Multi-source power distribution network fault section positioning method and system

Abstract

The invention relates to the technical field of power distribution network fault detection, in particular to a multi-source power distribution network fault section positioning method and system. According to the method, fault section positioning is carried out by fusing the electrical characteristics, the spatial-temporal characteristics and the power influence characteristics, that is, the electrical characteristics reflect the fault essence, the spatial-temporal characteristics constrain the propagation path, and the power influence characteristics correct interference, and multi-source power distribution network fault positioning is carried out through the three characteristics. A method for realizing fault positioning through operation by adopting a mathematical model or a physical model is abandoned, the operation is simple, and the interpretability is high.

Description

Technical Field

[0001] This invention relates to the field of power distribution network fault detection technology, specifically a method and system for locating fault sections in multi-source power distribution networks. Background Technology

[0002] With the large-scale integration of distributed power sources (such as photovoltaic and wind power), traditional single-source radial distribution networks are gradually evolving into complex distribution networks with multiple sources (including main grid power and distributed power), multiple interconnections (ring network / tandem structure), and multiple types of monitoring terminals (such as smart switches, fault indicators, and distribution terminals). These networks have the following characteristics: Topological complexity: multiple power supply paths exist, ring network interconnection switches dynamically switch on and off, and the direction and amplitude of fault current are significantly affected by the output of distributed power sources; Information heterogeneity: different monitoring terminals (such as fault indicators that only report fault overcurrent signals, smart switches that record circuit breaker action time and current waveforms, and distribution terminals that collect voltage / current timing data) have significant differences in data type, sampling frequency, and communication delay; Fault characteristic interference: distributed power sources may inject reverse current into the fault point during a fault (such as inverter-type power sources), causing the traditional "fault current direction consistency" location method to fail; Single-source information limitations: a single monitoring terminal may fail to cover the fault transient process, have communication anomalies, or sensor malfunctions that result in missed / false alarms, leading to unreliable location results.

[0003] Existing fault location methods for power distribution networks (such as the "head-to-end overcurrent method" based on fault indicators, the "time difference location method" based on smart switches, or machine learning models based on a single data source) are prone to problems such as "false location" (misclassifying non-faulty sections as faulty), "missed location" (faulty sections are not identified), or "multiple solutions" (multiple sections are suspected of being faulty). At the same time, there are existing technologies that use multi-source data and apply complex mathematical and physical models for fault location. However, such schemes are computationally complex and lack interpretability. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides a method and system for locating fault sections in multi-source distribution networks, which solves the problems existing in the prior art.

[0005] This invention provides a method for locating fault sections in a multi-source distribution network, comprising the following steps:

[0006] S1: Collect real-time operating data of the multi-source distribution network;

[0007] S2: Perform data preprocessing on the real-time running data to obtain preprocessed real-time running data;

[0008] S3: Extract multi-dimensional fault features from the preprocessed real-time operating data; the multi-dimensional fault features include electrical features, spatiotemporal features, and power supply influence features.

[0009] S4: Perform initial fault location based on the multi-dimensional fault characteristics;

[0010] Specifically: S4.1: Calculate the electrical feature matching degree C based on the electrical features in the multidimensional fault features. elec S4.2: Calculate the spatiotemporal feature matching degree C based on the spatiotemporal features in the multidimensional fault features. time S4.3: Calculate the power supply influence feature matching degree C based on the power supply influence feature in the multidimensional fault features. source S4.4: Calculate the comprehensive score S of the segment to be located. i S4.5: Perform initial location of the faulty section based on the comprehensive score of the section to be located;

[0011] S5: Fine-grained fault segment localization based on timing logic and topological constraints;

[0012] Specifically: S5.1: Eliminate non-faulty sections based on the fault propagation sequence; S5.2: Eliminate non-faulty sections based on the uniqueness of the fault current path.

[0013] Preferably, in S3, the electrical fault characteristics include current direction, current amplitude, and voltage drop; the spatiotemporal characteristics include the temporal correlation of the fault event and the spatial topology location of the monitoring terminal; and the power supply impact characteristics include the output status of the distributed power supply and the proportion of reverse current of the distributed power supply.

[0014] Preferably, the current direction extraction process is as follows: the current phase acquired by the intelligent switch is compared with the system reference voltage phase to calculate the current phase angle θ. i If θ i If θ ∈ [90°, 270°], then the current direction is determined to be the inflow section, and the fault point is located downstream of this section; if θ i If the value is in the range [-90°, 90°), it is determined to be an outflow section, and the fault point is located upstream of this section.

[0015] The current amplitude extraction process is as follows: Calculate the effective value I of the current at each segment end by sampling the current waveform. rms As the magnitude of the fault current;

[0016] The voltage sag extraction process is as follows: calculate the rate of change of voltage sag amplitude collected by the upstream and downstream power distribution terminals of the fault section.

[0017] Preferably, the extraction process of the time correlation of the fault event is as follows: recording the timestamps of key events, including: the overcurrent trigger time t of the fault indicator. fi The tripping time t of the intelligent switch trip , Voltage drop start time t of distribution terminal volt_drop ;

[0018] The process of extracting the spatial topology location of the monitoring terminal is as follows: construct a section-terminal adjacency matrix based on the physical connection relationship of the distribution network; the physical connection relationship of the distribution network includes the layout relationship of feeder section switches and tie switches.

