A method and system for quickly locating grounding faults in coal mine power supply systems

Abstract

A method and system for rapidly locating ground faults in a coal mine power supply system relates to the field of electrical digital data processing. The method comprises: obtaining real-time zero-sequence current values ​​at each junction box in the coal mine power supply system. When the zero-sequence current value at any junction box exceeds a preset current threshold, collecting voltage and current waveform data within a preset time period before and after the fault point; calculating the distortion rate and phase offset of each phase voltage based on the voltage and current waveform data; determining the fault type based on the changing trends of the distortion rate and phase offset; performing a time-series comparative analysis of the zero-sequence current values ​​collected at each junction box at the time of the fault to obtain the propagation time sequence characteristics of the zero-sequence current between each detection point; and locating the faulty cable section where the fault point is located based on the propagation time sequence characteristics and the fault type. Implementing this method can improve the accuracy of locating ground faults in coal mine power supply systems.

Description

Technical Field

[0001] The present application relates to the field of electrical digital data processing, and in particular to a method and system for quickly locating grounding faults in a coal mine power supply system. Background Art

[0002] With the continuous expansion of coal mining and the continuous upgrading of mining technology, the complexity of coal mine power supply systems is also increasing. As the energy guarantee for coal mine production, the safe and stable operation of coal mine power supply systems is crucial to the normal operation of coal mines. Among them, ground faults in cable networks are one of the most common types of faults in coal mine power supply systems. How to promptly detect and accurately locate ground fault points has become a key issue in ensuring the safe operation of coal mine power supply systems.

[0003] Currently, coal mine power supply systems often use zero-sequence current protection to detect ground faults. This solution installs zero-sequence current transformers at key nodes in the power supply system to detect zero-sequence current. When the zero-sequence current exceeds a set threshold, a ground fault is detected. Fault location primarily relies on segmented testing of suspected sections, using changes in zero-sequence current to determine the approximate location of the fault.

[0004] However, in practice, coal mine power supply systems have complex, branching cables, with some cables laid underground. In complex power supply networks, the zero-sequence current at each node often exhibits complex distribution characteristics when a fault occurs, making it difficult to accurately determine the specific location of the fault based solely on the zero-sequence current value. Summary of the Invention

[0005] The present application provides a method and system for quickly locating grounding faults in a coal mine power supply system, which are used to improve the accuracy of locating grounding faults in a coal mine power supply system.

[0006] In the first aspect, the present application provides a method for quickly locating grounding faults in a coal mine power supply system, which is applied to a coal mine power supply system. The method includes: obtaining the real-time zero-sequence current value at each junction box in the coal mine power supply system. When the real-time zero-sequence current value at any junction box exceeds a preset current threshold, the sampling point where the real-time zero-sequence current value first exceeds the preset current threshold is taken as the fault point, and the voltage waveform data and current waveform data within a preset time period before and after the fault point are collected. According to the voltage waveform data and the current waveform data, the distortion rate and phase offset of each phase voltage are calculated; the fault type is determined based on the changing trend of the distortion rate and the phase offset; a time-series comparison and analysis is performed on the zero-sequence current values ​​collected at each junction box at the moment of the fault to obtain the propagation time-series characteristics of the zero-sequence current between each detection point; and the fault cable section where the fault point is located is located according to the propagation time-series characteristics and the fault type.

[0007] In the above-described embodiment, zero-sequence current values ​​are collected in real time at each junction box, and voltage and current waveform data before and after the fault point is acquired when a fault occurs. The fault type is determined by calculating the distortion rate and phase offset of each phase voltage. This is combined with the propagation time characteristics of the zero-sequence current to perform location analysis. This establishes an active fault segment location method, enabling precise location of the cable segment where the fault point is located in complex power supply networks, improving the efficiency and accuracy of fault location.

[0008] In combination with some embodiments of the first aspect, in some embodiments, the step of locating the fault cable section where the fault point is located based on the propagation timing characteristics and the fault type specifically includes: obtaining a zero-sequence current propagation characteristic template corresponding to the fault type from a preset fault characteristic library, the zero-sequence current propagation characteristic template including a propagation delay reference value and an amplitude attenuation reference value of the zero-sequence current between adjacent detection points under different fault types; calculating the actual propagation delay value and the actual amplitude attenuation value between adjacent junction boxes based on the propagation timing characteristics, the actual propagation delay value is obtained by the zero-crossing time difference of the zero-sequence current waveform at the adjacent junction boxes, and the actual amplitude attenuation value is obtained by the ratio of the maximum amplitudes of the zero-sequence current at the adjacent junction boxes; calculating the matching degree of the actual propagation delay value and the actual amplitude attenuation value with the propagation delay reference value and the amplitude attenuation reference value in the zero-sequence current propagation characteristic template to obtain the matching degree value of each cable section; and determining the fault cable section where the fault point is located based on the matching degree value.

[0009] In the above embodiment, a zero-sequence current propagation characteristic template was established from a preset fault signature library. The actual propagation delay and amplitude attenuation values ​​were calculated to perform a match analysis with the template. Because faults in different locations can result in different propagation characteristics, calculating and comparing these match values ​​makes the fault location results more convincing and further improves the reliability of fault location.

[0010] In combination with some embodiments of the first aspect, in some embodiments, after the step of locating the fault cable section where the fault point is located according to the propagation timing characteristics and the fault type, the method also includes: respectively obtaining the waveform information of each cycle in the voltage waveform data and the current waveform data, and calculating the key characteristic parameters of each cycle according to the waveform information, the key characteristic parameters including the amplitude mutation point parameter, the phase mutation point parameter and the waveform distortion parameter; converting the key characteristic parameters into a key characteristic code, which is used to characterize the waveform information of each cycle; using a preset standard characteristic code as a reference characteristic code, comparing the key characteristic code with the reference characteristic code to obtain difference data; based on the difference data, screening out the cycle segment whose difference is greater than the preset difference threshold, and marking the cycle segment as the fault characteristic cycle segment; generating a fault characteristic data chain according to the time sequence arrangement of the fault characteristic cycle segment, the fault characteristic data chain including the start time, peak time and recovery time of each fault characteristic cycle segment.

[0011] In the above embodiment, waveform data is analyzed for cycles and key characteristic parameters are extracted to generate a fault characteristic data chain, establishing a complete record of the fault development process. This characteristic data chain contains key information such as the fault onset, peak value, and recovery time. It provides an important basis for accurately determining the fault type and analyzing the fault development trend, making fault location and analysis more comprehensive and in-depth.

[0012] In combination with some embodiments of the first aspect, in some embodiments, the step of converting the key feature parameters into a key feature code specifically includes: dividing the waveform data of each cycle into N equally spaced sampling segments; calculating the waveform change rate of adjacent sampling points in each sampling segment; marking the sampling points whose waveform change rate is greater than a preset change rate threshold as key sampling points; generating a position distribution sequence based on the distribution position of the key sampling points in the cycle; calculating the deviation between the actual waveform value at the key sampling point and the standard sinusoidal waveform value to generate a deviation value sequence; and combining and encoding the position distribution sequence and the deviation value sequence to obtain the key feature code.

[0013] In the above embodiment, waveform data for each cycle is sampled and divided at equal intervals. The waveform change rate is calculated and key sampling points are marked, generating a position distribution sequence and a deviation value sequence. By combining these codes to form key feature codes, digital representation of waveform characteristics is achieved. This encoding method preserves the key waveform change characteristics and deviation information, making fault feature extraction and comparison more accurate and efficient.

[0014] In combination with some embodiments of the first aspect, in some embodiments, after the step of locating the fault cable section where the fault point is located according to the propagation timing characteristics and the fault type, the method also includes: generating a difference waveform spectrum diagram based on the difference data, and the difference waveform spectrum diagram is used to characterize the waveform spectrum distribution during the fault process; according to the spectrum distribution characteristics in the difference waveform spectrum diagram, the fault feature data chain is divided into N sub-chains, and each sub-chain corresponds to an independent fault development stage; the change trend of the position distribution sequence and the deviation value sequence of the key sampling points in each sub-chain is counted to generate a fault evolution feature matrix; the fault evolution feature matrix is ​​matched with the preset standard fault feature library to obtain matching data; according to the matching data, K groups of historical fault data with the highest matching degree are selected from the preset standard fault feature library; the position distribution sequence and the deviation value sequence in the K groups of historical fault data are analyzed, common features are extracted, and a fault type feature template is generated; the fit of the fault feature data chain of the current fault and the fault type feature template is calculated, and when the fit is greater than the preset fit threshold, it is determined that the current fault type is the same as the fault type corresponding to the fault type feature template.

[0015] In the above embodiment, the fault signature data chain is segmented based on the difference waveform spectrum to generate a fault evolution signature matrix, which is then matched against a standard fault signature library. By extracting common features from multiple sets of historical fault data to generate feature templates, an adaptive fault type identification mechanism is established, enhancing the accuracy and adaptability of fault type determination.

