A method for low-voltage leakage protection in mines based on adaptive working conditions

Abstract

The present invention relates to the technical field of leakage protection, and provides a method for low-voltage leakage protection for mines based on working condition self-adaptation, comprising: detecting leakage resistance in real time and determining the working condition of a power grid; matching a target detection algorithm in a preset detection algorithm according to the working condition of the power grid; wherein the preset detection algorithm comprises: an integrated zero-sequence current mutation method, a zero-sequence current and zero-sequence voltage phase method, and a zero-sequence current magnitude phase comparison method; establishing a unified criterion according to the target detection algorithm to coordinate the action logic of a main switch and each branch switch, and determining the leakage degree of the leakage resistance; and executing switch locking on the leakage branch in sequence according to the leakage degree.

Description

Technical Field

[0001] The present invention relates to the field of coal mine leakage detection, and in particular to a working condition-based self-adaptive low-voltage leakage protection method for mines. Background Art

[0002] After the main feeder switch, the power is fed through several stages of low-voltage tiered feeder switches. Before closing and energizing, each stage uses an additional power supply DC detection principle to check the insulation level of cables and electrical equipment connected to the grid. If the insulation level falls below a set value, a leakage lockout signal is issued to prevent the switch from closing and energizing. After closing, the low-voltage main feeder switch uses DC without an additional power supply to check the grid insulation level, while each tiered feeder switch uses the zero-sequence power directional principle to check the grid insulation level. If the grid insulation level falls below the set value, the feeder switches at each stage trip and cut off power according to the inverse time principle.

[0003] (1) In order to ensure the vertical selectivity of the feeder switch, each time the grid adds or reduces a feeder switch, the leakage protection action time of the feeder switches above it needs to be reset, which is a cumbersome operation. Moreover, in order to achieve differential selectivity, the action time of the leakage protection of the feeder switch is artificially increased, which increases the risk of electric shock and leakage. In particular, the closer the leakage occurs to the low-voltage main feeder switch, the longer the waiting time for the protection action.

[0004] (2) The leakage resistance detection methods for the main switch and branch switches are not standardized. The main switch uses the "additional DC voltage method" to detect leakage resistance, while the branch switches use the "zero-sequence voltage method" to detect leakage resistance. The former has high accuracy but slow response speed, while the latter has a fast response speed, but the accuracy is poor because the zero-sequence voltage is affected by the system capacitance and grid voltage. Therefore, the leakage resistance detection criteria are not unified, and branch switches frequently refuse to operate and the main switch malfunctions frequently occur.

[0005] (3) The single “zero-sequence power direction” selective leakage protection principle will cause the short branch without leakage to be easily misoperated.

[0006] (4) It is easy to malfunction when there is a big difference in the length of the short line and the unbalanced current is large.

[0007] (5) It cannot solve the problem of decentralized leakage protection. Summary of the Invention

[0008] This application proposes a method for low-voltage leakage protection for mines based on working condition self-adaptation, which is used to actively switch algorithms based on the real-time working conditions of the power grid through three preset detection algorithms, thereby solving the technical problem that a single method has poor adaptability and is prone to errors under different working conditions.

[0009] In the first aspect, the present application proposes a method for low-voltage leakage protection for mines based on working condition self-adaptation, comprising:

[0010] By detecting leakage resistance in real time and determining the operating conditions of the power grid;

[0011] According to the grid operation conditions, the target detection algorithm is matched with the preset detection algorithm; wherein the preset detection algorithm includes: integrated zero-sequence current mutation method, zero-sequence current and zero-sequence voltage phase method, and zero-sequence current magnitude phase comparison method;

[0012] Based on the target detection algorithm, a unified criterion is established to coordinate the action logic of the main switch and each branch switch, and determine the leakage degree of the leakage resistor;

[0013] According to the degree of leakage, the switches of the leakage branches are locked in turn.

[0014] In this scheme, the operating condition of the power grid is determined by judging the resistance with leakage in the coal mine. Different from the traditional single zero-sequence current detection method to judge leakage, the main switch is simply tripped to achieve the distinction and control of different power grid conditions. This application adopts a method of matching the actual working condition and the detection algorithm. According to the actual working condition, the most suitable detection algorithm is determined, and the action logic of the main switch and branch of the entire power grid is constructed. In the case of multiple leakage resistors, different leakage resistors are located and the degree of leakage (the degree of affecting the safety and stability of the power grid) is calculated. Then, according to the degree of leakage, the corresponding leakage branch is controlled to be closed, thereby improving the power supply effect of the coal mine while minimizing the impact on the stability, safety and production of the coal mine.

[0015] Furthermore, when detecting the leakage resistance in real time, the current power grid is detected based on the zero-sequence voltage correction method to determine the power grid branch with the leakage resistance in the current power grid.

[0016] Furthermore, the method further includes, before matching the target detection algorithm in the preset detection algorithm according to the grid operation condition, the following steps:

[0017] Pre-set leakage scenarios; leakage scenarios include: sudden leakage, continuous leakage under steady-state operation, and interference leakage;

[0018] Configure corresponding detection matching rules based on the leakage scenario;

[0019] When the leakage scenario is sudden leakage, a first matching rule based on a dynamic threshold is set based on the current change rate of the zero-sequence current of each branch of the power grid;

[0020] When the leakage scenario is continuous leakage under steady-state operation, a second matching rule based on phase deviation is set based on the phase difference between each branch of the power grid;

[0021] When the leakage scenario is interference leakage, a third matching rule based on a dynamic threshold is set based on the zero-sequence current proportion and continuous power frequency cycle of each branch of the power grid.