[0019] Preferably, the extraction process for the reverse current ratio of the distributed power source is as follows: for the section of the distributed power source, calculate the amplitude I of the fault current injected by the distributed power source. dg and the total fault current I of the system total By comparison, the proportion of reverse current of distributed power sources is obtained.

[0020] Preferably, S4.1 specifically includes:

[0021] Current direction consistency C elec,方向 Assignment: If the fault current detected by all associated terminals within the segment to be located points in the same direction, then C elec,方向 =1; if a contradiction exists, then C elec,方向 =0.5;

[0022] Current amplitude anomaly C elec,幅值 Assignment: Calculate the ratio of the effective values ​​of the currents at the beginning and end of the section to be located; if r current ≥1.5 (indicating the fault point is close to the beginning and the current attenuation is not significant), then C elec,幅值 =1; if 1.2≤r current If C < 1.5, then C elec,幅值 =0.7; otherwise C elec,幅值 =0.3;

[0023] Voltage drop significance C elec,电压 Assignment: Calculate the voltage amplitude difference ΔU between the beginning and end of the section to be located; if ΔU ≥ 0.3U base Then C elec,电压 =1; if 0.1U base ≤ΔU<0.3U base Then C elec,电压 =0.6; otherwise C elec,电压 =0.2, the U base The reference voltage value for the section to be located;

[0024] The final electrical feature matching degree is the weighted score of each sub-item, calculated using the following formula:

[0025] C elec = ω1·C elec,方向 + ω2·C elec,幅值 + ω3·C elec,电压 ;

[0026] In the formula, ω1, ω2, and ω3 are coefficients.

[0027] Preferably, S4.2 is specifically:

[0028] Rationality of chronological order C time,顺序 Assignment: If the overcurrent time t of the fault indicator fi is earlier than the tripping time t of the associated intelligent switch trip by more than 2 ms, then C time,顺序 = 1; if the difference Δt between the tripping time of the associated intelligent switch and the overcurrent time of the fault indicator is less than 2 ms, then C time,顺序 = 0.4; if the overcurrent time t of the fault indicator fi is greater than the tripping time t of the associated intelligent switch trip , then C time,顺序 = 0;

[0029] Consistency of topological position C time,拓扑 Assignment: If the to-be-located section is located at the expected position of the fault propagation path, then C time,拓扑 = 1; otherwise Ctime, topology = 0.5;

[0030] The spatio-temporal feature matching degree is the weighted score of each sub-item, and the calculation formula is:

[0031] C time = ω4·C time,顺序 + ω5·C time,拓扑 ;

[0032] In the formula, ω4 and ω5 are coefficients.

[0033] Preferably, S4.3 is specifically:

[0034] Rationality of reverse current ratio C source,反向 Assignment: If the reverse current ratio r of the to-be-located section is ≤ 0.3, then C source,反向 = 1; if 0.3 < r ≤ 0.6, then C source,反向 = 0.6; if r > 0.6, then C source,反向 = 0.2;

[0035] Compatibility of power output status C source,出力 Assignment: If the distributed power source in the to-be-located section is in the power generation state and the reverse current ratio r > 0, then C source,出力 = 1; if the distributed power source in the to-be-located section is in the power generation state and the reverse current ratio r ≤ 0, C source,出力= 0.5; If the distributed power supply shuts down, then C source,出力 = 1;

[0036] The power influence feature matching degree is the weighted score of each sub - item, and the calculation formula is:

[0037] C source = ω6·C source,反向 + ω7·C source,出力 ;

[0038] In the formula, ω6 and ω7 are coefficients.

[0039] Preferably, in the S4.4:

[0040] The comprehensive score S of the to - be - located section i The calculation formula is:

[0041] Si ＝ ω8·C elec + ω9·C time + ω 10 ·C source ;

[0042] In the formula, ω8, ω9, ω 10 are coefficients.

[0043] Preferably, the S4.5 is specifically:

[0044] If the comprehensive score S of the unique to - be - located section i ≥ 0.8, directly determine that the to - be - located section is the fault section; if the scores of all sections are < 0.5, trigger a positioning failure alarm; if there are multiple sections with close comprehensive scores and 0.5 ≤ S < 0.8, and there are fault conflict sections at this time, then enter the next step.

[0045] According to another aspect of the present invention, a multi - source distribution network fault section positioning system is provided. The system adopts the above - mentioned multi - source distribution network fault section positioning method, and the system includes:

[0046] A data acquisition module for collecting the real - time operation data of the multi - source distribution network;

[0047] A pre - processing module for performing data pre - processing operations on the real - time operation data to obtain pre - processed real - time operation data;

[0048] A feature extraction module for performing multi - dimensional fault feature extraction on the pre - processed real - time operation data;

[0049] An initial positioning module for performing initial positioning of the fault section according to the multi - dimensional fault features;

[0050] The conflict section elimination module is used for fine-grained fault section location based on timing logic and topological constraints.

