A mine circuit fault self-diagnosis method and system

Abstract

The present application provides a method and system for self-diagnosis of mining circuit faults. Among them, the method collects the current, voltage waveform and three-dimensional vibration signal of the cable, generates anti-interference data through adaptive noise reduction and time synchronization; dynamically adjusts the waveform interception window based on the correlation between mechanical vibration intensity and current transient rate, extracts the peak slope change from the current transient segment, separates the high-frequency harmonic energy from the voltage transient segment, and performs time-frequency analysis on the vibration signal to extract the energy surge frequency point, and constructs a mechanical damage spectrum feature set; inputs the feature into the time-space correlation model, verifies the time-space consistency of current distortion and voltage anomaly to generate composite features; dynamically corrects the fault threshold based on environmental interference factors; and finally outputs the short-circuit spike or open circuit oscillation diagnosis result according to the correlation strength of electrical and mechanical features. The present application realizes accurate fault self-diagnosis of mining cables under complex working conditions.

Description

Technical Field

[0001] The present application relates to the technical field of intelligent fault diagnosis and prediction of mining equipment, and in particular to a method and system for self-diagnosis of mining circuit faults. Background Art

[0002] The underground environment of coal mines is humid, dusty, and subject to strong electromagnetic interference. Cables are prone to high-frequency transient faults such as short-circuit spikes and open-circuit oscillations due to mechanical damage or insulation aging. These faults are extremely short-lived and the signals are weak. Therefore, a diagnostic system is urgently needed to complete data collection, analysis, and alarm in a short period of time. At the same time, it must overcome the impact of complex underground working conditions on sensor accuracy and distinguish fault signals from electromagnetic noise.

[0003] The current mainstream solution uses a high-frequency traveling wave positioning and monitoring system. It captures the fault traveling wave signal through a high-frequency sensor installed on the cable grounding wire, calculates the fault distance using a two-end traveling wave ranging method, and uploads the waveform data to the cloud to trigger an alarm. The system relies on a high-speed sampling processor to achieve data acquisition and supports remote viewing of the fault waveform and location, but only uses a fixed threshold to filter high-frequency energy for diagnosis.

[0004] However, this solution has significant flaws. Its reliance on a single signal makes it impossible to distinguish between short-circuit spikes and electromagnetic interference, resulting in a high false alarm rate. The static feature analysis relies only on the amplitude and propagation time of the traveling wave, making it difficult to capture the subtle characteristics of early insulation degradation. Its dynamic adaptability is insufficient and the environmental parameter correction threshold is not integrated. It is prone to misjudgment during load fluctuations, and the response delay exceeds the transient fault blocking requirements. Summary of the Invention

[0005] The present application provides a mining circuit fault self-diagnosis method and system to solve the problems of easy misjudgment and response delay when load fluctuates in the prior art.

[0006] In a first aspect, the present application provides a method for self-diagnosis of mining circuit faults, comprising:

[0007] The cable's current waveform, voltage waveform, and three-dimensional vibration signal are synchronously collected through a multi-dimensional sensor group. Through adaptive noise reduction and time-series synchronous transmission in a humid and dusty environment, an anti-interference data set is generated.

[0008] Based on the correlation between the mechanical vibration intensity and the current transient rate in the anti-interference data set, the waveform clipping window is dynamically adjusted to capture the transient waveform segment, the peak slope change is extracted from the transient waveform segment of the current waveform, the harmonic frequency band energy of the short-circuit spike is separated from the transient waveform segment of the voltage waveform, and the three-dimensional vibration signal is simultaneously subjected to time-frequency analysis to generate a mechanical damage spectrum feature set;

[0009] Inputting the peak slope variation, harmonic frequency band energy and mechanical damage spectrum feature set into a spatiotemporal correlation model, and generating a composite feature vector through spatiotemporal correlation verification of current distortion features and voltage anomaly features;

[0010] Based on the environmental interference factors in the composite feature vector, combined with the cable surface temperature and dust data, the threshold value in the fault feature database is dynamically corrected;

[0011] A multi-dimensional matching calculation is performed on the composite feature vector and the corrected fault mode threshold, and a fault diagnosis result is determined based on the correlation strength between the electrical feature and the mechanical feature in the matching result.

[0012] Optionally, performing a multi-dimensional matching calculation on the composite feature vector and the corrected fault mode threshold, and determining a fault diagnosis result based on the correlation strength between the electrical feature and the mechanical feature in the matching result, including:

[0013] Decomposing the composite feature vector into an electrical feature sub-vector and a mechanical feature sub-vector, and performing Euclidean distance calculations on each of the sub-vectors with the corrected fault mode threshold to obtain electrical matching and mechanical matching;

[0014] Calculate the correlation strength value between the electrical matching degree and the mechanical matching degree. When the correlation strength value exceeds a preset fusion threshold, activate the fault type determination module to compare the deviation between the electrical matching degree and the short-circuit spike threshold with the deviation between the mechanical matching degree and the open-circuit oscillation threshold through the fault type determination module, and determine the fault diagnosis result based on the comparison result.

[0015] Optionally, determining a fault diagnosis result according to the comparison result includes:

[0016] If the deviation between the electrical matching degree and the short-circuit spike threshold is less than the deviation between the mechanical matching degree and the open circuit oscillation threshold, a short-circuit spike diagnosis result corresponding to cable damage is output;

[0017] If the deviation between the electrical matching degree and the short-circuit spike threshold is greater than the deviation between the mechanical matching degree and the open circuit oscillation threshold, a open circuit oscillation diagnosis result corresponding to the electrical anomaly is output.

[0018] Optionally, based on the environmental interference factor in the composite feature vector and in combination with the cable surface temperature and dust data, the threshold value in the fault feature database is dynamically corrected, including:

[0019] Extracting an environmental interference factor from the composite feature vector, wherein the environmental interference factor is calculated by multiplying a temperature value collected by a cable surface temperature sensor and a dust concentration monitoring value;

[0020] According to the magnitude of the environmental interference factor, a correction coefficient table corresponding to the temperature compensation coefficient and the dust attenuation coefficient is selected from the fault feature database;

[0021] The original threshold in the fault feature database is linearly combined with the temperature compensation coefficient and the dust attenuation coefficient in the correction coefficient table to generate a dynamically corrected threshold, wherein the temperature compensation coefficient has a piecewise linear relationship with the temperature value, and the dust attenuation coefficient has an exponential attenuation relationship with the dust concentration value.

[0022] Optionally, the peak slope variation, harmonic frequency band energy, and mechanical damage spectrum feature set are input into a spatiotemporal correlation model, and a composite feature vector is generated through spatiotemporal correlation verification of current distortion features and voltage anomaly features, including:

[0023] The peak slope variation, harmonic frequency band energy, and mechanical damage spectrum feature set are input into a feature fusion module, and data is reorganized according to a timestamp alignment rule to form a spatiotemporal matrix containing current distortion features, voltage anomaly features, and mechanical damage features;

[0024] Based on the current distortion characteristics and voltage anomaly characteristics in the space-time matrix, a time offset between the two within a preset time window is calculated by a cross-validation unit in the space-time correlation model. When the time offset is less than a preset threshold, a spatial correlation unit is triggered to compare the physical location corresponding to the mechanical damage characteristic with the spatial coordinates of the current and voltage anomaly area to generate a spatial coincidence degree.

[0025] According to the time offset and spatial overlap, correlation weight coefficients of current distortion characteristics, voltage abnormality characteristics and mechanical damage characteristics are determined, and the time offset, spatial overlap and correlation weight coefficients are weighted and summed to generate a composite feature vector.

[0026] Optionally, based on the current distortion features and voltage anomaly features in the space-time matrix, a time offset between the two within a preset time window is calculated by a cross-validation unit in the space-time correlation model. When the time offset is less than a preset threshold, a spatial correlation unit is triggered to compare the physical location corresponding to the mechanical damage feature with the spatial coordinates of the current and voltage anomaly area to generate a spatial coincidence, including:

[0027] Extracting a current distortion feature sequence and a voltage anomaly feature sequence from the space-time matrix, segmenting the current distortion feature sequence and the voltage anomaly feature sequence according to a preset time window sliding rule to form a current distortion feature segment and a voltage anomaly feature segment containing feature values ​​of a plurality of consecutive sampling points;

[0028] Performing point-by-point difference accumulation calculation on the current distortion characteristic segment and the voltage anomaly characteristic segment within the same time window, and dynamically adjusting the starting position of the voltage anomaly characteristic segment to minimize the phase difference between the two, thereby generating a time offset within the current time window;

[0029] determining whether the time offset is less than a preset threshold, and when the time offset is less than the preset threshold, sending a trigger instruction to the spatial association unit, and retrieving a physical position coordinate set corresponding to the mechanical damage feature and a spatial coordinate set of the current and voltage abnormal area through the trigger instruction;

[0030] Based on the physical position coordinate set and the spatial coordinate set of the current and voltage abnormal areas, with the minimum coverage area of ​​the mechanical damage feature as the benchmark, the coverage areas of all current and voltage abnormal areas are traversed, and the overlapping lengths of the two in the horizontal and vertical ranges are counted. The spatial overlap is generated according to the ratio of the product of the horizontal and vertical overlapping lengths to the coverage area of ​​the mechanical damage feature.

[0031] Optionally, dynamically adjusting the waveform capture window to capture transient waveform segments based on the correlation between the mechanical vibration intensity and the current transient rate in the anti-interference data set includes:

[0032] A vibration intensity detection module based on the three-dimensional vibration signal in the anti-interference data set calculates the root mean square value of the vibration amplitude and establishes a correlation relationship table between the vibration intensity and the current transient rate;

[0033] When it is detected that the current transient rate exceeds the rate threshold corresponding to the current vibration intensity interval in the association table, a slope mutation point or a local extreme point within a preset time range is selected from the current current waveform as the starting position of the waveform interception window, wherein the starting position is preferably a trough point;

[0034] Dynamically adjusting the length of the waveform capture window according to the ratio of the root mean square value to the current transient rate, wherein the window length is expanded when the ratio increases and contracted when the ratio decreases, and the window length is limited between a preset upper limit and a preset lower limit;

[0035] The amplitude mutation amplitude and spectral energy distribution of the current waveform in the adjusted window are analyzed. If the value of the amplitude mutation amplitude exceeds a preset threshold and the spectral energy is concentrated in the high frequency band, the current waveform is captured as a transient waveform segment.

[0036] Optionally, extracting a peak slope variation from a transient waveform segment of the current waveform, separating the harmonic frequency band energy of the short-circuit spike from the transient waveform segment of the voltage waveform, and simultaneously performing time-frequency analysis on the three-dimensional vibration signal to generate a mechanical damage spectrum feature set includes:

[0037] Extracting, from the captured transient waveform segment of the current waveform, a ratio of an amplitude increment to a time increment between adjacent peaks as a peak slope variation;

[0038] Performing frequency domain decomposition on the captured transient waveform segment of the voltage waveform, identifying a frequency band having a frequency three times higher than the fundamental frequency and having concentrated energy distribution, and calculating the energy integral of each sub-band within the selected frequency band as harmonic frequency band energy;

[0039] The three-dimensional vibration signal is subjected to time-frequency analysis using a multi-resolution decomposition method. The frequency point set with a sudden increase in vibration energy in each decomposition frequency band is extracted. The frequency position and energy increase of the frequency point set are recorded to generate a mechanical damage spectrum feature set consisting of frequency point distribution and energy change.

[0040] Optionally, a multi-dimensional sensor group is used to synchronously collect the current waveform, voltage waveform, and three-dimensional vibration signal of the cable. Through adaptive noise reduction and time-series synchronous transmission in a humid and dusty environment, an anti-interference data set is generated, including:

[0041] The sensor group consisting of current sensor, voltage sensor and three-dimensional vibration sensor collects the current waveform, voltage waveform and three-dimensional vibration signal of the cable in a synchronous triggering manner;

[0042] Dynamically selecting filtering parameters through a signal strength adaptive module based on the signal-to-noise ratio of the current waveform and the voltage waveform, and performing high-frequency noise suppression on the current waveform and the voltage waveform based on the filtering parameters, retaining only transient components in the signal whose amplitude fluctuation exceeds a preset threshold;

[0043] When transmitting the current waveform, voltage waveform, and three-dimensional vibration signal, the time stamps of the three axial components of the three-dimensional vibration signal, the current waveform, and the voltage waveform are aligned through a timing synchronization transmission module, and the signal transmission interval is dynamically adjusted according to the dust concentration monitoring result. When the dust concentration is higher than a preset critical value, the signal transmission interval is shortened to reduce signal loss caused by dust interference;

[0044] The current waveform, voltage waveform and three-dimensional vibration signal that have completed timing synchronization are superimposed and analyzed, and the intervals where the amplitudes of the three suddenly change within the same period are extracted to generate a transient signal set with strictly aligned timestamps as an anti-interference data set.

