Power communication fault analysis method and device, terminal equipment and medium

Abstract

The application discloses a power communication fault analysis method and device, terminal equipment and medium, and the method comprises the following steps: collecting historical fault information of power communication, determining fault causes and diagnosis strategies of the historical fault information, and constructing a database; using a clustering algorithm to classify the historical fault information; constructing an original fault diagnosis model, training the original fault diagnosis model by using a classification result, and generating a target fault diagnosis model; inputting power communication data to be analyzed into the target fault diagnosis model, matching output fault information with the database, and obtaining fault causes and diagnosis strategies of the output fault information. The application analyzes power communication faults based on an artificial neural network, avoids the complexity and subjectivity of manual analysis, can accurately and intelligently analyze power communication faults, and thus effectively improves the reliability of power communication.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power communication, and particularly relates to a power communication fault analysis method and device, a terminal equipment and a medium. BACKGROUND

[0002] The power communication network is an integral part of the modern power grid, and is one of the three pillars of safe, stable, economic and high-quality operation of the power grid. Improving the communication quality of the power communication network and increasing the reliability of the power communication network are consistent requirements of State Grid Corporation of China for the power communication network, and will be a continuous process throughout the life cycle of the power communication network. Among them, evaluating the reliability of the power communication network is usually closely related to its fault analysis. In the power communication network, the scope of fault measurement can be divided into two levels of device fault and network fault, and network fault is a deep reflection of device fault. However, the existing method for evaluating power communication faults usually relies on manual evaluation, which has strong subjectivity. Obviously, this way not only has a large workload and takes a long time, but also is prone to errors, and cannot effectively analyze the faults of the power communication network, so as to avoid the occurrence of faults. SUMMARY

[0003] The purpose of the present application is to provide a power communication fault analysis method, device, terminal equipment and medium to solve the problem of long time consumption and low accuracy of the existing power communication fault analysis.

[0004] To achieve the above-mentioned purpose, the present application provides a power communication fault analysis method, comprising:

[0005] Collecting historical fault information of power communication, determining fault causes and diagnosis strategies of the historical fault information, and constructing a database;

[0006] Classifying the historical fault information by using a clustering algorithm;

[0007] Constructing an original fault diagnosis model, training the original fault diagnosis model by using the classification result, and generating a target fault diagnosis model;

[0008] Inputting power communication data to be analyzed into the target fault diagnosis model, matching output fault information with the database, and obtaining fault causes and diagnosis strategies of the output fault information.

[0009] Further, as a preferred, the constructing an original fault diagnosis model comprises:

[0010] Constructing an original fault diagnosis model with a three-layer structure based on an artificial neural network, wherein,

[0011] The first layer is an input layer, the second layer is a fault diagnosis layer, used for detecting and diagnosing faults of known fault types, and the third layer is an adaptive detection and diagnosis layer, including identifying and diagnosing unknown fault types that the second layer fails to recognize in an adaptive manner.

[0012] Further, as preferably, the diagnostic algorithm of the fault diagnosis layer is:

[0013] The data feature vector to be detected is defined as S={s1, s2, s3,...s n}, and the i-th fault identifier in the fault diagnosis layer is Z={z i1 ,z i2 ,z i3 ,...z in};

[0014] The fusion degree of the data feature vector and the fault identifier is:

[0015]

[0016] In the formula, θ t represents the weight ratio of the k-th characteristic parameter, and satisfies 0<D(s,z i )<1;

[0017] After conversion, we get:

[0018]

[0019] In the formula, 0<f(s,z i )<1.

[0020] Further, as preferably, the standard for judging whether a power communication fault occurs is:

[0021] When f(s,z i )∈(0.67,0.89), it is judged that the power communication has a fault.

[0022] Further, as preferably, after collecting the historical fault information of the power communication, the historical fault information is further subjected to data preprocessing, including:

[0023] Judging whether the historical fault information meets the database case standard;

[0024] If yes, the historical fault information meeting the database case standard is subjected to data cleaning, denoising and normalization processing;

[0025] If no, the historical fault information not meeting the database case standard is eliminated.

[0026] Further, as preferred, the historical failure information meeting the database case standard is denoised, comprising:

[0027] Converting the historical failure information into an electrical signal;

[0028] Performing a multi-layer wavelet transform on the electrical signal to obtain low-frequency coefficients and high-frequency coefficients of the wavelet;

[0029] Performing a shrinkage processing on the high-frequency coefficients;

[0030] Performing an inverse transform on the wavelet after the shrinkage processing to obtain a signal estimation value;

[0031] Performing a signal median filtering processing on the signal estimation value to obtain the denoised historical failure information.

[0032] Further, as preferred, the historical failure information is classified by using a clustering algorithm, comprising:

[0033] Based on the historical failure information, determining an initial clustering cluster center, an iteration number threshold and a clustering number;

[0034] Calculating Manhattan distances of all data points in the historical failure information to the center, and distributing the data points to the center closest to the data points;

[0035] Using a mean function to obtain a new center of the clustering cluster, and calculating a shift amount of the new center;

[0036] Judging whether a difference between the shift amount and the iteration number threshold is less than a preset value;

[0037] If yes, the clustering ends;

[0038] If no, returning to execute the step of determining the initial clustering cluster center, the iteration number threshold and the clustering number.

