Relay protection outlet matrix dynamic detection system

Abstract

The present invention relates to the field of relay detection technology, and specifically to a relay protection outlet matrix dynamic detection system, which includes the following modules: a power acquisition module; a data processing module, which pre-processes the acquired operating parameters; a dynamic matrix generation module, which generates a dynamic detection matrix according to the extracted features; a fault detection module, which detects faults in the power system in real time according to the generated dynamic detection matrix, including identifying the fault type, fault location and fault degree; a protection action module, which generates corresponding protection action instructions after the fault detection module detects the fault, and transmits them to the relay protection device; a communication module; and a storage module. The present invention improves the accuracy and real-time performance of fault detection, so that the system can respond to various complex power fault situations in a timely and accurate manner.

Description

Technical Field

[0001] The invention relates to the technical field of relay detection, and in particular to a relay protection outlet matrix dynamic detection system. Background Art

[0002] The relay protection system is a vital part of the power system. Its main function is to detect and isolate the fault area in a timely and accurate manner when a fault or abnormality occurs in the power system, thereby ensuring the safety and stability of the power system. Traditional relay protection systems usually rely on a fixed output matrix for fault detection and protection actions, which has certain limitations when facing a complex and changing power system environment.

[0003] Specifically, the traditional relay protection system shows obvious shortcomings in dealing with the following situations:

[0004] The load in the power system often changes, and the traditional fixed matrix detection method has difficulty adapting to these changes, resulting in a decrease in detection accuracy and response speed.

[0005] Nonlinear faults (such as intermittent faults or asymmetric faults) are difficult to be effectively identified by a fixed detection matrix, and traditional methods are inefficient in detecting and handling such faults.

[0006] Faced with multi-point faults or complex fault scenarios, traditional relay protection systems are unable to quickly and accurately locate and identify all fault points, resulting in untimely fault isolation and processing.

[0007] With the rapid development of smart grids and distributed power sources, the complexity and uncertainty of power systems have further increased, which has put forward higher requirements on the detection accuracy and response speed of relay protection systems.

[0008] In view of the above problems, the prior art urgently needs a relay protection outlet matrix detection system that can dynamically adapt to changes in the power system. Summary of the invention

[0009] Based on the above objectives, the present invention provides a relay protection outlet matrix dynamic detection system.

[0010] Relay protection export matrix dynamic detection system, including the following modules:

[0011] The power collection module is used to collect the operating parameters of the power system in real time. The operating parameters include current, voltage, frequency and power, where power includes active power and reactive power;

[0012] The data processing module pre-processes the collected operating parameters, including data cleaning, filtering and feature extraction;

[0013] A dynamic matrix generation module generates a dynamic detection matrix according to the extracted features, and the dynamic detection matrix is ​​adjusted in real time according to the operating state of the power system;

[0014] A fault detection module detects faults in the power system in real time based on the generated dynamic detection matrix, including identifying the fault type, fault location and fault severity;

[0015] A protection action module is connected to the fault detection module, and after the fault detection module detects a fault, generates a corresponding protection action instruction and transmits it to the relay protection device;

[0016] Communication module, used to realize data communication and information interaction between modules;

[0017] The storage module is used to store the operation data, fault records and detection results of the power system.

[0018] Furthermore, the power collection module includes a plurality of sensor nodes and a data collection unit, specifically including:

[0019] Sensor nodes are arranged at each node of the power system. Each sensor node includes:

[0020] Current sensor, used to collect current parameters of the power system in real time;

[0021] Voltage sensor, used to collect voltage parameters of the power system in real time;

[0022] Frequency sensors, used to monitor frequency changes in power systems in real time;

[0023] Power sensors are used to measure the active and reactive power of power systems in real time;

[0024] The data acquisition unit is connected to all sensor nodes and is used to collect and summarize the power system operation parameters obtained from each sensor node. The data acquisition unit includes:

[0025] Signal conditioning circuit, used to perform preliminary conditioning on the data transmitted by the sensor node to ensure signal stability;

[0026] Analog-to-digital converter, which converts the analog signal collected by the sensor into a digital signal;

[0027] The data transmission interface transmits the processed digital signal to the data processing module via wired or wireless means.