[0022] Furthermore, matching the target detection algorithm in the preset detection algorithm includes:

[0023] Real-time acquisition of the three-phase voltage and zero-sequence voltage of the power grid, and simultaneous acquisition of the zero-sequence current of each branch and the total zero-sequence current;

[0024] Based on the preset ADC converter, the three-phase voltage and zero-sequence voltage of the power grid, the zero-sequence current of each branch and the total zero-sequence current are converted into detection digital signals;

[0025] Based on the detected digital signal, the insulation resistance value, the distribution coefficient of the distributed capacitance and the grounding method of each grounding point in the power grid are calculated, and the power grid status assessment is output;

[0026] Determine the time domain waveform characteristics of zero-sequence voltage and zero-sequence current according to the grid assessment status;

[0027] Determine the leakage type based on the time domain waveform characteristics; leakage types include high-resistance leakage, intermittent leakage, and multi-point scattered leakage;

[0028] Configure the matching rules of the target detection algorithm based on the leakage type.

[0029] Furthermore, the target detection algorithm is used to establish a unified criterion to coordinate the operation logic of the main switch and each branch switch, including:

[0030] When the target detection algorithm is the integrated zero-sequence current mutation method;

[0031] Set a dynamic time window, which includes a short-term detection window that captures sudden leakage through mutation rate, and a long-term integration window that identifies slowly changing high-resistance leakage through adaptive threshold calculation. The adaptive threshold calculation system performs capacitance compensation calculation and load current compensation calculation.

[0032] According to the dynamic time window, the leakage priority of each branch is sorted, and the main switch locking conditions and differential coordination conditions are set;

[0033] According to the main switch locking conditions and differential coordination conditions, set the pre-analog filter and digital sliding filter, determine the fault waveform of each branch, and configure the switch action logic of each branch in the preset load feature library according to the fault waveform.

[0034] Furthermore, the target detection algorithm is used to establish a unified criterion to coordinate the operation logic of the main switch and each branch switch, including:

[0035] When the target detection algorithm is the zero-sequence current and zero-sequence voltage phase method;

[0036] Set a windowed FFT algorithm to calculate the fundamental phase difference between the zero-sequence current and zero-sequence voltage of each branch, and set a dynamic phase threshold; the dynamic phase threshold also includes a capacitance compensation threshold and a reference offset threshold;

[0037] Determine the phase deviation of each leakage branch according to the dynamic phase threshold;

[0038] Configure the automatic locking conditions and switch action threshold of the main switch according to the phase deviation;

[0039] According to the automatic blocking conditions and the switching action threshold, the harmonic suppressor is set, the initial phase reference is determined, and the switching action logic of each branch is adaptively configured based on the initial phase reference and load fluctuation.

[0040] Furthermore, the target detection algorithm is used to establish a unified criterion to coordinate the operation logic of the main switch and each branch switch, including:

[0041] When the target detection algorithm is the zero-sequence current magnitude-phase method;

[0042] Set up a dynamic proportional threshold mechanism based on capacitance compensation mechanism and minimum guarantee threshold;

[0043] According to the dynamic proportional threshold mechanism, set the short-time judgment condition and the long-time integral judgment condition;

[0044] According to the short-time judgment conditions and long-time integral judgment conditions, configure the automatic blocking conditions of the main switch tripping function based on the branch scoring ratio;

[0045] According to the automatic blocking condition of the main switch tripping function based on the branch bisection ratio, the average filter of each branch is set, the priority action satisfied by each branch is determined, and the switch action logic of each branch is configured based on the priority action satisfied by each branch.

[0046] Furthermore, determining the leakage degree of the leakage resistor includes:

[0047] Obtain the power distribution network topology of the power grid and establish a leakage current conduction model;

[0048] According to the leakage current conduction model, calculate the propagation attenuation rate and path complexity corresponding to the current leakage resistance;

[0049] Determine the leakage inducing conditions based on the propagation attenuation rate;

[0050] Determine the conditional weight of leakage risk based on path complexity;

[0051] Based on the leakage inducing conditions and condition weights, a normalized risk calculation model is constructed to determine the multi-dimensional assessment value of the leakage degree.

[0052] Furthermore, the switching and locking of the leakage branches in sequence according to the degree of leakage may further include:

[0053] Setting a signal trigger mechanism; wherein the signal trigger mechanism includes a first trigger mechanism, a second trigger mechanism and a third trigger mechanism;

[0054] The first trigger mechanism is triggered when the leakage level is such that the branch zero-sequence current mutation amount is greater than the preset mutation value;

[0055] The second trigger mechanism is triggered when the leakage level is such that the phase offset between the zero-sequence current and the zero-sequence voltage is greater than a preset phase offset value;

[0056] The second trigger mechanism is triggered when the proportion of the zero-sequence current of the leakage branch is greater than the preset branch zero-sequence current reference value.

[0057] Furthermore, the switching and locking of the leakage branches in sequence according to the degree of leakage may further include:

[0058] According to the leakage degree, determine the drop rate of the insulation resistance of each branch, the ambient humidity and the historical leakage events;

[0059] Based on the drop rate of each branch insulation resistance, ambient humidity, and historical leakage events, a leakage prediction mechanism based on the LSTM neural network was established.

[0060] According to the leakage estimation mechanism, leakage fault prediction is carried out;

[0061] According to the leakage fault prediction, switchable branch paths are configured for target branches with high failure rates in turn.

[0062] Other features and advantages of the present invention will be described in the following description, and in part will become apparent from the description, or will be understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description and the accompanying drawings.

