier density change and the temperature change is used to calculate the difference between the normalized current carrier density change and the temperature change. The absolute value sign ensures that the differences of all segments are non-negative, thus obtaining an output term with clear statistical meaning. The benchmark comparison interval is set to three levels, where: The first paragraph is It is used to describe a periodic state in which temperature changes and current density changes are highly synchronized and the curve fit is tight; The second paragraph is It is used to identify periodic states where there are certain differences in the changing trends of temperature and current density, but the overall relationship still has coupling characteristics. The third paragraph is This indicates a significant inconsistency between the trends of temperature change and current density change, suggesting a periodic state with local decoupling or short-term abnormal disturbances.

[0036] Current period calculation results Located in the second paragraph Within this range, it indicates that there is a certain degree of deviation between temperature changes and current density fluctuations in the synchronous event segments within this cycle, but it does not deviate from the acceptable coupling range and has an overall consistent trend. Therefore, this result can be used to identify this cycle as a "moderately coupled state". Corresponding to the definition of the average difference factor of the intersection event, this factor, as an intermediate data item, quantitatively reflects the overall coupling level of the synchronous event set selected in the previous sub-step in terms of feature changes, and serves as the input basis for subsequent trend classification and cycle segment optimization. Its value directly affects the trend judgment of the cycle distribution trajectory and the setting of classification rules in the next stage.

[0037] S313: Compare the changing trends of the average difference factor of intersection events in multiple sampling periods, determine the changing pattern over time, identify the period segments with changing trends, optimize the data labeling of the period segments, and obtain the period intersection distribution characteristics. The number of valid synchronization events in each sampling period is counted, and the absolute difference sequence between the temperature sequence and the current density sequence is calculated for each event segment. The arithmetic mean of all points in the difference sequence is calculated as the average difference factor for that event. The average difference factor of multiple events in the same period is averaged as the representative value of that period, forming a periodic factor sequence arranged in time. Further, the difference between the average difference factors of every two adjacent periods in the periodic factor sequence is calculated, and the positive and negative directions of the difference are recorded. It is then determined whether the direction of change is consistent across multiple consecutive periods. For example, if the difference for three consecutive periods is +0.3, +0.5, and +0.4, the direction is consistent, and the periodic segment is identified as having a changing trend. Identification criteria are then set. To confirm a trend, three or more consecutive cycles must show the same direction. The trend cycle segments are then renumbered, replacing the original numbers with a new coding format. For example, if a continuous upward trend exists from cycle 5 to cycle 8, it is uniformly labeled "Trend Segment-A" to enhance subsequent identification efficiency. In the above operations, the trend identification threshold is set to a change in amplitude greater than 0.2 per cycle. If the cycle difference is 0.1, it is not included in the trend segment. In specific applications, if the monitoring cycle length is 10 milliseconds and the sampling frequency is 10 kHz, then each cycle contains 100 data points. If the average difference factors for cycles 2, 3, and 4 are 1.2, 1.5, and 1.8 respectively, and the differences are +0.3 and +0.3 respectively, the trend determination condition is met, and the segment is identified as a cycle intersection distribution characteristic.

[0038] Please see Figure 5 The specific steps for obtaining the trend starting point trajectory group are as follows: S411: Based on the periodic intersection distribution characteristics, filter periodic segments with a continuous progressive time distribution, determine the order of synchronous event nodes on the time axis in each period, optimize the period identification method, and obtain the continuous evolution period index. Extract the time indices corresponding to the identified intersection events in all sampling periods, sort them in ascending order by period number, and compare the distribution of intersection events on the time axis in adjacent periods to determine if the time indices exhibit a monotonically increasing trend. Calculate the difference between the first intersection node time of each pair of periods. If the node time of the later period is earlier than that of the previous period, the segment is marked as a non-progressive interval; otherwise, it is marked as a progressive segment. Set the judgment threshold that the time index difference is not less than the length of one sampling point. For example, if the sampling frequency is 10kHz, then each sampling point is 0.1ms. If the intersection node of period 1 occurs at 12.3ms and that of period 2 occurs at 12.4ms, the difference is 0.1ms, satisfying the progressive judgment. If the time difference between periods is negative or zero, then the continuous progressive segment is disconnected. During the tracking process, previously accumulated segments are marked as complete, continuously progressive cyclical segments. All cyclical segments that meet the conditions are collected, organized, and assigned new index numbers. To ensure the readability and structural distinctiveness of the index, a grouping numbering method is adopted. For example, the continuous segment with cycle numbers 4 to 7 is defined as "EvolSeg_01", and another segment from 8 to 10 is defined as "EvolSeg_02". Based on this, all cycles within each segment are renumbered as local sequences, such as "EvolSeg_01.1", "EvolSeg_01.2", etc. The corresponding markers are inserted into the original cycle data index field to replace the original global cycle numbers. A new cycle identifier sequence with continuous evolution attributes is output as a reference basis for subsequent trend node extraction and path tracking.