[0051] The embodiments of the present invention have the following technical effects:

[0052] This invention achieves fault location by integrating three types of features: electrical characteristics, spatiotemporal characteristics, and power source influence characteristics. Electrical characteristics reflect the nature of the fault, spatiotemporal characteristics constrain the propagation path, and power source influence characteristics correct for interference, resulting in the following synergistic effects: Complementarity: Electrical characteristics address the question of "whether a fault exists," spatiotemporal characteristics address the question of "which segment the fault occurs in," and power source influence characteristics address the question of "how to accurately determine the direction under multi-source interference." These three features complement each other. Anti-interference: Even if a single feature has uncertainties (such as a fault indicator failing to report), other features (such as voltage drop amplitude) can still provide effective information, reducing the risk of mislocation. Adaptability: It can handle both traditional single-source networks and complex scenarios with high-penetration distributed power source access, meeting the development needs of new distribution networks. In particular, this application uses the above three features for multi-source distribution network fault location, abandoning the method of using mathematical or physical models to achieve fault location through calculation. The calculation is simple and highly interpretable.

[0053] Meanwhile, in this invention, in the spatiotemporal features, the rationality of the time sequence can verify whether the fault propagation direction conforms to physical laws, and the consistency of the topological location further confirms the possibility of the fault. Through the above two sub-indicators, the spatiotemporal features verify the propagation direction and eliminate irrelevant segments by verifying the rationality of the time sequence and the topological constraints, and locate the fault to a specific physical segment, that is, to the feeder segment between two adjacent switches, avoiding regional misjudgments caused by relying solely on electrical features, such as misjudging the entire feeder as a fault segment. In the power supply influence features, the rationality of the reverse current ratio is used to correct the direction and re-determine the fault segment, and the compatibility of the power supply processing state is used to avoid power supply state interference. Through the above two sub-indicators, the problem of contradictory current directions caused by distributed power injection in multi-source networks is solved, and the positioning accuracy in complex scenarios is significantly improved.

[0054] Meanwhile, in this embodiment, through dual filtering of timing logic and topology constraints, the "multiple solutions" of the initial location are transformed into a uniquely determined fault segment, and the location accuracy is improved from 90% in step S3 to over 98%. It can handle complex situations such as dynamic switching of tie switches, reverse injection of distributed power sources, and load fluctuation interference, and meet the needs of high-penetration new energy distribution networks. Attached Figure Description

[0055] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0056] Figure 1 This is a flowchart of a method for locating fault sections in a multi-source distribution network provided by an embodiment of the present invention;

[0057] Figure 2 This is a flowchart of the initial fault location based on the multi-dimensional fault characteristics provided in the embodiments of the present invention;

[0058] Figure 3 This is a flowchart of fine-grained fault segment localization based on timing logic and topological constraints provided in an embodiment of the present invention;

[0059] Figure 4 This is a schematic diagram of eliminating conflict fault sections in a power distribution network, provided by an embodiment of the present invention. Detailed Implementation

[0060] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0061] Example 1, Appendix Figure 1 A flowchart of a method for locating fault sections in a multi-source distribution network is shown in the attached figure. Figure 1 As shown, a method for locating fault sections in a multi-source distribution network includes the following steps:

[0062] S1: Collect real-time operating data of the multi-source distribution network;

[0063] Real-time operation data of the multi-source distribution network is acquired synchronously through various monitoring terminals deployed on the multi-source distribution network.

[0064] The monitoring terminal includes a fault indicator, a smart switch, and a distribution terminal. The real-time operating data acquired by the fault indicator includes the time of fault occurrence, overcurrent waveform signal, and the section number corresponding to the installation location. The smart switch includes a circuit breaker and a load switch. The real-time operating data acquired by the circuit breaker includes the fault tripping time, opening and closing status, three-phase current / voltage waveforms before and after the fault, and the connection relationship of the section. The distribution terminal includes a DTU or an FTU. The real-time operating data acquired by the distribution terminal includes the voltage amplitude / phase angle upstream and downstream of the fault section, the effective value of the current / harmonic content (sampling frequency ≥ 100Hz), and the output power (active / reactive) and connection point voltage of the distributed power source.

[0065] S2: Perform data preprocessing on the real-time running data to obtain preprocessed real-time running data;

[0066] In multi-source distribution networks, monitoring terminals upload data to the data center via wireless communication, fiber optics, or power line carrier. Due to unstable communication channels, terminal hardware failures, or network congestion, data packet loss or outliers (such as sudden jumps in current values ​​beyond physical limits) may occur. Directly using such dirty data for fault location will severely interfere with the reliability of feature extraction and location results. Therefore, data preprocessing is necessary for the real-time operational data.

[0067] The data preprocessing operations include communication packet loss detection and processing, outlier detection and removal, normalization, and data time alignment.

[0068] Specifically, the communication packet loss detection and processing involves setting a fixed data reporting period for each monitoring terminal. If no data packet is received from a terminal within N consecutive periods, it is determined to be a communication packet loss.

[0069] Specifically, the data reporting cycle set for each monitoring terminal is as follows: the fault indicator reports its status every 10ms, and the smart switch reports its current sampling value every 1ms.

[0070] In this embodiment, the value of N is dynamically set according to the type of the monitoring terminal. When the monitoring terminal is a fault indicator, N=3, and when the monitoring terminal is a smart switch, N=5.

[0071] The communication packet loss handling process is as follows: Data points determined to be lost are marked as invalid data. If the packet loss occurs during a non-critical feature extraction period, the data point is directly removed. If the packet loss occurs during a critical fault period, the following interpolation method is used for compensation:

[0072] For continuous signals such as current / voltage, use linear interpolation or spline interpolation.