[0063] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0065] Figure 11 is an implementation diagram of a method for self-adaptive low-voltage leakage protection for mining in accordance with an embodiment of the present invention;

[0066] Figure 2 is a process flow chart of a method for determining a power grid branch in an embodiment of the present invention;

[0067] Figure 3 This is a flowchart of the steps for configuring corresponding detection matching rules according to the leakage scenario in an embodiment of the present invention;

[0068] Figure 4 A process diagram of target algorithm matching in an embodiment of the present invention;

[0069] Figure 5 A process diagram for determining the operation logic of the coordinated main switch and each branch switch in an embodiment of the present invention;

[0070] Figure 6 FIG2 is a diagram showing a process of determining a processing method in which a target detection algorithm is a zero-sequence current and zero-sequence voltage phase method in an embodiment of the present invention;

[0071] Figure 7 2 is a diagram illustrating an implementation process when the target detection algorithm in an embodiment of the present invention is a zero-sequence current magnitude-phase ratio algorithm;

[0072] Figure 8 1 is a diagram showing a calculation process of leakage level in an embodiment of the present invention;

[0073] Figure 9 A diagram showing the process of controlling the switch of a branch circuit in an embodiment of the present invention;

[0074] Figure 10 This is a branch switching operation diagram in an embodiment of the present invention. DETAILED DESCRIPTION

[0075] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0076] The traditional leakage detection method mainly uses a single detection method to determine whether there is leakage. In terms of leakage treatment, the main switch is usually cut off, which will cause a large-scale power outage. Based on the shortcomings of the traditional solution, this application proposes a working condition-based adaptive low-voltage leakage protection method for mines. The specific method is described in detail in the following. Figure 1 :

[0077] First, in step S01, leakage resistance detection is performed. Different from the direct detection of current and voltage in traditional technology, this application performs overall detection of the power grid through a pre-configured method. Then, according to the branches existing in the power grid, the value of the leakage resistance is measured in real time through the resistance detection circuit. According to the resistance value and the related current and voltage, the operating condition of the power grid is judged in step S02. The operating condition of the power grid will be analyzed according to different situations such as normal operation, slight leakage (sudden leakage, interference leakage and continuous leakage) and severe leakage (sudden leakage, interference leakage and continuous leakage). In actual implementation, the leakage condition of the resistance of different branches will be detected to determine the operating condition of the power grid, and the corresponding detection algorithm will be configured according to each branch.

[0078] In step S03, for different grid operating conditions, including main line operation and branch line operation, corresponding measurement algorithms are matched. In this application, three measurement methods are set: S031 integrated zero-sequence current mutation method, S032 zero-sequence current and zero-sequence voltage phase method, and S033 zero-sequence current magnitude phase ratio method;

[0079] In actual implementation, the three principles are intelligently converted according to the on-site working conditions, which fundamentally solves the problem that a single principle cannot be applied to all occasions and greatly improves the reliability of leakage protection.

[0080] In step S04, a unified judgment criterion is constructed based on the selected target detection algorithm, and an action logic network for the main switch and branch switches is established. In step S05, the leakage level is determined by sequentially analyzing the leakage level and impact on the power grid of different leakage resistors in the branch and central circuits. In actual implementation, centralized control of the branch switches and the main switch is implemented. The main switch and branch switches use a unified judgment criterion, eliminating false operations caused by inconsistent criteria. The leakage level of each branch is analyzed based on the comparative analysis of the capacitance current of each branch, achieving higher accuracy.

[0081] In S06, the leakage branch is locked and switched according to the leakage level, enabling the branch to be disconnected and the main switch to be closed. Specifically, the principle of prioritizing branches with greater leakage current addresses distributed leakage in branches: when multiple leakage points occur, the determined leakage level can determine the magnitude of the branch zero-sequence current and the change in that branch zero-sequence current. If multiple branches are detected meeting the tripping conditions, the priority branch to be tripped is determined based on the change in branch zero-sequence current, i.e., the degree of impact and risk to the overall power grid. The branch with the highest leakage level is tripped, and the main switch is never tripped by mistake.

[0082] In actual implementation, the grid's operating status is determined in real time using multiple electrical parameters. These parameters include: zero-sequence current mutation rate, which captures transient events such as equipment startup and shutdown, short-circuit shocks, and setting rules for identifying sudden operating conditions. Zero-sequence voltage stability, which sets stability per unit time, identifies steady-state conditions. If the proportion of zero-sequence current in two or more branches exceeds preset current and phase values, a multi-point leakage condition is identified.

[0083] Example 2:

[0084] This application provides a method for determining a power grid branch. Figure 2 The power grid in the coal mine is composed of multiple branches, each of which has load and insulation resistance. By configuring a zero-sequence voltage transformer, the zero-sequence voltage of the power grid is collected in real time and converted into corresponding electrical data. Then, the zero-sequence voltage correction method is used to calculate the leakage resistance and determine and locate the fault. Compared with the traditional zero-sequence voltage method, it can achieve higher measurement effects in the case of unbalanced power grid interference. During leakage detection, the accuracy of branch positioning can avoid misidentification caused by unbalanced power grid. The correction algorithm can eliminate power grid harmonics and unknown constant components.

[0085] When detecting leakage resistance in real time, the current power grid is detected based on the zero-sequence voltage correction method to determine the power grid branches with leakage resistance in the current power grid. Dynamic compensation algorithm.

