Distributed relay protection secondary AC loop abnormity diagnosis method

Abstract

The invention relates to the technical field of power system relay protection, and discloses a distributed relay protection secondary alternating current loop abnormity diagnosis method comprising the following steps: collecting current waveform data of each branch of a loop in real time through a distributed current transformer; respectively calculating a transverse difference degree between the branch current waveform and the reference current waveform and a longitudinal difference degree between the branch current waveform and the neutral current waveform; and inputting the transverse difference degree and the longitudinal difference degree into an artificial fish cluster clustering algorithm, and identifying and positioning a fault loop in real time through clustering analysis. Current waveform data are acquired in real time through the distributed current transformer, and the transverse and longitudinal difference degrees are calculated by adopting an artificial fish cluster clustering algorithm, so that the relay protection secondary circuit abnormity is efficiently diagnosed. The method can rapidly position a fault loop, improves the intelligent and automatic level of fault detection of a power system, reduces the cost of manual troubleshooting, improves the safety and stability of the system, and is higher in practicality and popularization value.

Description

Technical Field

[0001] The present invention relates to the technical field of relay protection in power systems, and specifically to a method for diagnosing abnormalities in the secondary AC circuit of distributed relay protection. Background Art

[0002] The relay protection device of the power system is an important part to ensure the safe operation of the power system. Especially in substations, the relay protection device needs to rely on voltage transformers (PTs) and current transformers (CTs) to provide accurate electrical parameters to ensure that the faulty circuit can be cut off in time when a fault occurs, avoiding the spread of the fault and causing more serious accidents. However, the grounding faults in the secondary circuit of relay protection, especially multi-point grounding faults, have become key problems to be solved urgently in the power system.

[0003] According to the regulations of the "Key Points for Anti-Accident Measures of Relay Protection and Safety Automatic Devices in Power Systems", the secondary circuit of the voltage transformer must ensure that there is only one grounding point. However, due to reasons such as improper design, construction defects, equipment aging or transformation, the multi-point grounding faults in the secondary circuit are gradually increasing. Multi-point grounding faults will cause the protection device to sense distorted voltage signals, such as changes in voltage amplitude, phase, frequency, etc., which may lead to misoperation or refusal to operate of the relay protection device, and in severe cases, it will cause greater accidents, affecting the safety and stability of the power system. Summary of the Invention

[0004] In view of the above problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by the present invention is: the problems of lagging detection, poor real-time performance, complex operation, and lack of continuity and accuracy in detecting multi-point grounding faults in the secondary circuit of relay protection in the prior art.

[0006] To solve the above technical problems, the present invention provides the following technical solution: a method for diagnosing abnormalities in the secondary AC circuit of distributed relay protection, including:

[0007] Real-time collecting the current waveform data of each branch of the circuit through a distributed current transformer;

[0008] Respectively calculating the horizontal difference degree between the branch current waveform and the reference current waveform, and the vertical difference degree between the branch current waveform and the neutral line current waveform;

[0009] Inputting the horizontal difference degree and the vertical difference degree into the artificial fish swarm clustering algorithm, and real-time identifying and locating the faulty circuit through clustering analysis.

[0010] As a preferred embodiment of the abnormal diagnosis method for the secondary AC circuit of distributed relay protection according to the present invention, the current waveform data of each branch in the acquisition circuit includes branch current waveform data, neutral line current waveform data, and reference current waveform data;

[0011] The branch current waveform data is the current waveform data of each branch in the secondary circuit of relay protection;

[0012] The neutral line current waveform data is the current waveform data of the neutral line in the secondary circuit of relay protection;

[0013] The reference current waveform data is the current waveform data of each branch under normal working conditions, serving as a reference benchmark waveform.