[0028] Furthermore, the data processing module includes the following units, specifically including:

[0029] The data cleaning unit is used to clean the collected operating parameters, including:

[0030] Noise filtering: removes random noise and outliers from sensor data;

[0031] Data correction: Correct sensor deviation and error to ensure data accuracy;

[0032] The filtering unit is used to filter the cleaned data, including:

[0033] Low-pass filter: removes high-frequency noise and retains low-frequency signal components;

[0034] High-pass filter: removes low-frequency interference and retains high-frequency signal components;

[0035] Bandpass filter: selects signal components in a specific frequency band as needed;

[0036] The feature extraction unit is used to extract key features from the filtered data, including:

[0037] Time domain feature extraction: extract the time domain features of the mean, variance, and peak of the data;

[0038] Frequency domain feature extraction: extract the spectral features of data through methods such as Fourier transform;

[0039] Statistical feature extraction: extract statistical features of data correlation and covariance.

[0040] Furthermore, the dynamic matrix generation module specifically includes:

[0041] The matrix initialization unit initializes the dynamic detection matrix according to the historical operation data of the power system and the extracted features. The initialization process includes:

[0042] Determine the dimensions and structure of the matrix;

[0043] Set initial thresholds and weights based on historical data and the statistical distribution of features;

[0044] The matrix update unit dynamically adjusts the detection matrix according to the feature data extracted in real time. The matrix update process includes:

[0045] Feature mapping: Map the feature data extracted in real time to the corresponding position in the dynamic detection matrix. Let the extracted feature be the F vector, which contains N feature data, and the dynamic detection matrix M is an N×N dimensional matrix;

[0046] Threshold adjustment: dynamically adjust the threshold in the matrix according to the changes in real-time data and historical data;

[0047] Weight update: Update the weights in the matrix according to the real-time power system status;

[0048] The matrix optimization unit is used to optimize the matrix during the matrix update process to ensure the accuracy and real-time performance of the detection.

[0049] Furthermore, the feature map is expressed as: ij (t) = g(F i (t),F j (t)); where M ij (t) is the dynamic check matrix element at time t, F i (t) and F j (t) are the i-th and j-th features extracted, respectively, g is the feature mapping function, which is a product method, for example: g(F i ,F j )=F i .F j ;

[0050] The threshold adjustment is expressed as: Among them, T ij (t) is the threshold at time t, is the historical mean of the characteristic, α is the adjustment coefficient;

[0051] The weight update is expressed as: W ij (t+1)=W ij (t)+β·(δ ij -W ij (t)), where W ij (t) is the weight at time t, δ ij is the deviation between the real-time detection result and the actual fault condition, and β is the learning rate.

[0052] Furthermore, the matrix optimization unit optimization process includes:

[0053] Error minimization: Minimize the detection error through an iterative algorithm, the formula is:

[0054] E=∑ i,j (M ij (t)-D ij (t)) 2 Where E is the total error, D ij (t) is the actual fault detection data;

[0055] Sparse processing: The sparse algorithm is used to reduce the redundant elements in the matrix and improve the detection efficiency, which is expressed as: S ij (t+1)=S ij (t)·(1-γ·‖S ij (t)‖0), where S ij (t) is the sparse matrix element at time t, γ is the sparsification coefficient, ‖S ij(t)‖0 is the number of non-zero elements in the matrix.

[0056] Furthermore, the fault detection module specifically includes:

[0057] The fault type identification unit is used to identify the fault type according to the generated dynamic detection matrix. The identification process includes converting the element M in the dynamic detection matrix ij (t) Compare with the preset threshold and identify the fault type through classification algorithm;

[0058] Eigenvector M(t)=[M 11 (t),M 12 (t),...,M NN (t)] as input, the classifier outputs fault type FaultType(t):

[0059] Use the support vector machine model (SVM): FaultType(t) = SVM(M(t));

[0060] A fault location identification unit is used to determine the fault location according to the dynamic detection matrix. The identification process includes determining the location of the fault according to the non-zero elements in the dynamic detection matrix and the corresponding characteristic data, and using a clustering algorithm to identify the fault location;

[0061] Eigenvector M(t)=[M 11 (t),M 12 (t),...,M NN (t)] is clustered into K locations:

[0062] FaultLocation(t)=Kmeans(M(t),K);

[0063] The fault severity identification unit is used to evaluate the severity of the fault. The evaluation process includes calculating the number and amplitude of elements exceeding the threshold in the dynamic detection matrix, and using weighted sum to calculate the fault severity, which is expressed as:

[0064] Among them, w ij is the matrix element M ij The weight, T ij is the corresponding threshold, max(0,M ij (t)-T ij ) represents the element M ij (t) The portion exceeding the threshold.