[0016] In combination with some embodiments of the first aspect, in some embodiments, after the step of locating the fault cable section where the fault point is located based on the propagation timing characteristics and the fault type, the method also includes: establishing a fault development state transition diagram based on the timing relationship of each sub-chain in the fault evolution characteristic matrix, and the state transition diagram is used to characterize the complete process from the occurrence to the development of the fault; calculating the transition probability between adjacent states in the fault development state transition diagram to generate a state transition probability matrix; and updating the fault type feature template according to the state transition probability matrix.

[0017] In the above embodiment, a state transition diagram reflecting the complete fault development process was established, a state transition probability matrix was calculated, and the fault type signature template was dynamically updated. This solution establishes a continuously optimized fault signature update mechanism, enabling the fault type signature template to adapt to changes in fault characteristics under different operating conditions, thereby improving the accuracy and reliability of fault type judgment in various complex environments.

[0018] In combination with some embodiments of the first aspect, in some embodiments, the step of updating the fault type feature template according to the state transition probability matrix specifically includes: detecting the degree of fit between each sub-chain feature and the corresponding stage feature in the fault type feature template; when the degree of fit is lower than a preset fitness threshold, calculating the transition probability of the sub-chain feature in the state transition probability matrix; updating the fault type feature template according to the transition probability; and performing real-time monitoring and fault warning of the waveform data based on the updated fault type feature template and state transition probability matrix.

[0019] In the above embodiment, the fit between subchain features and fault type signature templates is monitored in real time, and the signature templates are dynamically updated based on transition probabilities. By introducing a state transition probability matrix to calibrate and optimize the signature templates, a more precise correspondence between waveform features and fault types is established, improving the accuracy of fault type determination in complex power supply systems and reducing bias and misjudgment during the fault diagnosis process.

[0020] In a second aspect, an embodiment of the present application provides a coal mine power supply system, which includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to enable the coal mine power supply system to execute the method described in the first aspect and any possible implementation method of the first aspect.

[0021] In a third aspect, an embodiment of the present application provides a computer program product comprising instructions. When the computer program product is run on a coal mine power supply system, the coal mine power supply system executes the method described in the first aspect and any possible implementation method of the first aspect.

[0022] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium comprising instructions. When the instructions are executed on a coal mine power supply system, the coal mine power supply system executes the method described in the first aspect and any possible implementation method of the first aspect.

[0023] It is understandable that the coal mine power supply system provided in the second aspect, the computer program product provided in the third aspect, and the computer storage medium provided in the fourth aspect are all used to execute the methods provided in the embodiments of this application. Therefore, the beneficial effects that can be achieved can be referenced to the beneficial effects of the corresponding methods and will not be repeated here.

[0024] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages:

[0025] 1. This application collects zero-sequence current values ​​in real time at each junction box and obtains voltage and current waveform data before and after the fault point when a fault occurs. The fault type is determined by calculating the distortion rate and phase offset of each phase voltage. This is combined with the propagation timing characteristics of the zero-sequence current for location analysis. This establishes an active fault segment location method, enabling precise positioning of the cable segment where the fault point is located in complex power supply networks, improving the efficiency and accuracy of fault location.

[0026] 2. This application establishes a zero-sequence current propagation characteristic template from a preset fault signature library and analyzes the matching between the actual propagation delay and amplitude attenuation values ​​and the template. Because faults in different locations result in different propagation characteristics, calculating and comparing matching values ​​makes the fault location results more convincing and further improves the reliability of fault location.

[0027] 3. This application establishes a complete record of the fault development process by performing cycle analysis on waveform data and extracting key characteristic parameters to generate a fault characteristic data chain. This characteristic data chain contains key information such as the fault onset, peak value, and recovery time. It provides an important basis for accurately determining the fault type and analyzing the fault development trend, making fault location and analysis more comprehensive and in-depth. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 This is a flow chart of a method for quickly locating a grounding fault in a coal mine power supply system according to an embodiment of the present application;

[0029] Figure 2 This is a system deployment diagram of a method for quickly locating a grounding fault in a coal mine power supply system according to an embodiment of the present application;

[0030] Figure 3 This is a schematic diagram of module connections in a cable junction box in a method for quickly locating a grounding fault in a coal mine power supply system according to an embodiment of the present application;

[0031] Figure 4 This is a voltage change diagram of the method for quickly locating a ground fault in a coal mine power supply system according to an embodiment of the present application;

[0032] Figure 5 This is another flow chart of the method for quickly locating a grounding fault in a coal mine power supply system according to an embodiment of the present application;

[0033] Figure 6 This is a schematic diagram of the structure of a physical device of a coal mine power supply system in an embodiment of the present application. DETAILED DESCRIPTION

[0034] The terms used in the following examples of the present application are only for the purpose of describing specific embodiments and are not intended to limit the present application. As used in the specification of the present application, the singular expressions "a", "an", "above", "the", and "this" are intended to include plural expressions as well, unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used in the present application refers to any or all possible combinations of one or more of the listed items.

[0035] In the following, the terms "first" and "second" are used for descriptive purposes only and should not be understood to imply or suggest relative importance or implicitly indicate the number of the technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the embodiments of this application, unless otherwise specified, "plurality" means two or more.

[0036] For ease of understanding, the following describes the process of the method provided by this implementation. Figure 1 , which is a flow chart of a method for quickly locating grounding faults in a coal mine power supply system in an embodiment of the present application.

[0037] S101. Acquire a real-time zero-sequence current value at each junction box in a coal mine power supply system. The real-time zero-sequence current value is acquired by a zero-sequence current transformer provided at each junction box.

[0038] The coal mine power supply system refers to the overall power system network used to power underground coal mines, encompassing multiple distribution devices and power lines. A junction box is a sealed enclosure used to connect and distribute cables, protecting cable joints and enabling cable connections and distribution. The zero-sequence current value, the vector sum of the three-phase currents, indicates the degree of three-phase imbalance. A zero-sequence current transformer is an electromagnetic induction device used to detect zero-sequence current. It converts the primary current proportionally to the secondary current using the principle of electromagnetic induction.

[0039] During normal operation of a coal mine power supply system, continuous monitoring of the system's operating status is necessary to promptly detect potential faults. Specifically, the system installs zero-sequence current transformers at each cable junction box. These transformers utilize the principle of high-precision electromagnetic induction. A current signal proportional to the primary current is induced in the secondary coil via a cable passing through the transformer's primary coil. Through real-time sampling and signal processing, the zero-sequence current value at each junction box can be accurately acquired. Continuous monitoring of zero-sequence current provides fundamental data support for fault detection.

[0040] In some embodiments, zero-sequence current acquisition can be achieved through a variety of methods: Optionally, a high-precision zero-sequence transformer is used to acquire three-phase current in real time. The acquired analog signal is converted to a digital signal using digital signal processing technology. After filtering and amplification, the accurate zero-sequence current value is obtained. This solution first acquires the raw current signal through the transformer, then uses a high-speed ADC for analog-to-digital conversion. A digital filtering algorithm is then used to eliminate interference and noise. Finally, the final zero-sequence current value is obtained through data processing. Optionally, an intelligent current detection module is placed at the junction box, integrating zero-sequence current acquisition and data processing functions. The module includes components such as current sensors, signal conditioning circuits, and a data processing unit, and can automatically complete the entire process from signal acquisition to data output. It is understood that other methods can also be used to achieve the zero-sequence current acquisition function, such as using different types of sensors or different signal processing methods, which are not limited here.

[0041] S102. When the real-time zero-sequence current value at any junction box exceeds a preset current threshold, the sampling point where the real-time zero-sequence current value first exceeds the preset current threshold is taken as the fault point, and voltage waveform data and current waveform data within a preset time period before and after the fault point are collected.

[0042] The preset current threshold represents the zero-sequence current alarm value set by the system, used to determine whether a ground fault has occurred. The fault point refers to the specific location where the cable ground fault occurred. The voltage waveform data refers to the curve of system voltage changes over a period of time before and after the fault. The current waveform data refers to the curve of system current changes over a period of time before and after the fault. The sampling point represents the time point at which discrete samples are taken within the continuous waveform.

[0043] When an abnormal zero-sequence current is detected, the system immediately initiates a fault recording process to capture critical data from the fault process. Specifically, when the system detects that the zero-sequence current value at a junction box exceeds a preset threshold, it deems a ground fault to have occurred. At this point, the system immediately begins recording voltage and current waveform data for a specific period of time (typically 100ms-600ms) before and after the fault occurs. This data fully records the fault's generation, development, and resolution, and is invaluable for subsequent fault feature analysis and fault location.