[0073] Estimate the value at the time of packet loss in communication;

[0074] For discrete signals such as switch states, if the state is consistent before and after packet loss, it will be automatically filled back to the original state; if the state is inconsistent, it will be marked as an uncertain state.

[0075] Specifically, the outlier detection and removal involves: setting a reasonable range based on the physical characteristics of the power distribution network equipment, and removing values ​​that exceed the threshold; specifically:

[0076] Effective value of current: During normal operation, the amplitude of the three-phase current is usually ≤ 1.2 times the rated current. If the current value at a certain sampling point is > 2 times the rated current, it is judged as an abnormal value.

[0077] Voltage amplitude: within ±10% of the rated voltage is considered normal; values ​​exceeding this range will be rejected.

[0078] Switch status: The open / closed status must meet logical constraints, and contradictory data will be directly rejected; for example, meeting the logical constraints means that the same switch cannot report "open" and "closed" at the same time within 1ms.

[0079] The amplitude ranges of current / voltage signals collected by different monitoring terminals in the distribution network vary significantly (e.g., overcurrent signals detected by fault indicators may reach hundreds of amperes, while steady-state currents collected by distribution terminals are only tens of amperes; voltage signals may span ±20% of the rated value). Directly using raw values ​​for feature extraction (e.g., calculating the rate of change of current amplitude) will lead to an imbalance in numerical scale (e.g., current characteristics dominate model calculations, while voltage characteristics are ignored), affecting the accuracy of the fault location model. Therefore, it is necessary to normalize the current and voltage data. The specific normalization process is as follows:

[0080] Pre-set baseline values ​​based on the topology parameters of the distribution network;

[0081] The topology parameters include feeder type, transformer capacity, etc. For example, for the current reference value, the long-term allowable current carrying capacity of the feeder (e.g., 500A) is taken; for the voltage reference value, the rated line voltage (e.g., 10kV) is taken.

[0082] The raw current and voltage values ​​at each sampling point are normalized to per-unit current and per-unit voltage values.

[0083] The conversion formula for the original value of the current is as follows:

[0084]

[0085] In the formula, Ip.u.(t) is the per-unit value of the current at time t, I base I is the current reference value. rawThis is the original value of the current;

[0086] The conversion formula for the original voltage value is as follows:

[0087]

[0088] In the formula, Up.u.(t) is the per-unit value of the current at time t, U base U is the current reference value. raw This is the original value of the current.

[0089] Specifically, the data time alignment refers to:

[0090] A synchronous clock source is deployed at a key node of the power distribution network, and a time synchronization signal is broadcast to all monitoring terminals via wired or wireless means to ensure that the deviation between the local clock of each monitoring terminal and the reference time of the synchronous clock source is ≤1ms.

[0091] The key nodes include substations, feeder interconnection switches, etc.; the synchronization clock source includes GPS, Beidou time synchronization module, etc.; the wired method is fiber optic communication, and the wireless method is IEEE 1588 precision time protocol communication.

[0092] Each monitoring terminal synchronously records the timestamp of its local clock when collecting data;

[0093] Each terminal periodically receives a reference time broadcast by the synchronization clock source and calculates the deviation between the local clock and the reference time;

[0094] The timestamp of the data reported by each terminal is corrected to a unified time under the reference time based on the timestamp of the local clock and the deviation between the local clock and the reference time.

[0095] For example, a smart switch locally records the tripping time as t. local =10:00:00.000, the reference time for the synchronous clock source broadcast is t. ref =10:00:00.005, then the corrected timestamp is t. aligned =10:00:00.000+(10:00:00.005-10:00:00.000)=10:00:00.005.

[0096] S3: Extract multi-dimensional fault features from the preprocessed real-time running data;

[0097] The multi-dimensional fault characteristics include electrical characteristics, spatiotemporal characteristics, and power supply influence characteristics.

[0098] The electrical fault characteristics include current direction, current amplitude, and voltage sag.

[0099] The process for extracting the current direction is as follows: the current phase information collected by the smart switch is compared with the system reference voltage phase to calculate the current phase angle θ. i (The angle by which current leads / lags voltage), if θ i If θ ∈ [90°, 270°], then the current direction is determined to be the "inflow" section, meaning the fault point is located downstream of this section; if θ i If the value is in the range [-90°, 90°), it is determined to be an "outflow" segment, meaning the fault point is located upstream of this segment.

[0100] The current amplitude extraction process is as follows: Calculate the effective value I of the current at each segment end by sampling the current waveform. rms As the magnitude of the fault current.

[0101] The effective value of the terminal current of each section is obtained by the following calculation formula:

[0102]

[0103] In the formula, i(t) is the current sample value at time t, and T is the power frequency period.

[0104] It is worth emphasizing that the effective value of the end current of each section includes the effective value I of the end current of the section. rms,首 and the effective value of the end current of the section I rms,末 .

[0105] The voltage sag extraction process is as follows: calculate the rate of change of voltage amplitude collected by the upstream and downstream distribution terminals of the fault section; the calculation formula is:

[0106]

[0107] In the formula, U pre U is the steady-state voltage before the fault. fault This is the voltage after the fault.

[0108] The spatiotemporal characteristics include the temporal correlation of fault events and the spatial topological location of the monitoring terminal.

[0109] The extraction process of the time correlation of the fault event is as follows: record the timestamp of the key event, including: the overcurrent trigger time t of the fault indicator. fi The tripping time t of the intelligent switch trip , Voltage drop start time t of distribution terminal volt_drop .