[0086] The zero-sequence voltage correction method achieves accurate detection of leakage resistance by compensating for the interference of system distributed capacitance and voltage fluctuations on measurement. Its technical principles can be divided into the following three levels: establishing a parameter equivalent circuit model of the voltage pair; deducting the capacitive component from the total zero-sequence current to obtain the true leakage current; calculating the leakage resistance based on the corrected parameters; and finally realizing fault current vector analysis.

[0087] Example 3:

[0088] This application proposes a step to configure corresponding detection matching rules according to the leakage scenario, see Figure 3 ,

[0089] The present application first pre-sets the leakage scenario; wherein, the leakage scenario includes: sudden leakage scenarios, for example, fast current mutation and short duration, continuous leakage under steady-state operation, such as slow current change and stable phase difference caused by insulation aging, and interference leakage, such as zero-sequence current caused by equipment start-stop interference with a low proportion and periodic occurrence; the leakage data of each branch sensed by the zero-sequence current sensor will be divided in the preset leakage scenario, and then, according to the leakage scenario, the corresponding detection matching rules are configured, and different rules and weights are triggered for different scenarios. This is based on the detection rule configuration module, which can improve the accuracy and usage scenarios of leakage detection, prevent recognition errors caused by scene confusion, and also realize the linkage of dynamic switching of rules through scene presets.

[0090] In the specific detection rule configuration process, according to the dynamic threshold matching unit, when the leakage scenario is sudden leakage, a first matching rule based on the dynamic threshold is set based on the current change rate of the zero-sequence current of each branch of the power grid;

[0091] According to the phase deviation matching unit, when the leakage scenario is continuous leakage under steady-state operation, a second matching rule based on phase deviation is set based on the phase difference between each branch of the power grid;

[0092] When the leakage scenario is interfering, the power frequency cycle matching unit sets a third matching rule based on dynamic thresholds, taking into account the zero-sequence current ratio and the power frequency cycle duration of each grid branch. In interfering leakage scenarios, this simultaneous consideration of the zero-sequence current ratio and the power frequency cycle duration enables multi-dimensional judgment and higher accuracy.

[0093] I built it myself, building a scenario-rule mapping system:

[0094] For example, it can capture the differential characteristics of transient processes, determine the sudden change signal, and then determine the sudden fault through the steep increase characteristic of the current change rate;

[0095] For example, for steady-state continuous leakage, phase mismatch caused by insulation degradation can be detected based on branch phase deviation.

[0096] For example, for intermittent interference, combined with time-frequency domain joint analysis, the spatial distribution characteristics of the zero-sequence current ratio and the time continuity characteristics are collected to distinguish the authenticity.

[0097] Then, through the dynamic threshold algorithm, the grid operation status is tracked in real time, so that the threshold value of the detection rule is automatically adjusted according to the system operating conditions, ensuring a dynamic balance between detection sensitivity and reliability.

[0098] This application uses multi-dimensional electrical parameters such as current change rate, phase deviation, and zero-sequence current ratio for joint analysis, breaking through the limitations of single threshold detection. The combination of dynamic threshold and phase deviation covers transient, steady-state and intermittent leakage scenarios.

[0099] Combined with the introduction of dynamic thresholds, it can adapt to grid load fluctuations and reduce the risk of false alarms and missed alarms.

[0100] Duration criteria are set for interfering leakage, and continuous monitoring at the power frequency cycle level can effectively distinguish between instantaneous interference and real leakage. Phase difference analysis can eliminate the influence of steady-state interference such as harmonics and enhance system robustness.

[0101] Example 4:

[0102] This application proposes steps for how to perform target algorithm matching, see 4, for matching the target detection algorithm in the preset detection algorithm, specifically including:

[0103] First, the three-phase voltage and zero-sequence voltage of the power grid are collected in real time, and the zero-sequence current of each branch and the total zero-sequence current are collected simultaneously. By combining the total zero-sequence current, multi-parameter fusion analysis is realized. In actual operation, different currents and voltages need to be configured with corresponding sensors, such as Figure 4 In the specific implementation, the spatial distribution of branch zero-sequence current is combined to construct the grid topology parameter matrix and realize the distribution and positioning of each component in the grid.

[0104] Then, based on a preset ADC converter, the three-phase voltage and zero-sequence voltage of the power grid, as well as the zero-sequence current of each branch and the total zero-sequence current are converted into detection digital signals, preferably an ADC converter with a resolution of 16 bits or above.

[0105] The status assessment module detects digital signals, calculates the insulation resistance value, the distribution coefficient of distributed capacitance, and the grounding method of each grounding point in the power grid, and outputs a grid status assessment. The digital signal detection outputs the sensing parameters used to calculate the insulation resistance value and distributed capacitance. The calculated insulation resistance value and the distribution coefficient of distributed capacitance are the sensing coefficients of the insulation resistance value and the distributed capacitance in the power grid composed of each branch and main line, achieving an accurate assessment of the power grid status. The insulation resistance calculation can reflect the overall degree of insulation degradation, the distribution coefficient analysis and location can determine the source of capacitive coupling interference, and the grounding method identification can trace the source of leakage faults.

[0106] Based on the grid assessment status, the time-domain waveform characteristics of the zero-sequence voltage and current are then determined. These characteristics can be used to identify the type of leakage. Specifically, singular value decomposition (SWD) is performed on the zero-sequence voltage / current time-domain waveforms to determine high-frequency transient components, reflecting the breakdown characteristics of high-resistance leakage. Periodic fluctuation components are then determined to correspond to the intermittent conduction of intermittent leakage. Multimodal superposition components are then used to identify the parallel paths of multi-point leakage.