[0044] In some embodiments, the collection and recording of fault data can be achieved in a variety of ways: optionally, a high-speed data acquisition system is used, which includes a high-precision sampling circuit, a large-capacity memory and a real-time processing unit. The system can continuously collect voltage and current signals at a very high sampling rate. When a fault is detected, the historical data in the memory and the real-time data for a subsequent period of time are saved to form a complete fault record; optionally, an intelligent fault recording device is used, which has functions such as trigger acquisition, data caching, and waveform analysis. The device continuously samples but does not save data under normal conditions, and only starts data recording when a fault trigger condition is detected, which can greatly save storage space. It is understandable that other methods can also be used to achieve the collection and recording of fault data, such as using different trigger mechanisms or different storage strategies, which are not limited here.

[0045] S103 : Calculate the voltage distortion rate and phase offset of each phase voltage based on the voltage waveform data and the current waveform data.

[0046] Voltage waveform data refers to the time-varying data series of the three-phase voltage collected before and after a fault occurs. Current waveform data refers to the time-varying data series of the three-phase current collected before and after a fault occurs. The distortion rate indicates the deviation between the actual waveform and the standard sinusoidal waveform and is used to measure the degree of waveform distortion. Phase offset refers to the relative displacement between the actual waveform and the standard sinusoidal waveform on the time axis and is used to indicate the temporal variation characteristics of the waveform.

[0047] After acquiring waveform data from the fault process, this data needs to be thoroughly analyzed to extract key characteristic parameters. Specifically, the system first preprocesses the collected voltage and current waveform data, including data filtering and normalization. Then, using digital signal processing methods such as Fourier transform, the time domain waveforms are converted to the frequency domain for analysis. The waveform distortion rate is calculated by calculating the fundamental component and harmonic components of the waveform. Furthermore, the phase offset is calculated by analyzing the timing relationship between the waveform's zero crossings and peaks. These parameters effectively reflect changes in the system's electrical characteristics when a fault occurs.

[0048] In some embodiments, the calculation of waveform characteristic parameters can be achieved in a variety of ways: optionally, the waveform data is processed using a fast Fourier transform (FFT) algorithm, first the waveform data is processed by a window function to reduce spectrum leakage, then the FFT operation is performed to obtain the spectrum, the fundamental and harmonic amplitudes are calculated, and finally the distortion rate is calculated according to the amplitude ratio of each component, and the phase offset is calculated by the phase difference between adjacent periods; optionally, the waveform characteristics are analyzed using a wavelet transform method, first the appropriate wavelet basis function is selected to perform multi-scale decomposition on the waveform, then the waveform mutation characteristics and periodic characteristics are extracted at different scales, and finally the distortion rate and phase offset are calculated by combining the analysis results of each scale. It is understandable that other methods can also be used to calculate the waveform characteristic parameters, such as using other signal processing algorithms or feature extraction methods, which are not limited here.

[0049] S104: Determine the fault type based on the change trends of the distortion rate and the phase offset.

[0050] The fault type refers to the specific form of ground fault determined based on the changing patterns of the system's electrical characteristics. The change trend indicates how the distortion rate and phase offset change over time. The fault diagnosis model is a mathematical model or set of rules used to map waveform characteristic parameters to specific fault types.

[0051] After acquiring the waveform characteristic parameters, the specific fault type needs to be determined based on how these parameters change. Specifically, the system first analyzes the changing trends of the distortion rate and phase offset during the fault process, including characteristics such as the rate of change, magnitude, and duration. These changing characteristics are then matched against a pre-established fault signature library, and the most likely fault type is determined through pattern recognition or rule-based judgment. This waveform-based fault type determination method offers high accuracy and reliability.

[0052] In some embodiments, fault type determination can be achieved through a variety of methods: optionally, a rule-based expert system approach is used to first establish a knowledge base containing various fault characteristic rules, then substitute the waveform characteristic parameters obtained by actual calculation into the rules for reasoning, and obtain the final fault type judgment result through rule matching and confidence calculation; optionally, a machine learning approach is used to train a classification model using a large amount of historical fault data, establish a mapping relationship between waveform characteristic parameters and fault types, and then use the trained model to classify and judge new fault cases. It is understandable that other methods can also be used to determine fault types, such as using other artificial intelligence algorithms or hybrid methods, which are not limited here.

[0053] S105 , performing time sequence comparison analysis on the zero-sequence current values ​​collected at each junction box at the time of the fault, and obtaining the propagation time sequence characteristics of the zero-sequence current between each detection point.

[0054] Time series comparative analysis refers to the time series comparison of zero-sequence current values ​​collected at different time points and locations. Propagation time series characteristics represent the delay, attenuation, and other characteristics exhibited by the zero-sequence current signal during its propagation along the cable line. The detection point refers to the junction box location where the zero-sequence current transformer is installed. Zero-sequence current propagation characteristics refer to the physical properties exhibited by the zero-sequence current as it propagates along the cable line, including propagation speed and attenuation patterns.

[0055] After determining the fault type, the fault current propagation characteristics need to be analyzed to determine the fault location. Specifically, the system first collects zero-sequence current data from each junction box at the time of the fault, arranging this data in chronological order to form a complete record of the propagation process. It then analyzes the time difference and amplitude changes in the zero-sequence current waveforms between adjacent detection points, calculating characteristic parameters such as propagation delay and signal attenuation. These parameters reveal the propagation patterns of the fault current on the line, providing an important basis for fault location.

[0056] In some embodiments, the analysis of zero-sequence current propagation characteristics can be achieved in a variety of ways: optionally, using a waveform correlation analysis method, first time-synchronizing the zero-sequence current waveforms of each detection point, then calculating the cross-correlation function of the waveforms of adjacent detection points, determining the propagation delay by the time position of the cross-correlation peak, determining the degree of signal attenuation by the peak amplitude ratio, and finally combining the analysis results of multiple detection points to obtain a complete propagation characteristic; optionally, using a signal processing and feature extraction method, first performing wavelet decomposition on the zero-sequence current waveform, extracting the signal of the characteristic frequency band, and then analyzing the propagation law of the characteristic signal in time and space, and describing the propagation characteristics by establishing a mathematical model. It is understandable that other methods can also be used to achieve the analysis of zero-sequence current propagation characteristics, such as using other signal analysis methods or propagation models, which are not limited here.

[0057] S106. Locate the faulty cable section where the fault point is located according to the propagation time sequence characteristics and the fault type.

[0058] A cable segment refers to the cable route between two adjacent junction boxes. A propagation feature template is a set of reference values ​​for the propagation delay and amplitude attenuation of zero-sequence current between adjacent detection points under different fault types. A fault feature library is a database containing various typical fault cases and their characteristic parameters. The matching value refers to the degree of consistency between the actual propagation features and the feature template.

[0059] After obtaining the zero-sequence current propagation characteristics, the fault point needs to be precisely located in conjunction with the fault type information. Specifically, the system first retrieves the zero-sequence current propagation characteristic template corresponding to the current fault type from a preset fault signature library. The measured propagation delay and amplitude attenuation values ​​are then matched against the reference values ​​in the template to determine the matching degree for each cable segment. Finally, based on the matching degree, the most likely fault segment is determined. This method fully leverages the correlation between the zero-sequence current propagation characteristics and the fault location, enabling rapid and accurate fault location.

[0060] In some embodiments, fault location can be achieved through a variety of methods: optionally, a positioning algorithm based on propagation characteristics is used, first calculating the theoretical propagation delay based on the propagation speed of the zero-sequence current on the line, then comparing the actual measured propagation delay with the theoretical value, solving the fault location by establishing a mathematical equation, and finally verifying and correcting it by combining the amplitude attenuation characteristics; optionally, a pattern recognition method is used, first establishing a feature library containing a large number of historical fault cases, then matching the propagation characteristics of the current fault with the templates in the feature library, and determining the most matching fault mode through similarity calculation to obtain the fault location. It is understandable that other methods can also be used to achieve fault location, such as using other positioning algorithms or hybrid methods, which are not limited here.

[0061] This step specifically includes:

[0062] A zero-sequence current propagation characteristic template corresponding to the fault type is obtained from a preset fault feature library. The zero-sequence current propagation characteristic template includes a propagation delay reference value and an amplitude attenuation reference value of the zero-sequence current between adjacent detection points under different fault types.

[0063] The preset fault signature library is a data set that stores characteristic data for different types of faults. Zero-sequence current is one-third of the vector sum of the three-phase currents. The propagation characteristic template describes the propagation pattern of zero-sequence current in the cable. Propagation delay is the time required for zero-sequence current to propagate between adjacent detection points. Amplitude attenuation is the degree of amplitude reduction of the zero-sequence current during propagation. The detection point is the location where the current sensor is installed. Zero-sequence current propagation characteristic templates are extracted from the fault signature library. First, based on the identified fault type, the corresponding feature template is located in the signature library. Each feature template contains two key parameter sets: a propagation delay reference value matrix and an amplitude attenuation reference value matrix. The propagation delay reference value matrix records the propagation time of the zero-sequence current between any two adjacent detection points under different fault location conditions. The amplitude attenuation reference value matrix records the amplitude attenuation ratio of the zero-sequence current during propagation. These reference values ​​are standard values ​​derived from statistical analysis of a large number of historical fault cases. The extracted feature template serves as a reference for subsequent fault location.

