Power distribution network fault processing method

Abstract

The invention relates to the technical field of power grids, in particular to a power distribution network fault processing method. Comprising the following steps: carrying out fault detection by comprehensively considering various factors such as current, voltage, temperature and humidity through a multi-parameter detection method; through an intelligent algorithm, machine learning and a neural network, detection data are analyzed, and the accuracy and reliability of fault detection are improved; the single-phase earth fault positioning method mainly comprises the following steps of zero-mode traveling wave extraction, S transformation time-frequency matrix construction, fault feature extraction and a fault positioning algorithm. Through an automatic recovery system, according to a preset recovery strategy and remote control, rapid recovery of a fault is realized; through the intelligent scheduling system, the recovery path and sequence are optimized, and the recovery efficiency and safety are improved. The power distribution network fault processing method provided by the invention has the advantages of accurate fault detection, accurate fault positioning, high fault recovery efficiency and small influence of environmental factors.

Description

Technical Field

[0001] The invention relates to the technical field of power grids, and in particular to a method for handling distribution network faults. Background Art

[0002] The distribution network is an important part of the power system, responsible for delivering electricity from substations to end users. In actual operation, the distribution network often faces various faults, such as short circuits, line breaks, equipment failures, etc. These faults not only affect the stability and reliability of power supply, but may also cause serious safety accidents. The existing distribution network fault handling methods have the following shortcomings:

[0003] 1. Existing distribution network fault detection methods mainly rely on changes in current and voltage, but these changes may be affected by a variety of factors, resulting in inaccurate fault detection. For example, factors such as load changes and line interference may cause fluctuations in current and voltage, leading to false alarms or missed alarms.

[0004] 2. Existing distribution network fault location methods mainly rely on fault indicators and line inspections, but these methods have low positioning accuracy and are time-consuming. For example, the fault indicator can only roughly indicate the fault section and cannot accurately locate the fault point.

[0005] 3. Existing distribution network fault recovery methods mainly rely on manual operation, which has low recovery efficiency and is prone to errors. For example, when manually operating switches and load transfers, there may be operational errors, resulting in secondary faults.

[0006] 4. Traditional fault location methods are sensitive to environmental factors (such as noise and interference), which can easily lead to false alarms or missed alarms. For example, environmental factors such as lightning strikes and electromagnetic interference may affect the detection results of current and voltage.

[0007] Therefore, in view of this, the existing technology is studied and improved, and a distribution network fault handling method is proposed. Summary of the invention

[0008] The problem to be solved by the present invention is that fault detection is inaccurate, fault location is imprecise, fault recovery efficiency is low, and it is greatly affected by environmental factors.

[0009] The technical solution adopted by the present invention is: a method for handling distribution network faults, comprising the following steps:

[0010] Step 1: Use multi-parameter detection methods to comprehensively consider multiple factors such as current, voltage, temperature, and humidity to perform fault detection;

[0011] Step 2: Analyze the detection data through intelligent algorithms, machine learning and neural networks to improve the accuracy and reliability of fault detection;

[0012] Step 3: Single-phase grounding fault location method, the main steps are as follows:

[0013] S1. Zero-mode traveling wave extraction: Through the zero-mode traveling wave extraction technology, the zero-mode traveling wave component is extracted from the fault signal to reduce the interference of other mode traveling waves;

[0014] S2, S transform time-frequency matrix construction: Use S transform technology to construct the time-frequency matrix of zero-mode traveling waves to achieve multi-resolution analysis of fault signals;

[0015] S3, Fault feature extraction: By extracting features from the time-frequency matrix, the time and location of the fault are identified;

[0016] S4, Fault location algorithm: Based on the extracted features, a fault location algorithm is designed to achieve accurate location of single-phase grounding faults;

[0017] Step 4: Through the automated recovery system, based on the preset recovery strategy and remote control, rapid recovery of the fault is achieved;

[0018] Step 5: Optimize the recovery path and sequence through the intelligent scheduling system to improve recovery efficiency and safety.

[0019] As a further solution of the present invention: the multi-parameter detection method includes but is not limited to current detection, voltage detection, temperature detection and humidity detection;

[0020] Current detection monitors the current changes in real time through current transformers installed at each node of the distribution network; when the current suddenly exceeds the set threshold, fault detection is triggered;

[0021] Voltage detection monitors voltage changes in real time through voltage transformers installed at each node of the distribution network; when voltage mutations or fluctuations exceed the set threshold, fault detection is triggered;

[0022] Temperature detection uses temperature sensors installed on key equipment and nodes to monitor temperature changes in real time; when the temperature exceeds the set threshold, fault detection is triggered;

[0023] Humidity detection monitors humidity changes in real time through humidity sensors installed on key equipment and nodes; when the humidity exceeds the set threshold, fault detection is triggered.

[0024] As a further solution of the present invention: the intelligent algorithm includes data preprocessing, feature extraction, model training, and fault identification; data preprocessing is to normalize data of different parameters to the same dimension for subsequent processing; noise is removed by filtering technology to improve data quality; feature extraction is to extract the time domain characteristics of the mean, variance, and peak of current and voltage; frequency characteristics are extracted by Fourier transform; time-frequency characteristics are extracted by S transform; model training is to use support vector machines and random forest algorithms to train fault detection models; deep learning technology, convolutional neural networks, and recurrent neural networks are used to train fault detection models; fault identification is to classify real-time data through trained models and identify fault types; and the fault location is accurately located by combining geographic information systems and topological structures.