[0110] The process of extracting the spatial topology location of the monitoring terminal is as follows: Based on the physical connection relationship of the distribution network, a segment-terminal adjacency matrix is ​​constructed; specifically: if the fault indicator is installed in the middle of segment AB, then its associated segment is AB; if the smart switch is located at the beginning of segment BC, then its associated segment is BC; the segment range monitored by each terminal is determined through the topology relationship to avoid misjudgment across segments.

[0111] The physical connections of the power distribution network include the layout of feeder sectionalizing switches and tie switches.

[0112] The power supply impact characteristics include the output status of the distributed power source and the proportion of reverse current of the distributed power source.

[0113] The process of extracting the output status of the distributed power source is as follows: the active power P of the distributed power source is collected through the distribution terminal. dg and reactive power Q dg Determine the output status of the distributed power source: if P dg >0 and Q dg If P is close to 0, the distributed power source is in a generating state (potentially injecting current into the fault point); if P dg If the value is close to 0, the distributed power source is in an off-grid or shutdown state (which does not affect the fault current).

[0114] The extraction process for the reverse current ratio of the distributed power source is as follows: For the section of the distributed power source, calculate the amplitude I of the fault current injected by the distributed power source. dg and the total fault current I of the system total By comparison, the proportion of reverse current of distributed power sources is obtained.

[0115] This embodiment achieves the following synergistic effect by integrating the above three types of features: electrical features reflecting the nature of the fault, spatiotemporal features constraining the propagation path, and power supply influence features correcting interference: complementarity.

[0116] Electrical characteristics address the question of "whether a fault exists," spatiotemporal characteristics address the question of "where the fault occurs," and power source influence characteristics address the question of "how to accurately determine the direction under multi-source interference." These three characteristics complement each other. Interference resistance: Even if a single characteristic has uncertainties (such as a fault indicator failing to report), other characteristics (such as voltage drop amplitude) can still provide effective information, reducing the risk of mislocation. Adaptability: It can handle both traditional single-power-source networks and complex scenarios with high-penetration distributed power source access, meeting the development needs of new distribution networks.

[0117] S4: Perform initial fault location based on the multi-dimensional fault characteristics;

[0118] In this step, as shown in the appendix Figure 2As shown, S4 specifically includes:

[0119] S4.1: Calculate the electrical feature matching degree C based on the electrical features in the multidimensional fault features. elec ;

[0120] Specifically: Current direction consistency C elec,方向 Assignment: If the fault current detected by all associated terminals within the segment to be located points in the same direction, then C elec,方向 =1; if a contradiction exists, then C elec,方向 =0.5;

[0121] For example, if the detected fault currents all point in the same direction, they can all be determined to be "flowing" into the downstream of the section; if there is a contradiction, some terminals can be determined to be "flowing" into the section and some can be determined to be "flowing out".

[0122] Current amplitude anomaly C elec,幅值 Assignment: Calculate the ratio of the effective values ​​of the currents at the beginning and end of the section to be located; if r current ≥1.5 (indicating the fault point is close to the beginning and the current attenuation is not significant), then C elec,幅值 =1; if 1.2≤r current If C < 1.5, then C elec,幅值 =0.7; otherwise C elec,幅值 =0.3;

[0123] The formula for calculating the ratio of the effective values ​​of the first and last currents is:

[0124] r current =I rms,首 / I rms,末 ;

[0125] In the formula, I rms,首 The effective value of the current at the beginning of the section and I rms,末 This is the effective value of the current at the end of the section.

[0126] Voltage drop significance C elec,电压 Assignment: Calculate the voltage amplitude difference ΔU between the beginning and end of the section to be located; if ΔU ≥ 0.3U base Then C elec,电压 =1; if 0.1U base ≤ΔU<0.3U base Then C elec,电压 =0.6; otherwise C elec,电压 =0.2.

[0127] Wherein, the U base The reference voltage value is the voltage value of the section to be located.

[0128] The final electrical feature matching degree is the weighted score of each sub-item, calculated using the following formula:

[0129] C elec =ω1·C elec,方向 +ω2·C elec,幅值 +ω3·C elec,电压 ;

[0130] In the formula, ω1, ω2, and ω3 are coefficients.

[0131] In this embodiment, ω1 = 0.5, ω2 = 0.3, and ω3 = 0.2.

[0132] S4.2: Calculate the spatiotemporal feature matching degree C based on the spatiotemporal features in the multidimensional fault features. time ;

[0133] Specifically, this refers to: the rationality of the chronological order (C) time,顺序 Assignment: If the fault indicator overcurrent time t fi Earlier than the associated smart switch tripping time t trip If it is more than 2ms, then C time,顺序 =1; if the difference Δt between the tripping time of the associated smart switch and the overcurrent time of the fault indicator is less than 2ms, then C time,顺序 =0.4; if the fault indicator overcurrent time t fi Greater than the associated smart switch tripping time t trip Then C time,顺序 =0;

[0134] Topological location consistency C time,拓扑 Assignment: If the segment to be located is located at the expected position on the fault propagation path, then C time,拓扑 =1; otherwise Ctime,topology = 0.5.

[0135] For example, the expected location can be determined by the time difference between other segments, indicating that the fault point is located upstream of the segment.