[0045] In a second aspect, the present application provides a mining circuit fault self-diagnosis system, comprising:

[0046] The acquisition module uses a multi-dimensional sensor group to synchronously collect the current waveform, voltage waveform and three-dimensional vibration signal of the cable, and generates an anti-interference data set through adaptive noise reduction and time-series synchronous transmission in a humid and dusty environment;

[0047] an extraction module that dynamically adjusts a waveform capture window to capture transient waveform segments based on the correlation between mechanical vibration intensity and current transient rate in the anti-interference data set, extracts a peak slope variation from the transient waveform segments of the current waveform, separates harmonic frequency band energy of the short-circuit spike from the transient waveform segments of the voltage waveform, and simultaneously performs time-frequency analysis on the three-dimensional vibration signal to generate a mechanical damage spectrum feature set;

[0048] A correlation module inputs the peak slope variation, harmonic frequency band energy, and mechanical damage spectrum feature set into a spatiotemporal correlation model, generates a composite feature vector through spatiotemporal correlation verification of current distortion features and voltage anomaly features;

[0049] A correction module dynamically corrects the threshold value in the fault feature database based on the environmental interference factor in the composite feature vector and in combination with the cable surface temperature and dust data;

[0050] The matching module performs a multi-dimensional matching calculation on the composite feature vector and the corrected fault mode threshold, and determines the fault diagnosis result according to the correlation strength between the electrical feature and the mechanical feature in the matching result.

[0051] In an embodiment of the present application, a multi-dimensional sensor group synchronously collects the current waveform, voltage waveform, and three-dimensional vibration signal of the cable, and generates an anti-interference data set through adaptive noise reduction and time-series synchronous transmission in a humid and dusty environment. Based on the correlation between the mechanical vibration intensity and the current transient rate in the anti-interference data set, the waveform clipping window is dynamically adjusted to capture transient waveform segments, the peak slope change is extracted from the transient waveform segments of the current waveform, and the harmonic frequency band energy of the short-circuit spike is separated from the transient waveform segments of the voltage waveform. The three-dimensional vibration signal is simultaneously subjected to time-frequency analysis to generate a mechanical damage spectrum feature set. The peak slope change, harmonic frequency band energy, and mechanical damage spectrum feature set are input into a spatiotemporal correlation model, and a composite feature vector is generated through spatiotemporal correlation verification of the current distortion feature and the voltage anomaly feature. Based on the environmental interference factor in the composite feature vector and combined with the cable surface temperature and dust data, the threshold in the fault feature database is dynamically corrected. The composite feature vector is matched with the corrected fault mode threshold in a multi-dimensional manner, and the fault diagnosis result is determined based on the correlation strength between the electrical feature and the mechanical feature in the matching result.

[0052] This application has the following beneficial effects:

[0053] Through the collaborative collection of current, voltage and vibration signals by multi-dimensional sensors, combined with adaptive noise reduction and synchronous transmission technology, the interference of the underground environment can be effectively overcome to ensure data integrity and timing consistency; based on the dynamic correlation between mechanical vibration and current transient rate, the waveform is adaptively intercepted to accurately capture transient fault fragments; the current slope change, voltage harmonic energy and vibration time-frequency characteristics are extracted to construct a multi-dimensional fault feature set to improve the identification ability of cable insulation degradation, mechanical damage and short-circuit faults; the spatiotemporal synchronization of current, voltage and vibration characteristics is integrated using a spatiotemporal correlation model to generate a composite feature vector to avoid single signal misjudgment; the fault threshold is dynamically corrected in combination with environmental parameters to enhance the diagnostic robustness under complex working conditions; through multi-dimensional feature matching and correlation strength analysis, accurate diagnosis of high-frequency transient faults such as short-circuit spikes and open circuit oscillations is achieved, significantly reducing the false alarm rate.

[0054] Furthermore, by decomposing the composite feature vector into electrical feature sub-vectors and mechanical feature sub-vectors, the matching degree is calculated with the corrected fault mode threshold respectively. After the electrical matching degree and mechanical matching degree are generated, the correlation strength value of the two is calculated and the fault type judgment is triggered based on the preset fusion threshold. The deviation between the electrical matching degree and the short-circuit spike threshold, and the mechanical matching degree and the open circuit oscillation threshold are compared to determine the diagnosis result, thereby effectively distinguishing the fault modes of short-circuit spike and open circuit oscillation, improving the diagnostic resolution of composite faults, enhancing the diagnostic reliability through the dynamic correlation strength fusion and deviation comparison mechanism, and reducing the risk of false alarms caused by environmental interference.

[0055] These and other aspects of the present application will become more readily apparent from the description of the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0057] Figure 1 A flow chart of a method for self-diagnosis of mining circuit faults provided by the present application is shown;

[0058] Figure 2 A scenario diagram of a mining circuit fault self-diagnosis method provided by the present application is shown;

[0059] Figure 3 The schematic diagram shows the structure of a mining circuit fault self-diagnosis system provided by the present application. DETAILED DESCRIPTION

[0060] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0061] In some of the processes described in the specification and claims of this application and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this document or may be executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any order of execution. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this document are used to distinguish different messages, devices, modules, etc., and do not represent a sequential order, nor do they limit "first" and "second" to being different types.

[0062] Researchers have discovered significant technical bottlenecks in existing mining circuit fault diagnosis methods: Reliance on a single signal prevents effective differentiation between short-circuit spikes and electromagnetic interference signals, resulting in a high false alarm rate; static feature analysis relies solely on traveling wave amplitude and propagation time parameters, making it difficult to capture subtle characteristics of early insulation degradation; and insufficient dynamic adaptability results in thresholds that fail to incorporate environmental parameter corrections, making misjudgments prone to load fluctuations and response delays exceeding transient fault blocking requirements. Traditional methods struggle to accurately extract and dynamically verify fault features, particularly in the dusty and humid environments of underground coal mines. Therefore, an intelligent diagnostic method combining multimodal perception and dynamic correlation verification is urgently needed.

[0063] In response to the above problems, the present invention proposes a self-diagnosis method for mining circuit faults, the core of which is to construct a multi-dimensional spatiotemporal correlation model of sensor data and a dynamic threshold correction mechanism. Specifically, the current waveform, voltage waveform and three-dimensional vibration signal are collected synchronously by multiple sensors, and an anti-interference data set is generated in combination with an adaptive noise reduction algorithm; the waveform capture window is adjusted based on the dynamic correlation between mechanical vibration intensity and current transient rate, and the peak slope change, harmonic frequency band energy and time-frequency mechanical damage characteristics are extracted synchronously; the coupling relationship between electrical characteristics and mechanical characteristics is verified through a spatiotemporal correlation model, and the threshold database is dynamically corrected in combination with environmental parameters. This method significantly reduces the false alarm rate through multi-dimensional signal fusion, enhances the detection capability of early insulation degradation by using time-frequency dynamic analysis technology, and improves the response speed under complex working conditions through an adaptive threshold correction mechanism of environmental parameters, fundamentally solving the technical defects of traditional methods such as single feature dimension, static analysis limitations and insufficient environmental adaptability.

[0064] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without making creative efforts are within the scope of protection of this application.

[0065] Figure 1 A flowchart of a method for self-diagnosis of mining circuit faults is provided in the embodiment of the present application. Figure 1 As shown, the method includes:

[0066] 101. The current waveform, voltage waveform and three-dimensional vibration signal of the cable are synchronously collected by a multi-dimensional sensor group, and an anti-interference data set is generated through adaptive noise reduction and time-series synchronous transmission in a humid and dusty environment;

[0067] Optionally, step 101 may specifically include the following steps:

[0068] 1011. Using a sensor group consisting of a current sensor, a voltage sensor, and a three-dimensional vibration sensor, collect the current waveform, voltage waveform, and three-dimensional vibration signal of the cable in a synchronous triggering manner;

[0069] 1012. Dynamically select filtering parameters through a signal strength adaptive module based on the signal-to-noise ratio of the current waveform and the voltage waveform, and perform high-frequency noise suppression on the current waveform and the voltage waveform based on the filtering parameters, retaining only transient components in the signal whose amplitude fluctuation exceeds a preset threshold.

[0070] 1013. When transmitting the current waveform, voltage waveform, and three-dimensional vibration signal, align the timestamps of the three axial components of the three-dimensional vibration signal, the current waveform, and the voltage waveform through a timing synchronization transmission module, and dynamically adjust the signal transmission interval according to the dust concentration monitoring result. When the dust concentration is higher than a preset critical value, shorten the signal transmission interval to reduce signal loss caused by dust interference;

[0071] 1014. Perform superposition analysis on the current waveform, voltage waveform, and three-dimensional vibration signal that have completed timing synchronization, extract intervals where the amplitudes of the three waveforms suddenly change within the same time period, and generate a transient signal set with strictly aligned timestamps as an anti-interference data set.

[0072] In the above steps, the multi-dimensional sensor group refers to a combination of devices consisting of current sensors, voltage sensors, and three-dimensional vibration sensors. Adaptive noise reduction refers to a technology that dynamically adjusts filter parameters based on the signal-to-noise ratio of the current waveform and voltage waveform to suppress high-frequency noise. The time-series synchronization transmission module refers to a device that aligns the timestamps of the three axial components, current waveform, and voltage waveform of the three-dimensional vibration signal. The anti-interference data set refers to a set of transient signals generated by extracting the intervals of sudden amplitude changes of the current waveform, voltage waveform, and three-dimensional vibration signal within the same time period through superposition analysis. The transient component refers to the short-term variation in the signal where the amplitude fluctuation exceeds a preset threshold. The dust concentration monitoring result refers to the dynamic data of the ambient dust content obtained by the dust sensor.

[0073] In this embodiment, a combination of sensors (current sensor, voltage sensor, and 3D vibration sensor) first synchronously collects key cable signals. Specifically, synchronized signal acquisition is achieved through step 1011. The current sensor uses the Hall effect principle to collect the cable current waveform, the voltage sensor uses a voltage divider circuit to obtain the voltage waveform, and the 3D vibration sensor uses piezoelectric ceramic elements to synchronously measure vibration acceleration in the X, Y, and Z axes. The three sets of sensors are synchronously activated via a hardware trigger circuit. A unified trigger pulse signal is sent by the main control unit to ensure that the time starting points of the current waveform, voltage waveform, and 3D vibration signal are aligned, eliminating timing deviations caused by sensor response delays. To ensure that these different signals are fully aligned in time (to avoid data misalignment due to varying sensor response speeds), they do not start independently. Instead, a unified "start" electrical pulse signal is sent by the main controller to trigger all three sensors to begin data acquisition simultaneously, much like pressing the start button on multiple stopwatches simultaneously, ensuring that all signals have the same starting point.

[0074] Since the collected raw current and voltage waveforms are inevitably mixed with various high-frequency interference noises, adaptive noise reduction processing is performed in step 1012. The signal strength adaptive module (implemented by a software program embedded in the main control unit, the module runs the code logic in the adaptive filtering algorithm library in real time through the processor. Its core function is to automatically adjust subsequent filtering parameters based on the real-time calculated signal-to-noise ratio (SNR) of the signal quality indicator) uses a sliding window method to calculate the SNR of the current and voltage waveforms, updating the SNR value every 0.1 seconds. When the SNR is detected to be below 20dB, a Butterworth low-pass filter is automatically selected with a cutoff frequency of 500Hz. When the SNR is above 40dB, a Chebyshev filter is switched to with a cutoff frequency increased to 2kHz. Based on the dynamically selected filter parameters, the raw signal is digitally filtered, retaining only transient components with amplitude fluctuations exceeding 2% of the baseline value while eliminating power frequency interference and random high-frequency noise. In other words, the system automatically adjusts the noise reduction intensity based on the signal quality (SNR). The system continuously (every 0.1 seconds) calculates the signal-to-noise ratio (a measure of the ratio of useful to noisy components of a signal) of a short segment of the current or voltage waveform. A higher signal-to-noise ratio indicates better signal quality; a lower signal-to-noise ratio indicates greater noise pollution. The system has preset thresholds for determining the signal-to-noise ratio: when the calculated signal-to-noise ratio is below 20 decibels (dB, a unit of sound or signal strength; lower values ​​indicate greater noise), indicating excessive noise, the system automatically selects a Butterworth low-pass filter and sets it to allow only signal components with frequencies below 500 Hz to pass (retaining lower-frequency useful signals and filtering out high-frequency noise). Conversely, when the signal-to-noise ratio is above 40 decibels, indicating better signal quality, the system switches to a Chebyshev low-pass filter and raises the upper limit of the allowed signal frequency to 2000 Hz (2kHz), preserving more useful high-frequency details. Regardless of the filter and settings used, the ultimate goal is to eliminate those subtle, useless high-frequency fluctuations (power frequency interference and random noise) and retain only those short-term abnormal signals (transient components) with significant amplitude changes (fluctuations exceeding 2% of the normal baseline value).