[0014] As a preferred embodiment of the abnormal diagnosis method for the secondary AC circuit of distributed relay protection according to the present invention, inputting the horizontal difference degree and the vertical difference degree into the artificial fish swarm clustering algorithm includes constructing the horizontal difference degree and the vertical difference degree as input features, and the formula is expressed as:

[0015] d = [x1, x2, … x N ; y1, y2, … y N

[0016] Among them, x N represents the horizontal difference degree, and y N represents the vertical difference degree;

[0017] Defining the artificial fish swarm clustering algorithm includes defining the structure body and initializing the artificial fish model;

[0018] Defining the structure body includes defining a structure body named pattern. The pattern structure body consists of feature and category attributes. The feature attribute stores the sample set, and the category attribute stores the element category, with a default value of zero;

[0019] Initializing the artificial fish model includes defining each artificial fish as a structure body with location, fitness, and string attributes. The location attribute contains the clustering center, category, and the number of samples of each category. The fitness attribute is the fitness, and string is a 1*N vector, representing the classification result of the samples in the pattern vector, corresponding to the category attribute in pattern.

[0020] As a preferred embodiment of the abnormal diagnosis method for the secondary AC circuit of distributed relay protection according to the present invention, real-time identifying and locating the faulty circuit through clustering analysis includes setting the number of artificial fish as i and randomly generating i artificial fish; ​

[0021] Set the initial clustering center to zero and assign an initial class number to each class. The initial value of the class for each sample is a random value;

[0022] The string attribute of each fish in the artificial fish swarm is assigned a value of 1 or 2, representing the classification result of the sample;

[0023] Assign the first artificial fish as the optimal artificial fish, and initialize the clustering center and sample assignment through the class information and classification result in the pattern structure.

[0024] As a preferred solution of the distributed secondary AC circuit abnormal diagnosis method for relay protection described in the present invention, wherein: real-time identification and location of the faulty circuit through clustering analysis further includes fitness calculation and clustering behavior update;

[0025] The fitness calculation includes calculating the classification result of each fish according to the string attribute of each fish. The samples are divided into two classes according to the string, and the sum of the distances from each class of samples to the clustering center is calculated. The fitness is defined as the reciprocal of the sum of the distances;

[0026] The clustering behavior update includes updating the clustering center through the aggregation behavior, recording the information of the i-th artificial fish, and adding the clustering center values of the classes corresponding to the location attributes of n f fish in the field of vision of this artificial fish and then finding the average value as the clustering center value f of the location attribute of the fish swarm center c ;

[0027] Classify the element d in the pattern according to the clustering center value f c ;

[0028] Find the distance s c1 between the first element d1 in the pattern and the clustering center value f 1d1 of the first class of the fish swarm, and define the class of d1 as 1;

[0029] Calculate the distance s c2 between d1 and the clustering center value f 2d1 of the second class. When s 2d1 < s 1d1 , d1 is classified into the second class. When s 2d1 ≥ s 1d1 , d1 is classified into the first class;

[0030] Classify all elements in the pattern;

[0031] Assign the clustering center of each class to the fish swarm center as the clustering center value f c , and repeat the calculation of the fitness value fit c ;

[0032] When fit c > fit i and at this time, the artificial fish moves one step towards the center of the fish school, and the clustering center value f of the clustering operator is obtained s .

[0033] As a preferred solution of the distributed relay protection secondary AC circuit abnormal diagnosis method described in the present invention, wherein: real-time identifying and locating the faulty circuit through clustering analysis further includes performing a chasing behavior and restoring the information of the i-th artificial fish;

[0034] Find the fish with the maximum fitness value within the field of vision. When the fitness value fit of the maximum fish x > fit i and at this time, then the maximum artificial fish moves one step, and the clustering center value f of the chasing operator is obtained f ;

[0035] Evaluate the clustering operator and the chasing operator. The clustering operator and the chasing operator generate new clustering centers, and the fitness values when the elements in pattern are classified according to f s and f f are calculated respectively.