[0064] The actual propagation delay value and actual amplitude attenuation value between adjacent junction boxes are calculated based on the propagation timing characteristics. The actual propagation delay value is obtained by the zero-crossing time difference of the zero-sequence current waveform at adjacent junction boxes, and the actual amplitude attenuation value is obtained by the ratio of the maximum amplitudes of the zero-sequence current at adjacent junction boxes.

[0065] Among them, the propagation timing characteristics are characteristics that describe the time pattern of zero-sequence current propagation. Adjacent junction boxes are adjacent connection points on the cable line. The zero-crossing point is the moment when the zero-sequence current waveform crosses zero. The maximum amplitude is the peak value of the zero-sequence current waveform. The time difference is the interval between the occurrence of two events. The propagation parameters are calculated based on the collected zero-sequence current data. For each pair of adjacent junction boxes, their zero-sequence current waveform data is first extracted. By finding the zero-crossing point of the waveform, the time of the first positive zero-crossing point at the adjacent junction boxes is recorded, and the time difference Δt = t2-t1 is calculated to obtain the actual propagation delay value. Then, the maximum amplitudes A1 and A2 of the two waveforms are identified, and the amplitude ratio k = A2 / A1 is calculated to obtain the actual amplitude attenuation value. The above calculation process is repeated for all pairs of adjacent junction boxes on the cable line to obtain a complete set of actual propagation parameters.

[0066] The actual propagation delay value and the actual amplitude attenuation value are matched with the propagation delay reference value and the amplitude attenuation reference value in the zero-sequence current propagation characteristic template to obtain the matching value of each cable section.

[0067] The matching degree is a metric that measures the degree of similarity between actual propagation parameters and reference values. A cable segment is a cable segment between two adjacent junction boxes. The matching degree calculation rule defines the parameter comparison method. The matching degree value indicates the degree of parameter matching. The matching degree of the calculated actual propagation parameters is calculated against the reference values ​​in the feature template. For each cable segment, the matching degree of propagation delay and amplitude attenuation is calculated separately. The delay matching degree calculation formula is: η1 = 1 - |Δt - Δtref| / Δtref, where Δtref is the reference delay value. The amplitude attenuation matching degree calculation formula is: η2 = 1 - |k - kref| / kref, where kref is the reference attenuation value. The comprehensive matching degree η = w1·η1 + w2·η2, where w1 and w2 are weight coefficients, and w1 + w2 = 1. The comprehensive matching degree value of each cable segment is recorded.

[0068] The faulty cable section where the fault point is located is determined based on the matching value.

[0069] The fault point is the specific location where the fault occurred. The fault section is the cable segment containing the fault point. The location result is an estimate of the fault point's location. Location accuracy indicates how accurately the fault point has been located. The fault section is determined based on the calculated matching value. First, the matching values ​​of all cable sections are sorted. The section with the highest matching value is most likely to contain the fault point. When the matching value of a section is significantly higher than that of other sections (typically by more than 20%), it is determined to be the fault section. If there are multiple sections with similar matching values, a comprehensive judgment is required based on other characteristics (such as the magnitude of the fault current and the degree of waveform distortion). After the fault section is determined, a more precise location estimate can be made within the section based on the matching value. Once the faulty cable section is located, the location information is sent to the target client, which includes the mobile phones and computers used by operation and maintenance personnel.

[0070] The following describes the overall architecture of the coal mine power supply system ground fault rapid location system provided by this implementation. Figure 2 , which is a system deployment diagram of a method for quickly locating grounding faults in a coal mine power supply system in an embodiment of the present application.

[0071] This diagram illustrates the overall layout of the rapid ground fault location system for coal mine power supply systems. Multiple cable junction boxes are distributed along the coal mine power supply lines, each equipped with a ground current detection module. These detection modules communicate digitally with the power monitoring system at the monitoring center via power cables. Each ground current detection module at each junction box has a built-in zero-sequence current transformer (CCT) to obtain real-time zero-sequence current values ​​at each junction box in the coal mine power supply system. When a fault occurs, the detection module uploads this data to the power monitoring system via digital carrier communication, providing data support for subsequent fault analysis and location.

[0072] The following is a demonstration of the connection method of the ground current detection module provided by this embodiment in the cable junction box. Figure 3 , which is a schematic diagram of module connections in a cable junction box of a method for quickly locating grounding faults in a coal mine power supply system in an embodiment of the present application.

[0073] Figure 3 The figure shows the specific connection method of the ground current detection module in the cable junction box. The three-phase cable passes through the zero-sequence current transformer. The output side of the zero-sequence current transformer is connected to the module to detect the zero-sequence current in the three-phase cable. At the same time, the cable shield and the ground pile are connected to the module, forming a power carrier communication channel to ensure data transmission between the detection module and the power monitoring system. The zero-sequence current transformer is a key component for obtaining real-time zero-sequence current values. Its accurate acquisition lays the foundation for subsequent fault diagnosis and location, while the power carrier communication channel ensures the timely and accurate transmission of fault data.

[0074] The following is a presentation of the voltage changes during a single-phase ground fault provided by this implementation. Figure 4 , which is a voltage change diagram of the method for quickly locating grounding faults in a coal mine power supply system in an embodiment of the present application.

[0075] Figure 4 (a) shows the three-phase voltage changes during a single-phase ground fault. The horizontal axis is time (t / s), illustrating the time course before and after the fault occurs; the vertical axis is voltage (U), visually illustrating the changes in the three-phase voltage amplitudes. The three curves in the figure represent the voltages of phases A, B, and C, respectively. During normal operation, the three-phase voltages are relatively stable. After a single-phase ground fault occurs, the voltage in the fault phase drops significantly, while the voltage in the non-fault phase increases.

[0076] Figure 4 (b) shows the zero-sequence voltage variation during a single-phase ground fault. The horizontal axis is also time (t / s), and the vertical axis is zero-sequence voltage (U0). Under normal conditions, the zero-sequence voltage is close to zero. At the moment a single-phase ground fault occurs, the zero-sequence voltage rises rapidly and then maintains a specific fluctuation for a period of time. Analyzing the zero-sequence voltage variation can help determine the moment of fault occurrence.

[0077] The following is a more detailed description of the process of the method provided by this implementation. Figure 5 , is another flow chart of the method for quickly locating grounding faults in a coal mine power supply system in an embodiment of the present application.

[0078] S501. Locate the faulty cable section where the fault point is located according to the propagation time sequence characteristics and the fault type.

[0079] The propagation time sequence characteristics refer to the time delay and attenuation characteristics exhibited by zero-sequence current as it propagates along a cable line. The fault type represents the specific type of ground fault determined based on electrical characteristics. The cable section refers to the portion of the cable line between two adjacent junction boxes. The propagation characteristic template contains reference values ​​for the propagation time delay and amplitude attenuation of zero-sequence current between adjacent detection points for different fault types.

[0080] After obtaining the zero-sequence current propagation characteristic template corresponding to the current fault type from a preset fault signature library, the actual propagation delay and actual amplitude attenuation values ​​are calculated. The actual propagation delay is calculated by the time difference between the zero-crossing points of the zero-sequence current waveforms at adjacent junction boxes. This calculation method interpolates the waveform data to determine the precise zero-crossing instants, then calculates the time difference between the zero-crossing points at adjacent measuring points. The actual amplitude attenuation is calculated by taking the ratio of the maximum zero-sequence current amplitudes at adjacent junction boxes. This is done by dividing the maximum zero-sequence current detected at each measuring point during the fault period by the propagation direction. The calculated actual propagation delay and amplitude attenuation values ​​are then compared with the reference values ​​in the propagation characteristic template to determine the matching degree for each cable segment. This matching degree calculation uses a weighted Euclidean distance method, taking into account the matching degree of both the delay and attenuation characteristics. Finally, the faulty cable segment where the fault point is located is determined based on the matching degree, and the segment with the highest matching degree is selected as the fault location.

[0081] S502 , respectively obtaining waveform information of each cycle in the voltage waveform data and the current waveform data, and calculating key characteristic parameters of each cycle based on the waveform information, the key characteristic parameters including amplitude mutation point parameters, phase mutation point parameters, and waveform distortion parameters.

[0082] A cycle is a complete period of an AC electrical signal. Waveform information encompasses the complete signal change process within a cycle. The amplitude mutation point parameter indicates the time and degree of significant change in the waveform amplitude. The phase mutation point parameter indicates the time and degree of significant change in the waveform phase. Waveform distortion parameters describe the degree of deviation of the waveform from a standard sine wave.