[0025] As a further solution of the present invention: the single-phase grounding fault locating method includes a method for determining a faulty feeder in a distribution network and a method for accurately locating a traveling wave modulus time difference.

[0026] As a further solution of the present invention: the method for determining a faulty feeder in the distribution network has the following specific steps:

[0027] S1. Fault traveling wave characteristics: The high-voltage busbar of the substation is connected to n distribution network feeders, and the parameters of each feeder are the same, that is, the wave impedance is consistent; assuming that a single-phase grounding fault occurs in feeder 1, according to the refraction and reflection law of the traveling wave, the amplitude expression of the traveling wave is calculated as: In the formula, u F is the voltage traveling wave originating from the fault point; u A is the voltage traveling wave on the bus; i F is the current traveling wave of the fault feeder near the busbar; i N is the current traveling wave of other feeders except the fault feeder near the busbar; z is the wave impedance;

[0028] S2. Methods for determining faulty feeders: S-transformation principle, selection of traveling wave modulus, and establishment of a fault comparison library;

[0029] S-transformation principle: Assuming there is a waveform data signal h(t), the definition of its S-transformation is: In the formula, τ is time; f is frequency; i is the imaginary unit; the result of S transformation is a two-dimensional matrix, which contains the time-frequency characteristic information of the traveling wave, and the elements of the matrix are complex numbers;

[0030] Traveling wave modulus selection: When a single-phase grounding occurs in the power line of the distribution network, the traveling wave will propagate from the fault point to both ends of the line and even the entire network. Due to the mutual coupling of electricity and magnetism between the three-phase lines A, B, and C, it is difficult to analyze the propagation characteristics of the traveling wave of a certain phase alone. Therefore, it is necessary to decouple the traveling wave of each phase before analyzing it; Karenbauer transformation is selected for decoupling processing to decompose the three-phase voltage components of A, B, and C into independent zero-mode, 1-mode, and 2-mode components;

[0031]

[0032] Establishment of fault comparison library: A simulation model was built in PSCAD, and traveling wave detection devices were installed at the ends of feeders L1, L2, and L3, and then single-phase grounding faults were set at f1 to f8, and the intercepted traveling wave information was S-transformed to obtain the time-frequency matrix, so as to establish a data comparison library for different fault feeder branches. The comparison library contains the time-frequency matrix library E of each typical branch voltage zero-mode traveling wave after S-transformation. M×N ; M is the number of feeders, N is the case of different grounding resistances;

[0033]

[0034] According to the waveform similarity principle, when a single-phase fault occurs in the distribution network, the traveling wave waveforms measured by the traveling wave detection devices installed at the ends of each feeder are compared for similarity to determine the branch where the fault is located; the waveform similarity calculation formula is: Where E(i,j) is the element of the jth time sampling point of the i-th frequency of the time-frequency matrix in the comparison library; Q(i,j) is the element of the jth time sampling point of the i-th frequency of the time-frequency matrix after the S-transformation of the voltage zero-mode traveling wave measured at the feeder terminal when a fault occurs; ρ is the similarity between the two matrices, and its value is [0, 1]. The closer ρ is to 1, the higher the similarity of the two matrices.

[0035] As a further solution of the present invention: the precise positioning method of the traveling wave modulus time difference is mainly divided into the following three cases:

[0036] 1) Fault located on the main feeder: When it is determined that the ground fault occurs on the main feeder, the wave head arrival time of the line mode traveling wave is measured by the traveling wave detection device at the end of the feeder and the adjacent line, and the accurate location of the fault is located according to the double-end positioning method. The formula for double-end positioning is: Where d is the distance from the fault point to the detection point at the end of the i-th feeder; t i ,t j are the arrival times of the traveling waves recorded at the terminals of the i-th and j-th feeders respectively; L ijis the sum of the lengths of the i-th and j-th feeder lines; v is the speed of the traveling wave, and the line mode traveling wave is used for positioning in double-end positioning;

[0037] 2) Fault is located on the primary branch: When the fault occurs on the primary branch, the electrical distance between the fault location and the traveling wave measuring device at the end of the feeder is obtained based on the relationship between the zero-mode and line-mode traveling wave velocity difference and the fault location, thereby calculating the precise location of the fault on the primary branch; the improved double-end positioning method derived from the zero-mode and line-mode velocity difference principle is used to accurately locate the branch; the improved double-end positioning formula based on the zero-mode and line-mode velocity difference principle eliminates the disadvantage that the zero-mode velocity is difficult to determine, and the fault location is completed based on the velocity of the line-mode traveling wave and the arrival time of the zero-mode and line-mode traveling waves. The positioning formula is: Where d is the calculated electrical distance from the fault location to the detection point at the end of the i-th feeder; t i1 ,t j1 are the arrival times of the line mode traveling waves recorded by the traveling wave detection devices at the ends of the i-th and j-th feeder lines respectively; t i0 ,t j0 are the arrival times of the zero-mode traveling wave recorded by the traveling wave detection devices at the ends of the i-th and j-th feeder lines respectively; v is the velocity of the line-mode traveling wave;

[0038] 3) Fault is located in the secondary branch: When the fault occurs in the secondary branch, the traveling wave detection device at the end of the faulty line and the adjacent line is used to improve the double-end positioning calculation, and the traveling wave detection device at the end of the healthy line with fewer branches is introduced to perform positioning calculation using single-end positioning. When the error between the two is not greater than the given error, the fault point position is accurately measured. The single-end positioning formula is: Where v is the speed of the linear mode traveling wave; t i2 is the arrival time of the reflected traveling wave at the fault point measured at the end of the i-th feeder; t i1 is the arrival time of the initial fault traveling wave measured at the end of the i-th feeder.