[0036] As a preferred solution of the distributed relay protection secondary AC circuit abnormal diagnosis method described in the present invention, wherein: real-time identifying and locating the faulty circuit through clustering analysis further includes, when the clustering operator and the chasing operator have been evaluated, no longer performing the foraging behavior;

[0037] When the clustering operator and the chasing operator have not been evaluated, perform the foraging behavior;

[0038] Restore the state of the i-th artificial fish, randomly select a new position within the field of vision of the i-th artificial fish, update it, and calculate the clustering center and fitness of the new position;

[0039] The random behavior includes that, without considering the fitness improvement, the artificial fish randomly selects a new position for updating;

[0040] In each iteration, all artificial fish perform a random update once, update the clustering center of each fish, and complete one iteration;

[0041] After reaching the predetermined number of iteration settings, the recorded optimal artificial fish string value is assigned to the category attribute in the pattern structure, and according to the classification result, the category with the larger clustering center value is selected as the faulty category, and the circuit of the faulty category is the multi-point node circuit that has occurred.

[0042] A distributed secondary AC circuit abnormal diagnosis system for relay protection, wherein:

[0043] A data acquisition module that real - time collects the current waveform data of each branch of the circuit through a distributed current transformer;

[0044] A difference degree calculation module that calculates the horizontal difference degree between the branch current waveform and the reference current waveform, and the vertical difference degree between the branch current waveform and the neutral line current waveform respectively;

[0045] A fault location module that inputs the horizontal difference degree and the vertical difference degree into the artificial fish swarm clustering algorithm, and real - time identifies and locates the faulty circuit through clustering analysis.

[0046] A computer device, comprising: a memory and a processor; the memory stores a computer program, characterized in that: when the processor executes the computer program, the steps of the method described in any one of the present invention are realized.

[0047] A computer - readable storage medium, on which a computer program is stored, characterized in that: when the computer program is executed by a processor, the steps of the method described in any one of the present invention are realized.

[0048] Advantages of the present invention: The distributed secondary AC circuit abnormal diagnosis method provided by the present invention real - time collects the current waveform data of each branch through a distributed current transformer, and calculates the horizontal and vertical difference degrees. The present invention can accurately reflect the abnormal characteristics of the circuit, thereby improving the accuracy of fault location. The introduced artificial fish swarm clustering algorithm simulates the aggregation behavior of fish schools, optimizes the update mechanism of the clustering center, can adapt to complex fault scenarios, and further optimizes the clustering results through the following - up behavior and foraging behavior, enhancing the intelligent level of fault diagnosis. The system has a strong adaptive ability by dynamically adjusting the clustering center and classification information, without manual intervention, thus reducing the risk of human operation. In addition, the present invention adopts a multi - granularity time - series analysis method, which can analyze the current waveform data from multiple time dimensions, enhancing the diagnostic ability for complex fault patterns. By optimizing the fitness calculation and clustering behavior update, the present invention can more accurately classify and locate faults, with strong robustness. While reducing the calculation amount, it improves the processing efficiency of the system, has good scalability, and can adapt to the actual application of large - scale power systems. Brief Description of the Drawings

[0049] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0050] Figure 1 This is the overall flowchart of a distributed relay protection secondary AC circuit abnormal diagnosis method provided by the first embodiment of the present invention;

[0051] Figure 2 This is the flowchart of the artificial fish swarm algorithm for a distributed relay protection secondary AC circuit abnormal diagnosis method provided by the first embodiment of the present invention;

[0052] Figure 3 This is the insulation fault diagnosis process of the relay protection secondary AC circuit based on the artificial fish swarm algorithm for a distributed relay protection secondary AC circuit abnormal diagnosis method provided by the first embodiment of the present invention. Specific implementation manners

[0053] To make the above objects, features, and advantages of the present invention more obvious and understandable, the specific implementation manners of the present invention will be described in detail below with reference to the accompanying drawings of the specification. Obviously, the described embodiments are some embodiments of the present invention, rather than all embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0054] Embodiment 1, referring to Figure 1 , which is an embodiment of the present invention, provides a distributed relay protection secondary AC circuit abnormal diagnosis method, including:

[0055] S1: Real-time collect the current waveform data of each branch of the circuit through distributed current transformers.