[0083] The collected voltage and current waveform data are segmented into cycles, with each cycle serving as an independent analysis unit. Zero-point detection is used to determine the cycle start point, segmenting the continuous waveform data into multiple single-cycle waveforms. Feature parameters are extracted for each cycle waveform: First, the instantaneous amplitude sequence of the waveform is calculated. A threshold is set to detect amplitude mutation points, and the moment of the mutation and the amplitude change are recorded. The phase difference between adjacent sampling points is then calculated, and the phase mutation point is determined using threshold detection. The moment of the mutation and the phase change are recorded. Finally, the actual waveform is compared with a standard sine wave of the same frequency, and the root mean square error is calculated to obtain waveform distortion parameters. These characteristic parameters together provide a complete description of the fault waveform characteristics.

[0084] S503: Divide the waveform data of each cycle into N equally spaced sampling segments.

[0085] An equally spaced sampling segment is a waveform segment obtained by dividing a cycle into equal time intervals. A sampling point is a discrete sampling point in a continuous waveform. The N value is the pre-set number of sampling segments, which determines the sampling accuracy.

[0086] The complete periodic waveform is divided into N sampling segments in time. The time length of each sampling segment is equal, which is the period length divided by N. For an AC signal with an industrial frequency of 60Hz, one cycle is 20ms. If N is 100, the duration of each sampling segment is 0.2ms. Within each sampling segment, the number and position of sampling points are determined according to the sampling frequency. For example, when the sampling frequency is 10kHz, a 0.2ms sampling segment contains 2 sampling points. This equal-interval division method ensures uniform sampling of waveform features, providing a standardized data foundation for subsequent feature extraction and analysis. In practical applications, the selection of the N value needs to balance the requirements of computational complexity and analysis accuracy, and is usually selected between 60 and 200.

[0087] S504: Convert the key feature parameters into a key feature code, where the key feature code is used to characterize the waveform information of each cycle.

[0088] Key characteristic parameters include the amplitude and phase mutation point parameters, as well as waveform distortion parameters. Key characteristic codes are digital sequences obtained by encoding waveform characteristic parameters. The position distribution sequence represents the distribution of key sampling points within a cycle. The deviation value sequence represents the difference between the actual waveform value and the standard sine waveform value at the key sampling points.

[0089] The process of converting waveform features into signature codes first calculates the waveform change rate of adjacent sampling points within each sampling segment. The change rate is calculated by dividing the amplitude difference between adjacent sampling points by the sampling time interval. Sampling points with a waveform change rate greater than a preset change rate threshold are marked as key sampling points. The relative positions of these key sampling points in the cycle are recorded to generate a position distribution sequence. For each key sampling point, the deviation between its actual waveform value and the standard sinusoidal waveform value at the same time is calculated to generate a deviation value sequence. The position distribution sequence and the deviation value sequence are combined and encoded according to predetermined rules to obtain the final key signature code. The encoding rules use binary encoding, encoding the position information and deviation values ​​separately and then splicing them together to form a fixed-length binary signature code.

[0090] In some embodiments, this step specifically includes the following steps:

[0091] The waveform data of each cycle is divided into N equally spaced sampling segments.

[0092] In this step, a cycle refers to a complete signal period. Waveform data is a sequence of sampled values ​​that describes how a signal changes over time. Equally spaced sampling segments are time periods obtained by evenly dividing a cycle. The sampling segment length is the number of sampling points contained in each sampling segment. The sampling interval is the time interval between adjacent sampling points. The waveform data of each cycle is segmented into equally spaced segments. First, the starting and ending points of the cycle are determined, and the total number of sampling points M contained in the cycle is calculated. Based on the preset number of segments N, the length of each sampling segment is calculated as L = M / N (rounded down). Starting from the starting point of the cycle, a sampling segment is divided every L sampling points, and the length of the last sampling segment is determined by the actual number of remaining sampling points. Sampling segments are numbered from 0 to N-1, and each sampling segment records the index position of its starting sampling point. This equally spaced segmentation method ensures uniform sampling and analysis of the waveform.

[0093] Calculate the waveform change rate of adjacent sampling points in each sampling segment.

[0094] The waveform rate of change represents the rate of change of the waveform amplitude over time. Adjacent sampling points are two sampling points that are consecutive in time. Rate of change calculation is a numerical method based on the difference of the sampling point amplitudes. The difference interval is the time step used in the rate of change calculation. The local rate of change is calculated for the waveform data within each sampling segment. The rate of change is calculated using the first-order difference method. For each pair of adjacent sampling points (i, i+1) within the sampling segment, the calculation formula is: rate of change = (amplitude[i+1] - amplitude[i]) / sampling time interval. The unit of the rate of change is amplitude / time. To eliminate the influence of noise, the calculated rate of change sequence is subjected to a moving average filter with a filter window length of 3-5 sampling points. The filtered rate of change value at each sampling point is recorded to form the rate of change sequence for that sampling segment.

[0095] Sampling points whose waveform change rate is greater than a preset change rate threshold are marked as key sampling points.

[0096] Among them, the change rate threshold is the standard value for judging whether a sampling point is a key point. The key sampling point is the characteristic position where the waveform changes significantly. The marking process is to identify the attributes of the sampling point. The characteristic position refers to the time position of the key sampling point in the waveform. Traverse the change rate sequence of each sampling segment to identify the key sampling points. Compare the change rate value of each sampling point with the preset threshold. When the absolute value of the change rate is greater than the threshold, the sampling point is marked as a key sampling point. The marking uses a binary flag, 1 represents a key sampling point, and 0 represents an ordinary sampling point. At the same time, the time position and amplitude information of each key sampling point are recorded. In order to avoid the key points being too dense, when the adjacent sampling points all meet the threshold conditions, the point with the larger absolute value of the change rate is selected as the key sampling point.

[0097] A position distribution sequence is generated according to the distribution positions of the key sampling points in the cycle.

[0098] Among them, the position distribution sequence describes the spatial distribution of key sampling points within the cycle. The distribution position is the relative time position of the key sampling point. Serialization is the process of converting position information into a sequence form. Position encoding is a digital representation of position information. The position distribution sequence is generated based on the marked key sampling points. First, the number of key sampling points in each sampling segment is counted to obtain density distribution information. The phase angle of each key sampling point relative to the starting point of the cycle is calculated, and the phase angle range is 0-360 degrees. The phase angle is quantized into an 8-bit binary number to form a position code. The position codes of all key sampling points are combined in chronological order to generate a complete position distribution sequence. This encoding method not only retains the precise position information of the key points, but also achieves a compact representation of the data.

[0099] Calculate the deviation between the actual waveform value and the standard sine waveform value at the key sampling points and generate a deviation value sequence.

[0100] The standard sine waveform is the reference waveform under ideal conditions. The deviation value represents the difference between the actual waveform and the standard waveform. The actual waveform value is the actual signal amplitude obtained by sampling. The deviation value sequence records the deviation information of all key sampling points. The deviation between the waveform value and the standard value at the key sampling points is calculated. First, a standard sine waveform is generated based on the fundamental frequency and amplitude parameters of the cycle. For each key sampling point, the difference between its actual waveform value and the standard sine waveform value at the corresponding moment is calculated to obtain the deviation value. The deviation value is normalized, using the maximum amplitude of the actual waveform as the normalization reference. The normalized deviation values ​​of all key sampling points are organized in chronological order to generate a deviation value sequence. This deviation calculation method can reflect the degree and characteristics of waveform distortion.

[0101] The position distribution sequence and the deviation value sequence are combined and encoded to obtain the key feature code.

[0102] The signature code is a digital code that describes waveform characteristics. Combination coding is a coding method that combines multiple signature sequences. The coding rules define the specific method for combining sequences. Feature compression is a coding technique that reduces data redundancy. The position distribution sequence and the deviation value sequence are combined to generate a signature code. Using a segmented combination coding method, the code for each sampling segment consists of two parts: a position code and a deviation code. The position code uses 8 bits of binary to represent position information, while the deviation code uses 8 bits of binary to represent the normalized deviation value. The two codes are combined through bit splicing to form a 16-bit segment signature code. The signature codes of all sampling segments are sequentially concatenated to generate a complete key signature code. This combination coding method achieves efficient representation and storage of waveform characteristics.

[0103] S505: Using a preset standard feature code as a reference feature code, performing a difference comparison between the key feature code and the reference feature code to obtain difference data.

[0104] The preset standard signature code refers to the signature code corresponding to the standard waveform under normal operating conditions. The baseline signature code serves as a reference for comparison. The difference data indicates the degree of difference between the actual signature code and the baseline signature code. The coding distance refers to the number of bits that differ between two binary signature codes.