[0039] As a further solution of the present invention: the automated recovery system includes a recovery strategy and remote control. The recovery strategy is to preset a variety of recovery strategies based on the topological structure and historical data of the distribution network, including load transfer and backup power supply switching; the recovery strategy is dynamically adjusted according to the real-time fault situation to ensure the optimal recovery path; the remote control is to monitor the operating status of the distribution network in real time through the SCADA system and detect faults in time; and the recovery operations of switch operation and load transfer are realized through remote control technology.

[0040] As a further solution of the present invention: the intelligent scheduling system includes recovery path optimization, recovery sequence optimization, and security assurance; recovery path optimization is to use graph theory algorithms, Dijkstra algorithm, A* algorithm to optimize the recovery path; evaluate the recovery time and cost of different paths and select the optimal path; recovery sequence optimization is to set the recovery sequence according to the user's priority and importance; dynamically adjust the recovery sequence according to the real-time fault situation and recovery progress to ensure efficient recovery; security assurance is to perform a safety check before executing the recovery operation to ensure the safety of the operation; during the recovery process, the fault area is isolated in time to prevent the fault from spreading.

[0041] Beneficial effects of the present invention:

[0042] The present invention monitors the changes of current, voltage, temperature and humidity in real time by installing current transformers, voltage transformers, temperature sensors and humidity sensors at each node of the distribution network. When the changes of these parameters exceed the set threshold, fault detection is triggered. By comprehensively considering multiple parameters, the uncertainty of single parameter detection is reduced and the accuracy of fault detection is improved. The multi-parameter detection method can more comprehensively reflect the operating status of the distribution network and reduce the occurrence of false alarms and missed alarms.

[0043] The present invention uses intelligent algorithms such as support vector machines, random forests, convolutional neural networks, and recurrent neural networks to analyze detection data and improve the accuracy and reliability of fault detection through data preprocessing, feature extraction, model training, and fault identification. Intelligent algorithms can effectively process complex data and improve the reliability and accuracy of fault detection. The automated processing flow reduces human intervention and improves the efficiency of fault handling.

[0044] The present invention uses zero-mode traveling wave extraction technology to extract zero-mode traveling wave components from fault signals and reduce interference from other mode traveling waves. The S transform technology is used to construct a time-frequency matrix of zero-mode traveling waves to achieve multi-resolution analysis of fault signals. By extracting features from the time-frequency matrix, the time and location of the fault are identified. Based on the extracted features, a fault location algorithm is designed to achieve accurate location of single-phase grounding faults. The S transform technology has high resolution and can accurately extract the time-frequency characteristics of the fault signal to improve the accuracy of fault location. The S transform technology has good noise resistance and robustness, and can effectively reduce the impact of environmental factors on fault location.

[0045] The present invention realizes rapid recovery of faults through preset recovery strategies and remote control. The recovery strategies include load transfer, backup power supply switching, etc., and the recovery strategies are dynamically adjusted according to the real-time fault conditions. Remote control monitors the operating status of the distribution network in real time through the SCADA system, detects faults in time and performs recovery operations. The automated recovery system can respond to faults quickly, shorten recovery time and improve recovery efficiency. Remote control technology reduces manual operations, reduces operational risks and improves safety.

[0046] The present invention optimizes the recovery path and sequence to ensure efficient recovery through recovery path optimization, recovery sequence optimization and security assurance. Recovery path optimization uses graph theory algorithms, such as Dijkstra algorithm and A* algorithm, to evaluate the recovery time and cost of different paths and select the optimal path. Recovery sequence optimization sets the recovery sequence according to the user's priority and importance, and dynamically adjusts the recovery sequence. Security assurance performs a safety check before executing the recovery operation to ensure the safety of the operation, isolate the fault area in time, and prevent the fault from spreading. The intelligent scheduling system can optimize the recovery path and sequence to ensure efficient recovery, while ensuring the safety of the operation and preventing the fault from spreading. By optimizing the recovery path and sequence, priority recovery of important users is ensured, improving user experience. DETAILED DESCRIPTION

[0047] The present invention will be further described below.

[0048] Example 1

[0049] A method for handling a distribution network fault comprises the following steps:

[0050] Step 1: Perform fault detection by comprehensively considering multiple factors such as current, voltage, temperature, and humidity through a multi-parameter detection method; the multi-parameter detection method includes but is not limited to current detection, voltage detection, temperature detection, and humidity detection;

[0051] Current detection monitors the current changes in real time through current transformers installed at each node of the distribution network; when the current suddenly exceeds the set threshold, fault detection is triggered;

[0052] Voltage detection monitors voltage changes in real time through voltage transformers installed at each node of the distribution network; when voltage mutations or fluctuations exceed the set threshold, fault detection is triggered;

[0053] Temperature detection uses temperature sensors installed on key equipment and nodes to monitor temperature changes in real time; when the temperature exceeds the set threshold, fault detection is triggered;

[0054] Humidity detection monitors humidity changes in real time through humidity sensors installed on key equipment and nodes; when the humidity exceeds the set threshold, fault detection is triggered.

[0055] Step 2: Analyze detection data through intelligent algorithms, machine learning and neural networks to improve the accuracy and reliability of fault detection; intelligent algorithms include data preprocessing, feature extraction, model training, and fault identification; data preprocessing is to normalize data of different parameters to the same dimension for subsequent processing; remove noise through filtering technology to improve data quality; feature extraction is to extract the time domain characteristics of the mean, variance, and peak of current and voltage; extract frequency characteristics through Fourier transform; extract time-frequency characteristics through S transform; model training is to use support vector machines and random forest algorithms to train fault detection models; use deep learning technology, convolutional neural networks, and recurrent neural networks to train fault detection models; fault identification is to classify real-time data through trained models and identify fault types; combine geographic information systems and topological structures to accurately locate fault locations.