[0039] S412: Based on the continuous evolution cycle index, compare the synchronous event sequence of each cycle, determine the trend of the change slope of the current density corresponding to the first time node of each cycle and the change trend of adjacent cycles, identify the nodes with different change slopes, and obtain the starting node identification sequence. Data records for all cycles within each evolutionary period are read sequentially. The time index and corresponding current density value of the first synchronization event node in each cycle are extracted. The difference in current density between two sampling points before and after the starting node's time point is calculated to obtain an approximate slope, which is used as the slope of the first node in that cycle. The slope differences of the first nodes in two adjacent cycles are then compared to determine whether the difference exceeds a preset threshold. This threshold is set as the difference in density change rate of 1 ampere per square millimeter. If the slope of the first node in cycle A is 3.2 A / mm² / ms and that in cycle B is 5.1 A / mm² / ms, then the difference is 1.9, which meets the judgment condition. The point is marked as a slope difference node. The corresponding period number and time index of the node are added to the starting node identifier list. The cycle data in all continuous evolution cycle segments are traversed. The number of starting nodes marked as slope difference nodes in each cycle segment is counted. Each segment is required to contain at least one starting node. Otherwise, the segment is considered an evolution failure segment and is removed from the index sequence. A unique identifier is added to the valid starting nodes. For example, in the "EvolSeg_01" segment, if the starting node of period number 5 is identified as a node with a prominent change, it is identified as "EvolSeg_01.StartNode_5". After sorting out all valid identifiers, the output is the starting node identifier sequence.