[0105] The difference between the key feature code obtained by converting the actual waveform and the preset standard feature code is calculated. The difference is calculated using the Hamming distance method, which counts the number of bits with different values ​​at the same position in the two feature codes. The specific calculation process is to perform an XOR operation on the two feature codes, and the number of 1s in the result is the Hamming distance. At the same time, considering the varying importance of different feature parameters, a weight coefficient is introduced when calculating the overall difference. Different weights are assigned to the differences in the position distribution sequence and the differences in the deviation value sequence, and the weighted sum is used to obtain the final difference data. This weighted calculation method highlights the influence of important feature parameters and improves the accuracy of the difference calculation.

[0106] S506 , based on the difference data, filter out frequency bands whose difference is greater than a preset difference threshold, and mark the frequency bands as fault characteristic frequency bands.

[0107] The difference threshold is a pre-set difference judgment standard. The fault characteristic frequency band refers to a continuous frequency sequence that exhibits obvious fault characteristics. The difference statistic represents the cumulative effect of the difference data. The marker information includes the starting and ending positions of the fault characteristic frequency band.

[0108] The calculated difference data is screened and analyzed, and the cycles with differences greater than the preset threshold are marked. In order to avoid misjudgment caused by sudden interference, the sliding window method is used to smooth the difference data. The window length is set to 3 cycles, and the average value of the difference within the window is calculated as the difference statistic of the central cycle. When the difference statistic of a certain cycle exceeds the preset threshold and the adjacent cycles also show a large difference, the cycle and its adjacent abnormal cycles are collectively marked as a fault characteristic cycle segment. For each identified fault characteristic cycle segment, its starting cycle position, ending cycle position and maximum difference value are recorded. This information is used for subsequent fault analysis.

[0109] S507 : Generate a fault feature data chain according to the time sequence of the fault feature cycle segments, where the fault feature data chain includes the start time, peak time, and recovery time of each fault feature cycle segment.

[0110] The fault signature data chain is an ordered data structure containing information about multiple fault signature cycles. The starting time is the time when the fault signature begins to appear. The peak time is the time when the fault signature reaches its maximum level. The recovery time is the time when the fault signature disappears. A time series arrangement represents a data sequence organized in chronological order.

[0111] The identified fault feature cycles are analyzed and organized in time series. First, all fault feature cycles are arranged in chronological order to form a preliminary time series chain. For each cycle, three key moments are determined by analyzing its difference change curve: the moment when the difference begins to exceed the threshold is recorded as the starting moment, the moment when the difference reaches the local maximum is recorded as the peak moment, and the moment when the difference returns to below the threshold is recorded as the recovery moment. These key moment information are organized in chronological order to construct a complete fault feature data chain. Each node in the data chain contains moment information and the corresponding difference value. The nodes are connected through a time series relationship to form a complete data structure that describes the fault development process.

[0112] S508 : Generate a difference waveform spectrum diagram according to the difference degree data, where the difference waveform spectrum diagram is used to characterize the waveform spectrum distribution during the fault process.

[0113] The difference waveform spectrogram visualizes the distribution characteristics of difference data in the frequency domain. The spectrum distribution represents the energy distribution of a signal across different frequency components. Spectral components refer to the amplitude of a signal at specific frequencies. Spectral features refer to the characteristic patterns exhibited in the spectrogram.

[0114] The process of converting the difference data into a spectral representation uses the short-time Fourier transform (SFT) method. First, the difference data sequence is segmented into fixed-length segments, and each segment is multiplied by a Hanning window to reduce spectral leakage. A fast Fourier transform is then performed on each windowed segment to obtain the spectral components within that time segment. The spectra of adjacent time segments are arranged in chronological order to form a two-dimensional graph of the spectrum over time, known as the difference waveform spectrogram. The horizontal axis of the spectrogram represents time, and the vertical axis represents frequency. The color depth in the graph indicates the energy level of different frequency components. This time-frequency analysis method can simultaneously reflect the changing patterns of fault characteristics in both the time and frequency domains.

[0115] S509 , dividing the fault feature data chain into N sub-chains according to the spectrum distribution characteristics in the difference waveform spectrum diagram, where each sub-chain corresponds to an independent fault development stage.

[0116] A subchain is a continuous segment within a fault signature data chain that exhibits similar spectral characteristics. A fault development stage refers to each time period within the fault process that exhibits distinct characteristics. Spectral distribution characteristics include the location and energy of the primary frequency components. The classification is based on similarity criteria for spectral characteristics.

[0117] The fault signature data chain is segmented by analyzing the spectral distribution characteristics in the difference waveform spectrum diagram. First, the main frequency components in the spectrum diagram are extracted, including the fundamental component and significant harmonic components at each moment. The similarity of the spectral distribution between adjacent time periods is calculated using the cosine similarity method, which divides the inner product of the two spectral vectors by the product of their moduli. When the similarity value changes significantly, the moment is marked as a segmentation point. Based on the segmentation points, the fault signature data chain is divided into multiple sub-chains, each of which corresponds to an independent stage in the fault development process. This spectral feature-based segmentation method can accurately identify characteristic changes during the fault development process.

[0118] S510 , counting the changing trends of the position distribution sequence and the deviation value sequence of the key sampling points in each sub-chain, and generating a fault evolution feature matrix.

[0119] Key sampling points are sampling locations within a waveform that exhibit significant characteristics. The position distribution sequence represents the temporal distribution of key sampling points within a cycle. The deviation value sequence represents the difference between the actual value and the baseline value at the key sampling point. The trend of change refers to the temporal variation of these sequences. The fault evolution feature matrix is ​​a two-dimensional data structure that describes the changes in various characteristic parameters during the fault development process.

[0120] Statistical analysis is performed on the data in each subchain to extract fault evolution characteristics. First, the position distribution of key sampling points within each subchain is calculated. The position distribution sequence is obtained by counting the number of key sampling points in each sampling segment. The deviation values ​​at these key sampling points are then calculated, and the deviation value sequence is obtained by subtracting the actual waveform value from the standard waveform value. Trend analysis is performed on the position distribution sequence and the deviation value sequence, and statistical characteristics such as the mean, variance, and rate of change of the sequence are calculated. These statistical characteristics are organized into a matrix according to a predetermined format. The rows of the matrix represent different characteristic parameters, the columns represent different time periods, and the matrix element values ​​represent the statistical values ​​of the corresponding characteristic parameters in a specific time period. This matrix representation method comprehensively describes the characteristic changes exhibited by the fault during its development.

[0121] S511 , performing matching calculation on the fault evolution feature matrix and the preset standard fault feature library to obtain matching data.

[0122] The pre-set standard fault signature library contains characteristic data from various typical fault cases. Matching calculation involves comparing the current fault signature with the standard signatures in the signature library. The matching data indicates the degree of similarity between the current fault and each standard fault case. The feature vector is a data representation of the feature matrix converted into a one-dimensional array.

[0123] The generated fault evolution feature matrix is ​​matched against the standard features in the feature library. First, the feature matrix is ​​converted into a feature vector by expanding the matrix row-wise and normalizing it. For each standard fault case in the feature library, its feature vector is extracted and normalized in the same way. The cosine similarity method is used to calculate the similarity between the current fault feature vector and each standard feature vector. The calculation formula is the inner product of the two vectors divided by the product of their moduli. The matching value for each standard case is recorded to form a complete matching data sequence. This matching method based on vector similarity can effectively measure the similarity between fault features.

[0124] S512 : Select K groups of historical fault data with the highest matching degree from a preset standard fault feature library based on the matching degree data.

[0125] The K groups of historical fault data refer to the K fault cases with the highest matching scores selected from the feature library. The matching threshold is used to select fault cases with a sufficiently high matching score. Historical fault signatures contain complete information such as fault type, development process, and treatment method. Similarity sorting arranges fault cases in descending order based on matching scores.

[0126] Sort and filter the calculated matching data. First, sort all standard fault cases in descending order according to their matching values ​​to obtain a sorted fault case sequence. Starting from the head of the sequence, select the K fault cases with the highest matching values, where the K value is usually set to 3 to 5. For each selected fault case, extract its complete feature information, including the time, location, type, development process, and treatment method of the fault. This historical fault data provides an important reference for the analysis and treatment of the current fault. Selecting multiple sets of high-matching historical fault data can analyze the characteristics of the current fault from multiple perspectives and improve the accuracy of fault diagnosis.

[0127] S513 , analyzing the position distribution sequence and the deviation value sequence in the K groups of historical fault data, extracting common features, and generating a fault type feature template.

[0128] Common features refer to identical or similar characteristics exhibited across multiple historical fault cases. Fault type feature templates are standard feature descriptions constructed based on these common features. Position distribution features describe the distribution of key sampling points along the time axis. Deviation features describe the distribution of waveform deviations from normal conditions. Feature clustering is the process of grouping similar features.