[0136] The spatiotemporal feature matching degree is the weighted score of each sub-item, calculated using the following formula:

[0137] C time =ω4·C time,顺序 +ω5·C time,拓扑 ;

[0138] In the formula, ω4 and ω5 are coefficients.

[0139] In this embodiment, ω4 = 0.7 and ω5 = 0.3.

[0140] S4.3: Calculate the power supply influence feature matching degree C based on the power supply influence feature in the multidimensional fault features. source ;

[0141] Specifically: Rationality C of the reverse current ratio source,反向 Assignment: If the reverse current ratio r of the to-be-located section ≤ 0.3, then C source,反向 = 1; if 0.3 < r ≤ 0.6, then C source,反向 = 0.6; if r > 0.6, then C source,反向 = 0.2;

[0142] Compatibility C of the power output state source,出力 Assignment: If the distributed power source in the to-be-located section is in the power generation state and the reverse current ratio r > 0, then C source,出力 = 1; if the distributed power source in the to-be-located section is in the power generation state and the reverse current ratio r ≤ 0, C source,出力 = 0.5; if the distributed power source is shut down, then C source,出力 = 1.

[0143] The matching degree of the power source influence characteristics is the weighted score of each sub-item, and the calculation formula is:

[0144] C source = ω6·C source,反向 + ω7·C source,出力 ;

[0145] In the formula, ω6, ω7 are coefficients,

[0146] In this embodiment, ω6 = 0.4, ω7 = 0.6. <…>S4.4: Calculate the comprehensive score S of the to-be-located section i ;

[0148] Among them, the comprehensive score S of the to-be-located section i The calculation formula is:

[0149] S i = ω8·C elec + ω9·C time + ω 10 ·C source ; [[ID=6​​​​​​​​​​​​​​​If it is ≥ 0.8, directly determine that the to-be-located section is a fault section; if the scores of all sections are < 0.5, trigger a positioning failure alarm; if there are multiple sections with similar comprehensive scores and the comprehensive score 0.5 ≤ S < 0.8, and there are fault conflict sections at this time, then proceed to the next step.

[0154] Among them, the fault conflict section refers to the situation where multiple fault sections appear in the initial positioning due to the complexity of the multi-source distribution network and the limitations of monitoring data.

[0155] In this step, the electrical characteristics solve whether there is a fault current, that is, the physical evidence of the existence of the fault; the spatio-temporal characteristics solve where the fault occurs, that is, the spatial anchoring of the fault location, and the power influence characteristics solve the misjudgment of the direction under multi-source interference, that is, the anti-interference correction in complex scenarios; in the spatio-temporal characteristics, the rationality of the time sequence can verify whether the fault propagation direction conforms to the physical law, and the consistency of the topological position further confirms the possibility of the fault. Through the above two sub-indicators, the spatio-temporal characteristics exclude irrelevant sections through the rationality of the time sequence to verify the propagation direction and topological constraints, and locate the fault to a specific physical section, that is, accurate to the feeder section between two adjacent switches, avoiding regional misjudgments caused by relying solely on electrical characteristics, such as misjudging the entire feeder as a fault section; in the power influence characteristics, the rationality of the reverse current ratio is used to correct the direction and re-determine the fault section, and the compatibility of the power processing state is used to avoid power state interference. Through the above two sub-indicators, the problem of current direction contradiction caused by distributed power injection in the multi-source network is solved, and the positioning accuracy in complex scenarios is significantly improved.

[0156] S5: Conduct fine positioning of the fault section based on sequential logic and topological constraints;

[0157] In the initial positioning of step S4, although high-probability fault sections can be screened out through multi-feature collaborative analysis, due to the complexity of the distribution network (such as multi-source power supply, dynamic switching of tie switches) and the limitations of monitoring data (such as single-terminal missed reports, communication delays), the following conflict scenarios may occur: multi-section high scores: the comprehensive scores S of multiple sections i are close and the comprehensive score 0.5 ≤ S < 0.8, and the unique fault section cannot be directly determined. Therefore, it is necessary to locate the fault section for the situation where the fault section cannot be located in the initial positioning.

[0158] Specifically, as shown in the appendix Figure 3 it is shown that S5 is specifically as follows:

[0159] S5.1: Exclude non-fault sections based on the fault propagation order;

[0160] Specifically, extract the fault event timestamps of the monitoring terminals in the fault conflict section;

[0161] The fault event timestamp is obtained by extracting the fault event timestamps of all monitoring terminals associated with the fault conflict section.

[0162] The key event timestamps include: the fault indicator overcurrent trigger time t. fi The tripping time t of the intelligent switch trip 2. Voltage drop start time t at the distribution terminal volt_drop .

[0163] Calculate the time difference based on the timestamp of the fault event;

[0164] The time difference includes: the time difference Δt between the tripping time of the intelligent switch and the overcurrent triggering time of the fault indicator. fi-trip The time difference Δt between the voltage drop start time of the distribution terminal and the tripping time of the smart switch trip-volt ;

[0165] Compare the time difference with a threshold to exclude non-faulty sections;

[0166] If Δt fi-trip If the time sequence of the segment to be located is contradictory (<2ms), then the segment to be located is a non-faulty segment and is excluded from the search; if Δt trip-volt If the time interval is greater than 10ms, the time sequence of the segment is determined to be abnormal, and the segment to be located is a non-faulty segment and is excluded.