[0040] S413: Based on the starting node identifier sequence, calculate the number of synchronization events for each node in the corresponding period, the time interval between adjacent nodes, and the temperature fluctuation amplitude of the node sequence, using the following formula: ; Obtain the trend direction preservation parameters for each node to obtain the trend starting point trajectory group, where, Indicates the first The trend direction parameter remains constant at each trend starting node. Indicates the first The absolute difference in temperature fluctuation amplitude within the period in which each trend's starting point is located. Indicates the first The trend starting point is at the... The event quantity factor of each synchronization event location. This indicates the total number of synchronization events within that period. Indicates the first The time interval data between the starting node of a trend and its adjacent nodes; The trend direction persistence parameter is a comprehensive measure used to quantify the persistence of the trend evolution direction of each trend initiation node in the power cable fault monitoring sequence. This parameter reflects the ability of a trend initiation node to maintain the subsequent trend trajectory under the combined effects of multiple factors such as temperature fluctuation amplitude, number of synchronous events, and time intervals within the same period. A higher value indicates that the trend initiation node is not only located at a position with significant temperature fluctuations, but also accumulates more synchronous events within the corresponding period, and the time distribution between adjacent initiation nodes is more compact, which helps to determine the degree of dominance and continuity of the node in the overall trend evolution direction. The monitoring data for each trend starting node within its corresponding period are retrieved sequentially, and three types of parameters are extracted: the number of synchronized events, the time interval between adjacent trend nodes, and the temperature fluctuation amplitude within the period. Let the number of parameters be... There is a trend starting node, within which 6 synchronization events are detected, with event times of 0.8s, 1.3s, 1.9s, 2.5s, 3.1s, and 3.7s respectively. The counting result is... The temperature monitoring values ​​within the 3rd cycle of the trend node were 41.2℃, 43.5℃, 44.8℃, and 46℃, indicating a temperature fluctuation range of... ℃, the time interval between the start event of this node and the adjacent trend node 4 is 3.4s, corresponding to The participating items are subjected to min-max normalization, and the following settings are applied to the data samples in this batch: The minimum number of synchronization events is 2, and the maximum number is 10. The minimum temperature fluctuation range is 2.5℃, and the maximum temperature fluctuation range is 6℃. The minimum time interval is 2s, and the maximum time interval is 6s. The normalized calculations are as follows: ; ; ; Substitute the above data into the formula to calculate the numerator: ; Calculate the denominator term again: ; The trend direction maintenance parameter is: ; Maintain the trend direction parameter The normalized values ​​are divided into the following three intervals: when When the trend remains weak, it is defined as a weak trend interval. The temperature fluctuation of such nodes is low, or the number of synchronous events is insufficient, or the time interval between them and their neighboring nodes is large, resulting in poor overall trend continuity. when When the trend remains in the middle range, these nodes are defined as having a certain consistency in terms of temperature changes, event density, and temporal connectivity, and are suitable as candidate nodes to participate in trend trajectory construction. when When the trend is strong, the node is defined as a region where the temperature fluctuation is significant, the synchronous events are concentrated, and the node distribution is compact. It has obvious trend-dominant characteristics and is given priority to be included in the main trajectory path. Calculation results Substituting the above intervals for comparison, it can be seen that it falls into the medium trend range. This result indicates that the trend starting node 3 has a certain degree of trend direction coordination in terms of parameters such as temperature fluctuation, number of events, and node time density. Its directional change does not show obvious abrupt changes in the spatial-temporal dimension, nor does it show significant aggregation characteristics.

[0041] Please see Figure 6 The specific steps for obtaining the trend direction factor are as follows: S511: Based on the trend starting point trajectory group, determine the correspondence between the voltage change sequence and the conductor flow change sequence in the target interval within the sampling boundary, optimize the time axis sorting of voltage change nodes and conductor flow state change nodes, identify change segments that occur synchronously or sequentially, and obtain the current-voltage alignment feature set. Extract the voltage change sequence and the conductor flow change sequence recorded by all monitoring nodes within the target interval. Align the two sequences on the time axis to unify the sampling boundaries of each time node. Construct a change record table based on the same time index. Read the voltage value and conductor current density value at each sampling time point and calculate the increase / decrease direction of adjacent time points. By judging whether the directional signs of the voltage change value and current density change value before and after the same time point are consistent, construct a preliminary list of change segments for the voltage change node sequence and the conductor flow state change node sequence. Further extract the time index of the starting point of change in the two sequences and judge the order of the starting points of change in the two sequences. Mark the segment where the voltage change time is less than the current change time as "voltage first", and the other segment as "current first". In the judgment process, the error tolerance threshold is set to the time interval corresponding to one sampling point. For example, if the sampling frequency is 5kHz, the error tolerance is 0.2 milliseconds. If the time difference between two change nodes is less than this value, it is judged to occur synchronously, and the segment is marked as "synchronous change". Then, all segments are divided into time periods. If the same type of change occurs continuously for more than three sampling periods, they are merged into the same change segment. At the same time, the start index, end index, and change type of each change segment are recorded. For example, in the sequence of a certain monitoring node, the voltage change point appears at 5.6 milliseconds and the current change point appears at 5.8 milliseconds. The time difference is 0.2 milliseconds, which is equal to the tolerance threshold, so it is classified as "synchronous change". All the identified change types, corresponding time indices and sequence types are sorted out to form a current-voltage aligned feature set.