[0075] Next, the processed current, voltage, and three-dimensional vibration signals (including data in the X, Y, and Z directions) need to be transmitted to the data processing center. In a humid and dusty environment, transmitted signals are susceptible to interference or even loss. To address this issue and ensure precise time synchronization of all signals, the system utilizes time-series synchronous transmission technology. Specifically, time-series synchronous transmission is implemented in step 1013. The time-series synchronous transmission module uses the Network Time Protocol to add nanosecond-level timestamps to the X / Y / Z components of the three-dimensional vibration signal, as well as the filtered current and voltage waveforms. An abnormal timestamp detection algorithm identifies and corrects time deviations between signal channels, ensuring that the time alignment error of multiple signals is less than 1ms. The dust concentration monitoring unit collects dust data using light scattering. When the dust concentration exceeds 200 mg / m³, the transmission interval adjustment algorithm is triggered, dynamically shortening the signal transmission interval from the default 100ms to 50ms. This increases the frequency of data packet transmission and reduces signal loss caused by dust. In other words, the system uses a high-precision time protocol (such as a simplified version of the Network Time Protocol (NTP) or hardware timestamps) to assign a nanosecond (billionth of a second) time stamp (timestamp) to each collected data point (including current, voltage, and X / Y / Z vibration values). The system also runs an algorithm to check that the time stamps on different signal channels are strictly aligned. If a signal's time stamp is found to have a slight deviation (for example, 0.5 milliseconds late), the algorithm automatically corrects it to ensure that the time error between all signals is less than 1 millisecond. Simultaneously, the dust concentration in the environment is continuously monitored by a dust sensor (typically by emitting a beam of light and measuring the intensity of the light scattered back by dust particles to calculate the concentration). This dust concentration data is used to dynamically adjust the speed of signal transmission: when the dust concentration is normal (less than 200 milligrams per cubic meter - mg / m³), the system defaults to sending data in packages every 100 milliseconds (ms); when the dust sensor detects a sharp increase in concentration exceeding 200 mg / m³ (the preset critical value), the system will immediately start adjusting and shorten the interval between data package transmission to once every 50 milliseconds.

[0076] Finally, all current, voltage, and X / Y / Z 3D vibration signals, which have undergone precise time alignment (consistent timestamps) and noise reduction, are combined for comprehensive analysis, generating an anti-interference dataset in step 1014. Time-domain superposition analysis is performed on the synchronized current, voltage, and 3D vibration waveforms, using a sliding window algorithm with a 10ms step size to traverse all signal data. A threshold comparison module detects amplitude mutations exceeding 30% of the baseline value within the same time window for all three signals. Wavelet transforms are then used to extract the time-frequency characteristics of each signal segment. The selected transient signal segments are reassembled according to nanosecond timestamps to form a multidimensional anti-interference dataset containing waveform amplitude, vibration vector, and ambient dust concentration, providing high-precision input for subsequent fault diagnosis. Specifically, the system sets a 10ms "time window" and, starting from the signal's inception point, slides through the entire time period in 10ms steps. Within each 10ms window, the system simultaneously examines the values ​​of the current, voltage, and 3D vibration signals (typically, the composite vibration intensity in three directions). Only when the amplitudes of all three signals within this small window show a significant and abnormal increase simultaneously (the amplitude change exceeds 30% of their respective normal baseline values) does the system deem this a "transient event" worthy of attention, caused by a potential problem (such as a short circuit or mechanical shock). For example, at a certain point in time, if the current suddenly soars from 100A to 150A (exceeding the baseline of 100A by 30%, or 130A), while the voltage drops sharply from 220V to 150V (below the baseline of 220V by 30%, or 154V), and the vibration intensity suddenly increases by more than 30% of the normal level, and all three occur simultaneously within a 10-millisecond window, they will be captured. For these captured transient event intervals, the system uses wavelet transform technology (a tool that simultaneously analyzes the time and frequency characteristics of a signal) to extract more detailed characteristic information of the signal waveform. Ultimately, the system reassembles and packages all the selected transient event fragments, along with their precise nanosecond time tags, corresponding waveform amplitudes, three-dimensional vibration direction data, and the dust concentration values ​​at the time, to form a multi-dimensional, time-aligned, and interference-resistant dataset that highlights abnormal transient events. This high-quality dataset provides a reliable foundation for subsequent accurate determination of cable faults and their nature.

[0077] In practical applications, for example, a coal mine's main haul tunnel cable intelligent monitoring project deployed an acquisition system consisting of HX-10C high-frequency current sensors, PT-8K voltage sensors, and a triaxial MEMS vibration array. This system uses a GPS-synchronized clock trigger mechanism to achieve 5ms-level time synchronization. The system continuously samples the cable's three-phase current, ground voltage, and X / Y / Z axial vibration signals at a 50kHz sampling rate. When the signal-to-noise ratio (SNR) of the phase A current drops to 25dB, the adaptive module switches to a sixth-order Chebyshev filter with a passband cutoff frequency of 12kHz and a stopband attenuation of 60dB. This filter removes high-frequency harmonics generated by the start and stop of the tunnel boring machine, retaining only transient current fluctuations lasting more than 3ms and with an amplitude exceeding 20% ​​of the rated value. When the environmental monitoring unit reports tunnel humidity reaching 95% RH, the timing synchronization module activates a redundant check mechanism, improving the time alignment accuracy of the triaxial vibration signal and the electrical signal to ±0.2ms. Furthermore, by adding CRC check bits, the packet retransmission rate under dust interference is reduced to one-third the industry standard. The data analysis platform performed correlation mining on 3.6TB of raw data sampled continuously for 8 hours, identified the spatiotemporal coupling events of 132nd current harmonic distortion and Z-axis 5-8kHz vibration spectrum mutation, and constructed an anti-interference database containing millisecond-level fault feature fragments, providing high-confidence samples for subsequent insulation layer damage prediction.

[0078] In the overall solution of the above step 101, a multi-dimensional sensor group is used to collaboratively collect the current and voltage waveforms and three-dimensional vibration signals of the cable during operation, and a signal strength adaptive module is used to dynamically match the filter parameters to achieve high-frequency noise suppression. The time stamps of multi-source signals are accurately aligned in combination with the timing synchronous transmission mechanism, and the transmission interval is dynamically compressed based on the dust concentration to reduce the signal loss rate. The synchronous transient feature interval is extracted using multi-modal signal superposition analysis to construct a full-time domain fault feature data set with strong anti-interference ability, effectively overcoming the dust and moisture interference and electromagnetic noise coupling problems under complex working conditions in mines, and significantly improving the identification accuracy and integrity of cable fault features. At the same time, the adaptive transmission strategy ensures the accuracy and stability of signal acquisition, providing highly reliable data support for early warning and precise positioning of faults in underground circuit systems, greatly reducing the risk of misjudgment and missed detection due to environmental interference, and enhancing the active safety protection capability and operation and maintenance efficiency of the mine power supply system.

[0079] 102. Based on the correlation between the mechanical vibration intensity and the current transient rate in the anti-interference data set, dynamically adjust the waveform capture window to capture the transient waveform segment, extract the peak slope change from the transient waveform segment of the current waveform, separate the harmonic frequency band energy of the short-circuit spike from the transient waveform segment of the voltage waveform, and simultaneously perform time-frequency analysis on the three-dimensional vibration signal to generate a mechanical damage spectrum feature set;

[0080] Optionally, step 102 may include the following specific steps:

[0081] 1021. A vibration intensity detection module based on the three-dimensional vibration signal in the anti-interference data set calculates a root mean square value of the vibration amplitude and establishes a correlation relationship table between vibration intensity and current transient rate;

[0082] 1022. When it is detected that the current transient rate exceeds the rate threshold corresponding to the current vibration intensity interval in the association table, a slope mutation point or a local extreme point within a preset time range is selected from the current current waveform as the starting position of the waveform interception window, wherein the starting position is preferably a trough point;

[0083] 1023. Dynamically adjust the length of the waveform capture window according to the ratio of the root mean square value to the current transient rate, wherein the window length is expanded when the ratio increases and is contracted when the ratio decreases, and the window length is limited between a preset upper limit and a preset lower limit;

[0084] 1024. Analyze the amplitude mutation amplitude and spectrum energy distribution of the current waveform in the adjusted window. If the amplitude mutation amplitude exceeds a preset threshold and the spectrum energy is concentrated in the high frequency band, capture the current waveform as a transient waveform segment.

[0085] 1025. Extracting, from the captured transient waveform segment of the current waveform, a ratio of an amplitude increment to a time increment between adjacent peaks as a peak slope variation;

[0086] 1026. Perform frequency domain decomposition on the captured transient waveform segment of the voltage waveform, identify a frequency band having a frequency three times higher than the fundamental frequency and having concentrated energy distribution, and calculate the energy integral of each sub-band in the selected frequency band as harmonic frequency band energy;

[0087] 1027. Perform time-frequency analysis on the three-dimensional vibration signal using a multi-resolution decomposition method, extract a set of frequency points where vibration energy suddenly increases within each decomposition frequency band, record the frequency position and energy amplification of the frequency point set, and generate a mechanical damage spectrum feature set consisting of frequency point distribution and energy change.

[0088] In the above steps, mechanical vibration intensity refers to the vibration energy level quantified by the root mean square value of the three-dimensional vibration signal. Current transient rate refers to the number of sudden changes in the current waveform amplitude exceeding a preset threshold per unit time. The vibration intensity detection module is the processing unit that calculates the root mean square value of the three-dimensional vibration signal. The association table is a two-dimensional data table that stores the current transient rate thresholds corresponding to different vibration intensity intervals. The waveform capture window is the time range defined according to dynamic rules for capturing transient waveforms. A trough point is the location of the lowest amplitude between adjacent peaks in the current waveform. A local extreme point is a sampling point where the amplitude reaches a local maximum or minimum in the waveform. The peak slope change is the ratio of the amplitude increment to the time increment between adjacent peaks in the current transient segment. Frequency domain decomposition is the process of converting a time domain signal into a frequency domain energy distribution. Harmonic frequency band energy is the integrated energy value of frequency bands with frequencies that are integer multiples of the fundamental frequency in the voltage transient segment. Multi-resolution decomposition is an analysis technique that decomposes a signal into different time-frequency resolution levels using wavelet packet transform. The energy surge frequency point refers to the frequency domain location where the energy increase exceeds 40% at a specific frequency of the vibration signal. The mechanical damage spectrum feature set refers to the set matrix containing the damage characteristic frequencies and their energy increase parameters.

[0089] In the present embodiment, a correlation pattern was discovered between the mechanical vibration intensity of the cable (representing the intensity of the physical shaking) and the transient change rate of the current signal (representing the frequency of abnormal current changes in a short period of time). This pattern is formed into a correlation table through data accumulation. The system uses this pattern to intelligently "focus" observations and dynamically adjust the analysis time window (i.e., which small time period to focus on) to more accurately capture transient signal segments that reflect problems. First, a correlation table is constructed in step 1021. The vibration intensity detection module uses a moving root mean square (RMS) algorithm to calculate the RMS value of the composite vector of the three-dimensional vibration signal (X / Y / Z) using a 100ms window. The transient rate is simultaneously counted as the number of sudden changes in the current waveform amplitude exceeding 15% of the baseline value every 5 seconds. The RMS vibration intensity value is divided into 20 intensity intervals at 0.5g intervals. The maximum transient rate threshold allowed in each interval is recorded, forming a correlation table with the intensity interval number and the corresponding speed limit value. In other words, the system calculates the combined vibration intensity of the three-dimensional vibration signal (X / Y / Z directions) using the root mean square (RMS) mathematical method. Furthermore, by analyzing extensive historical data, the system divides vibration intensity into 20 levels (intensity intervals) at intervals of 0.5g (unit of gravity acceleration). Each vibration intensity level is assigned a maximum allowable current transient rate threshold (for example, when the vibration intensity is between 1.0g and 1.5g, a current transient rate exceeding 3 per second may indicate a problem). This creates a correlation table.

[0090] Secondly, the system monitors the current transient rate in real time to see if it exceeds the safety threshold corresponding to the current vibration intensity. Once the current transient rate exceeds the safety threshold (for example, the actual rate reaches 3.5 times per second, exceeding the 2.8 times per second corresponding to 1.2g), the system deems it likely that an abnormal event requiring special attention has occurred. Step 1022 determines the window start position. When the current transient rate exceeds the associated threshold corresponding to the current vibration intensity, extreme value detection is performed on the current waveform. The sliding difference method is used to calculate the first-order derivative between adjacent sampling points. A trough is marked when the derivative sign changes from negative to positive. If no trough is found within 10 consecutive sampling cycles, a local extreme value detection algorithm is used to find the sampling point with the largest amplitude change rate as the window start position, ensuring that the window captures the onset of the sudden change. In other words, to capture the key signal segment (the transient waveform segment) of this event, the system needs to determine at which point in the current waveform to begin "focused observation." The system prioritizes finding a "trough" (the lowest point between two adjacent peaks) on the current waveform, which is typically the starting point of an abnormal event (much like the small dip before the arrival of an earthquake wave). The system identifies current troughs by calculating the rate of change (slope) between adjacent points on the current waveform. The turning point where the rate of change suddenly changes from negative (indicating a decrease) to positive (indicating the beginning of an increase) is marked as a trough. If no such trough is found within 10 consecutive sampling points (representing a very short period of time, such as 0.1 milliseconds), the system will then look for the point on the current waveform with the most dramatic change (the largest absolute value of the rate of change) as the starting point for observation.