[0107] Determine the leakage type based on the time domain waveform characteristics; leakage types include high-resistance leakage, intermittent leakage, and multi-point scattered leakage;

[0108] Configure the matching rules of the target detection algorithm based on the leakage type.

[0109] In the specific implementation, the algorithm matching rules are combined with wavelet transform to enhance the weak fault characteristics of the leakage signal, thereby solving the problem that the high-resistance leakage signal is easily submerged by noise in specific scenarios.

[0110] Then, a mapping knowledge base is built to map leakage types and detection algorithms to each other;

[0111] In the case of high-resistance leakage, the adaptive threshold tracking algorithm can be triggered to automatically adjust the detection sensitivity based on the insulation resistance change rate;

[0112] The intermittent leakage current activates the time series matching algorithm, the fault repetition pattern is determined by waveform similarity, distributed collaborative diagnosis is initiated by multi-point leakage current, and the fault branch is located by using the branch current phase relationship.

[0113] Example 5:

[0114] This application proposes a method for determining the operation logic of the coordinated main switch and each branch switch. Figure 5 The purpose is to establish a unified criterion to coordinate the action logic of the main switch and each branch switch based on the target detection algorithm. In the specific implementation, if the target detection algorithm is the integrated zero-sequence current mutation method.

[0115] In traditional technology, the fixed time window leakage protection method cannot distinguish between short-term sudden leakage and long-term slowly changing high-resistance leakage, so a dynamic time window will be configured.

[0116] Set a dynamic time window, which includes a short-term detection window that captures sudden leakage through mutation rate, and a long-term integration window that identifies slowly changing high-resistance leakage through adaptive threshold calculation. The adaptive threshold calculation system performs capacitance compensation calculation and load current compensation calculation.

[0117] In actual implementation, the zero-sequence current mutation rate can be used to characterize the transient energy of the fault based on the short-time detection window. Then, a differential operation is used to determine the steep front feature of the sudden leakage, and the mutation rate threshold is combined to quickly trigger the protection.

[0118] In actual implementation, based on the long-time window integration window, in the slowly varying leakage scenario, the energy accumulation effect is evaluated based on the current time domain integral value, and the integral threshold is dynamically corrected in combination with the insulation resistance degradation model, thereby avoiding the fire risk caused by the continuous release of leakage energy. The two provide corresponding scenario inputs for different working conditions, forming a closed loop.

[0119] Based on a dynamic time window, each branch's leakage priority is ranked, and the main switch lockout conditions and differential coordination conditions are set. In specific implementation, priority sorting constructs a three-dimensional feature vector based on the branch leakage current amplitude, rate of change, and duration. The severity of leakage in different branches is calculated using the entropy weight method to generate corresponding priority queues. This then determines the main switch lockout and differential coordination conditions. During differential coordination, if a valid branch-level protection action is detected, the main switch delays the activation of the lockout logic. Based on status feedback, the main switch trips only if the branch protection fails, forming a master-slave protection hierarchy.

[0120] Based on the master switch locking conditions and differential coordination, pre-analog filtering and digital sliding filtering are configured. The fault waveforms of each branch are determined, and the switch action logic for each branch is configured based on the fault waveforms in the preset load signature library. In specific implementation, an active filter with adjustable dielectric frequency automatically configures the passband range based on the load characteristics, eliminating interference signals generated in specific frequency bands, such as inverter harmonics.

[0121] In the actual implementation process, a digital sliding filter technology solution will also be adopted. Based on the improved weighted moving average algorithm, the high-frequency components of the fault waveform will be retained in the actual processing process, while the random noise will be smoothed and the filter coefficient will be optimized in real time through residual analysis.

[0122] For example, a database containing typical load waveform characteristics for motors, LED lighting, and variable-frequency equipment can be built to dynamically time-warp fault waveforms collected in real time. If the waveform characteristics exceed a threshold similarity with the startup / switching patterns of normal loads in the database, the protection output is automatically suppressed, enabling intelligent identification of operating conditions.

[0123] The capacitive leakage current component is calculated in real time based on the distributed capacitance parameters of the power grid and stripped from the total zero-sequence current to eliminate false alarms. The zero-sequence current generated by load asymmetry is inferred by the three-phase imbalance, and the reference threshold for the action is dynamically corrected to ensure that the sensitivity of high-resistance leakage detection is not affected by load changes, improving detection effectiveness in complex scenarios.

[0124] Example 6:

[0125] This application provides a target detection algorithm for the zero sequence current and zero sequence voltage phase method, see Figure 6 Through the zero-sequence current and zero-sequence voltage phase method, the capacitance load fluctuation caused by the fixed phase threshold and the measurement error caused by harmonic interference are handled.

[0126] If the target detection algorithm is the zero-sequence current and zero-sequence voltage phase method;

[0127] A windowed FFT algorithm is used to improve the phase extraction accuracy of the fundamental wave. The fundamental phase difference between the zero-sequence current and zero-sequence voltage of each branch is calculated, and a dynamic phase threshold is set. The dynamic phase threshold also includes a capacitance compensation threshold and a reference offset threshold. This addresses the phase jump problem caused by harmonics in traditional methods. The windowed FFT algorithm is used to refine the spectrum of the zero-sequence current / voltage signal. The windowed FFT algorithm suppresses spectrum leakage based on the Hanning window function, accurately extracts the amplitude and phase of the fundamental component, and dynamically adjusts the window length to ensure the time accuracy and precision of the phase difference calculation by balancing the requirements of frequency resolution and real-time performance. Furthermore, based on the inherent phase offset caused by capacitive current in the calculation of distributed capacitance parameters in the power grid, the offset can be deducted from the measured phase in actual measurements, thus resolving false alarms caused by capacitive coupling.