[0056] Step 3: Single-phase grounding fault location method, the main steps are as follows:

[0057] S1. Zero-mode traveling wave extraction: Through the zero-mode traveling wave extraction technology, the zero-mode traveling wave component is extracted from the fault signal to reduce the interference of other mode traveling waves;

[0058] S2, S transform time-frequency matrix construction: Use S transform technology to construct the time-frequency matrix of zero-mode traveling waves to achieve multi-resolution analysis of fault signals;

[0059] S3, Fault feature extraction: By extracting features from the time-frequency matrix, the time and location of the fault are identified;

[0060] S4, Fault location algorithm: Based on the extracted features, a fault location algorithm is designed to achieve accurate location of single-phase grounding faults;

[0061] The single-phase grounding fault location method includes a method for determining the faulty feeder of the distribution network and a method for accurately locating the traveling wave modulus time difference.

[0062] The specific steps of the method for determining the faulty feeder in the distribution network are as follows:

[0063] S1. Fault traveling wave characteristics: The high-voltage busbar of the substation is connected to n distribution network feeders, and the parameters of each feeder are the same, that is, the wave impedance is consistent; assuming that a single-phase grounding fault occurs in feeder 1, according to the refraction and reflection law of the traveling wave, the amplitude expression of the traveling wave is calculated as: In the formula, u F is the voltage traveling wave originating from the fault point; u A is the voltage traveling wave on the bus; i F is the current traveling wave of the fault feeder near the busbar; i N is the current traveling wave of other feeders except the fault feeder near the busbar; z is the wave impedance;

[0064] According to the analysis of the amplitude expression of the traveling wave, the voltage and current traveling waves measured at the head end of the faulty feeder have opposite polarities, while the voltage and current traveling waves measured at other normal feeders have the same polarity. Therefore, the faulty feeder can be identified by the polarity of the traveling wave; however, when the grounding resistance is large and the fault occurs when the initial phase angle is close to 0, the amplitudes of the voltage and current traveling waves decrease, making it difficult to accurately distinguish the polarities of the voltage and current traveling waves, and thus it is impossible to accurately identify the faulty feeder. Therefore, it is necessary to use the time-frequency matrix after the traveling wave S transform to select the fault line;

[0065] S2. Methods for determining faulty feeders: S-transformation principle, selection of traveling wave modulus, and establishment of a fault comparison library;

[0066] S transform principle: The S transform has high resolution and the window can change with the frequency. It has the advantages of both wavelet transform and short-time Fourier transform, and has good resolution in both time domain and frequency domain. Therefore, it has a good effect on analyzing non-stationary signals such as distribution network fault traveling waves. Assuming that there is a waveform data signal h(t), the definition of its S transform is: In the formula, τ is time; f is frequency; i is the imaginary unit; the result of S transformation is a two-dimensional matrix, which contains the time-frequency characteristic information of the traveling wave, and the elements of the matrix are complex numbers; the amplitude and phase information of the traveling wave can be obtained from it, and the elements of each column represent the information at different times, and the elements of each row represent the information at different frequencies;

[0067] Traveling wave modulus selection: When a single-phase grounding occurs in the power line of the distribution network, the traveling wave will propagate from the fault point to both ends of the line and even the entire network. Due to the mutual coupling of electricity and magnetism between the three-phase lines A, B, and C, it is difficult to analyze the propagation characteristics of the traveling wave of a certain phase alone. Therefore, it is necessary to decouple the traveling wave of each phase before analyzing it; Karenbauer transformation is selected for decoupling processing to decompose the three-phase voltage components of A, B, and C into independent zero-mode, 1-mode, and 2-mode components;

[0068]

[0069] Because the zero-mode component of the traveling wave has a larger amplitude and a higher attenuation degree than the line-mode component, it contains a large amount of fault characteristic information that affects the attenuation process of the traveling wave and is easy to extract. Therefore, this paper selects the zero-mode component of the traveling wave for in-depth analysis and forms a time-frequency matrix through S-transformation to extract the time-frequency characteristics of the fault traveling wave.

[0070] Establishment of fault comparison library: A simulation model was built in PSCAD, and traveling wave detection devices were installed at the ends of feeders L1, L2, and L3, and then single-phase grounding faults were set at f1 to f8, and the intercepted traveling wave information was S-transformed to obtain the time-frequency matrix, so as to establish a data comparison library for different fault feeder branches. The comparison library contains the time-frequency matrix library E of each typical branch voltage zero-mode traveling wave after S-transformation. M×N ; M is the number of feeders, N is the case of different grounding resistances;

[0071]

[0072] According to the waveform similarity principle, when a single-phase fault occurs in the distribution network, the traveling wave waveforms measured by the traveling wave detection devices installed at the ends of each feeder are compared for similarity to determine the branch where the fault is located; the waveform similarity calculation formula is: In the formula, E(i,j) is the element of the jth time sampling point of the i-th frequency of the time-frequency matrix in the comparison library; Q(i,j) is the element of the jth time sampling point of the i-th frequency of the time-frequency matrix after the voltage zero-mode traveling wave S transformation measured at the feeder terminal when the fault occurs; ρ is the similarity between the two matrices, and its value is [0, 1]. The closer ρ is to 1, the higher the similarity of the two matrices. The higher the similarity with the corresponding branch time-frequency matrix in the comparison library, the greater the possibility that the fault occurs in this branch.