[0091] After determining the observation starting point, the system needs to determine the duration of this "focused observation" time window. The window length is adjusted in step 1023. Based on the ratio of the vibration RMS value to the current transient rate, a linear interpolation algorithm is used to dynamically adjust the window length within a preset range of 50-300ms. The window length is set to extend by 10ms for every 0.1 increase in the ratio. When the dust concentration exceeds 200mg / m³, a length compensation mechanism is activated to extend the window by an additional 5ms. The upper limit of the window length is also set to not exceed the maximum duration of the fault signature, 300ms, and the lower limit is set to not be less than the minimum valid signature duration, 50ms. In other words, the window length is not fixed but dynamically adjusted based on the ratio of the current RMS value to the current transient rate. This ratio can be understood as the "vibration intensity per unit current sudden change rate." If this ratio increases (indicating strong vibration but relatively few current sudden changes, possibly indicating a problem primarily related to mechanical shock), the system will appropriately extend the observation window (by 10ms for each 0.1 increase in the ratio) to provide a more comprehensive view and capture potentially longer vibration events. Conversely, if the ratio decreases (meaning frequent current fluctuations but relatively weak oscillations, potentially indicating an electrical problem), the system shortens the observation window to focus more closely on brief current changes. Furthermore, given the actual duration of fault signatures, the window length is strictly limited to between 50 milliseconds (minimum) and 300 milliseconds (maximum). Furthermore, if ambient dust concentrations are high (over 200 mg / m³), the system extends the window by an additional 5 milliseconds to compensate for potential loss of signal detail due to dust interference.

[0092] Then, within this dynamically adjusted time window, the system carefully examines the current waveform to determine whether it contains a valuable transient event segment. The transient segment is captured in step 1024. A fast Fourier transform is performed on the current waveform within the window to calculate the energy proportion in the high-frequency band between 500Hz and 2000Hz. The maximum amplitude change within the window is simultaneously detected. When the high-frequency energy proportion exceeds 60% and the amplitude change reaches 25% of the baseline value, the dual-threshold comparator is triggered to output a capture command. Current and voltage waveforms within the time window that meet these criteria are marked as transient segments, and their start timestamps are recorded. Specifically, the system primarily examines two criteria: first, whether the maximum instantaneous current change within the window is sufficiently large (exceeding 25% of the baseline value); and second, whether the current signal within this window contains a significant amount of high-frequency components (whether the energy in the 500Hz to 2000Hz range accounts for more than 60% of the total signal energy). Only when both conditions are met does the system confirm that the current and voltage signals captured within the window are valid transient waveform segments. For example, if the normal current baseline is 100A, and the current in the window suddenly surges to over 125A (more than 25%), and the energy of the high-frequency hissing sound in the window current signal accounts for a high proportion (>60%), then the capture is successful.

[0093] The system then extracts a key feature from the captured current transient segment: the peak slope change. The peak slope is calculated in step 1025. The captured current transient segment undergoes cubic spline interpolation, and the adaptive threshold method is used to identify adjacent peak locations. The amplitude difference and time difference between each pair of peaks in the first three complete peak cycles are calculated, and the median of these three values ​​is taken as the peak slope change. For asymmetric waveforms, an area compensation algorithm is used to eliminate measurement errors, retaining four decimal places. In other words, this describes how quickly the current rises from one peak to the next. The system first accurately locates the locations of several consecutive peaks within the segment (using smooth interpolation and an adaptive threshold method). It then calculates the height difference (amplitude increment) and time difference (time increment) between each pair of adjacent peaks in the first three complete peak cycles, and then calculates the ratio of these two increments (slope = height difference / time difference). The median of these three slope values ​​(the middle value after sorting) is taken as the final peak slope change, which can avoid the influence of individual outliers.

[0094] Next, the system focuses on the "short-circuit spike" characteristics of the captured voltage transient, particularly those high-frequency harmonic components. Harmonic energy is extracted in step 1026. Six-layer wavelet packet decomposition is performed on the voltage transient, reconstructing 32 sub-bands with frequencies three times or higher than the fundamental frequency (150-1000 Hz). A modified Singer function is used to calculate the energy integral of each sub-band within a 20ms time window, selecting characteristic frequency bands whose integral exceeds three times the background noise level. The energy integrals of the selected frequency bands are summed to form the harmonic frequency band energy characteristic value of the transient. In other words, the system decomposes the voltage signal into different frequency bands (frequency domain decomposition), focusing on frequency bands with frequencies at least three times higher than the fundamental frequency (typically 50Hz or 60Hz) (i.e., above 150Hz). Six-layer wavelet packet decomposition is performed, dividing the range from 150Hz to 1000Hz into 32 finer sub-bands. For each sub-band, the system calculates its energy integral within a 20ms time window. Then, the system only selects those characteristic sub-bands whose energy is more than three times higher than the usual background noise and whose energy is relatively concentrated. By adding up the energy integral values ​​of these selected sub-bands, the "harmonic frequency band energy" representing the short-circuit spike characteristics is obtained.

[0095] Finally, the system performs a more detailed time-frequency analysis of the three-dimensional vibration signal (X / Y / Z), simultaneously examining the temporal variation of vibration intensity and its distribution across different frequencies. The goal is to extract unique spectral "fingerprints" that characterize mechanical damage (such as insulation wear and structural looseness). A feature set is generated in step 1027. The three-dimensional vibration signal undergoes eight-layer wavelet packet decomposition, calculating the Hilbert-Huang marginal spectrum for each of the 24 subbands in each layer. A sliding energy detection method is used to identify frequencies within each subband where the energy increase exceeds 40%, recording their center frequency and percentage increase. All characteristic damage frequencies are grouped into 50Hz intervals to form a mechanical damage spectral feature set matrix containing frequency distribution, increase intensity, and duration. In other words, the system uses an eight-layer wavelet packet decomposition method to decompose the vibration signal into components with varying frequency resolutions. Within each decomposed subband, the system calculates the signal's marginal spectrum (a spectrum that reflects the energy intensity of each frequency point) and uses a sliding window method to detect specific frequency points (or bins) where the vibration energy suddenly and significantly increases (increases exceeding 40%) within a short period of time. The system records the specific frequency locations of these energy surges and the energy increase (percentage increase). Finally, all detected characteristic frequency points are grouped into 50Hz intervals (for example, all damage characteristic points between 50-100Hz, 100-150Hz, etc. are grouped separately), forming a detailed "Mechanical Damage Spectrum Feature Set." This feature set records the frequency points at which abnormal and severe vibrations were detected (possibly corresponding to specific types of mechanical damage) and the severity of these vibration anomalies.

[0096] In practical applications, for example, in a coal mine intelligent cable monitoring system, when a coal mining machine cuts gangue and causes cable abnormalities, the system tracks the vibration signal at a 2000Hz sampling rate and detects that the X-axis vibration root mean square value suddenly increases from the daily 1.2g to 3.8g, and the associated current transient rate exceeds 120 Threshold. At this time, the analysis module locks the deep valley point in the current waveform, such as the instantaneous 280A depression point under the 380A reference current, as the starting mark of the waveform interception window. According to the dynamic ratio of vibration intensity to current change rate (such as when 3.8g vibration corresponds to 0.85g / ( ), the window length is automatically extended from the default 20ms to 35ms to fully capture the current waveform segment containing three consecutive peak mutations. For the intercepted current transient segment, the system accurately calculates the characteristics of adjacent peaks: for example, the first and second peaks jump from 420A to 680A within 2ms, generating 130 The slope of the third wave suddenly decays to 510A in the next 3ms, forming -56.7 The negative rate of change of the voltage waveform fragments analyzed simultaneously was detected after a 2048-point FFT transform in the high-frequency region of 1500-4500Hz. The energy integrated value in the 2350Hz frequency band reached 17 times that of the standard operating condition, clearly indicating a short-circuit discharge characteristic. The three-dimensional vibration signal was decomposed using a five-layer wavelet packet. The third decomposition layer (560-1200Hz frequency band) captured an energy surge event: the vibration energy at 850Hz on the Z axis soared from the normal 5m / s² to 22m / s², accompanied by a 9-fold increase in the energy of the 720Hz component on the X axis. This formed a damage fingerprint that included frequency coordinates, duration, and energy gradient. These characteristic data were mapped in the three-dimensional tunnel model, helping engineers accurately locate the risk of mechanical damage to the insulation layer of the cable segment near the E15 support.

[0097] In the overall solution of the above step 102, the waveform interception window position and length are adaptively adjusted through the dynamic correlation model of mechanical vibration intensity and current transient rate, the trough point is preferentially selected as the starting position and the window boundary is elastically expanded or contracted according to the ratio, and the transient waveform fragments with amplitude mutation and high-frequency energy concentration are accurately captured. Combined with the extraction of peak slope change and the integral calculation of harmonic frequency band energy, the multi-resolution decomposition method is simultaneously used to perform time-frequency analysis on the vibration signal to locate the energy surge frequency point, and construct a composite fault characterization system that integrates electrical transient characteristics and mechanical damage spectrum. It effectively solves the transient feature omission problem caused by fixed window interception in traditional methods, strengthens the correlation mining ability of weak short-circuit spikes and vibration energy mutations, and significantly improves the coupled diagnosis accuracy of early mechanical damage and electrical faults of cables. At the same time, the interpretability of fault types is enhanced through the quantitative description of frequency distribution and energy amplification, providing multi-dimensional feature support for the accurate identification and fault tracing of hidden defects of cables under complex working conditions.

[0098] 103. Input the peak slope variation, harmonic frequency band energy, and mechanical damage spectrum feature set into a spatiotemporal correlation model, and generate a composite feature vector through spatiotemporal correlation verification of current distortion features and voltage anomaly features;

[0099] Optionally, step 103 may specifically include the following steps:

[0100] 1031. Input the peak slope variation, harmonic frequency band energy, and mechanical damage spectrum feature set into a feature fusion module, reorganize the data according to a timestamp alignment rule, and form a spatiotemporal matrix containing current distortion features, voltage anomaly features, and mechanical damage features;

[0101] 1032. Based on the current distortion features and voltage anomaly features in the space-time matrix, a time offset between the two within a preset time window is calculated by a cross-validation unit in the space-time correlation model. When the time offset is less than a preset threshold, a spatial correlation unit is triggered to compare the physical location corresponding to the mechanical damage feature with the spatial coordinates of the current and voltage anomaly region to generate a spatial coincidence.

[0102] Among them, step 1032 may specifically include the following processes: extracting the current distortion feature sequence and the voltage anomaly feature sequence from the time-space matrix, segmenting and intercepting the current distortion feature sequence and the voltage anomaly feature sequence according to the preset time window sliding rule to form a current distortion feature segment and a voltage anomaly feature segment containing characteristic values ​​of multiple consecutive sampling points; performing point-by-point difference accumulation calculation on the current distortion feature segment and the voltage anomaly feature segment in the same time window, and dynamically adjusting the starting position of the voltage anomaly feature segment to minimize the phase difference between the two, thereby generating a time offset in the current time window; judging the time Whether the offset is less than a preset threshold. When the time offset is less than the preset threshold, a trigger instruction is sent to the spatial association unit, and the physical position coordinate set corresponding to the mechanical damage feature and the spatial coordinate set of the current and voltage abnormal area are retrieved through the trigger instruction; based on the physical position coordinate set and the spatial coordinate set of the current and voltage abnormal area, with the minimum coverage area of ​​the mechanical damage feature as the benchmark, the coverage areas of all current and voltage abnormal areas are traversed, and the overlapping lengths of the two in the horizontal and vertical ranges are counted, and the spatial overlap is generated according to the ratio of the product of the horizontal and vertical overlapping lengths to the coverage area of ​​the mechanical damage feature.

[0103] 1033. Determine the correlation weight coefficients of the current distortion feature, the voltage anomaly feature, and the mechanical damage feature based on the time offset and the spatial overlap, perform weighted summation on the time offset, the spatial overlap, and the correlation weight coefficients, and generate a composite feature vector.