[0056] Collecting the current waveform data of each branch of the circuit includes branch current waveform data, neutral line current waveform data, and reference current waveform data;

[0057] The branch current waveform data is the current waveform data of each branch of the relay protection secondary circuit;

[0058] The neutral line current waveform data is the current waveform data of the neutral line of the relay protection secondary circuit;

[0059] The reference current waveform data is the current waveform data of each branch under normal working conditions, serving as a reference benchmark waveform.

[0060] Furthermore, real-time acquisition of the current waveform data of each branch of the loop by the distributed current transformer provides basic data support for the multi-point grounding fault diagnosis of the secondary circuit of the relay protection. The branch current waveform data is used to reflect the current characteristics of each branch under the actual operating conditions, while the neutral line current waveform data serves as an important reference for comparison with the branch current, and is used to judge the correlation and differences between the current waveforms. The reference current waveform data records the current characteristics of the system under normal operating conditions and provides a reference basis for diagnosing abnormalities and faults. Through the real-time acquisition of these data, the operating state of the power system can be comprehensively reflected, providing an accurate and efficient data source for subsequent difference calculation, fault identification and early warning, so as to achieve fast and real-time fault detection and improve the safety and stability of the system.

[0061] S2: Calculate the horizontal difference degree between the branch current waveform and the reference current waveform, and the vertical difference degree between the branch current waveform and the neutral line current waveform respectively.

[0062] For each branch current waveform data, an elastic metric method is used to compare it with the reference current waveform. The horizontal difference degree is used to quantify the deviation degree of the branch current waveform from the current waveform in the normal (reference) state on the time axis.

[0063] The elastic metric method can effectively ignore the differences caused by frequency fluctuations or small offsets, and focuses on evaluating the differences between the overall shape and main features of the waveforms. This helps to avoid errors caused by system noise and frequency fluctuations.

[0064] Compare the current waveform of each branch with the current waveform of the neutral line and calculate the vertical difference degree.

[0065] The vertical difference degree is used to evaluate the similarity of the branch current and the neutral line current in terms of waveform, focusing on the trends and amplitude differences of the current changes, so as to discover waveform changes caused by grounding faults or abnormal conditions.

[0066] Furthermore, through the calculation of the horizontal difference degree and the vertical difference degree, the deviation between the branch current waveform and the normal reference waveform and the neutral line current waveform can be accurately evaluated, providing accurate data support for fault diagnosis. The horizontal difference degree effectively ignores frequency fluctuations and small offsets through the elastic metric method, focuses on the overall shape and characteristics of the current waveform, and reduces error interference; the vertical difference degree focuses on detecting the differences in waveform trends and amplitudes between the branch current and the neutral line current, helps to identify current changes caused by grounding faults or other abnormal conditions, and enhances the accuracy and reliability of fault location.

[0067] S3: Input the horizontal difference degree and the vertical difference degree into the artificial fish swarm clustering algorithm, and real-time identify and locate the fault loop through clustering analysis.

[0068] Analyze and process the loop current waveform data of each loop such as the neutral line of PT / CT collected distributively and the grounding wire of the N600 grounding point, elastically measure the horizontal difference degree between the branch circuit current and the normal line current waveform, and the vertical difference degree between the branch circuit and the neutral line circuit current waveform. Use the artificial fish swarm algorithm to perform clustering analysis on the two-dimensional matrix composed of the horizontal and vertical difference degrees. The specific analysis steps are as follows:

[0069] Step 1: Perform clustering analysis on the target N lines. Their horizontal and vertical difference degrees are respectively: d = [x1, x2,... x N ; y1, y2,... y N , where x N is the horizontal difference degree, and y N is the vertical difference degree. This matrix is used as the input of the artificial fish swarm clustering algorithm.

[0070] Step 2: Define a structure named pattern. The pattern structure consists of feature and category attributes. The feature attribute stores the sample set, and the category attribute stores the element category, with a default value of zero.

[0071] Step 3: Initialize the artificial fish model. Each artificial fish is defined as a structure with location, fitness, and string attributes. The location attribute contains the clustering center of each category, the category, and the number of samples in the category. The fitness attribute is the fitness, and string is a 1*N vector representing the classification result of the samples in the pattern vector, corresponding to the category attribute in pattern.