[0129] The K selected groups of historical fault data were subjected to in-depth analysis and feature extraction. First, the position distribution sequences of each group of historical fault data were aligned, and the dynamic time warping algorithm was used to eliminate differences in time scales. The mean and standard deviation of the aligned sequences were calculated to obtain the statistical characteristics of the position distribution. The same alignment and statistical analysis process was performed on the deviation value sequence. The obtained statistical features were clustered, and the K-means clustering method was used to identify representative feature centers. A fault type feature template was constructed based on the feature center. The template contains feature descriptions in two dimensions: position distribution and deviation value. This template generation method based on multi-case analysis can extract the essential characteristics of the fault type.

[0130] S514 , calculating the degree of fit between the fault feature data link of the current fault and the fault type feature template. When the degree of fit is greater than a preset degree of fit threshold, determining that the current fault type is the same as the fault type corresponding to the fault type feature template.

[0131] The degree of fit refers to the degree of agreement between the current fault signature and the template signature. The degree of fit threshold is used to determine whether the degree of fit reaches an acceptable level. The fault signature data link contains complete signature information for the current fault. Feature matching is the process of comparing the current signature with the template signature.

[0132] The degree of fit between the feature data chain of the current fault and the feature template of the fault type is calculated. First, the current fault feature data chain is preprocessed, including data standardization and time alignment. The Euclidean distance between the current feature and the template feature is calculated. The smaller the distance, the higher the degree of fit. The specific calculation method is to calculate the distance between the position distribution feature and the deviation value feature separately, and then weighted sum them to obtain the overall degree of fit. The weight coefficient is determined according to the importance of the two types of features. Generally, the weight of the position distribution feature is higher. The calculated degree of fit is compared with the preset threshold. When the degree of fit exceeds the threshold, it is determined that the current fault is of the same type as the fault type corresponding to the template.

[0133] S515. Establish a fault development state transition diagram based on the time sequence relationship of each sub-chain in the fault evolution feature matrix. The state transition diagram is used to represent the complete process from fault occurrence to development.

[0134] A state transition diagram is a directed graph structure that describes system state changes. Nodes represent different fault states. Edges represent transitions between states. Transition probability refers to the probability of transitioning from one state to another. State duration refers to the length of time the system remains in a certain state.

[0135] A complete fault development state transition diagram is constructed based on the fault evolution feature matrix. Each subchain is first treated as an independent fault state, and the characteristic change relationships between adjacent subchains are analyzed. The direction and strength of state transitions are determined by calculating the characteristic differences between adjacent states. Transitions between states are represented by directed edges, with edge weights determined by the magnitude of the characteristic differences. For each state node, the statistical values ​​and duration of its characteristic parameters are recorded. All state nodes and transition edges are organized into a directed graph structure to form a complete state transition diagram. This graphical representation intuitively illustrates the fault development process and the characteristic changes at each stage.

[0136] S516 , calculating the transition probabilities between adjacent states in the fault development state transition diagram, and generating a state transition probability matrix.

[0137] The state transition probability represents the probability of a system transitioning from one state to another. The state transition probability matrix is ​​a two-dimensional matrix that describes the transition relationships between all states. Adjacent states refer to states that occur one after another during the fault development process. The transition interval refers to the time required for a state transition. The state duration indicates how long the system remains in a certain state.

[0138] The probability of state transition relationships in the fault development state transition diagram is calculated. First, the number of occurrences and duration of each state node are counted, and the stability index of the state is calculated. For any two adjacent states, the number of transitions between them is counted, and divided by the total number of occurrences of the starting state to obtain the transition probability between the two states. The transition probabilities between all state pairs are organized into a matrix form according to the state number sequence. The rows and columns of the matrix represent the starting state and the target state, respectively, and the matrix element values ​​represent the corresponding transition probabilities. The diagonal elements represent the probability that the state remains in the current state. This statistical-based probability calculation method can quantitatively describe the state transition law during the fault development process.

[0139] S517: Update the fault type feature template according to the state transition probability matrix.

[0140] The fault type feature template is a standard model that describes the characteristics of a specific fault type. The state transition feature refers to the characteristic changes exhibited during the state transition process. The template update rule defines how to adjust the template parameters based on the new transition probability. The feature weight refers to the importance of different feature parameters in the template.

[0141] The fault type feature template is updated based on the calculated state transition probability matrix. First, the main transition paths in the state transition probability matrix are analyzed to identify the critical state sequence during the fault development process. The weighted average of each state feature in the critical state sequence is calculated, with the weight coefficient determined by both the state stability and the transition probability. The calculated weighted features are fused with the original template features, and a recursive update method is used. The new feature value is equal to the weighted sum of the original feature value and the weighted feature. The updated template retains the stability of the original features while incorporating the new state transition features, improving the template's adaptability and accuracy.

[0142] S518: Detect the degree of fit between each sub-chain feature and the corresponding stage feature in the fault type feature template.

[0143] The subchain features represent the set of features at a specific stage in the fault development process. The corresponding stage features are the standard features in the template that correspond to the subchain's temporal position. The fit calculation rules define how feature similarity is calculated. The feature matching process compares actual features with standard features.

[0144] The fit of each subchain's features to the fault type feature template is tested. First, the subchain's temporal position in the fault development process is determined, and the standard features for the corresponding stage are extracted from the feature template. The subchain and standard features are preprocessed, including data normalization and time alignment. The Euclidean distance and cosine similarity between the two sets of features are calculated to comprehensively derive the fit index. The fit calculation takes into account both feature amplitude differences and distribution morphology differences. The same fit calculation process is performed for each subchain to obtain a complete fit sequence. This segment-by-segment fit test method enables a detailed assessment of the degree of feature fit at each stage of fault development.

[0145] S519: When the degree of fit is lower than a preset degree of fit threshold, the transition probability of the sub-chain feature in the state transition probability matrix is ​​calculated.

[0146] The fit threshold is a standard value for determining the degree of feature matching. Subchain features contain complete feature information at a specific stage of fault development. State transition probability refers to the probability of transitioning from one state to another. Feature deviation refers to the difference between the actual feature and the standard feature. Abnormal state transitions indicate unexpected state changes.

[0147] When the fit of a subchain is detected to be below a preset threshold, the transition characteristics of this abnormal state need to be analyzed. First, the difference between the subchain's characteristics and those of its preceding and succeeding subchains is calculated, and the degree of characteristic change is quantified using the Euclidean distance method. Based on this difference, similar transition patterns are searched in the state transition probability matrix, and the state transition probability corresponding to this subchain's characteristics is calculated. The specific calculation process matches the subchain's characteristics with the characteristics of each state in the matrix, determines the most similar state, and then extracts the transition probability row vector of this state in the matrix. This probability-based analysis method can assess the development trend of abnormal states.

[0148] S520: Update the fault type feature template according to the transition probability.

[0149] Transition probability represents the likelihood of a state transition. Template updating refers to adjusting template parameters based on new feature information. Feature fusion combines new features with existing ones. Update weights determine the influence of new features during the update process. Dynamic features refer to feature parameters that change over time.

[0150] The calculated transition probabilities are used to update the fault type feature template. First, the feature update weight coefficient is determined based on the transition probability. Features with higher transition probabilities are assigned greater update weights. For feature parameters that require updating, a weighted average method is used for feature fusion, with the new feature value being equal to the weighted sum of the original feature value and the current feature. The update process considers the temporal correlation of features, and feature updates in adjacent time periods are recursively propagated. The updated template maintains the basic structure of the original features while incorporating the newly identified abnormal feature information.

[0151] S521. Perform real-time monitoring and fault warning on waveform data based on the updated fault type feature template and state transition probability matrix.

[0152] Real-time monitoring refers to the continuous analysis and evaluation of waveform data. Fault early warning is the issuance of predictive warnings before a fault occurs. Warning thresholds are used to determine whether a warning signal is necessary. Feature trends refer to the changing trends of feature parameters over time. Warning levels indicate the severity of the fault risk.

[0153] Real-time monitoring and early warning are performed based on the updated feature template and transition probability matrix. First, feature extraction is performed on the collected waveform data, and the degree of fit between the current features and the template features is calculated. Simultaneously, the state transition probability matrix is ​​used to predict the feature development trend and calculate the probability of transitioning to a faulty state. When the degree of fit falls below the warning threshold or the transition probability exceeds the warning threshold, the early warning mechanism is triggered. The early warning information includes the current state assessment results, the predicted development trend, and recommended measures. This monitoring and early warning method based on feature templates and probabilistic models can promptly detect potential fault risks.

[0154] The coal mine power supply system in the embodiment of the present invention is described below from the perspective of hardware processing. Figure 6 , which is a schematic diagram of the structure of a physical device of a coal mine power supply system in an embodiment of the present application.

[0155] It should be noted that Figure 6 The structure of the coal mine power supply system shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0156] like Figure 6 As shown, the coal mine power supply system includes a central processing unit (CPU) 601, which can perform various appropriate actions and processes based on programs stored in a read-only memory (ROM) 602 or programs loaded from a storage unit 608 into a random access memory (RAM) 603, such as the methods described in the above embodiments. RAM 603 also stores various programs and data required for system operation. CPU 601, ROM 602, and RAM 603 are interconnected via a bus 604. An input / output (I / O) interface 605 is also connected to bus 604.