[0128] Determine the phase deviation of each leakage branch according to the dynamic phase threshold;

[0129] Configure the automatic locking conditions and switch action threshold of the main switch according to the phase deviation;

[0130] According to the automatic blocking conditions and the switching action threshold, the harmonic suppressor is set, the initial phase reference is determined, and the switching action logic of each branch is adaptively configured based on the initial phase reference and load fluctuation.

[0131] During operation, a phase baseline reference value is established through a sliding average of historical data. If the load causes the baseline reference value to drift, the boundary can be dynamically adjusted to prevent false leakage failures. In actual operation, the phase deviation is the normalized deviation between the measured phase difference and the baseline value. The severity of the leakage can be mapped based on the deviation, achieving a graded response. The harmonic suppressor uses a multi-order band-stop filter bank, with notch frequencies designed for characteristic harmonics such as the 3rd, 5th, and 7th harmonics to suppress harmonic contamination of the fundamental phase measurement.

[0132] Automatically collect multi-cycle phase data during fault-free periods and generate steady-state reference values using the Kalman filter algorithm to provide a reliable reference for fault detection. Build a load fluctuation feature library to correlate different load types (such as impact loads and periodic loads) with phase fluctuation patterns.

[0133] When load fluctuation characteristics are detected, the phase deviation judgment threshold is automatically relaxed and the action delay is extended to avoid false action triggered by normal operating condition fluctuations;

[0134] During steady-state operation, the high-sensitivity mode is restored to ensure timely capture of small leakage signals. When the phase deviation of any branch reaches the action threshold and lasts for more than a preset time (such as two power frequency cycles), the main switch locking logic is activated, leaving a priority action time window for branch protection;

[0135] The branch switches adopt inverse time characteristics, the main switch is set with a fixed delay, and protection selectivity is achieved through time and space coordination to ensure that the fault removal range is minimized.

[0136] Example 7:

[0137] When the target detection algorithm of this application is the zero-sequence current magnitude-phase ratio algorithm, refer to Figure 7 This application calculates the branch leakage score ratio based on the short-term and long-term judgment results, performs a sliding average filter on the score ratio to suppress transient interference, and sorts the branch priorities based on the score ratio to achieve differentiated switching actions, thereby achieving coordinated locking of the main switch and branch switches. By using the phase comparison method for high-resistance leakage, multi-point leakage can be sorted by score ratio.

[0138] If the target detection algorithm is the zero-sequence current magnitude-phase method;

[0139] A dynamic proportional threshold mechanism based on the capacitance compensation mechanism and the minimum guarantee threshold is set up; the capacitive leakage current component generated by the distributed capacitance of the power grid is calculated in real time, and this component is dynamically deducted from the measured zero-sequence current value to eliminate false alarm signals caused by capacitance coupling.

[0140] Based on the dynamic proportional threshold mechanism, short-term and long-term integral judgment conditions are set. For judgment condition settings, a minimum operating current threshold is set based on historical fault data to ensure protection is triggered even when the compensation algorithm is ineffective or when there is extremely low resistance. To address the high-complexity transient characteristics of sudden leakage, the corresponding triggering criteria are configured based on the sudden change in the current rate of change within the sliding time window. For persistent leakage, zero-sequence voltage is used, and a scoring function is constructed that combines current amplitude weights, duration, and trend slope to determine the severity of leakage in different branches.

[0141] According to the short-time judgment conditions and long-time integral judgment conditions, configure the automatic blocking conditions of the main switch tripping function based on the branch scoring ratio;

[0142] According to the automatic blocking condition of the main switch tripping function based on the branch bisection ratio, the average filter of each branch is set, the priority action satisfied by each branch is determined, and the switch action logic of each branch is configured based on the priority action satisfied by each branch.

[0143] The main switch lockout condition is dynamically generated based on the difference between the highest-rated branch and its secondary branches. When this difference exceeds a set ratio, the lockout logic is activated, ensuring that the most severe faults are cleared first. A composite structure of sliding mean and median filtering is used to suppress pulse interference and periodic noise while preserving the key mutation characteristics of the fault waveform. Multi-branch zero-sequence current baseline data is automatically collected during leakage fault periods, and a dynamic benchmark library is established through cluster analysis, providing a reliable reference for score ratio calculations.

[0144] Priority-operating branches are ranked by rating to create an action queue. High-rated branches utilize rapid inverse-time action, while low-rated branches utilize a delayed review mechanism. When branch protection successfully clears the fault, the main switch automatically locks and resets. If branch protection fails, the main switch activates backup protection after a set delay.

[0145] Example 8:

[0146] In the calculation process of leakage degree in this application, please refer to Figure 8 , which will realize multi-dimensional evaluation of leakage.

[0147] Obtain the distribution network topology of the power grid and establish a leakage current conduction model. In this process, the node admittance matrix is used to simulate the distribution characteristics of leakage current in different branches in real time and quantify the current diffusion path of different grounding fault points.

[0148] Based on the leakage current conduction model, the propagation attenuation rate and path complexity corresponding to the current leakage resistance are calculated. Then, through a grid analysis method, the power grid is decomposed into minimum impedance path units to analyze the transmission direction and attenuation characteristics of the leakage current. In terms of path complexity, the line impedance and distributed capacitance parameters are integrated to calculate the amplitude reduction characteristics of the leakage resistance from the fault point to the first transmission node serving as the transfer node as the loss characteristic during the transmission process. In addition, the number of branch nodes, line hierarchy, and parallel loop characteristics on the leakage current conduction path are used to implement path complexity evaluation and calculation based on the entropy method.