[0167] S5.2: Eliminate non-faulty sections based on the uniqueness of the fault current path;

[0168] The physical topology of the distribution network (such as the status of feeder sectionalizing switches and tie switches) determines the propagation path of fault current. If a section is located on an isolated feeder section formed by the disconnection of the tie switch, or on a path that has been ruled out by other evidence, its possibility of fault can be directly ruled out.

[0169] Specifically, this involves obtaining the real-time topology status of the multi-source distribution network;

[0170] The current on / off status of the tie switch and the status of the feeder section switch are obtained by the status of the smart switch in the multi-source distribution network.

[0171] Based on the real-time topology state, the possible propagation paths of the fault current are determined:

[0172] For example, if the tie switch is closed, the fault may involve multiple feeders controlled by the tie switch; if the tie switch is open, the fault is isolated to a single feeder.

[0173] Verify whether the conflict section is located on a possible propagation path. This eliminates non-faulty sections.

[0174] For example, as shown in the appendix Figure 4 As shown, AF represents a section. If the conflicting section is located on feeder L1 and the tie switch S1 of L1 is open (the fault has not spread to other feeders), it meets the topology constraints; if the conflicting section scores are close (S... i =0.72), but it is located on feeder L2, and the tie switch S1 of L2 is open (the fault does not involve L2), so this section is excluded.

[0175] In this embodiment, through dual filtering of timing logic and topological constraints, the "multiple solutions" of the initial location are transformed into a uniquely determined fault segment, and the location accuracy is improved from 90% in step 3 to over 98%. It can handle complex situations such as dynamic switching of tie switches, reverse injection of distributed power sources, and load fluctuation interference, and meet the needs of high-penetration new energy distribution networks.

[0176] Example 2: The present invention also provides a multi-source distribution network fault section location system, wherein the system adopts the multi-source distribution network fault section location method of Example 1, and the system includes:

[0177] The data acquisition module is used to collect real-time operating data of the multi-source distribution network;

[0178] The preprocessing module is used to perform data preprocessing operations on the real-time running data to obtain preprocessed real-time running data;

[0179] The feature extraction module is used to extract multi-dimensional fault features from the preprocessed real-time running data;

[0180] The initial location module is used to perform initial location of the fault section based on the multi-dimensional fault characteristics.

[0181] The conflict section elimination module is used for fine-grained fault section location based on timing logic and topological constraints.

[0182] Example 3: The present invention also provides an electronic device, including one or more processors and a memory.

[0183] A processor can be a central processing unit (CPU) or other form of processing unit with data processing and / or instruction execution capabilities, and can control other components in an electronic device to perform desired functions.

[0184] The memory may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, random access memory (RAM) and / or cache memory. The non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and a processor may execute the program instructions to implement a multi-source distribution network fault location method according to any embodiment of this application described above, and / or other desired functions. Various contents such as initial external parameters and thresholds may also be stored in the computer-readable storage medium.

[0185] Furthermore, embodiments of this application may also be computer-readable storage media storing computer program instructions thereon, which, when executed by a processor, cause the processor to implement the multi-source distribution network fault location method provided in any embodiment of this application.

[0186] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the technical solutions of the embodiments of the present invention.

Claims

1. A method for locating fault sections in a multi-source distribution network, characterized in that, Includes the following steps: S1: Collect real-time operating data of the multi-source distribution network; S2: Perform data preprocessing on the real-time running data to obtain preprocessed real-time running data; S3: Extract multi-dimensional fault features from the preprocessed real-time running data; The multi-dimensional fault characteristics include electrical characteristics, spatiotemporal characteristics, and power supply influence characteristics; S4: Perform initial fault location based on the multi-dimensional fault characteristics; Specifically: S4.1: Calculate the electrical feature matching degree C based on the electrical features in the multidimensional fault features. elec S4.2: Calculate the spatiotemporal feature matching degree C based on the spatiotemporal features in the multidimensional fault features. time S4.3: Calculate the power supply influence feature matching degree C based on the power supply influence feature in the multidimensional fault features. source ; S4.4: Calculate the comprehensive score S of the segment to be located. i S4.5: Perform initial location of the faulty section based on the comprehensive score of the section to be located; S5: Fine-grained fault segment localization based on timing logic and topological constraints; Specifically: S5.1: Eliminate non-faulty sections based on the fault propagation sequence; S5.2: Eliminate non-faulty sections based on the uniqueness of the fault current path.

2. The method for locating fault sections in a multi-source distribution network according to claim 1, characterized in that: In S3, the electrical fault characteristics include current direction, current amplitude, and voltage drop; the spatiotemporal characteristics include the temporal correlation of the fault event and the spatial topology location of the monitoring terminal; and the power supply impact characteristics include the output status of the distributed power supply and the proportion of reverse current of the distributed power supply.

3. The method for locating fault sections in a multi-source distribution network according to claim 2, characterized in that: The process for extracting the current direction is as follows: the current phase acquired by the smart switch is compared with the system reference voltage phase to calculate the current phase angle θ. i If θ i If θ ∈ [90°, 270°], then the current direction is determined to be the inflow section, and the fault point is located downstream of this section; if θ i If the value is in the range [-90°, 90°), it is determined to be an outflow section, and the fault point is located upstream of this section. The current amplitude extraction process is as follows: Calculate the effective value I of the current at each segment end by sampling the current waveform. rms As the magnitude of the fault current; The voltage sag extraction process is as follows: calculate the rate of change of voltage sag amplitude collected by the upstream and downstream distribution terminals of the fault section.