[0042] S512: Based on the current-voltage aligned feature set, analyze the time sequence of voltage change sequence and conductor flow sequence, determine the order of each change segment, filter the sequence segment that changes first, and mark the start and end index information to obtain the feature change sorting sequence. For each change segment, its start and end time indices and change type label are read. First, all segments marked "voltage first change" or "current first change" are extracted, and "synchronous change" segments are removed. Then, list structures are created for the "voltage first change" and "current first change" segments respectively, and the time index of each segment in the list is sorted to ensure the consistency of the segment order on the time axis. Subsequently, the time sequence of each segment in the two lists is compared, and which sequence is marked as the first event more frequently is identified. Statistical analysis is performed on multiple change segments within the same evolutionary period, and a screening threshold is set to determine that sequences that occur at least 3 times in 5 consecutive segments are considered the first events. For example, if voltage changes 4 times and current changes 1 time in periods 10 to 15, then the voltage sequence is considered the priority sequence within that period. All its change segments are recorded as the main sequence of significant changes. At the same time, the start index and end index of each selected change segment are recorded, and its change direction and numerical range are saved as additional information. For example, in a voltage change segment, the start index is 210, the end index is 250, and the change value rises from 120V to 160V. Then this segment is recorded as the rising segment and given the complete index identifier "EV_210_250". All selected segments are arranged in ascending order of time to construct the feature change sorting sequence.

[0043] S513: Based on the characteristic change sorting sequence, determine the correlation between the voltage change direction and the conductor flow change direction in continuous segments, compare the directional labels of each time segment, identify sequence segments with consistent trends, and obtain the trend directionality factor. The time range and direction of each voltage change segment and its corresponding conductor flow change segment are read segment by segment. Tags are extracted from the change directions in both sequences to determine whether the voltage change is increasing, decreasing, or stable, and the conductor flow change is similarly determined as increasing, decreasing, or remaining constant in current density. Directional tag strings are constructed for each consecutive time period. For example, if the voltage drops from 140V to 100V and the conductor current density drops from 9A / mm² to 7A / mm² within a certain time period, the voltage direction tag is "decreasing" and the current direction tag is "weakening." The two directional tags are further compared; if the tags are identical or similar, the segments are considered to have a consistent trend. The directional consistency judgment condition is set as follows: two... Directional labels belong to the same-direction change group, i.e., "rising-increasing" or "falling-weakening". Those not belonging to the same group, such as "rising-weakening", are judged as inconsistent trend segments. All segments with consistent trends are assigned a unified trend factor code and added to the trend directionality factor list. For each item, the start and end time index, voltage direction, current carrying direction and trend consistency identifier are recorded. For example, in a certain segment, the voltage rises from 200V to 230V, the conductor current rises from 8.5A / mm² to 10.0A / mm², and the time period is from 420 to 470 sampling points, it is marked as "positive correlation trend segment" and identified as "TF_Pos_420_470". All sequence segments with consistent trend directions are output to form the trend directionality factor.

[0044] A power cable fault monitoring system based on big data, the system includes: The data synchronization module is based on cable monitoring nodes. It analyzes the spatial order of each cable monitoring node and the middle of the structure, adjusts the sampling start time, unifies the acquisition process of each node, corrects the difference in signal response time, and obtains the timing-aligned voltage characteristics. The trend separation module, based on the time-aligned voltage characteristics, determines the occurrence process of current spike signals within a period, analyzes the changing trend of voltage phase angle over time, identifies data segments with opposite changing directions, and obtains the interference separation feature sequence. The hot load extraction module compares the changes in conductor current density in each period based on the interference separation feature sequence, analyzes the synchronicity between current density and temperature during temperature fluctuations, and statistically analyzes the time distribution of synchronization events within the sampling period to obtain the periodic intersection distribution characteristics. The trajectory filtering module, based on the periodic intersection distribution characteristics, filters continuously progressive periods in time distribution, analyzes the starting nodes of trend trajectories, determines whether they meet the trend evolution requirements, optimizes the starting node selection method, and obtains the trend starting point trajectory group. The directional factor module, based on the trend starting point trajectory group, adjusts and analyzes the voltage and conductor flow change characteristics within the target interval, sorts out their respective time series changes, determines the order of characteristic changes, calculates the trend directionality, and obtains the trend direction factor.