[0104] In the above steps, the spatiotemporal correlation model refers to an analytical model that integrates the spatiotemporal relationships between current distortion features, voltage anomaly features, and mechanical damage features. The feature fusion module is a processor that reorganizes multidimensional features according to timestamp alignment rules. The spatiotemporal matrix is ​​a three-dimensional data structure consisting of time, space, and feature dimensions. The cross-validation unit is a processing unit that verifies the temporal relationship between current and voltage features. The time offset is the absolute value of the time difference between the appearance of current distortion features and voltage anomaly features. The spatial correlation unit is an analytical module that verifies the spatial positional relationship between mechanical damage and electrical anomalies. Spatial overlap refers to the overlap ratio between the mechanical damage area and the electrical anomaly area in spatial coordinates. The correlation weight coefficient is a feature importance coefficient calculated based on the spatiotemporal matching degree. The composite feature vector is the multidimensional feature set obtained by weighted fusion of spatiotemporal features.

[0105] In this embodiment, the key features extracted in the previous steps (current peak slope change, voltage harmonic frequency band energy, and mechanical damage spectrum feature set) are comprehensively analyzed to verify their temporal and spatial correlations. Ultimately, these features are fused to generate a "composite feature vector" that more comprehensively reflects the cable's abnormal condition. Relying solely on the features of individual signals is insufficiently reliable. It is necessary to examine whether the electrical anomaly features of current and voltage occur nearly simultaneously in time (temporal correlation) and whether they coincide spatially with the location of mechanical damage detected by vibration (spatial correlation). This dual temporal and spatial verification significantly improves the accuracy of fault diagnosis. First, a spatiotemporal matrix is ​​constructed in step 1031. The feature fusion module uses a timestamp matching algorithm to align the peak slope change, harmonic frequency band energy, and mechanical damage spectrum feature set with millisecond-level timestamps. Feature data with missing timestamps is compensated using cubic spline interpolation, forming a three-dimensional matrix consisting of a time axis (0-1000ms), a spatial axis (cable coordinates X / Y / Z), and a feature axis (current / voltage / vibration features). Each data cell stores the feature value corresponding to a spatiotemporal point. First, the system inputs feature data from these three sources (peak slope change, harmonic frequency band energy, and mechanical damage spectrum signatures) into a feature fusion module. This module's core task is to reorganize and align this data according to precise timestamps (millisecond level), ensuring that current, voltage, and vibration features at the same time are grouped together. For rare time points where feature data is missing, the system uses mathematical interpolation methods (cubic spline interpolation) to accurately estimate and fill in the missing data. Ultimately, all of this time-aligned feature data, along with its corresponding spatial location information (e.g., the point on the cable where current and voltage anomalies occurred, and the location where mechanical damage was detected), is organized into a structured three-dimensional data table—a space-time matrix. The three dimensions of this matrix are: the time axis (e.g., from 0 milliseconds to 1000 milliseconds), the spatial axis (specific cable location coordinates, such as X / Y / Z or markers along the cable length), and the feature axis (containing the current distortion feature values, voltage anomaly feature values, and mechanical damage feature values).

[0106] Secondly, the system uses the constructed space-time matrix to verify the temporal synchronization of the current distortion characteristics and the voltage anomaly characteristics through the internal "space-time correlation model". The space-time correlation model refers to an analytical model that integrates the space-time relationship of current distortion characteristics, voltage anomaly characteristics and mechanical damage characteristics. The space-time correlation verification is performed through step 1032. The cross-validation unit performs a cross-correlation analysis on the current distortion feature sequence and the voltage anomaly feature sequence in the space-time matrix, and intercepts the feature segment with a sliding window step of 10ms. The phase alignment algorithm is used to dynamically adjust the starting position of the voltage feature segment, and the cumulative value of the difference between the two feature segments is minimized. Determine the optimal time offset. When the offset is detected to be less than 5ms, the spatial association unit activates the position matching algorithm. It spatially projects and compares the three-dimensional coordinates corresponding to the mechanical damage signature, located using the vibration signal propagation time difference, with the coordinates of the current and voltage anomaly areas located using the impedance method. The geometric mean of the horizontal overlap ratio (Lx / Ltotal) and the vertical overlap ratio (Ly / Ltotal) is calculated as the spatial overlap. Specifically, the "cross-validation unit" extracts the time-varying sequence of current distortion eigenvalues ​​(current distortion feature sequence) and the time-varying sequence of voltage anomaly eigenvalues ​​(voltage anomaly feature sequence) from the matrix. It then sets a sliding time window (for example, 100 milliseconds wide) and slides this window step by step along the time axis (for example, 10 milliseconds at a time). Within each window position, a small segment of the current and voltage feature sequences is extracted. The system then attempts to fine-tune the starting position of the voltage feature sequence within this window. The system calculates the absolute difference between the eigenvalues ​​of the current and (shifted) voltage sequences at each corresponding time point and accumulates these differences. The system searches for the offset (Δt) at the start of the voltage sequence that minimizes the cumulative difference D. This optimal offset Δt is the "time offset" at which the current distortion and voltage anomaly characteristics appear within the current window. When temporal synchronization is verified (Δt < a preset threshold), the spatial correlation unit is activated. The system checks whether the specific physical location of the detected mechanical damage (localized by analyzing the time difference between vibration signals propagating to different sensors, resulting in a set of three-dimensional coordinates) spatially overlaps with the area where the current and voltage anomalies are detected (localized by measuring cable impedance changes, for example, to locate the electrical anomaly, also resulting in a set of spatial coordinates). The system first obtains the coordinate set corresponding to the mechanical damage characteristics (representing the damage area) and the spatial coordinate set of the current and voltage anomaly areas (representing the electrical anomaly areas). Then, using the minimum coverage area of ​​the mechanical damage area (e.g., a rectangular area) as a reference, it traverses the coverage area of ​​all current and voltage anomaly areas. The system calculates the overlap length (Lx and Ly) between the mechanical damage area and each electrical anomaly area along the cable length (transverse) and cross-sectional (longitudinal) directions.

[0107] Finally, the system needs to comprehensively consider the degree of synchronization in time (time offset Δt) and the degree of spatial coincidence (spatial overlap S) to evaluate the overall correlation strength of the three features of current, voltage, and vibration in this event. Generate a composite feature vector through step 1033. Based on the numerical range of time offset and spatial overlap, a fuzzy inference system is used to determine the correlation weight coefficient: when the time offset is <3ms and the spatial overlap is >80%, the current feature weight is set to 0.5, voltage 0.3, and vibration 0.2; when the time offset is 3-5ms and the overlap is 60-80%, the weights are adjusted to 0.4, 0.3, and 0.3. Through the weighted summation formula , where α, β, and γ are weight coefficients. This generates a composite feature vector containing 12-dimensional feature parameters, with each eigenvalue retained to three significant decimal places. This correlation strength is converted into a "weight coefficient" when the three features are finally fused. A larger weight indicates a more important and reliable feature in the current event. The weight coefficients are determined based on a pre-set rule (a fuzzy inference system): if the time offset is very small (Δt < 3ms) and the spatial overlap is very high (S > 80%), the three are considered highly correlated. Using these weight coefficients, the system performs a weighted summation of the time offset Δt, the spatial overlap S, and the original three eigenvalues. This weighted summation results in a numerical combination containing information from multiple dimensions, called a "composite feature vector." The resulting composite feature vector is a 12-dimensional data set (containing 12 values) that combines the original eigenvalues ​​with verified spatiotemporal correlation information, providing a highly concentrated and reliable feature input for accurate cable fault diagnosis.

[0108] In practical applications, for example, in a coal mine underground intelligent diagnosis system, when an abnormality occurs in the E12 bracket section cable, the system captures the current peak slope change at a sampling frequency of 2000 times per second: within a 15ms time window, it detects three times that exceed 80 The positive slope mutation and 2- The voltage waveform experienced a steep negative dip, while high-frequency harmonics with energy levels 18 times greater than those under standard operating conditions appeared in the 2350Hz frequency band. The spatiotemporal correlation model aligned the current distortion and voltage anomaly signatures, including the timestamp "2023-09-15 14:23:35.782," with the 32m / s² energy surge detected by the vibration sensor at the 850Hz frequency on the X-axis. Phase calibration was performed using a sliding 50ms time window. The model found a 3.2ms delay between the voltage anomaly signature and the current distortion. Dynamic adjustment reduced the time offset to 0.8ms, below the preset 5ms threshold. The spatial correlation unit then retrieved the spatial coordinate data of the cables within a 3-meter radius of the E12 support. The 1.2-meter diameter mechanical damage area was compared to the 1.5-meter diameter area covered by the current and voltage anomaly signals. The model measured 0.9 meters of horizontal overlap and 1.1 meters of vertical overlap, resulting in a calculated spatial overlap of 83%. The system automatically assigns a time correlation weight of 0.4, a spatial weight of 0.5, and a feature intensity weight of 0.1, generating a composite feature vector containing 12-dimensional parameters such as slope mutation, harmonic energy integral, and spectral energy gradient. This vector is visually displayed through the red warning area of ​​the three-dimensional cable model, guiding inspection personnel to accurately locate the damaged point of the cable armor layer 1.8 meters east of the E12 bracket.

[0109] In the overall solution of the above step 103, cross-dimensional data fusion of current distortion characteristics, voltage anomaly characteristics and mechanical damage characteristics is performed through a spatiotemporal correlation model, and the time offset of the current and voltage anomaly characteristics is accurately calculated using a dynamic phase alignment algorithm. The multi-dimensional feature correlation weight coefficient is generated by combining the quantitative analysis of the overlapping areas of the spatial coordinate set. The weighted fusion mechanism is used to integrate time synchronization, spatial overlap and feature contribution into a composite feature vector, which effectively solves the misjudgment problem caused by the isolated analysis of electrical and mechanical characteristics in traditional methods. The isolated abnormal signals caused by local interference are eliminated through the spatiotemporal dual verification mechanism, and the credibility of the fault characteristics under the conditions of multi-physical field coupling is enhanced. At the same time, based on the dynamic weight allocation strategy, adaptive adjustment of the feature importance under different working conditions is realized, which significantly improves the characterization capability and diagnostic robustness of the composite fault characteristics, and provides a strong correlation evidence chain across spatiotemporal dimensions for the accurate identification of early hidden cable faults.

[0110] 104. Based on the environmental interference factor in the composite feature vector and in combination with the cable surface temperature and dust data, dynamically correct the threshold value in the fault feature database;

[0111] Optionally, step 104 may specifically include the following steps:

[0112] 1041. Extracting an environmental interference factor from the composite feature vector, where the environmental interference factor is calculated by multiplying a temperature value collected by a cable surface temperature sensor and a dust concentration monitoring value;

[0113] 1042. Selecting a correction coefficient table corresponding to a temperature compensation coefficient and a dust attenuation coefficient from a fault feature database according to the magnitude of the environmental interference factor;

[0114] 1043 linearly combines the original threshold value in the fault feature database with the temperature compensation coefficient and the dust attenuation coefficient in the correction coefficient table to generate a dynamically corrected threshold value, wherein the temperature compensation coefficient is in a piecewise linear relationship with the temperature value, and the dust attenuation coefficient is in an exponential attenuation relationship with the dust concentration value.

[0115] In the above steps, the environmental interference factor refers to the interference intensity index formed by the product of the cable surface temperature and the dust concentration. The temperature compensation coefficient refers to the linear correction parameter that adjusts the characteristic threshold according to temperature changes. The dust attenuation coefficient refers to the exponential correction parameter that reflects the impact of dust concentration on signal transmission quality. The correction coefficient table refers to a two-dimensional query table that stores the temperature compensation coefficient and dust attenuation coefficient corresponding to different environmental interference intervals. The piecewise linear relationship refers to the calculation method of the temperature compensation coefficient using different linear slopes in different temperature intervals. The exponential attenuation relationship refers to the dust attenuation coefficient increasing according to the dust concentration. The mathematical relationship of a function's regular decreasing behavior.

[0116] In the embodiment of the present application, in order to make the judgment criteria (thresholds) of fault diagnosis more intelligent and adaptable to the actual environment, it is necessary to avoid misjudgment or missed judgment due to environmental interference (mainly temperature and dust). The environmental conditions of the cable (such as surface temperature and dust concentration) will directly affect the signal quality collected by the sensor and the intensity of the fault characteristics. Therefore, the system needs to dynamically adjust the "standard lines" (thresholds) stored in the fault characteristic database for judging whether a fault has occurred according to the real-time environmental interference level. First, the environmental interference factor is calculated through step 1041. The cable surface temperature sensor uses an infrared temperature measurement module to collect temperature data at a frequency of 10 times per second, and the dust concentration monitor outputs a minute-level average based on the principle of light scattering. The temperature value (in °C) and the dust concentration value (in mg / m³) are normalized to eliminate dimensional differences and then calculated using the product formula EIF = 0.7T + 0.3D, where T is the normalized temperature value and D is the normalized dust concentration value. The coefficients 0.7 and 0.3 are empirical weights, indicating that the impact of temperature (70%) on interference is greater than that of dust (30%). In other words, the system needs to quantify the current level of environmental interference, which is achieved through a metric called the "environmental interference factor." This factor is calculated from two key environmental parameters: the real-time temperature of the cable surface (collected by an infrared temperature sensor, for example, measuring 10 times per second) and the dust concentration in the environment (monitored by a light scattering dust sensor, typically outputting an average value every minute). To fairly combine these two quantities (temperature in °C and dust in mg / m³), the system first normalizes them (think of them as scaling them to between 0 and 1 or some other common range). For example, if the cable surface temperature is measured at 60°C, the normalized T is 0.6 (assuming the normalized range is 0-100°C); and the dust concentration is 150 mg / m³, the normalized D is 0.3 (assuming the normalized range is 0-500 mg / m³). Therefore, the environmental interference factor (EIF) is 0.7*0.6+0.3*0.3=0.42+0.09=0.51.