[0072] Step 4: Randomly generate i artificial fish. Set the clustering center in location to zero, assign the category number according to the number of categories, and set the number of elements in each category to zero. Set the fitness value in fitness to zero. string is 1*N, and the element values are random vectors of 1 or 2. Assign the first artificial fish to the bulletin board and record it as the optimal artificial fish. Assign the string values of the i artificial fish to the category attribute of the pattern structure in sequence. Calculate the initial clustering center according to the classification and assign the number of elements in each category to the location attribute of the i-th artificial fish.

[0073] Step 5: Calculate the fitness. According to the different string attributes of each fish, different classification results are obtained. Taking the i-th artificial fish as an example, assign the values in the string of the i-th artificial fish to the category attribute in the pattern. The elements in the pattern are divided into two categories, and the distances between the elements in class 1 and class 2 and the cluster centers of class 1 and class 2 stored in the location attribute of the i-th artificial fish are accumulated. The fitness is defined as the reciprocal of this accumulated value.

[0074] Step 6: Update the artificial fish information through the schooling behavior of the artificial fish. First, record the information of the i-th artificial fish. Calculate the average value of the cluster center values corresponding to the location attributes of n f fish in the field of vision of this artificial fish as the cluster center value f of the location attribute of the fish school center. c According to this value, divide the elements d in the pattern into two categories. The classification method is as follows: Calculate the distance s c1 between the first element d1 in the pattern and the cluster center value f 1d1 of class 1 of the fish school center. Define the category of d1 as 1. Calculate the distance s c2 between d1 and the cluster center value f 2d1 of class 2. Compare the sizes of s 2d1 and s 1d1 . If s 2d1 is less than s 1d1 , divide d1 into the second category, otherwise it remains in the first category. The remaining elements in the pattern are classified in the same way. The cluster center of each category in this classification situation is assigned to the fish school center as the true cluster center value f c . Repeat the method in Step 5 to calculate the fitness value fit c at this time. If it satisfies fit c > fit i and the artificial fish moves one step towards the fish school center to obtain the cluster center value f s of the schooling operator.

[0075] Step 7: Execute the following behavior. Restore the information of the i-th artificial fish, find the fish with the maximum fitness value in the field of vision. If the fitness value fit x > fit i and then the largest artificial fish moves one step to obtain the cluster center value f f of the following operator.

[0076] Step 8: Evaluate the schooling operator and the following operator. The schooling operator and the following operator generate new cluster centers. Calculate the elements in the pattern according to f s and f fFitness value during classification.

[0077] Step 9: If the clustering operator and the following operator have been executed, the foraging behavior is not executed anymore; otherwise, the foraging behavior is executed. Restore the state of the i-th artificial fish, randomly select a new position within the field of view of the i-th artificial fish, update it, and calculate the clustering center and fitness of the new position. If the fitness is higher at this position, execute the foraging behavior; otherwise, execute the random behavior, that is, randomly update once regardless of the fitness change. All artificial fish execute once, update the clustering center of each fish, and complete one iteration.

[0078] Step 10: After reaching the number of iterations, assign the recorded optimal artificial fish string value to the category attribute in the pattern structure. Select the category with the larger clustering center value as the fault category according to the classification result, and the circuit of the fault category is the multi-point node loop that has occurred.

[0079] Furthermore, through the artificial fish swarm algorithm, the horizontal and vertical difference degrees of the current waveform are clustered and analyzed, so as to realize the identification and location of the real-time fault loop. Specifically, the artificial fish swarm algorithm simulates the process of fish schools in nature to find the optimal solution through clustering and following behaviors. Through multiple iterations, the algorithm can effectively identify the abnormal patterns in the current waveform data, and determine which circuits have faults by calculating the clustering center and fitness value. Finally, the identification of the fault circuit helps to quickly locate the multi-point node loop, improve the fault detection and location efficiency of the system, and provide support for the safe and stable operation of the power system.