4. The method for locating fault sections in a multi-source distribution network according to claim 2, characterized in that: The process for extracting the time correlation of the fault events is as follows: record the timestamps of key events, including: the overcurrent trigger time t of the fault indicator. fi The tripping time t of the intelligent switch trip , Voltage drop start time t of distribution terminal volt_drop ; The process of extracting the spatial topology location of the monitoring terminal is as follows: construct a section-terminal adjacency matrix based on the physical connection relationship of the distribution network; the physical connection relationship of the distribution network includes the layout relationship of feeder section switches and tie switches.

5. The method for locating fault sections in a multi-source distribution network according to claim 2, characterized in that: The extraction process for the reverse current ratio of the distributed power source is as follows: For the section of the distributed power source, calculate the amplitude I of the fault current injected by the distributed power source. dg and the total fault current I of the system total By comparison, the proportion of reverse current of distributed power sources is obtained.

6. The method for locating fault sections in a multi-source distribution network according to claim 1, characterized in that: Specifically, S4.1 is as follows: Current direction consistency C elec,方向 Assignment: If the fault current detected by all associated terminals within the segment to be located points in the same direction, then C elec,方向 =1; if a contradiction exists, then C elec,方向 =0.5; Current amplitude anomaly C elec,幅值 Assignment: Calculate the ratio of the effective values ​​of the currents at the beginning and end of the section to be located; If r current ≥1.5 (indicating the fault point is close to the beginning and the current attenuation is not significant), then C elec,幅值 =1; If 1.2≤r current If C < 1.5, then C elec,幅值 =0.7; otherwise C elec,幅值 =0.3; Voltage drop significance C elec,电压 Assignment: Calculate the voltage amplitude difference ΔU between the beginning and end of the section to be located; if ΔU ≥ 0.3U base Then C elec,电压 =1; if 0.1U base ≤ΔU<0.3U base Then C elec,电压 =0.6; otherwise C elec,电压 =0.2, the U base The reference voltage value for the section to be located; The final electrical feature matching degree is the weighted score of each sub-item, calculated using the following formula: C elec =w1·C elec,方向 +w2·C elec,幅值 +w3·C elec,电压 ; In the formula, ω1, ω2, and ω3 are coefficients.

7. The method for locating fault sections in a multi-source distribution network according to claim 6, characterized in that: Specifically, S4.2 is as follows: Time sequence rationality C time,顺序 Assignment: If the fault indicator overcurrent time t fi Earlier than the associated smart switch tripping time t trip If it is more than 2ms, then C time,顺序 =1; if the difference Δt between the tripping time of the associated smart switch and the overcurrent time of the fault indicator is less than 2ms, then C time,顺序 =0.4; if the fault indicator overcurrent time t fi Greater than the associated smart switch tripping time t trip Then C time,顺序 =0; Topological location consistency C time,拓扑 Assignment: If the segment to be located is located at the expected position on the fault propagation path, then C time,拓扑 =1; otherwise Ctime,topology = 0.5; The spatiotemporal feature matching degree is the weighted score of each sub-item, calculated using the following formula: C time <w4·C time,顺序 +w5·C time,拓扑 ; In the formula, ω4 and ω5 are coefficients.

8. The method for locating fault sections in a multi-source distribution network according to claim 7, characterized in that: Specifically, S4.3 is as follows: Rationality C of reverse current ratio source,反向 Assignment: If the reverse current ratio r of the to-be-located section ≤ 0.3, then C source,反向 = 1; if 0.3 < r ≤ 0.6, then C source,反向 = 0.6; if r > 0.6, then C source,反向 = 0.2; Power output status compatibility C source,出力 Assignment: If the distributed power source in the section to be located is in a generating state and the reverse current ratio r > 0, then C source,出力 =1; If the distributed power source in the section to be located is in the power generation state and the reverse current ratio r≤0, C source,出力 =0.5; if the distributed power source shuts down, then C source,出力 =1; The matching degree of the power supply influence features is the weighted score of each sub-item, calculated using the following formula: C source =w6·C source，反向 +w7·C source，出力 ; In the formula, ω6 and ω7 are coefficients.

9. The method for locating fault sections in a multi-source distribution network according to claim 7, characterized in that: Specifically, S4.5 is as follows: If there is a comprehensive score S of the only section to be located i ≥ 0.8, directly determine that the section to be located is a faulty section; if the scores of all sections are < 0.5, trigger a positioning failure alarm; if there are multiple sections with close comprehensive scores and the comprehensive score 0.5 ≤ S < 0.8, and there are faulty conflict sections at this time, then proceed to the next step.

10. A fault section location system for a multi-source distribution network, characterized in that, The system employs the multi-source distribution network fault section location method according to any one of claims 1-9, and the system comprises: The data acquisition module is used to collect real-time operating data of the multi-source distribution network; The preprocessing module is used to perform data preprocessing operations on the real-time running data to obtain preprocessed real-time running data; The feature extraction module is used to extract multi-dimensional fault features from the preprocessed real-time running data; The initial location module is used to perform initial location of the fault section based on the multi-dimensional fault characteristics. The conflict section elimination module is used for fine-grained fault section location based on timing logic and topological constraints.

Images