[0065] Furthermore, the support vector machine model is trained based on historical fault data:

[0066]

[0067] subject to you i (w·M i +b)≥1-ξ i ,ξ i ≥0;

[0068] Among them, w is the weight vector, b is the bias, ξ i is the relaxation variable, C is the penalty parameter, y i For the fault type label, the real-time feature vector is classified into the fault type:

[0069] FaultType(t)=sign(w·M(t)+b).

[0070] Furthermore, the clustering algorithm specifically includes:

[0071] Initialization: Select K initial cluster centers;

[0072] Iterative updates: Among them, c k is the kth cluster center, S k is the set of feature vectors belonging to the kth cluster;

[0073] Assign data points:

[0074] Clustering results: Determine the fault location based on the distance between the feature vector and the cluster center:

[0075] FaultLocation(t)=argmin k ‖M(t)-c k ‖ 2 .

[0076] Furthermore, the protection action module includes:

[0077] The protection strategy generation unit is used to generate corresponding protection action instructions according to the fault information provided by the fault detection module. The generation process includes:

[0078] Receive information about the fault type, fault location and fault severity:

[0079] FaultType(t),FaultLocation(t),FaultSeverity(t);

[0080] Generate protection action instructions based on preset protection strategies and rules, specifically:

[0081] ProtectionCommand(t)=f(FaultType(t), FaultLocation(t), FaultSeverity(t)), where f is the protection strategy function, including the action type (such as circuit breaker, alarm, etc.) and specific parameters;

[0082] The protection instruction sending unit is used to transmit the generated protection action instruction to the relay protection device. The sending process includes:

[0083] The protection action instructions are encoded into command format and the communication module is used to transmit the instructions, including:

[0084] Data encoding: EncodedCommand(t)=encode(ProtectionCommand(t), where encode is an encoding function that encodes the protection action command into a transmission format;

[0085] Data transmission: send(EncodedCommand(t)), where send is a transmission function that sends the encoded command to the relay protection device via wireless communication.

[0086] Beneficial effects of the present invention:

[0087] The present invention collects multiple operating parameters of the power system such as current, voltage, frequency and power (including active power and reactive power) in real time, and performs data cleaning, filtering and feature extraction on these parameters, thereby generating a dynamic detection matrix, which can be adjusted in real time according to the operating state of the power system, and accurately identifies the fault type, fault location and fault degree through support vector machine (SVM) and K-means clustering algorithm. This detection method based on multi-parameter features and dynamic matrix significantly improves the accuracy and real-time performance of fault detection, allowing the system to respond to various complex power fault situations in a timely and accurate manner.

[0088] The present invention can dynamically adapt to the real-time changes of the power system through the combination of dynamic detection matrix and multiple characteristic parameters. The matrix update unit dynamically adjusts the detection matrix according to the characteristic data extracted in real time, and performs error minimization and sparse processing through the matrix optimization unit, thereby improving the detection efficiency and accuracy of the system. At the same time, the protection strategy generation unit can generate corresponding protection action instructions according to different fault types, fault locations and fault degrees, and transmit them to the relay protection device in time through the communication module, ensuring the efficient protection and response of the system under various complex situations. This dynamic adaptation mechanism greatly improves the flexibility and adaptability of the system.