[0080] Embodiment 2 is an embodiment of the present invention, which provides a distributed secondary AC circuit abnormal diagnosis system for relay protection, including:

[0081] A data acquisition module that collects the current waveform data of each branch of the circuit in real time through a distributed current transformer.

[0082] A difference degree calculation module that calculates the horizontal difference degree between the branch current waveform and the reference current waveform, and the vertical difference degree between the branch current waveform and the neutral line current waveform respectively.

[0083] A fault location module that inputs the horizontal difference degree and the vertical difference degree into the artificial fish swarm clustering algorithm, and real-time identifies and locates the fault loop through clustering analysis.

[0084] Embodiment 3 is an embodiment of the present invention, which is different from the previous two embodiments in that:

[0085] If the above-described functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROMs, Read-Only Memories), random access memories (RAMs, Random Access Memories), magnetic disks, or optical discs.

[0086] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a predefined sequence list of executable instructions for implementing logical functions, and can be specifically implemented in any computer-readable medium for use by an instruction execution system, apparatus, or device (such as a computer-based system, a system including a processor, or other systems that can fetch instructions from the instruction execution system, apparatus, or device and execute the instructions), or in combination with these instruction execution systems, apparatuses, or devices. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device.

[0087] More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection portion with one or more wirings (electronic device), a portable computer diskette case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other media, then editing, interpreting, or otherwise processing it as appropriate, and then storing it in a computer memory.

[0088] It should be understood that each part of the present invention can be implemented by hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, any one or a combination of the following techniques well known in the art can be used: discrete logic circuits having logic gate circuits for implementing logic functions on data signals, application specific integrated circuits having appropriate combinational logic gate circuits, programmable gate arrays (PGAs), field programmable gate arrays (FPGAs), etc.

[0089] Example 4: According to the secondary circuit diagram of the simulated substation, install five branches and one neutral line, and configure corresponding current transformers respectively. Set the N600 grounding point as the only grounding point to ensure that there is no multi-point grounding fault in the initial state. Set the reference current waveforms of each branch as the reference waveforms under normal operating conditions.

[0090] Real-time collect the current waveform data of each branch and the neutral line through the distributed current transformers, and record them in the system data acquisition module.

[0091] Under the normal operating state of the system, record the reference current waveform data of each branch.

[0092] During the experiment, simulate different types of multi-point grounding faults, such as double-point grounding, three-point grounding, etc., and apply faults on different branches respectively.

[0093] Compare the detection time, accuracy, and false alarm rate of the traditional detection methods (such as resistance method, voltage method) and the method of the present invention after the occurrence of the fault.

[0094] Calculate the horizontal difference degree between the branch current waveform and the reference current waveform, and the vertical difference degree between the branch current waveform and the neutral line current waveform respectively. Input the calculated difference degree data into the artificial fish swarm clustering algorithm for real-time clustering analysis to identify and locate the fault loop. The experimental results are shown in Table 1.

[0095] Table 1 Experimental result table

[0096]

[0097] Through the above experimental data table, it can be clearly seen the significant advantages of the "distributed relay protection secondary AC circuit abnormal diagnosis method" of the present invention in multi-point grounding fault detection.

[0098] When detecting multi-point grounding faults using traditional resistance and voltage methods, the required time is relatively long. Especially in the case of complex faults (such as five-point grounding), the detection time increases significantly. For example, in the case of a five-point grounding fault, the resistance method and the voltage method respectively require 30 seconds and 28 seconds, while the method of the present invention only requires 10 seconds.

[0099] The method of the present invention can quickly process a large amount of current waveform data through real-time acquisition and an efficient artificial fish swarm algorithm, significantly shortening the fault detection time and improving the response speed.

[0100] Traditional methods can achieve an accuracy rate of 100% in the case of single-point grounding faults, but the accuracy rate decreases in the case of multi-point grounding faults. For example, the accuracy rate of the resistance method in the case of a five-point grounding fault is only 90%.

[0101] The method of the present invention maintains a high accuracy rate in different types of multi-point grounding faults, with a minimum of 90%. Even in complex fault scenarios, it can still effectively identify the fault loop, ensuring the safety and stability of the system.