[0117] Secondly, with the environmental interference factor EIF, the system will know the "level" of the current environmental interference. Next, the system will search the fault feature database for the corresponding "correction coefficient table" based on the size of this EIF value. Select the correction coefficient table through step 1042. The fault feature database is divided into five intervals according to the numerical range of the environmental interference factor: when EIF<0.3, select the low temperature and low dust correction table; when 0.3≤EIF<0.6, select the normal temperature conventional table; when 0.6≤EIF<0.9, select the high temperature and low dust table; when 0.9≤EIF<1.2, select the high dust compensation table; when EIF≥1.2, select the extreme environment table. Each correction coefficient table contains the temperature compensation coefficient and dust attenuation coefficient The database is a two-dimensional matrix, with rows corresponding to every 5°C temperature compensation node in the 20°C to 80°C range, and columns corresponding to every 50 mg / m³ dust concentration node in the 0-500 mg / m³ range. In other words, the database is pre-divided into five intervals based on interference levels, with a dedicated correction coefficient table prepared for each interval: when the EIF is very low (<0.3), the "Low Temperature, Low Dust Correction Table" is used; when the EIF is relatively low (0.3 <= EIF < 0.6), the "Normal Temperature Conventional Table" is used; when the EIF is moderate (0.6 <= EIF < 0.9), the "High Temperature, Low Dust Table" is used; when the EIF is high (0.9 <= EIF < 1.2), the "High Dust Compensation Table" is used; and when the EIF is very high (>= 1.2), the "Extreme Environment Table" is used. These correction coefficient tables store two very important correction coefficients: the "temperature compensation coefficient" α(T) and the "dust attenuation coefficient" β(D). Each table is a two-dimensional matrix, with rows corresponding to temperature intervals (e.g., every 5°C from 20°C to 80°C) and columns corresponding to dust concentration intervals (e.g., every 50 mg / m³ from 0 mg / m³ to 500 mg / m³). Based on the current measured temperature and dust concentration, the system finds the nearest temperature and dust concentration intervals in the selected correction factor table and retrieves the corresponding α(T) and β(D) values. For example, if the current EIF = 0.51 falls within the [0.3, 0.6) interval, select the "Normal Temperature Conventional Table." If the measured temperature is 45°C, the table will search for the rows corresponding to the temperature intervals of 40°C or 50°C (interpolation may be required). If the measured dust concentration is 120 mg / m³, the table will search for the columns corresponding to the concentration intervals of 100 mg / m³ or 150 mg / m³ (interpolation may be required), ultimately yielding the specific values ​​for α(T) and β(D).

[0118] Finally, the temperature compensation coefficient α(T) and dust attenuation coefficient β(D) obtained by table lookup are used to dynamically modify the original threshold θ0 in the fault feature database to obtain a new threshold that adapts to the current environment. The threshold is dynamically modified through step 1043. The original fault threshold Perform double coefficient correction calculation: . Temperature compensation coefficient In the range of 20-50℃, the temperature increases linearly by 0.02 / ℃, and in the range of 50-80℃, the temperature increases linearly by 0.05 / ℃. use The exponential function calculation shows that when D=200mg / m³, the attenuation coefficient reaches 0.632. The corrected threshold While retaining the positive offset introduced by temperature compensation and the negative compensation caused by dust attenuation, this ultimately generates a dynamic threshold parameter set adapted to actual environmental conditions. In other words, rising temperature typically makes certain fault characteristics (such as resistance and certain signal amplitudes) more pronounced. Therefore, a positive compensation (+α(T)) is required to raise the threshold (slightly loosening the diagnostic criteria to prevent temperature increases themselves from being misidentified as faults). The temperature compensation coefficient α(T) increases at different rates (slope) across different temperature ranges: in the mild temperature range (e.g., 20°C to 50°C), α(T) increases by a smaller percentage (e.g., 0.02) for every 1°C increase in temperature. In the high temperature range (e.g., 50°C to 80°C), α(T) increases by a larger percentage (e.g., 0.05) for every 1°C increase in temperature. This reflects the fact that temperature's impact on the system may not be linear, with greater impact at higher temperatures. For example, in the range below 50°C, α(T) = 0.02*(T - 20) (assuming α = 0 at 20°C). If T = 40°C, then α(T) = 0.02*(40-20) = 0.4. On the other hand, increased dust concentration typically interferes with signal transmission, weakening the strength of the fault signature detected by the sensor. Therefore, a negative attenuation (-β(D)) is required to lower the threshold (making the diagnostic criteria slightly stricter and preventing missed detections due to dust obstruction that weakens the true fault signal). The dust attenuation coefficient β(D) increases exponentially with increasing dust concentration D: β(D) = 1-e^(-0.005*D). The characteristic of an exponential function is that at very low dust concentrations, the attenuation is minimal; as the concentration increases, the attenuation increases rapidly, approaching 1 (i.e., a maximum attenuation of 100%). For example, when D = 0 mg / m³, β(D) = 1-e^0 = 1-1 = 0; when D = 100 mg / m³, β(D) = 1-e^(-0.005*100) = 1-e^(-0.5)≈1-0.6065=0.3935; when D = 200 mg / m³, β(D) = 1-e^(-1)≈1-0.3679=0.6321. Ultimately, the original threshold θ0 is amplified by (1+α(T)) (temperature compensation) and reduced by (1-β(D)) (dust attenuation), generating a dynamic new threshold that adapts to the current specific environmental conditions (temperature T and dust concentration D). .

[0119] In practical applications, for example, in a coal mine's underground intelligent diagnostic system, when the cable surface temperature sensor in section E12 detected a high temperature of 65°C and the dust concentration monitor displayed 220 mg / m³, the system automatically calculated the environmental interference factor to be 14,300 (65 × 220). The fault signature database invoked a preset correction factor table: when the interference factor exceeded 12,000, the second correction scheme was activated, with a temperature compensation factor of 0.8 corresponding to a temperature range of 60-70°C, a dust attenuation coefficient of 0.7, and an exponential attenuation curve based on a concentration of 200-250 mg / m³. The system corrected the current sudden change threshold of 85A and the voltage harmonic threshold of 50mV from the original fault threshold library to 85 × 0.8 × 0.7 = 47.6A and 50 × 0.8 × 0.7 = 28mV, respectively. The vibration energy threshold was also adjusted from 25m / s² to 25 × 0.7 = 17.5m / s². This dynamic threshold mechanism enables the system to accurately distinguish between normal load fluctuations and real short-circuit faults when the coal mining machine generates an 80A current fluctuation, avoiding false alarms caused by changes in cable impedance due to high temperature.

[0120] In the overall solution of the above step 104, the environmental interference factor is constructed by integrating the monitoring data of the cable surface temperature and dust concentration, and the fault feature threshold is dynamically corrected by combining the piecewise linear adjustment mechanism of the temperature compensation coefficient and the exponential decay model of the dust attenuation coefficient. The product calculation method is used to quantify the environmental interference intensity and adaptively match the correction coefficient table. The linear combination algorithm is used to dynamically couple the environmental parameters with the original threshold, effectively solving the problem of feature misjudgment caused by temperature drift and dust shielding in complex working conditions of traditional fixed thresholds. The compensation mechanism is used to improve the adaptability of the fault diagnosis system to changes in the cable surface state, significantly reducing the false alarm rate and missed detection probability in high temperature and high dust environments, and at the same time enhancing the robustness and self-healing ability of the feature database under different environmental interference intensities, ensuring the reliability of cable fault diagnosis results and the accuracy of maintenance decisions.

[0121] 105. Perform multi-dimensional matching calculation on the composite feature vector and the corrected fault mode threshold, and determine the fault diagnosis result according to the correlation strength between the electrical feature and the mechanical feature in the matching result.

[0122] Optionally, step 105 may specifically include the following steps:

[0123] 1051. Decompose the composite feature vector into electrical feature sub-vectors and mechanical feature sub-vectors, and perform Euclidean distance calculations on each of the sub-vectors with the corrected fault mode thresholds to obtain electrical matching and mechanical matching.

[0124] 1052. Calculate the correlation strength value between the electrical matching degree and the mechanical matching degree. When the correlation strength value exceeds a preset fusion threshold, activate the fault type determination module to compare the deviation between the electrical matching degree and the short-circuit spike threshold with the deviation between the mechanical matching degree and the open circuit oscillation threshold through the fault type determination module, and determine the fault diagnosis result based on the comparison result.

[0125] Among them, step 1052 may specifically include the following process: if the deviation between the electrical matching degree and the short-circuit spike threshold is smaller than the deviation between the mechanical matching degree and the open circuit oscillation threshold, then the short-circuit spike diagnosis result corresponding to the cable damage is output; if the deviation between the electrical matching degree and the short-circuit spike threshold is greater than the deviation between the mechanical matching degree and the open circuit oscillation threshold, then the open circuit oscillation diagnosis result corresponding to the electrical abnormality is output.

[0126] In the above steps, the composite feature vector refers to a multidimensional data set containing current distortion, voltage anomaly, and mechanical damage characteristics. The electrical matching degree refers to the similarity measure between the electrical feature subvector in the composite feature vector and the fault threshold. The mechanical matching degree refers to the similarity measure between the mechanical feature subvector and the fault threshold. The correlation strength value refers to the product or weighted calculation result of the electrical matching degree and the mechanical matching degree. The fusion threshold refers to the correlation strength critical value that triggers the fault type determination. The fault type determination module refers to the decision unit that determines the fault type based on the comparison of electrical and mechanical feature deviations. The short-circuit spike threshold refers to the current and voltage feature matching threshold corresponding to the cable short-circuit fault. The circuit breaker oscillation threshold refers to the mechanical vibration feature matching threshold corresponding to the line circuit breaker fault.

[0127] In this embodiment, the "composite feature vector" generated in the previous step, which incorporates spatiotemporal correlation information, is comprehensively compared with the "fault mode threshold" dynamically adjusted based on environmental conditions to determine the specific cable fault type (e.g., short circuit or open circuit) and its location. The composite feature vector is decomposed into a portion representing electrical anomalies (an electrical feature subvector) and a portion representing mechanical damage (a mechanical feature subvector). Similarity matching calculations are performed against the corresponding fault criteria (corrected thresholds) to obtain "electrical matching" and "mechanical matching." The system then pays special attention to the correlation strength between these two matching degrees (i.e., whether they are both high or one high and the other low). Only when the correlation strength is sufficiently strong can a reliable fault type determination be made. First, a feature matching calculation is performed in step 1051. The composite feature vector is decomposed, based on feature type, into an electrical feature subvector containing peak slope variation and harmonic frequency band energy, and a mechanical feature subvector containing mechanical damage spectral characteristics. The geometric distance between the electrical feature subvector and the corrected short-circuit spike threshold is calculated using the Euclidean distance formula. The calculated result is normalized to obtain an electrical matching degree in the range of 0-1. The geometric distance between the mechanical signature subvector and the corrected short-circuit oscillation threshold is simultaneously calculated, and the mechanical matching degree is obtained through the same normalization process. Specifically, the system splits the composite feature vector (F), containing multiple numerical values, into two parts based on the characteristics of the features: an "electrical signature subvector" (E), which primarily includes the previously calculated current peak slope change and voltage harmonic energy, two indicators reflecting electrical anomalies; and a "mechanical signature subvector" (M), which primarily includes the frequency points and their amplitude information of energy surges in the previously generated mechanical damage spectrum signature set, reflecting physical structural damage. The system then compares the electrical signature subvector (E) with standard threshold vectors (such as current slope thresholds and voltage harmonic energy thresholds) predefined and context-corrected for "short-circuit spike faults" in the fault signature database. The "Euclidean distance" (a mathematical measure of the proximity of two points in multidimensional space, with smaller distances indicating greater similarity) is calculated between them. For easier understanding, this distance is converted into an "electrical matching degree" ranging from 0% to 100% (a higher percentage indicates a more typical short-circuit fault profile). Similarly, the system compares the mechanical characteristic subvector (M) with a predefined and environmentally corrected standard threshold vector for "circuit oscillation faults" (such as vibration energy thresholds at specific frequencies), calculates the Euclidean distance, and converts it into a "mechanical match" between 0% and 100% (a higher percentage indicates that the current mechanical vibration characteristics more closely resemble a typical circuit oscillation fault). For example, the composite characteristic vector F is broken down into its electrical subvector E, which contains the current slope change of 82A / ms and the voltage harmonic energy of 480mV² / Hz; the mechanical subvector M contains the vibration energy at the 850Hz frequency of 28m / s².The revised short-circuit spike threshold is: current slope 70A / ms, harmonic energy 350mV² / Hz. The Euclidean distance between E and the short-circuit threshold is calculated to be 15.2, assuming a normalized electrical matching of 78%. The revised open-circuit oscillation threshold is: 850Hz vibration energy = 20m / s². The Euclidean distance between M and the open-circuit threshold is calculated to be 8, assuming a normalized mechanical matching of 60%.