[0149] Determine the leakage inducing conditions based on the propagation attenuation rate;

[0150] Determine the conditional weight of leakage risk based on path complexity;

[0151] Based on leakage-inducing conditions and condition weights, a normalized risk calculation model is constructed to determine a multi-dimensional assessment value for the leakage level. During implementation, a correlation model is established between the attenuation rate, the external environment, and the device status. Because high attenuation rates are associated with hidden leakage, it is necessary to combine changes in environmental parameters to achieve leakage warnings and leakage level calculations. In the multi-dimensional assessment of leakage level, higher weights are assigned due to the high uncertainty of current conduction paths in complex circuits. Parameters such as environmental parameters, propagation attenuation rates, and path complexity are then normalized into dimensionless indicators. A weighted summation is then used to generate an assessment model based on the multi-dimensional assessment of leakage level to determine the leakage level.

[0152] Example 9:

[0153] This application proposes how to control the switch of the branch circuit, see Figure 9 First, a signal trigger mechanism is set up to solve the problem of multiple mechanisms triggering in coordination and untimely leakage response, forming a closed-loop trigger mechanism to optimize the trigger result; wherein, the signal trigger mechanism includes the first trigger mechanism, the second trigger mechanism and the third trigger mechanism;

[0154] The first trigger mechanism is triggered when the leakage level is such that the branch zero-sequence current mutation amount is greater than the preset mutation value. In actual implementation, the differential change rate of the branch zero-sequence current is monitored in real time. If the mutation rate exceeds the preset threshold, it is determined to be a sudden ground fault, such as lightning strike and short circuit, and locking control is implemented to cut off the fault path.

[0155] The second trigger mechanism is triggered when the leakage level is such that the phase deviation between the zero-sequence current and the zero-sequence voltage is greater than the preset phase deviation value. In actual implementation, the fundamental phase difference between the zero-sequence current and voltage will be extracted through FFT. If the deviation exceeds the threshold, it will be determined that the corresponding branch has a high-resistance leakage phenomenon caused by insulation deterioration, triggering a delayed lockout.

[0156] The third trigger mechanism is triggered when the zero-sequence current in the leakage branch exceeds a preset branch zero-sequence current reference value. In practice, the ratio of each branch's zero-sequence current to the total zero-sequence current is calculated. If one or more branches exceed a reference value (e.g., 70%), they are identified as the primary leakage path, triggering a selective lockout to prevent over-tripping of the main switch.

[0157] Example 10:

[0158] This application refers to the switching of branch paths. Figure 10 Because traditional technologies cannot predict possible future failures, they rely on current leakage parameters to make circuit adjustments. They cannot adjust the power grid based on the predicted power grid topology dynamics, achieve early warning and active defense of leakage failures, and reduce losses from sudden power outages.

[0159] This application determines the rate of decrease of the insulation resistance of each branch, the ambient humidity, and historical leakage events based on the degree of leakage. Specifically, through the online detection device, the time series data of the branch insulation resistance is determined, and the slope of decrease per unit time is calculated to quantify the aging rate of the equipment. In addition, a humidity sensor network is deployed to collect the micro-environment data of the branch in real time and build a fitness-insulation performance degradation correlation model. That is, according to the rate of decrease of the insulation resistance of each branch, the ambient humidity, and historical leakage events, a leakage prediction mechanism based on the LSTM neural network is built; in this process, by integrating the fault records and extracting the periodic correlation features, all these data are input into the LSTM network to realize dynamic switching of the power grid topology and isolate high-risk branches.

[0160] According to the leakage estimation mechanism, leakage fault prediction is carried out;

[0161] Based on leakage fault prediction, switchable branch paths are configured for target branches with high failure rates. In the path switchable configuration, the switchable path must be compatible with the grid topology. For example, backup path switching can be achieved through an intelligent circuit breaker matrix to ensure power supply stability and continuity in non-fault areas.

[0162] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

Claims

1. A method for low-voltage leakage protection for mines based on working condition self-adaptation, characterized in that: include: By detecting leakage resistance in real time and determining the operating conditions of the power grid; According to the operating conditions of the power grid, the target detection algorithm is matched with the preset detection algorithm; Among them, the preset detection algorithms include: integrated zero-sequence current mutation method, zero-sequence current and zero-sequence voltage phase method, and zero-sequence current magnitude phase comparison method; The method further includes, before matching the target detection algorithm in the preset detection algorithm according to the grid operation condition, the following steps: Pre-set leakage scenarios; leakage scenarios include: sudden leakage, continuous leakage under steady-state operation, and interference leakage; Configure corresponding detection matching rules based on the leakage scenario; When the leakage scenario is sudden leakage, a first matching rule based on a dynamic threshold is set based on the current change rate of the zero-sequence current of each branch of the power grid; When the leakage scenario is continuous leakage under steady-state operation, a second matching rule based on phase deviation is set based on the phase difference between each branch of the power grid; When the leakage scenario is interference leakage, a third matching rule based on dynamic threshold is set based on the zero-sequence current proportion of each branch of the power grid and the continuous power frequency cycle; The matching of the target detection algorithm in the preset detection algorithm includes: Real-time acquisition of the three-phase voltage and zero-sequence voltage of the power grid, and simultaneous acquisition of the zero-sequence current of each branch and the total zero-sequence current; Based on the preset ADC converter, the three-phase voltage and zero-sequence voltage of the power grid, the zero-sequence current of each branch and the total zero-sequence current are converted into detection digital signals; Based on the detected digital signal, the insulation resistance value, the distribution coefficient of the distributed capacitance and the grounding method of each grounding point in the power grid are calculated, and the power grid status assessment is output; Determine the time domain waveform characteristics of zero-sequence voltage and zero-sequence current according to the grid assessment status; Determine the leakage type based on the time domain waveform characteristics; leakage types include high-resistance leakage, intermittent leakage, and multi-point scattered leakage; Configure the matching rules of the target detection algorithm based on the leakage type; Based on the target detection algorithm, a unified criterion is established to coordinate the action logic of the main switch and each branch switch, and determine the leakage degree of the leakage resistor; Determining the leakage degree of the leakage resistance further includes: Obtain the power distribution network topology of the power grid and establish a leakage current conduction model; According to the leakage current conduction model, calculate the propagation attenuation rate and path complexity corresponding to the current leakage resistance; Determine the leakage inducing conditions based on the propagation attenuation rate; Determine the conditional weight of leakage risk based on path complexity; Based on the leakage inducing conditions and condition weights, a normalized risk calculation model is constructed to determine the multi-dimensional assessment value of the leakage degree; According to the degree of leakage, the switches of the leakage branches are locked in turn.