[0045] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A method for monitoring power cable faults based on big data, characterized in that, Includes the following steps: S1: Based on cable monitoring nodes, analyze the spatial order of each cable monitoring node and the middle of the structure, adjust the sampling start time, unify the acquisition process of each node, correct the difference in signal response time, and obtain the timing-aligned voltage characteristics. S2: Based on the time-aligned voltage characteristics, determine the occurrence process of current spike signals within the period, analyze the changing trend of voltage phase angle with time, identify data segments with opposite changing directions, and obtain interference separation characteristic sequences; S3: Based on the interference separation feature sequence, compare the changes in conductor current density in each period, analyze the synchronicity between current density and temperature during temperature fluctuations, and statistically analyze the time distribution of synchronization events within the sampling period to obtain the periodic intersection distribution characteristics. S4: Based on the periodic intersection distribution characteristics, filter the periods with continuously progressive time distribution, analyze the starting nodes of the trend trajectory, determine whether the trend evolution requirements are met, optimize the starting node selection method, and obtain the trend starting point trajectory group. S5: Based on the trend starting point trajectory group, adjust and analyze the voltage and conductor flow change characteristics within the target interval, sort out their respective time series changes, determine the order of characteristic changes and calculate the trend directionality to obtain the trend directionality factor.

2. The power cable fault monitoring method based on big data according to claim 1, characterized in that, The time-aligned voltage features include synchronous acquisition identifiers, node response sequences, and alignment baseline parameters; the interference separation feature sequences include peak discrimination results, reverse trend segments, and interference label indexes; the periodic intersection distribution characteristics include intersection time nodes, distribution characteristic indicators, and synchronous event sequences; the trend starting point trajectory group includes trend starting point indexes, progressive period numbers, and trajectory grouping information; and the trend directionality factor includes directional labels, sequence feature codes, and trend indication parameters.

3. The power cable fault monitoring method based on big data according to claim 1, characterized in that, The specific steps for obtaining the timing-aligned voltage characteristics are as follows: S111: Based on cable monitoring nodes, analyze the layout of cable monitoring nodes and the middle of the structure. Combine the spatial sorting and configuration parameters of the monitoring nodes to optimize the sampling start time of each monitoring node. By adjusting the time order to correspond to the physical distribution, obtain the sampling synchronization parameter set. S112: Based on the sampling synchronization parameter set, analyze the synchronization of the voltage data acquisition start point of each channel, compare the voltage signal response time differences within the same sampling period, integrate and reassemble the response timing of the channels, and obtain the response alignment sequence. S113: Based on the response alignment sequence, compare the synchronization characteristics of each channel, identify the corresponding time index, call the middle part of the structure as a unified reference point, optimize the synchronization process of the voltage fluctuation sequence of each channel, reorganize the response data between channels, and obtain the time-aligned voltage characteristics.

4. The power cable fault monitoring method based on big data according to claim 1, characterized in that, The specific steps for obtaining the interference separation feature sequence are as follows: S211: Based on the time-aligned voltage characteristics, determine the abrupt change position in each current data frame within the period, identify the time segment of amplitude change in the current signal sequence, and combine the time index when the signal abrupt change occurs to aggregate the identified response intervals to obtain the current peak response interval. S212: Based on the current peak response interval, determine the direction of change of voltage phase angle with time in each response interval, compare the rising and falling process of phase angle sequence under time index, make a continuous judgment on the direction of change, identify segments with trends, and obtain voltage phase trend sequence. S213: Based on the voltage phase trend sequence, filter the time segments within the period where the current spike signal is opposite to the phase angle trend direction, determine the signal data characteristics of the segments, and mark the corresponding time index to obtain the interference separation feature sequence.

5. The power cable fault monitoring method based on big data according to claim 1, characterized in that, The specific steps for obtaining the periodic intersection distribution characteristics are as follows: S311: Based on the interference separation feature sequence, analyze the current carrying density sequence and temperature sequence collected at each monitoring point along the cable, synchronously compare the changing trends of the two sets of sequences at the same time node, identify continuous data segments with consistent trends, and obtain a set of periodic synchronization events. S312: Based on the set of periodic synchronization events, calculate the change in current density and temperature of all consecutive segments in each period to obtain the average difference factor of the intersection events; S313: Compare the changing trends of the average difference factor of the intersection events in multiple sampling periods, determine the changing pattern over time, identify the periodic segments with changing trends, optimize the data labeling of the periodic segments, and obtain the periodic intersection distribution characteristics.