[0157] The following components are connected to the I / O interface 605: an input section 606 including an audio input device, push button switches, and the like; an output section 607 including a liquid crystal display (LCD), an audio output device, indicator lights, and the like; a storage section 608 including a hard disk and the like; and a communication section 609 including a network interface card such as a LAN (Local Area Network) card or a modem. The communication section 609 performs communication processing via a network such as the Internet. A drive 610 is also connected to the I / O interface 605 as needed. Removable media 611, such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, is installed in the drive 610 as needed, so that computer programs read from the removable media can be installed in the storage section 608 as needed.

[0158] In particular, according to an embodiment of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, an embodiment of the present invention includes a computer program product comprising a computer program carried on a computer-readable medium, the computer program including a computer program for executing the methods illustrated in the flowcharts. In such an embodiment, the computer program can be downloaded and installed from a network via the communication section 609 and / or installed from removable media 611. When executed by the central processing unit (CPU) 601, the computer program performs the various functions defined in the present invention.

[0159] It should be noted that specific examples of computer-readable storage media may include, but are not limited to, an electrical connection having one or more conductors, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, portable compact disc read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0160] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present invention. Each box in the flowchart or block diagram can represent a module, program segment, or part of the code, and the above-mentioned module, program segment, or part of the code contains one or more executable instructions for implementing the specified logical functions. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings.

[0161] Specifically, the coal mine power supply system of this embodiment includes a processor and a memory. The memory stores a computer program. When the computer program is executed by the processor, the method for quickly locating a grounding fault in a coal mine power supply system provided in the above embodiment is implemented.

[0162] As another aspect, the present invention further provides a computer-readable storage medium, which may be included in the coal mine power supply system described in the above embodiments, or may exist independently and not be incorporated into the coal mine power supply system. The storage medium carries one or more computer programs, and when executed by a processor of the coal mine power supply system, the coal mine power supply system implements the method for rapidly locating a ground fault in a coal mine power supply system provided in the above embodiments.

[0163] As described above, 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 above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.

[0164] As used in the above embodiments, the term “when” may be interpreted to mean “if” or “after” or “in response to determining that” or “in response to detecting that”, depending on the context. Similarly, the phrases “upon determining that” or “if (stated condition or event) is detected” may be interpreted to mean “if determining that” or “in response to determining that” or “upon detecting (stated condition or event)” or “in response to detecting (stated condition or event)”, depending on the context.

[0165] Those skilled in the art will appreciate that all or part of the process steps in the above-described method embodiments can be implemented by a computer program instructing the relevant hardware. The program can be stored in a computer-readable storage medium, and when executed, the program can include the process steps in the above-described method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM or random access memory (RAM), magnetic disks, or optical disks.

Claims

1. A method for quickly locating grounding faults in a coal mine power supply system, characterized in that: Applied to a coal mine power supply system, the method includes: Obtaining real-time zero-sequence current values ​​at each junction box in the coal mine power supply system; When the real-time zero-sequence current value at any junction box exceeds the preset current threshold, the sampling point where the real-time zero-sequence current value first exceeds the preset current threshold is taken as the fault point, and the voltage waveform data and current waveform data within a preset time period before and after the fault point are collected; Calculating the voltage distortion rate and the phase shift of each phase voltage based on the voltage waveform data and the current waveform data; determining a fault type based on a change trend of the distortion rate and the phase offset; Perform time-sequence comparison and analysis on the zero-sequence current values ​​collected at each junction box at the time of the fault to obtain the time-sequence characteristics of the zero-sequence current propagation between each detection point; Acquire a zero-sequence current propagation characteristic template corresponding to the fault type from a preset fault feature library, wherein the zero-sequence current propagation characteristic template includes a propagation delay reference value and an amplitude attenuation reference value of the zero-sequence current between adjacent detection points under different fault types; The actual propagation delay value and the actual amplitude attenuation value between adjacent junction boxes are calculated based on the propagation timing characteristics, wherein the actual propagation delay value is obtained by the zero-crossing time difference of the zero-sequence current waveforms at adjacent junction boxes, and the actual amplitude attenuation value is obtained by the ratio of the maximum amplitudes of the zero-sequence currents at adjacent junction boxes; Calculating the matching degree of the actual propagation delay value and the actual amplitude attenuation value with the propagation delay reference value and the amplitude attenuation reference value in the zero-sequence current propagation characteristic template to obtain a matching degree value for each cable section; The faulty cable section where the fault point is located is determined according to the matching value.

2. The method according to claim 1, characterized in that After the step of locating the faulty cable section where the fault point is located according to the propagation time sequence characteristics and the fault type, the method further includes: Obtaining waveform information of each cycle in the voltage waveform data and the current waveform data respectively, and calculating key characteristic parameters of each cycle based on the waveform information, wherein the key characteristic parameters include an amplitude mutation point parameter, a phase mutation point parameter, and a waveform distortion parameter; Converting the key feature parameters into key feature codes, wherein the key feature codes are used to characterize the waveform information of each cycle; Using a preset standard feature code as a reference feature code, performing a difference comparison between the key feature code and the reference feature code to obtain difference data; According to the difference data, a frequency band having a difference greater than a preset difference threshold is screened out, and the frequency band is marked as a fault characteristic frequency band; According to the time sequence arrangement of the fault characteristic cycle segments, a fault characteristic data chain is generated, and the fault characteristic data chain includes the starting time, peak time and recovery time of each fault characteristic cycle segment.

3. The method according to claim 2, characterized in that The step of converting the key feature parameter into a key feature code specifically includes: Divide the waveform data of each cycle into N equally spaced sampling segments; Calculating the waveform change rate of adjacent sampling points in each sampling segment; Marking sampling points where the waveform change rate is greater than a preset change rate threshold as key sampling points; Generate a position distribution sequence according to the distribution positions of the key sampling points in the cycle; Calculating the deviation between the actual waveform value at the key sampling point and the standard sinusoidal waveform value to generate a deviation value sequence; The position distribution sequence and the deviation value sequence are combined and encoded to obtain the key feature code.

4. The method according to claim 2, characterized in that After the step of locating the faulty cable section where the fault point is located according to the propagation time sequence characteristics and the fault type, the method further includes: generating a difference waveform spectrum diagram according to the difference degree data, wherein the difference waveform spectrum diagram is used to characterize the waveform spectrum distribution during the fault process; Dividing the fault feature data chain into N subchains according to the spectrum distribution characteristics in the difference waveform spectrum diagram, each subchain corresponding to an independent fault development stage; Counting the changing trends of the position distribution sequence and the deviation value sequence of the key sampling points in each of the sub-chains to generate a fault evolution feature matrix; Calculating the matching degree between the fault evolution feature matrix and a preset standard fault feature library to obtain matching degree data; According to the matching degree data, K groups of historical fault data with the highest matching degree are selected from the preset standard fault feature library; Analyze the position distribution sequence and deviation value sequence in the K groups of historical fault data, extract common features, and generate a fault type feature template; The degree of fit between the fault feature data chain of the current fault and the fault type feature template is calculated, and when the degree of fit is greater than a preset degree of fit threshold, it is determined that the current fault type is the same as the fault type corresponding to the fault type feature template.

5. The method according to claim 4, characterized in that After the step of locating the faulty cable section where the fault point is located according to the propagation time sequence characteristics and the fault type, the method further includes: According to the time sequence relationship of each sub-chain in the fault evolution characteristic matrix, a fault development state transition diagram is established, wherein the state transition diagram is used to represent the complete process from the occurrence to the development of the fault; Calculate the transition probability between adjacent states in the fault development state transition diagram to generate a state transition probability matrix; The fault type feature template is updated according to the state transition probability matrix.

6. The method according to claim 5, characterized in that The step of updating the fault type feature template according to the state transition probability matrix specifically includes: Detecting the degree of fit between each sub-chain feature and the corresponding stage feature in the fault type feature template; When the fitness is lower than the preset fitness threshold, calculating the transition probability of the sub-chain feature in the state transition probability matrix; Updating the fault type feature template according to the transition probability; Real-time monitoring and fault warning of waveform data are carried out based on the updated fault type feature template and state transition probability matrix.

7. A coal mine power supply system, characterized in that: The coal mine power supply system includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to enable the coal mine power supply system to execute the method described in any one of claims 1-6.

8. A computer-readable storage medium comprising instructions, characterized in that: When the instruction is executed on a coal mine power supply system, the coal mine power supply system is caused to execute the method according to any one of claims 1 to 6.

9. A computer program product, characterized in that When the computer program product is run on a coal mine power supply system, the coal mine power supply system is enabled to execute the method according to any one of claims 1 to 6.

Images