[0089] The fault detection module of the present invention can not only accurately identify the fault type, fault location and fault degree, but also generate detailed protection action instructions. Through the support vector machine (SVM) classification algorithm and the K-means clustering algorithm, the system can comprehensively analyze the fault information and generate optimized protection strategies. These strategies include multiple protection actions such as issuing alarm signals, local disconnection, cutting off the power supply of related lines, and comprehensive power outages and emergency shutdowns. This comprehensive fault handling and protection strategy not only ensures the safe operation of the power system, but also provides maintenance personnel with detailed fault information, facilitates the rapid investigation and handling of faults, and significantly improves the overall stability and safety of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0090] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0091] Figure 1 A schematic diagram of functional modules of a detection system according to an embodiment of the present invention;

[0092] Figure 2 Schematic diagram of a fault detection module according to an embodiment of the present invention. DETAILED DESCRIPTION

[0093] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with specific embodiments.

[0094] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0095] like Figure 1-Figure 2 As shown, the relay protection export matrix dynamic detection system includes the following modules:

[0096] The power collection module is used to collect the operating parameters of the power system in real time. The operating parameters include current, voltage, frequency and power, where power includes active power and reactive power;

[0097] The data processing module pre-processes the collected operating parameters, including data cleaning, filtering and feature extraction;

[0098] A dynamic matrix generation module generates a dynamic detection matrix based on the extracted features, and the dynamic detection matrix is ​​adjusted in real time according to the operating status of the power system;

[0099] A fault detection module detects faults in the power system in real time based on the generated dynamic detection matrix, including identifying the fault type, fault location and fault severity;

[0100] The protection action module is connected to the fault detection module. After the fault detection module detects a fault, it generates a corresponding protection action instruction and transmits it to the relay protection device;

[0101] Communication module, used to realize data communication and information interaction between modules;

[0102] The storage module is used to store the operation data, fault records and detection results of the power system.

[0103] The power operation parameters are the basis for generating a dynamic detection matrix, and the relay protection output uses the detection results of this matrix to respond to faults and perform protection actions. This process ensures that the relay protection system can dynamically adapt to real-time changes in the power system, accurately identify faults and take timely protection actions.

[0104] The power collection module includes multiple sensor nodes and data collection units, including:

[0105] Sensor nodes are arranged at each node of the power system. Each sensor node includes:

[0106] Current sensor, used to collect current parameters of the power system in real time;

[0107] Voltage sensor, used to collect voltage parameters of the power system in real time;

[0108] Frequency sensors, used to monitor frequency changes in power systems in real time;

[0109] Power sensors are used to measure the active and reactive power of power systems in real time;

[0110] The data acquisition unit is connected to all sensor nodes and is used to collect and summarize the power system operation parameters obtained from each sensor node. The data acquisition unit includes:

[0111] Signal conditioning circuit, used to perform preliminary conditioning on the data transmitted by the sensor node to ensure signal stability;

[0112] Analog-to-digital converter, which converts the analog signal collected by the sensor into a digital signal;

[0113] The data transmission interface transmits the processed digital signal to the data processing module via wired or wireless means.

[0114] Through the collaborative work of the above-mentioned sensor nodes and data acquisition units, the power acquisition module can obtain the operating parameters of the power system such as current, voltage, frequency, active power and reactive power in real time and accurately, providing basic data support for subsequent data processing and the generation of dynamic detection matrices.

[0115] The data processing module includes the following units, specifically:

[0116] The data cleaning unit is used to clean the collected operating parameters, including:

[0117] Noise filtering: removes random noise and outliers from sensor data;

[0118] Data correction: Correct sensor deviation and error to ensure data accuracy;

[0119] The filtering unit is used to filter the cleaned data, including:

[0120] Low-pass filter: removes high-frequency noise and retains low-frequency signal components;

[0121] High-pass filter: removes low-frequency interference and retains high-frequency signal components;

[0122] Bandpass filter: selects signal components in a specific frequency band as needed;

[0123] The feature extraction unit is used to extract key features from the filtered data, including:

[0124] Time domain feature extraction: extract the time domain features of the mean, variance, and peak of the data;

[0125] Frequency domain feature extraction: extract the spectral features of data through methods such as Fourier transform;

[0126] Statistical feature extraction: extract statistical features of data correlation and covariance.

[0127] The details are as follows:

[0128] 1. Data cleaning unit, used to clean the collected operating parameters, including:

[0129] Noise filtering: Use a mean filter to remove random noise:

[0130] Among them, x[n] is the original data, x clean [n] is the cleaned data, and M is the half-width of the filter window.