[0073] The precise positioning method of the traveling wave modulus time difference is mainly divided into the following three cases:

[0074] 1) Fault located on the main feeder: When it is determined that the ground fault occurs on the main feeder, the wave head arrival time of the line mode traveling wave is measured by the traveling wave detection device at the end of the feeder and the adjacent line, and the accurate location of the fault is located according to the double-end positioning method. The formula for double-end positioning is: Where d is the distance from the fault point to the detection point at the end of the i-th feeder; t i ,t j are the arrival times of the traveling waves recorded at the terminals of the i-th and j-th feeders respectively; L ij is the sum of the lengths of the i-th and j-th feeder lines; v is the speed of the traveling wave, and the line mode traveling wave is used for positioning in double-end positioning;

[0075] 2) Fault is located in the primary branch: When the fault occurs in the primary branch of the feeder, if it is calculated according to the double-end positioning formula, only the branch node where the fault occurs can be obtained, but the specific location of the fault on the branch cannot be accurately obtained. In addition, due to the large number of branches, it is difficult for single-end positioning to identify the first reflected wave head of the fault point. Therefore, when the fault occurs on the primary branch, the electrical distance between the fault location and the traveling wave measuring device at the end of the feeder is obtained according to the relationship between the zero-mode and line-mode traveling wave velocity difference and the fault location, so as to calculate the precise location of the fault on the primary branch. The improved double-end positioning method derived from the zero-mode and line-mode wave velocity difference principle is used to accurately locate the branch. The double-end positioning formula based on the improved zero-mode and line-mode wave velocity difference principle eliminates the disadvantage that the zero-mode wave velocity is difficult to determine, and the fault location is completed according to the wave velocity of the line-mode traveling wave and the arrival time of the zero-mode and line-mode traveling waves. The positioning formula is: Where d is the calculated electrical distance from the fault location to the detection point at the end of the i-th feeder; t i1 ,t j1 are the arrival times of the line mode traveling waves recorded by the traveling wave detection devices at the ends of the i-th and j-th feeder lines respectively; t i0 ,t j0 are the arrival times of the zero-mode traveling wave recorded by the traveling wave detection devices at the ends of the i-th and j-th feeder lines respectively; v is the velocity of the line-mode traveling wave;

[0076] 3) Fault located at the secondary branch: When the fault occurs in the secondary branch of the feeder, due to the large number of branches, the refraction and reflection are complex, which may affect the accurate calibration of the arrival time of the zero mode and line mode traveling waves measured by the traveling wave detection device at the end of the feeder; therefore, when the fault occurs in the secondary branch, the traveling wave detection device at the end of the faulty line and the adjacent line is used to improve the double-end positioning calculation, and the traveling wave detection device at the end of the healthy line with few branches is introduced to perform positioning calculation using single-end positioning. When the error between the two is not greater than the given error, the fault point position is accurately measured. The single-end positioning formula is: Where v is the speed of the linear mode traveling wave; t i2 is the arrival time of the reflected traveling wave at the fault point measured at the end of the i-th feeder; t i1 is the arrival time of the initial fault traveling wave measured at the end of the i-th feeder.

[0077] Step 4: Through the automated recovery system, according to the preset recovery strategy and remote control, rapid recovery of faults can be achieved; the automated recovery system includes recovery strategy and remote control. The recovery strategy is to preset multiple recovery strategies based on the topology and historical data of the distribution network, including load transfer and backup power supply switching; according to the real-time fault situation, the recovery strategy is dynamically adjusted to ensure the optimal recovery path; remote control is to monitor the operating status of the distribution network in real time through the SCADA system and detect faults in time; through remote control technology, the recovery operation of switch operation and load transfer can be achieved.

[0078] Step 5: Through the intelligent scheduling system, optimize the recovery path and sequence to improve recovery efficiency and safety. The intelligent scheduling system includes recovery path optimization, recovery sequence optimization, and safety assurance; recovery path optimization uses graph theory algorithms, Dijkstra algorithm, and A* algorithm to optimize the recovery path; evaluate the recovery time and cost of different paths and select the optimal path; recovery sequence optimization sets the recovery sequence based on the user's priority and importance; dynamically adjusts the recovery sequence based on the real-time fault situation and recovery progress to ensure efficient recovery; safety assurance is to conduct a safety check before performing the recovery operation to ensure the safety of the operation; during the recovery process, isolate the fault area in time to prevent the fault from spreading.

[0079] Through the above improvements, the distribution network fault handling method of the present invention has achieved significant improvements in fault detection, location and recovery. The method not only improves the accuracy and reliability of fault detection, but also improves the accuracy of fault location and recovery efficiency, reduces the impact of environmental factors, and ensures the safety of operation.

[0080] Example 2

[0081] Application scenario of multi-parameter detection method: In the hot summer weather, a city's distribution network is prone to single-phase grounding faults due to increased load. It is necessary to monitor the operating status of the distribution network in real time to detect and handle faults in a timely manner.

[0082] Implementation steps:

[0083] 1. Install sensors: Install current transformers, voltage transformers, temperature sensors, and humidity sensors at key nodes of the distribution network.

[0084] Current transformer: installed at each node to monitor the current change in real time. Voltage transformer: installed at each node to monitor the voltage change in real time.

[0085] Temperature sensor: installed in key equipment and nodes to monitor temperature changes in real time.