[0128] Next, after obtaining the electrical and mechanical matching, the system needs to determine whether these two results "synergistically" point to the same fault event. This is achieved by calculating their "correlation strength value." When the correlation strength is sufficiently high, the fault type determination module is triggered. This module is a decision-making unit that determines the fault type based on the deviation between the electrical and mechanical characteristics. This module compares the deviation between the electrical matching and the short-circuit fault standard (short-circuit spike threshold) and the deviation between the mechanical matching and the open-circuit fault standard (open-circuit oscillation threshold). The "deviation" here refers to the original Euclidean distance value (or its absolute value) obtained during the matching calculation. The correlation strength determination is performed in step 1052. The electrical and mechanical matching are weighted and multiplied to calculate the correlation strength value. The weight coefficient is dynamically adjusted based on environmental interference factors. When the correlation strength value exceeds the preset fusion threshold of 0.75, the fault type determination module is activated. This module calculates the absolute deviation ΔE between the electrical matching and the short-circuit spike threshold, and the absolute deviation ΔM between the mechanical matching and the open-circuit oscillation threshold. The final fault type is determined by comparing ΔE and ΔM. If ΔE is less than ΔM, it is identified as a short-circuit spike fault caused by cable insulation damage; if ΔE is greater than ΔM, it is identified as a circuit oscillation fault caused by mechanical fracture. A diagnostic report is generated, including the fault location coordinates and a confidence rating. The system presets a "fusion threshold" (e.g., 75%). Only when the calculated correlation strength exceeds this threshold (for example, in one case, the correlation strength reaches 85% > 75%) does the system deem the electrical anomaly and mechanical damage to be sufficiently strongly correlated (likely caused by the same fault source) and activate the final "fault type determination module" for accurate classification. If the correlation strength falls below the threshold, the electrical and mechanical features may be out of sync or unrelated, and the system may need to collect more data or issue an uncertainty warning. The system compares the deviation between the electrical matching and the short-circuit spike threshold (ΔE) and the deviation between the mechanical matching and the circuit oscillation threshold (ΔM). If the electrical deviation is smaller (ΔE < ΔM), the current electrical signature more closely matches the typical pattern of a short-circuit fault (even if the mechanical signature has a certain degree of match, it's not as typical as the electrical signature). The system then identifies the fault as a "short-circuit spike fault caused by cable insulation damage." Conversely, if the mechanical deviation is smaller (ΔE > ΔM), the current mechanical vibration signature more closely matches the typical pattern of a short-circuit fault. The system then identifies the fault as a "short-circuit oscillation fault caused by mechanical breakage or looseness." Continuing with the previous example, an electrical match of 78% corresponds to a raw Euclidean distance of ΔE = 15.2 (the distance from the short-circuit threshold), while a mechanical match of 60% corresponds to a raw Euclidean distance of ΔM = 8 (the distance from the short-circuit threshold). Comparing ΔE = 15.2 and ΔM = 8, we find that ΔE > ΔM (15.2 > 8), indicating that the mechanical signature deviates less from its standard (the short-circuit threshold) and is more typical.Therefore, while a high correlation strength indicates a problem, the system ultimately determines it as a "circuit oscillation diagnosis result corresponding to an electrical anomaly" based on the deviation comparison rules. (Note: This determination requires specific threshold definitions and scenarios; the original rules are used as examples here.) Ultimately, the system generates a comprehensive report that includes the identified fault type (such as "short circuit spike" or "circuit oscillation"), the precise location of the fault (3D coordinates), and the confidence level of the diagnostic result (confidence level).

[0129] In practical applications, for example, in a coal mine's underground cable intelligent diagnosis system, when the system detected an anomaly in the E12 support section of the cable, the system collected a composite feature vector containing a current peak slope change of 82A / ms, a voltage harmonic energy value of 480mV² / Hz in the 2350Hz frequency band, and an energy value of 28m / s² at the 850Hz frequency point in the vibration spectrum. The system decomposed this vector into an electrical subvector (including the first two items) and a mechanical subvector (the third item), and then matched each with the dynamically corrected fault thresholds. The Euclidean distance between the electrical subvector and the short-circuit spike threshold (current slope threshold of 70A / ms and harmonic energy threshold of 350mV² / Hz) was 15.2, achieving a 78% match. The Euclidean distance between the mechanical subvector and the open circuit oscillation threshold (vibration energy threshold of 20m / s²) was 8, achieving a 60% match. When the system calculated the correlation strength between the two, it discovered that the vibration signal generated by the shearer pick impacting the gangue and the sudden change in current exhibited a 0.5ms time synchronization. The correlation strength reached 0.85, exceeding the preset fusion threshold of 0.75. The fault diagnosis module then initiated a comparison: the deviation between the electrical matching and the short-circuit threshold was 15.2 - 12.8 = 2.4, and the deviation between the mechanical matching and the open-circuit threshold was 8 - 5 = 3. Due to the smaller electrical deviation, the system determined that the short circuit was caused by damaged cable insulation. The 3D model pinpointed a 3cm-long crack in the cable sheath 1.8 meters east of the E12 support. This diagnostic result enabled maintenance personnel to directly locate the fault with a partial discharge detector and pinpoint the fault within 10 minutes, saving four hours compared to a traditional full-line inspection.

[0130] In the overall solution of the above step 105, accurate cable fault identification is achieved through multi-dimensional decomposition of composite feature vectors and matching calculation of dynamic correction thresholds. Independent matching degree calculation and correlation strength value fusion analysis of electrical feature sub-vectors and mechanical feature sub-vectors are adopted. Combined with the deviation comparison mechanism of the fault type determination module, differentiated diagnostic conclusions are output based on the deviation magnitude difference between the short-circuit spike threshold and the open circuit oscillation threshold. This effectively solves the problem of fault type confusion caused by feature coupling in traditional methods, eliminates the risk of false triggering of a single feature through collaborative verification of electrical and mechanical features, and significantly improves the ability to distinguish between short-circuit spikes and mechanical damage oscillations. At the same time, the adaptability of the diagnostic system to composite faults and the decision-making credibility are enhanced based on multi-dimensional matching of dynamic correction thresholds, providing mine cable operation and maintenance with an intelligent diagnostic solution that is both sensitive and specific.

[0131] The following is a complete embodiment of steps 101 to 105:

[0132] like Figure 2As shown, in a coal mine's underground power supply system, intelligent monitoring devices deployed along the coal mining face collect cable operation data. When a shearer cuts gangue, the system first uses a multi-dimensional sensor array (step 101) to synchronously collect cable operation data. A high-frequency current sensor detects a sudden increase in phase C current from 520A to 720A within 0.5 seconds, while the voltage drops from 380V to 260V. A vibration sensor captures a transient vertical shock of 3.2g. Facing the humid and dusty underground environment (dust concentrations as high as 280 milligrams per cubic meter - mg / m³), the system activates adaptive noise reduction and time-synchronized transmission. A Butterworth filter with an 8kHz cutoff frequency filters out the 12kHz high-frequency noise generated by the inverter. When the dust concentration reaches 280mg / m³, redundant transmission mode is activated, compressing the signal interval to 15ms, ensuring that the deviation between the vibration and electrical signal timestamps is less than ±0.8ms. These noise-reduced and strictly time-aligned signals are then superimposed and analyzed to generate a high-quality, interference-resistant dataset. Based on this dataset, the system entered the feature extraction phase. It discovered a significant correlation between the cable's severe vibration (high intensity) and abnormally rapid current changes (transient rates as high as 150 amperes per millisecond). Based on the correlation between vibration intensity and current transient rate (150 A / ms), the dynamic capture window was expanded to 40 ms, capturing three peak transitions. The system then accurately calculated the maximum slope change of 86.7 A / ms (describing the steepness of the current rise from one peak to the next). Frequency domain analysis of the voltage waveform within the same time window extracted the maximum slope change of 86.7 A / ms and detected a harmonic energy peak of 450 mV² / Hz in the 2350 Hz band of the voltage waveform. Wavelet packet decomposition of the vibration signal revealed that the energy at the 850 Hz frequency point surged from 5 m / s² to 25 m / s², forming a spectral signature of mechanical damage.

[0133] Next, the system input these key features (current peak slope of 86.7A / ms, voltage harmonic energy of 450mV² / Hz at 2350Hz, and vibration energy peak of 25m / s² at 850Hz) into the spatiotemporal correlation model. The model verified a 1.2ms time offset between the current distortion and voltage anomalies, as well as an 85% spatial overlap within a 2m radius around the E12 support. Combining the cable surface temperature of 68°C and the dust concentration of 280mg / m³, the system calculated an environmental interference factor of 19040. The model then dynamically adjusted the short-circuit current threshold to 52A and the harmonic energy threshold to 210mV² / Hz. The model first verified the temporal synchronization of the electrical anomaly features: it confirmed that the time difference between the current distortion and voltage anomalies was only 1.2 milliseconds, far less than the preset tolerance threshold of 5ms, indicating a high correlation between the two. Subsequently, spatial comparison revealed that the detected mechanical damage signature (located by vibration source) and the electrical anomaly (located by electrical signal signature) were located within a 2-meter radius of the E12 hydraulic support, with a spatial overlap of up to 85%. Based on this high spatiotemporal correlation, the model generated a composite feature vector that integrated the electrical and mechanical anomaly information. Simultaneously, the system monitored environmental conditions in real time: the cable surface temperature sensor measured 68 degrees Celsius (°C), and the dust concentration remained at a high level of 280 mg / m³. Combining these two environmental parameters, the calculated environmental interference factor (EIF) reached a high of 19040 (quantifying the degree of interference from the harsh environment). Based on this high interference factor, the system dynamically adjusted the thresholds in the fault signature database: the current sudden change threshold for short-circuit faults was lowered to 52 amperes (A), and the harmonic energy threshold was lowered to 210 millivolts squared per hertz (mV² / Hz). These adjustments account for the fact that high temperatures and dust can amplify certain electrical signal features and attenuate sensor signals, respectively, making the diagnostic criteria more adaptable to the harsh underground environment. Finally, the system performed a multi-dimensional matching calculation between the generated composite feature vector and these dynamically modified thresholds. The results showed that the electrical features (current slope and voltage harmonics) matched the modified short-circuit fault thresholds by 82%, and the mechanical features (850Hz vibration energy) matched the modified open-circuit fault thresholds by 73%. More importantly, the correlation strength (which comprehensively reflects the likelihood of a coordinated electrical and mechanical anomaly) reached 0.89, significantly exceeding the preset fusion trigger threshold of 0.75. Based on the judgment rules, the system further compared the deviation between the electrical feature matching and the short-circuit threshold (which was relatively small) and the mechanical feature matching and the open-circuit threshold (which was relatively large). The system ultimately determined the fault type as an intermittent short circuit caused by mechanical damage to the cable insulation (such as extrusion or scraping). Integrating spatiotemporal positioning information, the system accurately located the fault point in the 3D model as a cable segment 2.3 meters east of the E12 hydraulic support.Based on this diagnostic result, on-site maintenance personnel went directly to the target location and indeed discovered a crack in the cable insulation, approximately 5 centimeters long. The accuracy of the diagnosis was verified through insulation resistance testing. Multi-dimensional matching revealed an 82% electrical signature match and a 73% mechanical match. The correlation strength value of 0.89 exceeded the threshold, indicating that the intermittent short circuit was caused by mechanical damage to the cable insulation. The 3D model located the fault point 2.3 meters east of the E12 support. On-site maintenance personnel discovered a 5-centimeter insulation crack. Insulation testing confirmed the accuracy of the diagnosis, significantly reducing overall troubleshooting time and effectively avoiding unplanned downtime. This precise positioning significantly shortened the overall troubleshooting time, effectively avoiding unplanned downtime caused by cable faults, and ensuring continuous and safe production in the coal mine.