2. A method for low-voltage leakage protection for mining based on working condition self-adaptation according to claim 1, characterized in that: When detecting the leakage resistance in real time, the current power grid is detected based on the zero-sequence voltage correction method to determine the power grid branch with the leakage resistance in the current power grid.

3. The method for low-voltage leakage protection for mining based on working condition self-adaptation according to claim 1, characterized in that: The target detection algorithm is used to establish a unified criterion to coordinate the operation logic of the main switch and each branch switch, including: When the target detection algorithm is the integrated zero-sequence current mutation method; Set a dynamic time window, which includes a short-term detection window that captures sudden leakage through mutation rate, and a long-term integration window that identifies slowly changing high-resistance leakage through adaptive threshold calculation. The adaptive threshold calculation system performs capacitance compensation calculation and load current compensation calculation. According to the dynamic time window, the leakage priority of each branch is sorted, and the main switch locking conditions and differential coordination conditions are set; According to the main switch locking conditions and differential coordination conditions, set the pre-analog filter and digital sliding filter, determine the fault waveform of each branch, and configure the switch action logic of each branch in the preset load feature library according to the fault waveform.

4. The method for low-voltage leakage protection for mining based on working condition self-adaptation according to claim 1, characterized in that: The method of establishing a unified criterion for coordinating the operation logic of the main switch and each branch switch according to the target detection algorithm also includes: When the target detection algorithm is the zero-sequence current and zero-sequence voltage phase method; Set a windowed FFT algorithm to calculate the fundamental phase difference between the zero-sequence current and zero-sequence voltage of each branch, and set a dynamic phase threshold; the dynamic phase threshold also includes a capacitance compensation threshold and a reference offset threshold; Determine the phase deviation of each leakage branch according to the dynamic phase threshold; Configure the automatic locking conditions and switch action threshold of the main switch according to the phase deviation; According to the automatic blocking conditions and the switching action threshold, the harmonic suppressor is set, the initial phase reference is determined, and the switching action logic of each branch is adaptively configured based on the initial phase reference and load fluctuation.

5. The method for low-voltage leakage protection for mining based on working condition self-adaptation according to claim 1, characterized in that: The method of establishing a unified criterion for coordinating the operation logic of the main switch and each branch switch according to the target detection algorithm also includes: When the target detection algorithm is the zero-sequence current magnitude-phase method; Set up a dynamic proportional threshold mechanism based on capacitance compensation mechanism and minimum guarantee threshold; According to the dynamic proportional threshold mechanism, set the short-time judgment condition and the long-time integral judgment condition; According to the short-time judgment conditions and long-time integral judgment conditions, configure the automatic blocking conditions of the main switch tripping function based on the branch scoring ratio; According to the automatic blocking condition of the main switch tripping function based on the branch scoring ratio, the average filter of each branch is set, the priority action satisfied by each branch is determined, and the switch action logic of each branch is configured based on the priority action satisfied by each branch.

6. The method for low-voltage leakage protection for mining based on working condition self-adaptation according to claim 1, characterized in that: The step of sequentially locking the switches of the leakage branches according to the degree of leakage further includes: Setting a signal trigger mechanism; wherein the signal trigger mechanism includes a first trigger mechanism, a second trigger mechanism and a third trigger mechanism; The first trigger mechanism is triggered when the leakage level is such that the branch zero-sequence current mutation amount is greater than the preset mutation value; The second trigger mechanism is triggered when the leakage level is such that the phase offset between the zero-sequence current and the zero-sequence voltage is greater than a preset phase offset value; The second trigger mechanism is triggered when the proportion of the zero-sequence current of the leakage branch is greater than the preset branch zero-sequence current reference value.

7. The method for low-voltage leakage protection for mining based on working condition self-adaptation according to claim 1, characterized in that: The step of sequentially locking the switches of the leakage branches according to the degree of leakage further includes: According to the leakage degree, determine the drop rate of the insulation resistance of each branch, the ambient humidity and the historical leakage events; Based on the drop rate of each branch insulation resistance, ambient humidity, and historical leakage events, a leakage prediction mechanism based on the LSTM neural network was established. According to the leakage estimation mechanism, leakage fault prediction is carried out; According to the leakage fault prediction, switchable branch paths are configured for target branches with high failure rates in turn.

Images