[0086] Humidity sensor: installed in key equipment and nodes to monitor humidity changes in real time.

[0087] 2. Data collection: The current, voltage, temperature and humidity data are collected in real time through sensors and transmitted to the central monitoring system.

[0088] 3. Data processing:

[0089] Current detection: When the current suddenly changes beyond the set threshold (for example, more than 10% of the normal value), fault detection is triggered.

[0090] Voltage detection: When the voltage suddenly changes or fluctuates beyond a set threshold (for example, more than 5% of the normal value), fault detection is triggered.

[0091] Temperature detection: When the temperature exceeds the set threshold (for example, over 40°C), fault detection is triggered.

[0092] Humidity detection: When the humidity exceeds a set threshold (for example, more than 80%), fault detection is triggered.

[0093] 4. Fault detection: The central monitoring system comprehensively determines whether there is a fault based on the collected data. If multiple parameters exceed the threshold at the same time, the possibility of fault is further confirmed.

[0094] The multi-parameter detection method can more comprehensively reflect the operating status of the distribution network and reduce the occurrence of false alarms and missed alarms. Real-time measurement and data processing can detect potential faults in time, take measures in advance, and avoid the expansion of accidents.

[0095] Example 3

[0096] Application scenario of intelligent algorithms in fault detection: During the operation of the distribution network of an industrial park, it is necessary to accurately detect and classify complex fault conditions.

[0097] Implementation steps:

[0098] 1. Data preprocessing: Normalize the collected current, voltage, temperature and humidity data to eliminate dimensional differences. Use filtering technology to remove noise and improve data quality.

[0099] 2. Feature extraction:

[0100] Time domain features: Extract time domain features such as mean, variance, and peak of current and voltage.

[0101] Frequency characteristics: The frequency characteristics of current and voltage are extracted through Fourier transform.

[0102] Time-frequency characteristics: The time-frequency characteristics of current and voltage are extracted through S transform.

[0103] 3. Model training:

[0104] Support Vector Machine (SVM): Use SVM to train the fault detection model, train with historical data, and identify the fault type. Random Forest: Use the random forest algorithm to train the fault detection model, and make comprehensive judgments through multiple decision trees.

[0105] Deep learning: Use convolutional neural networks (CNN) and recurrent neural networks (RNN) to train fault detection models, and train them with large amounts of data to improve the accuracy and generalization ability of the models.

[0106] 4. Fault identification: Classify real-time data through trained models to identify fault types. Combined with geographic information systems and topological structures, accurately locate fault locations.

[0107] Intelligent algorithms can effectively process complex data and improve the reliability and accuracy of fault detection. Automated processing reduces human intervention and improves the efficiency of fault handling.

[0108] Example 4

[0109] Application scenario of the single-phase grounding fault location method: During the operation of a city's distribution network, a single-phase grounding fault occurred, and the fault position needed to be located quickly and accurately.

[0110] Implementation steps:

[0111] 1. Zero-mode traveling wave extraction: Through zero-mode traveling wave extraction technology, the zero-mode traveling wave component is extracted from the fault signal to reduce the interference of other mode traveling waves.

[0112] 2. S-transform time-frequency matrix construction: Use S-transform technology to construct the time-frequency matrix of zero-mode traveling waves to achieve multi-resolution analysis of fault signals.

[0113] 3. Fault feature extraction: By extracting features from the time-frequency matrix, the time and location of the fault can be identified.

[0114] 4. Fault location algorithm:

[0115] Determination of faulty feeder in distribution network:

[0116] Traveling wave characteristics: According to the refraction and reflection law of traveling waves, the amplitude expression of traveling waves is calculated, and the faulty feeder is identified by the polarity of traveling waves.

[0117] Time-frequency matrix after S transformation: The time-frequency matrix after S transformation is used to select the fault line and improve the accuracy of fault feeder determination.

[0118] Precise positioning of the time difference of the traveling wave modulus:

[0119] The fault is located on the main feeder: Through the double-end positioning method, the distance from the fault point to the detection point at the end of the feeder is calculated according to the arrival time of the traveling wave head.

[0120] The fault is located at the first-level branch: Based on the relationship between the speed difference of zero-mode and line-mode traveling waves and the fault location, the double-end positioning method is improved to accurately locate the fault location.

[0121] The fault is located at the secondary branch: the traveling wave detection device at the end of the faulty line and the adjacent line is used in combination with the single-ended positioning method to accurately locate the fault position.

[0122] The S transform technology has high resolution and can accurately extract the time-frequency characteristics of the fault signal to improve the accuracy of fault location. The S transform technology has good noise resistance and robustness, and can effectively reduce the impact of environmental factors on fault location.

[0123] Example 5

[0124] Application scenario of the automatic restoration system: After a single-phase grounding fault occurs in a city's distribution network, power supply needs to be restored quickly to ensure normal power supply for users.

[0125] Implementation steps:

[0126] 1. Recovery strategy:

[0127] Load transfer: Based on the topology and historical data of the distribution network, a load transfer strategy is preset to transfer the load in the fault area to the normally operating feeder.

[0128] Backup power supply switching: preset backup power supply switching strategy, when the main power supply fails, it automatically switches to the backup power supply to ensure the continuity of power supply.

[0129] 2. Dynamic adjustment: Dynamically adjust the recovery strategy based on the real-time fault situation to ensure the optimal recovery path.

[0130] 3. Remote control:

[0131] SCADA system: Through the SCADA system, the operating status of the distribution network can be monitored in real time and faults can be discovered in time.

[0132] Switch operation: Through remote control technology, switch operation can be realized to quickly isolate the fault area and restore power supply.