[0134] Figure 3 The present invention provides a schematic diagram of a system for self-diagnosing faults in a mining circuit. Figure 2 As shown, the system includes:

[0135] The acquisition module 31 uses a multi-dimensional sensor group to synchronously collect the current waveform, voltage waveform and three-dimensional vibration signal of the cable, and generates an anti-interference data set through adaptive noise reduction and time-series synchronous transmission in a humid and dusty environment;

[0136] an extraction module 32 that dynamically adjusts a waveform capture window to capture transient waveform segments based on the correlation between the mechanical vibration intensity and the current transient rate in the anti-interference data set, extracts a peak slope variation from the transient waveform segments of the current waveform, separates the harmonic frequency band energy of the short-circuit spike from the transient waveform segments of the voltage waveform, and simultaneously performs time-frequency analysis on the three-dimensional vibration signal to generate a mechanical damage spectrum feature set;

[0137] The correlation module 33 inputs the peak slope variation, harmonic frequency band energy and mechanical damage spectrum feature set into a spatiotemporal correlation model, and generates a composite feature vector by performing spatiotemporal correlation verification between the current distortion feature and the voltage anomaly feature;

[0138] A correction module 34 dynamically corrects the threshold value in the fault feature database based on the environmental interference factor in the composite feature vector and in combination with the cable surface temperature and dust data;

[0139] The matching module 35 performs a multi-dimensional matching calculation on the composite feature vector and the corrected fault mode threshold, and determines a fault diagnosis result according to the correlation strength between the electrical feature and the mechanical feature in the matching result.

[0140] Figure 3 The mine circuit fault self-diagnosis system can perform Figure 1The implementation principle and technical effects of the mining circuit fault self-diagnosis method described in the illustrated embodiment will not be described in detail. The specific manner in which each module and unit performs operations in the mining circuit fault self-diagnosis system in the above embodiment has been described in detail in the embodiment of the method, and will not be elaborated on here.

[0141] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A mine circuit fault self-diagnosis method, characterized in that: include: The cable's current waveform, voltage waveform, and three-dimensional vibration signal are synchronously collected through a multi-dimensional sensor group. Through adaptive noise reduction and time-series synchronous transmission in a humid and dusty environment, an anti-interference data set is generated. Based on the correlation between the mechanical vibration intensity and the current transient rate in the anti-interference data set, the waveform clipping window is dynamically adjusted to capture the transient waveform segment, the peak slope change is extracted from the transient waveform segment of the current waveform, the harmonic frequency band energy of the short-circuit spike is separated from the transient waveform segment of the voltage waveform, and the three-dimensional vibration signal is simultaneously subjected to time-frequency analysis to generate a mechanical damage spectrum feature set; Inputting the peak slope variation, harmonic frequency band energy and mechanical damage spectrum feature set into a spatiotemporal correlation model, and generating a composite feature vector through spatiotemporal correlation verification of current distortion features and voltage anomaly features; Based on the environmental interference factors in the composite feature vector, combined with the cable surface temperature and dust data, the threshold value in the fault feature database is dynamically corrected; Performing a multi-dimensional matching calculation on the composite feature vector and the corrected fault mode threshold, and determining a fault diagnosis result based on the correlation strength between the electrical feature and the mechanical feature in the matching result; The method includes dynamically adjusting the waveform capture window to capture the transient waveform segment based on the correlation between the mechanical vibration intensity and the current transient rate in the anti-interference data set, including: A vibration intensity detection module based on the three-dimensional vibration signal in the anti-interference data set calculates the root mean square value of the vibration amplitude, and establishes a correlation relationship table between the vibration intensity and the current transient rate; when it is detected that the current transient rate exceeds the rate threshold corresponding to the current vibration intensity interval in the correlation relationship table, a slope mutation point or a local extreme point within a preset time range is selected from the current current waveform as the starting position of the waveform interception window, wherein the starting position is preferably a trough point; according to the ratio of the root mean square value to the current transient rate, the length of the waveform interception window is dynamically adjusted, wherein the window length is expanded when the ratio increases and contracted when the ratio decreases, and the window length is limited between a preset upper limit and a lower limit; the amplitude mutation amplitude and the spectral energy distribution of the current waveform in the adjusted window are analyzed, and if the value of the amplitude mutation amplitude exceeds the preset threshold and the spectral energy is concentrated in the high frequency band, the current waveform is captured as a transient waveform segment; The peak slope variation is extracted from the transient waveform segment of the current waveform, the harmonic frequency band energy of the short-circuit spike is separated from the transient waveform segment of the voltage waveform, and the three-dimensional vibration signal is subjected to time-frequency analysis to generate a mechanical damage spectrum feature set, including: From the captured transient waveform segments of the current waveform, the ratio of the amplitude increment to the time increment between adjacent peaks is extracted as the peak slope variation. The captured transient waveform segments of the voltage waveform are subjected to frequency domain decomposition to identify frequency bands with a frequency three times higher than the fundamental frequency and concentrated energy distribution. The energy integral of each sub-band within the selected frequency band is calculated as the harmonic frequency band energy. The three-dimensional vibration signal is subjected to time-frequency analysis using a multi-resolution decomposition method to extract a set of frequency points with sudden increases in vibration energy within each decomposed frequency band. The frequency positions and energy increases of the frequency point sets are recorded to generate a mechanical damage spectrum feature set consisting of frequency point distribution and energy changes. The peak slope variation, harmonic frequency band energy, and mechanical damage spectrum feature set are input into the spatiotemporal correlation model, and the spatiotemporal correlation verification of the current distortion feature and the voltage anomaly feature is performed to generate a composite feature vector, which includes: The peak slope variation, harmonic frequency band energy and mechanical damage spectrum feature set are input into a feature fusion module, and data is reorganized according to the timestamp alignment rule to form a time-space matrix containing current distortion features, voltage anomaly features and mechanical damage features; the current distortion feature sequence and the voltage anomaly feature sequence are extracted from the time-space matrix, and the current distortion feature sequence and the voltage anomaly feature sequence are segmented and intercepted according to a preset time window sliding rule to form current distortion feature segments and voltage anomaly feature segments containing feature values ​​of multiple consecutive sampling points; the current distortion feature segment and the voltage anomaly feature segment in the same time window are cumulatively calculated point by point, and the phase difference between the two is minimized by dynamically adjusting the starting position of the voltage anomaly feature segment to generate a time offset in the current time window; The threshold value in the fault feature database is dynamically corrected based on the environmental interference factor in the composite feature vector and combined with the cable surface temperature and dust data, including: An environmental interference factor is extracted from the composite feature vector, where the environmental interference factor is calculated by multiplying the temperature value collected by the cable surface temperature sensor and the dust concentration monitoring value; according to the size of the environmental interference factor, a correction coefficient table corresponding to the temperature compensation coefficient and the dust attenuation coefficient is selected from the fault feature database; the original threshold in the fault feature database is linearly combined with the temperature compensation coefficient and the dust attenuation coefficient in the correction coefficient table to generate a dynamically corrected threshold, wherein the temperature compensation coefficient and the temperature value have a piecewise linear relationship, and the dust attenuation coefficient and the dust concentration value have an exponential attenuation relationship.

2. The method according to claim 1, characterized in that Performing a multi-dimensional matching calculation on the composite feature vector and the corrected fault mode threshold, and determining a fault diagnosis result based on the correlation strength between the electrical feature and the mechanical feature in the matching result, including: Decomposing the composite feature vector into an electrical feature sub-vector and a mechanical feature sub-vector, and performing Euclidean distance calculations on each of the sub-vectors with the corrected fault mode threshold to obtain electrical matching and mechanical matching; Calculate the correlation strength value between the electrical matching degree and the mechanical matching degree. When the correlation strength value exceeds a preset fusion threshold, activate the fault type determination module to compare the deviation between the electrical matching degree and the short-circuit spike threshold with the deviation between the mechanical matching degree and the open-circuit oscillation threshold through the fault type determination module, and determine the fault diagnosis result based on the comparison result.

3. The method according to claim 2, characterized in that Determining the fault diagnosis result according to the comparison result includes: If the deviation between the electrical matching degree and the short-circuit spike threshold is less than the deviation between the mechanical matching degree and the open circuit oscillation threshold, a short-circuit spike diagnosis result corresponding to cable damage is output; If the deviation between the electrical matching degree and the short-circuit spike threshold is greater than the deviation between the mechanical matching degree and the open circuit oscillation threshold, a open circuit oscillation diagnosis result corresponding to the electrical anomaly is output.

4. The method according to claim 1, wherein The peak slope variation, harmonic frequency band energy, and mechanical damage spectrum feature set are input into the spatiotemporal correlation model, and a composite feature vector is generated through spatiotemporal correlation verification of current distortion features and voltage anomaly features, including: The peak slope variation, harmonic frequency band energy, and mechanical damage spectrum feature set are input into a feature fusion module, and data is reorganized according to a timestamp alignment rule to form a spatiotemporal matrix containing current distortion features, voltage anomaly features, and mechanical damage features; Based on the current distortion characteristics and voltage anomaly characteristics in the space-time matrix, a time offset between the two within a preset time window is calculated by a cross-validation unit in the space-time correlation model. When the time offset is less than a preset threshold, a spatial correlation unit is triggered to compare the physical location corresponding to the mechanical damage characteristic with the spatial coordinates of the current and voltage anomaly area to generate a spatial coincidence degree. According to the time offset and spatial overlap, correlation weight coefficients of current distortion characteristics, voltage abnormality characteristics and mechanical damage characteristics are determined, and the time offset, spatial overlap and correlation weight coefficients are weighted and summed to generate a composite feature vector.

5. The method according to claim 4, characterized in that When the time offset is less than a preset threshold, the spatial correlation unit is triggered to compare the physical location corresponding to the mechanical damage feature with the spatial coordinates of the current and voltage abnormal area to generate a spatial coincidence, including: determining whether the time offset is less than a preset threshold, and when the time offset is less than the preset threshold, sending a trigger instruction to the spatial association unit, and retrieving a physical position coordinate set corresponding to the mechanical damage feature and a spatial coordinate set of the current and voltage abnormal area through the trigger instruction; Based on the physical position coordinate set and the spatial coordinate set of the current and voltage abnormal areas, with the minimum coverage area of ​​the mechanical damage feature as the benchmark, the coverage areas of all current and voltage abnormal areas are traversed, and the overlapping lengths of the two in the horizontal and vertical ranges are counted. The spatial overlap is generated according to the ratio of the product of the horizontal and vertical overlapping lengths to the coverage area of ​​the mechanical damage feature.

6. The method according to claim 1, characterized in that The multi-dimensional sensor group synchronously collects the cable's current waveform, voltage waveform, and three-dimensional vibration signal. Through adaptive noise reduction and time-series synchronous transmission in a humid and dusty environment, an anti-interference data set is generated, including: The sensor group consisting of current sensor, voltage sensor and three-dimensional vibration sensor collects the current waveform, voltage waveform and three-dimensional vibration signal of the cable in a synchronous triggering manner; Dynamically selecting filtering parameters through a signal strength adaptive module based on the signal-to-noise ratio of the current waveform and the voltage waveform, and performing high-frequency noise suppression on the current waveform and the voltage waveform based on the filtering parameters, retaining only transient components in the signal whose amplitude fluctuation exceeds a preset threshold; When transmitting the current waveform, voltage waveform, and three-dimensional vibration signal, the time stamps of the three axial components of the three-dimensional vibration signal, the current waveform, and the voltage waveform are aligned through a timing synchronization transmission module, and the signal transmission interval is dynamically adjusted according to the dust concentration monitoring result. When the dust concentration is higher than a preset critical value, the signal transmission interval is shortened to reduce signal loss caused by dust interference; The current waveform, voltage waveform and three-dimensional vibration signal that have completed timing synchronization are superimposed and analyzed, and the intervals where the amplitudes of the three suddenly change within the same period are extracted to generate a transient signal set with strictly aligned timestamps as an anti-interference data set.

7. A mine circuit fault self-diagnosis system, used to execute the mine circuit fault self-diagnosis method according to any one of claims 1 to 6, characterized in that: include: The acquisition module uses a multi-dimensional sensor group to synchronously collect the current waveform, voltage waveform and three-dimensional vibration signal of the cable, and generates an anti-interference data set through adaptive noise reduction and time-series synchronous transmission in a humid and dusty environment; an extraction module that dynamically adjusts a waveform capture window to capture transient waveform segments based on the correlation between mechanical vibration intensity and current transient rate in the anti-interference data set, extracts a peak slope variation from the transient waveform segments of the current waveform, separates harmonic frequency band energy of the short-circuit spike from the transient waveform segments of the voltage waveform, and simultaneously performs time-frequency analysis on the three-dimensional vibration signal to generate a mechanical damage spectrum feature set; A correlation module inputs the peak slope variation, harmonic frequency band energy, and mechanical damage spectrum feature set into a spatiotemporal correlation model, generates a composite feature vector through spatiotemporal correlation verification of current distortion features and voltage anomaly features; A correction module dynamically corrects the threshold value in the fault feature database based on the environmental interference factor in the composite feature vector and in combination with the cable surface temperature and dust data; The matching module performs a multi-dimensional matching calculation on the composite feature vector and the corrected fault mode threshold, and determines the fault diagnosis result according to the correlation strength between the electrical feature and the mechanical feature in the matching result.

Images