[0102] Traditional methods have a certain false alarm rate in multi-point grounding fault detection. Especially in the case of double-point and three-point grounding faults, the false alarm rates reach 2% and 3% respectively.

[0103] The method of the present invention has a low false alarm rate, with a maximum false alarm rate of 5%. However, in practical applications, by optimizing the algorithm parameters and introducing an elastic measurement method, the false alarm rate can be further reduced, enhancing the reliability of detection. The artificial fish swarm algorithm plays a key role in the method of the present invention. By simulating the clustering and chasing behaviors of fish schools, it realizes efficient clustering analysis of the difference degree of current waveforms. Compared with the threshold setting and manual operation relied on by traditional methods, the method of the present invention can automatically adjust the clustering center to adapt to different fault types, with stronger adaptability and flexibility. The introduction of the elastic measurement algorithm enables the method of the present invention to effectively ignore power grid frequency fluctuations and noise interference, enhancing the robustness of data analysis and ensuring the accuracy of fault detection.

[0104] In practical applications, the method of the present invention not only improves the speed and accuracy of fault detection, but also reduces the workload of manual troubleshooting and lowers the operation and maintenance costs. Through real-time online monitoring, it can promptly detect and locate the fault loop, avoiding larger accidents caused by fault spread and ensuring the safe and stable operation of the power system.

[0105] Compared with traditional methods, the method of the present invention shows higher efficiency and reliability when dealing with multi-point grounding faults. Especially in a complex power grid environment, it has stronger practicality and promotion value.

[0106] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered within the scope of the claims of the present invention.

Claims

1. A method for diagnosing abnormalities in the secondary AC circuit of distributed relay protection, characterized in that, Including: Real-time collect the current waveform data of each branch of the loop through a distributed current transformer; Calculate the horizontal difference degree between the branch current waveform and the reference current waveform, and the vertical difference degree between the branch current waveform and the neutral line current waveform respectively; Input the horizontal difference degree and the vertical difference degree into the artificial fish swarm clustering algorithm, and identify and locate the faulty loop in real time through cluster analysis.

2. The abnormal diagnosis method for the secondary AC circuit of distributed relay protection according to claim 1, characterized in that: Collecting the current waveform data of each branch of the loop includes branch current waveform data, neutral line current waveform data, and reference current waveform data; The branch current waveform data is the current waveform data of each branch of the secondary relay protection loop; The neutral line current waveform data is the current waveform data of the neutral line of the secondary relay protection loop; The reference current waveform data is the current waveform data of each branch under normal working conditions, serving as a reference benchmark waveform.

3. The abnormal diagnosis method for the secondary AC circuit of distributed relay protection according to claim 2, wherein: Inputting the horizontal difference degree and the vertical difference degree into the artificial fish swarm clustering algorithm includes constructing the horizontal difference degree and the vertical difference degree as input features, and the formula is expressed as: d=[x1,x2,…x N ;y1,y2,…y N ] Among them, x N represents the horizontal difference degree, and y N represents the vertical difference degree; Defining the artificial fish swarm clustering algorithm includes defining a structure and initializing the artificial fish model; Defining the structure includes defining a structure named pattern. The pattern structure consists of feature and category attributes. The feature attribute stores the sample set, and the category attribute stores the element category, with a default value of zero; Initializing the artificial fish model includes defining each artificial fish as a structure with location, fitness, and string attributes. The location attribute contains the clustering center of each category, the category, and the number of samples in the category. The fitness attribute is the fitness, and string is a 1*N vector representing the classification result of the samples in the pattern vector, corresponding to the category attribute in pattern.

4. The abnormal diagnosis method for the secondary AC circuit of distributed relay protection according to claim 3, characterized in that: Identifying and locating the faulty loop in real time through cluster analysis includes setting the number of artificial fish as i and randomly generating i artificial fish; Set the initial clustering center to zero, and assign an initial category number to each category. The initial value of the category of each sample is a random value; The string attribute of each fish in the artificial fish swarm is assigned a value of 1 or 2, representing the classification result of the sample; Assign the first artificial fish as the optimal artificial fish, and initialize the clustering center and sample allocation through the category information and classification result in the pattern structure.