[0131] Data correction: Use the linear correction formula to correct the sensor deviation:

[0132] x corrected =x measured ×α+β, where x measured is the measured data, x corrected is the corrected data, α and β are the correction coefficients.

[0133] 2. A filtering unit is used to filter the cleaned data, including:

[0134] Low-pass filter: Use UR low-pass filter to remove high-frequency noise, the transfer function is:

[0135] Among them, b0, b1, b2 and a1, a2 are filter coefficients.

[0136] High-pass filter: Use an IIR high-pass filter to remove low-frequency interference. The transfer function is:

[0137] Among them, b0, b1, b2 and a1, a2 are filter coefficients.

[0138] Bandpass filter: The bandpass filter selects the signal components in a specific frequency band, and the transfer function is: Among them, b0, b1 and a1, a2 are filter coefficients, and N and M are the number of delay steps.

[0139] 3. Feature extraction unit, used to extract key features from the filtered data, including:

[0140] Time domain feature extraction:

[0141] Calculate the mean:

[0142] Calculate the variance:

[0143] Calculate peak value: x peak =max(x[n]);

[0144] Frequency domain feature extraction: Extract spectrum features through fast Fourier transform (FFT):

[0145] Among them, X[k] is the frequency domain signal, and x[n] is the time domain signal;

[0146] Statistical feature extraction: Calculate the correlation coefficient: Where Cov(x,y) is the covariance of x and y, σ x and σ y are the standard deviations of x and y respectively.

[0147] The dynamic matrix generation module specifically includes:

[0148] The matrix initialization unit initializes the dynamic detection matrix according to the historical operation data of the power system and the extracted features. The initialization process includes:

[0149] Determine the dimensions and structure of the matrix;

[0150] Set initial thresholds and weights based on historical data and the statistical distribution of features;

[0151] The matrix update unit dynamically adjusts the detection matrix according to the feature data extracted in real time. The matrix update process includes:

[0152] Feature mapping: Map the feature data extracted in real time to the corresponding position in the dynamic detection matrix. Let the extracted feature be the F vector, which contains N feature data, and the dynamic detection matrix M is an N×N dimensional matrix;

[0153] Threshold adjustment: dynamically adjust the threshold in the matrix according to the changes in real-time data and historical data;

[0154] Weight update: Update the weights in the matrix according to the real-time power system status;

[0155] The matrix optimization unit is used to optimize the matrix during the matrix update process to ensure the accuracy and real-time performance of the detection.

[0156] The feature map is represented as: M ij (t) = g(F i (t),F j (t)); where M ij (t) is the dynamic check matrix element at time t, F i (t) and F j (t) are the i-th and j-th features extracted, respectively, g is the feature mapping function, which is a product method, for example: g(F i ,F j )=F i .F j ;

[0157] The threshold adjustment is expressed as: Among them, T ij (t) is the threshold at time t, is the historical mean of the characteristic, α is the adjustment coefficient;

[0158] The weight update is expressed as: W ij (t+1)=W ij (t)+β·(δ ij -W ij (t)), where W ij (t) is the weight at time t, δ ij is the deviation between the real-time detection result and the actual fault condition, and β is the learning rate.

[0159] The matrix optimization unit optimization process includes:

[0160] Error minimization: Minimize the detection error through an iterative algorithm, the formula is:

[0161] E=∑ i,j (M ij (t)-D ij (t)) 2 Where E is the total error, D ij (t) is the actual fault detection data;

[0162] Sparse processing: The sparse algorithm is used to reduce the redundant elements in the matrix and improve the detection efficiency, which is expressed as: S ij (t+1)=S ij (t)·(1-γ·‖S ij (t)‖0), where S ij (t) is the sparse matrix element at time t, γ is the sparsification coefficient, ‖S ij (t)‖0 is the number of non-zero elements in the matrix.