[0133] Load transfer: Through remote control technology, load transfer is achieved to ensure normal power consumption for users.

[0134] The automated recovery system can respond to failures quickly, shorten recovery time, and improve recovery efficiency. Remote control technology reduces manual operations, reduces operational risks, and improves safety.

[0135] Example 6

[0136] Application scenario of intelligent dispatching system: After multiple faults occur in a city's distribution network, it is necessary to optimize the recovery path and sequence to ensure efficient restoration of power supply.

[0137] Implementation steps:

[0138] 1. Recovery path optimization:

[0139] Graph theory algorithms: Use graph theory algorithms, such as Dijkstra algorithm and A* algorithm, to optimize the recovery path.

[0140] Evaluate paths: Evaluate the recovery time and cost of different paths and select the optimal path.

[0141] 2. Optimize the recovery order:

[0142] User priority: Set the recovery order based on user priority and importance.

[0143] Dynamic adjustment: Dynamically adjust the recovery order based on real-time fault conditions and recovery progress to ensure efficient recovery.

[0144] 3. Security assurance:

[0145] Safety check: Before performing a recovery operation, perform a safety check to ensure the safety of the operation.

[0146] Isolate the faulty area: During the recovery process, isolate the faulty area in a timely manner to prevent the fault from spreading.

[0147] The intelligent scheduling system can optimize the recovery path and sequence to ensure efficient recovery, while ensuring the safety of operations and preventing the spread of faults. By optimizing the recovery path and sequence, priority recovery is ensured for important users, improving user experience.

[0148] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for handling distribution network faults, characterized in that: The following steps are involved: Step 1: Use multi-parameter detection method to comprehensively consider multiple factors such as current, voltage, temperature and humidity to perform fault detection; Step 2: Analyze detection data through intelligent algorithms, machine learning and neural networks to improve the accuracy and reliability of fault detection; Step 3: Single-phase grounding fault location method, the main steps are as follows: S1. Zero-mode traveling wave extraction: Through the zero-mode traveling wave extraction technology, the zero-mode traveling wave component is extracted from the fault signal to reduce the interference of other mode traveling waves; S2, S transform time-frequency matrix construction: Use S transform technology to construct the time-frequency matrix of zero-mode traveling waves to achieve multi-resolution analysis of fault signals; S3, Fault feature extraction: By extracting features from the time-frequency matrix, the time and location of the fault are identified; S4, Fault location algorithm: Based on the extracted features, a fault location algorithm is designed to achieve accurate location of single-phase grounding faults; Step 4: Through the automated recovery system, based on the preset recovery strategy and remote control, rapid recovery of the fault is achieved; Step 5: Optimize the recovery path and sequence through the intelligent scheduling system to improve recovery efficiency and safety.

2. A distribution network fault processing method according to claim 1, characterized in that: The multi-parameter detection method includes but is not limited to current detection, voltage detection, temperature detection and humidity detection; Current detection monitors the current changes in real time through current transformers installed at each node of the distribution network; when the current suddenly exceeds the set threshold, fault detection is triggered; Voltage detection monitors voltage changes in real time through voltage transformers installed at each node of the distribution network; when voltage mutations or fluctuations exceed the set threshold, fault detection is triggered; Temperature detection uses temperature sensors installed on key equipment and nodes to monitor temperature changes in real time; When the temperature exceeds the set threshold, fault detection is triggered; Humidity detection monitors humidity changes in real time through humidity sensors installed on key equipment and nodes; when the humidity exceeds the set threshold, fault detection is triggered.

3. A distribution network fault processing method according to claim 1, characterized in that: The intelligent algorithm includes data preprocessing, feature extraction, model training, and fault identification; Data preprocessing is to normalize the data of different parameters to the same dimension to facilitate subsequent processing; Remove noise through filtering technology to improve data quality; feature extraction is to extract the time domain characteristics of the mean, variance, and peak value of current and voltage; extract frequency characteristics through Fourier transform; extract time-frequency characteristics through S transform; Model training uses support vector machines and random forest algorithms to train fault detection models. Deep learning technology, convolutional neural networks, and recurrent neural networks are used to train fault detection models. Fault identification uses trained models to classify real-time data and identify fault types. The fault location is accurately located by combining geographic information systems and topological structures.

4. A distribution network fault processing method according to claim 1, characterized in that: The single-phase grounding fault locating method comprises a method for determining a faulty feeder of a distribution network and a method for accurately locating a traveling wave modulus time difference.