5. The abnormal diagnosis method for the secondary AC circuit of distributed relay protection according to claim 4, characterized in that: Identifying and locating the faulty loop in real time through cluster analysis also includes fitness calculation and cluster behavior update; The fitness calculation includes calculating the classification result of each fish according to the string attribute of each fish; Divide the samples into two categories according to string, calculate the sum of the distances from each category of samples to the clustering center, and the fitness is defined as the reciprocal of the sum of the distances; The clustering behavior update includes updating the clustering center through the swarm behavior, recording the information of the i-th artificial fish, and calculating the average value by adding the clustering center values of the categories corresponding to the location attributes of n f fish in the field of vision of the i-th artificial fish, and using this average value as the clustering center value f of the location attribute of the fish swarm center c ; According to the clustering center value f c Classify the element d in the pattern; Find the distance s between the first element d1 in the pattern and the cluster center value f of the fish school center class 1 c1 ; 1d1 Define the class of d1 as 1; Calculate the distance s between d1 and the clustering center value f of class 2 c2 When s 2d1 satisfies 2d1 <s 1d1 , d1 is classified into class 2. When s 2d1 ≥s 1d1 , d1 is classified into class 1; Classify all elements in pattern; The cluster center of each class is assigned to the fish school center as the cluster center value f c , and the fitness value fit is recalculated c ; When fit c >fit i and at this time, the artificial fish moves one step towards the center of the fish school to obtain the clustering center value f of the clustering operator s .

6. The abnormal diagnosis method for the secondary AC circuit of distributed relay protection according to claim 5, characterized in that: Identifying and locating the faulty loop in real time through cluster analysis also includes performing the following behavior and restoring the information of the i-th artificial fish; Find the fish with the largest fitness in the field of view. When the fitness value of the largest fish is fit x >fit i and When , the ith artificial fish moves one step and obtains the cluster center value f of the tail-chasing operator f ; Evaluate the clustering operator and the following operator. The clustering operator and the following operator generate new clustering centers, and respectively calculate the fitness values when the elements in the pattern are classified according to f s and f f classification.

7. The abnormal diagnosis method for the secondary AC circuit of distributed relay protection according to claim 6, characterized in that: Identifying and locating the faulty loop in real time through cluster analysis also includes not performing the foraging behavior when the clustering operator and the following operator have been evaluated. When the clustering operator and the following operator are not evaluated, foraging behavior is executed; Restore the state of the i-th artificial fish, randomly select a new position within the field of view of the i-th artificial fish, perform an update, and calculate the clustering center and fitness of the new position; The random behavior includes that, without considering fitness improvement, the artificial fish randomly selects a new position for update; In each iteration, all artificial fish perform a random update once, update the clustering center of each fish, and complete one iteration; After reaching the preset number of iterations, the recorded optimal artificial fish string value is assigned to the category attribute in the pattern structure, and according to the classification result, the category with a larger clustering center value is selected as the fault category, and the circuit of the fault category is the multi-point node loop that occurs.

8. A distributed relay protection secondary AC circuit abnormal diagnosis system using the method according to any one of claims 1-7, characterized in that: A data acquisition module that collects the current waveform data of each branch of the circuit in real time through a distributed current transformer; A difference calculation module that calculates the horizontal difference between the branch current waveform and the reference current waveform, and the vertical difference between the branch current waveform and the neutral line current waveform respectively; A fault location module that inputs the horizontal difference and the vertical difference into the artificial fish swarm clustering algorithm, and real-time identifies and locates the fault circuit through clustering analysis.

9. A computer device, comprising a memory and a processor, the memory storing a computer program, characterized in that: When the processor executes the computer program, the steps of the distributed relay protection secondary AC circuit abnormal diagnosis method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, the steps of the distributed relay protection secondary AC circuit abnormal diagnosis method according to any one of claims 1 to 7 are implemented.

Images