[0163] The fault detection module specifically includes:

[0164] The fault type identification unit is used to identify the fault type according to the generated dynamic detection matrix. The identification process includes converting the element M in the dynamic detection matrix ij (t) Compare with the preset threshold and identify the fault type through classification algorithm;

[0165] Eigenvector M(t)=[M 11 (t),M 12 (t),...,M NN (t)] as input, the classifier outputs fault type FaultType(t):

[0166] Use the support vector machine model (SVM): FaultType(t) = SVM(M(t));

[0167] A fault location identification unit is used to determine the fault location according to the dynamic detection matrix. The identification process includes determining the location of the fault according to the non-zero elements in the dynamic detection matrix and the corresponding characteristic data, and using a clustering algorithm to identify the fault location;

[0168] Eigenvector M(t)=[M 11 (t),M 12 (t),...,M NN (t)] is clustered into K locations:

[0169] FaultLocation(t)=Kmeans(M(t),K);

[0170] The fault severity identification unit is used to evaluate the severity of the fault. The evaluation process includes calculating the number and amplitude of elements exceeding the threshold in the dynamic detection matrix, and using weighted sum to calculate the fault severity, which is expressed as:

[0171] Among them, w ij is the matrix element M ij The weight, T ij is the corresponding threshold, max(0,M ij (t)-T ij ) represents the element M ij (t) The portion exceeding the threshold.

[0172] The support vector machine model is trained based on historical failure data:

[0173]

[0174] subject to you i (w·M i +b)≥1-ξ i ,ξ i ≥0;

[0175] Among them, w is the weight vector, b is the bias, ξ i is the relaxation variable, C is the penalty parameter, y i For the fault type label, the real-time feature vector is classified into the fault type:

[0176] FaultType(t)=sign(w·M(t)+b).

[0177] Clustering algorithms include:

[0178] Initialization: Select K initial cluster centers;

[0179] Iterative updates: Among them, c k is the kth cluster center, Sk is the set of feature vectors belonging to the kth cluster;

[0180] Assign data points:

[0181] Clustering results: Determine the fault location based on the distance between the feature vector and the cluster center:

[0182] FaultLocation(t)=argmin k ‖M(t)-c k ‖ 2 .

[0183] The protection action modules include:

[0184] The protection strategy generation unit is used to generate corresponding protection action instructions according to the fault information provided by the fault detection module. The generation process includes:

[0185] Receive information about the fault type, fault location and fault severity:

[0186] FaultType(t),FaultLocation(t),FaultSeverity(t);

[0187] Generate protection action instructions based on preset protection strategies and rules, specifically:

[0188] ProtectionCommand(t)=f(FaultType(t), FaultLocation(t), FaultSeverity(t)), where f is the protection strategy function, including the action type (such as circuit breaker, alarm, etc.) and specific parameters;

[0189] The protection instruction sending unit is used to transmit the generated protection action instruction to the relay protection device. The sending process includes:

[0190] The protection action instructions are encoded into command format and the communication module is used to transmit the instructions, including:

[0191] Data encoding: EncodedCommand(t)=encode(ProtectionCommand(t), where encode is an encoding function that encodes the protection action command into a transmission format;

[0192] Data transmission: send(EncodedCommand(t)), where send is a transmission function that sends the encoded command to the relay protection device via wireless communication.

[0193] The fault type, fault location, fault degree and protection strategy mainly include the following

[0194] Failure type:

[0195] Common fault types in power systems include:

[0196] Short circuit fault, overcurrent fault, undervoltage fault, overvoltage fault, ground fault, and frequency abnormality.

[0197] Fault location:

[0198] Fault locations usually include the following parts in the power system:

[0199] Transformers, transmission lines, distribution lines, motors, busbars.

[0200] Failure degree:

[0201] The degree of failure can be divided into the following levels:

[0202] Minor fault, medium fault, severe fault, critical fault, protection strategy,

[0203] According to different fault types, fault locations and fault severity, the following protection strategies can be adopted:

[0204] Short circuit fault:

[0205] Minor: send out an alarm signal;

[0206] Medium: local circuit breaking, isolating the fault area;

[0207] Serious: Cut off the power supply of related lines;

[0208] Critical: Total power outage, emergency shutdown;

[0209] Overcurrent fault:

[0210] Minor: send out an alarm signal;

[0211] Medium: limit the current and balance the current by adjusting the load;

[0212] Serious: Partial circuit breaker, isolate the fault area;

[0213] Emergency: Cut off the power supply of related lines and shut down the machine immediately;

[0214] Undervoltage fault:

[0215] Minor: send out an alarm signal;

[0216] Medium: adjust the voltage and stabilize the voltage through the voltage regulator;

[0217] Serious: Partial circuit breaker, isolate the fault area;

[0218] Emergency: Cut off the power supply of related lines and shut down the machine immediately;

[0219] Overvoltage fault:

[0220] Minor: send out an alarm signal;

[0221] Medium: adjust the voltage and stabilize the voltage through the voltage regulator;

[0222] Serious: Partial circuit breaker, isolate the fault area;

[0223] Emergency: Cut off the power supply of related lines and shut down the machine immediately;

[0224] Ground fault:

[0225] Minor: send out an alarm signal;

[0226] Moderate: Check the grounding system and repair it;

[0227] Serious: Partial circuit breaker, isolate the fault area;

[0228] Emergency: Cut off the power supply of related lines and shut down the machine immediately;

[0229] Frequency anomaly:

[0230] Minor: send out an alarm signal;

[0231] Medium: Adjust load and generation to stabilize frequency;

[0232] Serious: Partial circuit breaker, isolate the fault area;

[0233] Emergency: Cut off the power supply of related lines and shut down the machine immediately.

[0234] Those skilled in the art should understand that the discussion of any of the above embodiments is merely illustrative and is not intended to imply that the scope of the present invention is limited to these examples. Under the concept of the present invention, the technical features in the above embodiments or different embodiments may be combined, the steps may be implemented in any order, and there are many other variations of the different aspects of the present invention as described above, which are not provided in detail for the sake of simplicity.

[0235] The present invention is intended to cover all such substitutions, modifications and variations that fall within the broad scope of the claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. Relay protection outlet matrix dynamic detection system, characterized in that: Includes the following modules: The power collection module is used to collect the operating parameters of the power system in real time. The operating parameters include current, voltage, frequency and power, where power includes active power and reactive power; The data processing module pre-processes the collected operating parameters, including data cleaning, filtering and feature extraction; The dynamic matrix generation module generates a dynamic detection matrix according to the extracted features. The dynamic detection matrix is ​​adjusted in real time according to the operating state of the power system, specifically including: The matrix initialization unit initializes the dynamic detection matrix according to the historical operation data of the power system and the extracted features. The initialization process includes: Determine the dimensions and structure of the matrix; Set initial thresholds and weights based on historical data and the statistical distribution of features; The matrix update unit dynamically adjusts the detection matrix according to the feature data extracted in real time. The matrix update process includes: Feature mapping: Map the feature data extracted in real time to the corresponding position in the dynamic detection matrix. Let the extracted feature be A vector containing feature data, dynamic detection matrix for Dimensional matrix; Threshold adjustment: dynamically adjust the threshold in the matrix according to the changes in real-time data and historical data; Weight update: Update the weights in the matrix according to the real-time power system status; The matrix optimization unit is used to optimize the matrix during the matrix update process to ensure the accuracy and real-time performance of the detection. The optimization process of the matrix optimization unit includes: Error minimization: Minimize the detection error through an iterative algorithm, the formula is: in, is the total error, is the actual fault detection data; Sparse processing: The sparse algorithm is used to reduce the redundant elements in the matrix and improve the detection efficiency, which is expressed as: ,in, For the moment The sparse matrix elements of is the sparsification coefficient, is the number of non-zero elements in the matrix; The feature map is expressed as: ;in, For the moment The dynamic check matrix elements of and The extracted and Features, is the feature mapping function; The threshold adjustment is expressed as: ,in, For the moment The threshold value, is the historical mean of the feature, is the adjustment factor; The weight update is expressed as: ,in, For the moment The weight of To detect the deviation between the real-time detection result and the actual fault condition, is the learning rate; A fault detection module detects faults in the power system in real time based on the generated dynamic detection matrix, including identifying the fault type, fault location and fault severity; A protection action module is connected to the fault detection module, and after the fault detection module detects a fault, generates a corresponding protection action instruction and transmits it to the relay protection device; Communication module, used to realize data communication and information interaction between modules; The storage module is used to store the operation data, fault records and detection results of the power system.