5. A distribution network fault processing method according to claim 4, characterized in that: The specific steps of the method for determining a faulty feeder in a distribution network are as follows: S1. Fault traveling wave characteristics: The high-voltage busbar of the substation is connected to n distribution network feeders, and the parameters of each feeder are the same, that is, the wave impedance is consistent; assuming that a single-phase grounding fault occurs in feeder 1, according to the refraction and reflection law of the traveling wave, the amplitude expression of the traveling wave is calculated as: In the formula, u F is the voltage traveling wave originating from the fault point; u A is the voltage traveling wave on the bus; i F is the current traveling wave of the fault feeder near the busbar; i N is the current traveling wave of other feeders except the fault feeder near the busbar; z is the wave impedance; S2. Methods for determining faulty feeders: S-transformation principle, selection of traveling wave modulus, and establishment of a fault comparison library; S-transformation principle: Assuming there is a waveform data signal h(t), the definition of its S-transformation is: In the formula, τ is time; f is frequency; i is the imaginary unit; the result of S transformation is a two-dimensional matrix, which contains the time-frequency characteristic information of the traveling wave, and the elements of the matrix are complex numbers; Traveling wave modulus selection: When a single-phase grounding occurs in the power line of the distribution network, the traveling wave will propagate from the fault point to both ends of the line and even the entire network. Due to the mutual coupling of electricity and magnetism between the three-phase lines A, B, and C, it is difficult to analyze the propagation characteristics of the traveling wave of a certain phase alone. Therefore, it is necessary to decouple the traveling wave of each phase before analyzing it; Karenbauer transformation is selected for decoupling processing to decompose the three-phase voltage components of A, B, and C into independent zero-mode, 1-mode, and 2-mode components; Establishment of fault comparison library: A simulation model was built in PSCAD, and traveling wave detection devices were installed at the ends of feeders L1, L2, and L3, and then single-phase grounding faults were set at f1 to f8, and the intercepted traveling wave information was S-transformed to obtain the time-frequency matrix, so as to establish a data comparison library for different fault feeder branches. The comparison library contains the time-frequency matrix library E of each typical branch voltage zero-mode traveling wave after S-transformation. M×n ; M is the number of feeders, N is the case of different grounding resistances; According to the waveform similarity principle, when a single-phase fault occurs in the distribution network, the traveling wave waveforms measured by the traveling wave detection devices installed at the ends of each feeder are compared for similarity to determine the branch where the fault is located; The waveform similarity calculation formula is: Where E(i,j) is the element of the jth time sampling point of the i-th frequency of the time-frequency matrix in the comparison library; Q(i,j) is the element of the jth time sampling point of the i-th frequency of the time-frequency matrix after the S-transformation of the voltage zero-mode traveling wave measured at the feeder terminal when a fault occurs; ρ is the similarity between the two matrices, and its value is [0, 1]. The closer ρ is to 1, the higher the similarity of the two matrices.

6. A distribution network fault processing method according to claim 5, characterized in that: The precise positioning method of the traveling wave modulus time difference is mainly divided into the following three cases: 1) Fault located on the main feeder: When it is determined that the ground fault occurs on the main feeder, the wave head arrival time of the line mode traveling wave is measured by the traveling wave detection device at the end of the feeder and the adjacent line, and the accurate location of the fault is located according to the double-end positioning method. The formula for double-end positioning is: Where d is the distance from the fault point to the detection point at the end of the i-th feeder; t i ,t j are the arrival times of the traveling waves recorded at the terminals of the i-th and j-th feeders respectively; L ij is the sum of the lengths of the i-th and j-th feeders; v is the speed of the traveling wave, and the line mode traveling wave is used for positioning in double-end positioning; 2) Fault located on the primary branch: When the fault occurs on the primary branch, the electrical distance between the fault location and the traveling wave measuring device at the end of the feeder is obtained based on the relationship between the zero-mode and line-mode traveling wave velocity difference and the fault location, thereby calculating the precise location of the fault on the primary branch; the improved double-end positioning method derived from the zero-mode and line-mode velocity difference principle is used to accurately locate the branch; the improved double-end positioning formula based on the zero-mode and line-mode velocity difference principle eliminates the disadvantage that the zero-mode velocity is difficult to determine, and the fault location is completed based on the velocity of the line-mode traveling wave and the arrival time of the zero-mode and line-mode traveling waves. The positioning formula is: Where d is the calculated electrical distance from the fault location to the detection point at the end of the i-th feeder; t i1 ,t j1 are the arrival times of the line mode traveling waves recorded by the traveling wave detection devices at the ends of the i-th and j-th feeder lines respectively; t i0 ,t j0 are the arrival times of the zero-mode traveling wave recorded by the traveling wave detection devices at the ends of the i-th and j-th feeder lines respectively; v is the velocity of the line-mode traveling wave; 3) Fault is located in the secondary branch: When the fault occurs in the secondary branch, the traveling wave detection device at the end of the faulty line and the adjacent line is used to improve the double-end positioning calculation, and the traveling wave detection device at the end of the healthy line with fewer branches is introduced to perform positioning calculation using single-end positioning. When the error between the two is not greater than the given error, the fault point position is accurately measured. The single-end positioning formula is: Where, v is the speed of the line mode traveling wave; t i2 is the arrival time of the reflected traveling wave at the fault point measured at the end of the i-th feeder; t i1 is the arrival time of the initial fault traveling wave measured at the end of the i-th feeder.

7. A distribution network fault processing method according to claim 1, characterized in that: The automatic recovery system includes recovery strategies and remote control. The recovery strategies are preset according to the topological structure and historical data of the distribution network, including load transfer and backup power supply switching; According to the real-time fault situation, the recovery strategy is adjusted dynamically to ensure the optimal recovery path; remote control is to monitor the operating status of the distribution network in real time through the SCADA system to detect faults in time; Through remote control technology, the recovery operation of switch operation and load transfer is realized.

8. A distribution network fault processing method according to claim 1, characterized in that: The intelligent scheduling system includes recovery path optimization, recovery sequence optimization, and security assurance; recovery path optimization uses graph theory algorithms, Dijkstra algorithm, and A* algorithm to optimize the recovery path; evaluates the recovery time and cost of different paths and selects the optimal path; recovery sequence optimization sets the recovery sequence based on the user's priority and importance; dynamically adjusts the recovery sequence based on the real-time fault situation and recovery progress to ensure efficient recovery; Safety assurance is to conduct a safety check before executing the recovery operation to ensure the safety of the operation; during the recovery process, the fault area is isolated in time to prevent the fault from spreading.

Images