6. The power cable fault monitoring method based on big data according to claim 1, characterized in that, The specific steps for obtaining the trend starting point trajectory group are as follows: S411: Based on the periodic intersection distribution characteristics, filter periodic segments with a continuously progressive time distribution, determine the order of synchronous event nodes on the time axis in each period, optimize the period identification method, and obtain the continuous evolution period index. S412: Based on the continuous evolution cycle index, compare the synchronization event sequence of each cycle, determine the change slope of the current density corresponding to the first time node of each cycle and the change trend of the adjacent cycle, identify the nodes with different change slopes, and obtain the starting node identification sequence. S413: Based on the starting node identifier sequence, calculate the number of synchronization events for each node in the corresponding period, the time interval between adjacent nodes, and the temperature fluctuation amplitude of the node sequence, using the formula: ； Obtain the trend direction preservation parameters for each node to obtain the trend starting point trajectory group, where, Indicates the first The trend direction parameter remains constant at each trend starting node. Indicates the first The absolute difference in temperature fluctuation amplitude within the period in which each trend's starting point is located. Indicates the first The trend starting point is at the... The event quantity factor of each synchronization event location. This indicates the total number of synchronization events within the period. Indicates the first The time interval data between the starting node of a trend and its adjacent nodes.

7. The power cable fault monitoring method based on big data according to claim 1, characterized in that, The specific steps for obtaining the trend direction factor are as follows: S511: Based on the trend starting point trajectory group, determine the correspondence between the target interval voltage change sequence and the conductor flow change sequence within the sampling boundary, optimize the time axis sorting of voltage change nodes and conductor flow state change nodes, identify change segments that occur synchronously or sequentially, and obtain the current-voltage alignment feature set. S512: Based on the current-voltage alignment feature set, analyze the time sequence of voltage change sequence and conductor flow sequence, determine the order of each change segment, filter the sequence segment that changes first, and mark the start and end index information to obtain the feature change sorting sequence. S513: Based on the sorted sequence of the characteristic changes, determine the correlation between the voltage change direction and the conductor flow change direction in a continuous segment, compare the directional labels of each time segment, identify sequence segments with a consistent trend, and obtain the trend directionality factor.

8. The power cable fault monitoring method based on big data according to claim 1, characterized in that, The monitoring node refers to a data acquisition point set up along the power cable to collect electrical parameters in real time. The signal response time difference refers to the time sequence in which different nodes detect the same electrical event. The current spike signal refers to the instantaneous current signal with a sudden change in amplitude that appears in the monitoring data.

9. A power cable fault monitoring system based on big data, characterized in that, The system is used to implement the power cable fault monitoring method based on big data as described in any one of claims 1-8, and the system includes: The data synchronization module is based on cable monitoring nodes. It analyzes the spatial order of each cable monitoring node and the middle of the structure, adjusts the sampling start time, unifies the acquisition process of each node, corrects the difference in signal response time, and obtains the timing-aligned voltage characteristics. Based on the time-aligned voltage characteristics, the trend separation module determines the occurrence process of current spike signals within the period, analyzes the changing trend of voltage phase angle over time, identifies data segments with opposite changing directions, and obtains the interference separation feature sequence. Based on the interference separation feature sequence, the hot load extraction module compares the changes in conductor current density in each period, analyzes the synchronicity between current density and temperature during temperature fluctuations, and statistically analyzes the time distribution of synchronization events within the sampling period to obtain the periodic intersection distribution characteristics. Based on the periodic intersection distribution characteristics, the trajectory filtering module filters periods with continuously progressive time distribution, analyzes the starting nodes of trend trajectories, determines whether they meet the trend evolution requirements, optimizes the starting node selection method, and obtains the trend starting point trajectory group. Based on the trend starting point trajectory group, the directional factor module adjusts and analyzes the voltage and conductor flow change characteristics within the target interval, sorts out their respective time series changes, determines the order of characteristic changes and calculates the trend directionality to obtain the trend directionality factor.

Images