2. The relay protection outlet matrix dynamic detection system according to claim 1 is characterized in that: The power collection module includes multiple sensor nodes and a data collection unit, specifically including: Sensor nodes are arranged at each node of the power system. Each sensor node includes: Current sensor, used to collect current parameters of the power system in real time; Voltage sensor, used to collect voltage parameters of the power system in real time; Frequency sensors, used to monitor frequency changes in power systems in real time; Power sensors are used to measure the active and reactive power of power systems in real time; The data acquisition unit is connected to all sensor nodes and is used to collect and summarize the power system operation parameters obtained from each sensor node. The data acquisition unit includes: Signal conditioning circuit, used to perform preliminary conditioning on the data transmitted by the sensor node to ensure signal stability; Analog-to-digital converter, which converts the analog signal collected by the sensor into a digital signal; The data transmission interface transmits the processed digital signal to the data processing module via wired or wireless means.

3. The relay protection outlet matrix dynamic detection system according to claim 1 is characterized in that: The data processing module includes the following units, specifically: The data cleaning unit is used to clean the collected operating parameters, including: Noise filtering: removes random noise and outliers from sensor data; Data correction: Correct sensor deviation and error to ensure data accuracy; The filtering unit is used to filter the cleaned data, including: Low-pass filter: removes high-frequency noise and retains low-frequency signal components; High-pass filter: removes low-frequency interference and retains high-frequency signal components; Bandpass filter: selects signal components in a specific frequency band as needed; The feature extraction unit is used to extract key features from the filtered data, including: Time domain feature extraction: extract the time domain features of the mean, variance, and peak of the data; Frequency domain feature extraction: extract the spectral features of data through methods such as Fourier transform; Statistical feature extraction: extract statistical features of data correlation and covariance.

4. The relay protection outlet matrix dynamic detection system according to claim 1, characterized in that: The fault detection module specifically includes: The fault type identification unit is used to identify the fault type according to the generated dynamic detection matrix. The identification process includes converting the elements in the dynamic detection matrix into Compare with the preset threshold and identify the fault type through classification algorithm; Eigenvector As input, the classifier outputs the fault type FaultType : Using the Support Vector Machine Model: FaultType ; A fault location identification unit is used to determine the fault location according to the dynamic detection matrix. The identification process includes determining the location of the fault according to the non-zero elements in the dynamic detection matrix and the corresponding characteristic data, and using a clustering algorithm to identify the fault location; Eigenvector Cluster to Locations: FaultLocation =Kmeans ; The fault severity identification unit is used to evaluate the severity of the fault. The evaluation process includes calculating the number and amplitude of elements exceeding the threshold in the dynamic detection matrix, and using weighted sum to calculate the fault severity, which is expressed as: ,in, is the matrix element The weight of is the corresponding threshold, Representation elements The portion that exceeds the threshold.

5. The relay protection outlet matrix dynamic detection system according to claim 4 is characterized in that: The support vector machine model is trained based on historical fault data: ； ； in, is the weight vector, is the bias, is the loose variable, is the penalty parameter, For the fault type label, the real-time feature vector is classified into the fault type: 。 6. The relay protection outlet matrix dynamic detection system according to claim 5, characterized in that: The clustering algorithm specifically includes: Initialization: Select Initial cluster centers; Iterative updates: ,in, For the Cluster centers, For the The set of feature vectors of clusters; Assign data points: ; Clustering results: Determine the fault location based on the distance between the feature vector and the cluster center: 。 7. The relay protection outlet matrix dynamic detection system according to claim 1, characterized in that: The protection action module includes: The protection strategy generation unit is used to generate corresponding protection action instructions according to the fault information provided by the fault detection module. The generation process includes: Receive information about the fault type, fault location and fault severity: FaultType ,FaultLocation ,FaultSeverity ; Generate protection action instructions based on preset protection strategies and rules, specifically: ,in, To protect the strategy function, including action type and specific parameters; The protection instruction sending unit is used to transmit the generated protection action instruction to the relay protection device. The sending process includes: The protection action instructions are encoded into command format and the communication module is used to transmit the instructions, including: Data encoding: , where encode is an encoding function that encodes the protection action instruction into a transmission format; Data transmission: , where send is the transmission function, which sends the encoded instructions to the relay protection device through wireless communication.

Images