[0039] The application further provides a power communication failure analysis device, comprising:

[0040] An information collection unit, configured to collect historical failure information of power communication, determine failure causes and diagnosis strategies of the historical failure information, and construct a database;

[0041] A clustering analysis unit, configured to classify the historical failure information by using a clustering algorithm;

[0042] A model construction unit, configured to construct an original failure diagnosis model, train the original failure diagnosis model by using a classification result, and generate a target failure diagnosis model;

[0043] The fault analysis unit inputs power communication data to be analyzed into the target fault diagnosis model, matches output fault information with the database, and obtains fault causes and diagnosis strategies of the output fault information.

[0044] The application also provides a terminal device, comprising:

[0045] one or more processors;

[0046] a memory coupled to the processors, storing one or more programs;

[0047] When the one or more programs are executed by the one or more processors, the one or more processors implement the power communication fault analysis method according to any one of the preceding embodiments.

[0048] The application also provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the power communication fault analysis method according to any one of the preceding embodiments.

[0049] Compared with the prior art, the application has the following beneficial effects:

[0050] The application discloses a power communication fault analysis method, device, terminal device and medium, the method comprising: collecting historical fault information of power communication, determining fault causes and diagnosis strategies of the historical fault information, and constructing a database; using a clustering algorithm to classify the historical fault information; constructing an original fault diagnosis model, training the original fault diagnosis model using the classification results, and generating a target fault diagnosis model; inputting power communication data to be analyzed into the target fault diagnosis model, matching output fault information with the database, and obtaining fault causes and diagnosis strategies of the output fault information. The application analyzes power communication faults based on artificial neural networks, avoids the complexity and subjectivity of manual analysis, can accurately and intelligently analyze power communication faults, and thus effectively improves the reliability of power communication. BRIEF DESCRIPTION OF DRAWINGS

[0051] In order to more clearly illustrate the technical solutions of the application, the following will briefly introduce the drawings needed in the embodiments. Obviously, the drawings in the following description are only some embodiments of the application, and other drawings can be obtained by those skilled in the art without creative effort on the basis of these drawings.

[0052] Figure 1 is a flowchart of a power communication fault analysis method provided by an embodiment of the application;

[0053] Figure 2is a schematic diagram of an artificial neural network provided by some embodiments of the present application;

[0054] Figure 3 is a structural schematic diagram of a power communication fault analysis device provided by some embodiments of the present application;

[0055] Figure 4 is a structural schematic diagram of a terminal device provided by some embodiments of the present application. DETAILED DESCRIPTION

[0056] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application but not all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of the present application.

[0057] It should be understood that the step numbers used herein are only for the convenience of description and are not limited to the execution sequence of the steps.

[0058] It should be understood that the terms used in the present application are only for the purpose of describing specific embodiments and are not intended to limit the present application. As used in the present application and the appended claims, unless otherwise clearly indicated by the context, the singular forms "a", "an" and "the" are intended to include the plural forms.

[0059] The terms "comprise" and "include" indicate the presence of described features, whole, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, whole, steps, operations, elements, components and / or sets thereof.

[0060] The term "and / or" means any combination of one or more of the associated listed terms and all possible combinations thereof, and includes these combinations.

[0061] First embodiment:

[0062] Referring to Figure 1 , some embodiments of the present application provide a power communication fault analysis method. As shown in Figure 1 , the power communication fault analysis method comprises steps S10 to S40. Each step is as follows:

[0063] S10, collect historical fault information of power communication, determine the fault cause and diagnosis strategy of the historical fault information, and construct a database.

[0064] In this step, first, the historical fault information of power communication is collected, then the fault cause and diagnosis strategy corresponding to the fault information are determined, and then the database is constructed according to the fault information, the fault cause and the diagnosis strategy. Specifically, in the embodiment, the historical fault information can be poor transmission quality, the corresponding fault cause can be weak network line anti-interference ability or poor shielding effect, and the diagnosis strategy can be to replace the multi-core wire.

[0065] As a preferred embodiment, preferably, the fault symptoms of the historical fault information and the fault cause have an m→n mapping relationship, that is, a plurality of fault symptoms can point to the same fault cause, or one fault symptom is the result of the simultaneous action of a plurality of fault causes.

[0066] In a specific embodiment, after collecting the historical fault information of power communication, the historical fault information is further subjected to data preprocessing, including:

[0067] determining whether the historical fault information meets the database case standard;

[0068] If yes, the historical fault information meeting the database case standard is subjected to data cleaning, denoising and normalization processing;

[0069] If no, the historical fault information not meeting the database case standard is rejected.

[0070] In the embodiment, in order to unify the data format in the constructed database, data preprocessing is usually required, and a corresponding database case standard is set, for example, a normalization processing method is used, that is, a maximum and minimum value method is used to normalize each information to 0 to 100.

[0071] In an embodiment, the historical fault information meeting the database case standard is subjected to denoising, including the following steps:

[0072] 1.1) converting the historical fault information into an electrical signal;

[0073] 1.2) subjecting the electrical signal to multi-layer wavelet transform, decomposing it into 3 layers, and then obtaining low-frequency coefficients and high-frequency coefficients of the wavelet;

[0074] 1.3) subjecting the obtained high-frequency coefficients of each layer to shrinkage processing, and the function used for shrinkage processing is:

[0075]

[0076] In the formula, T j represents a threshold value, γ jk is the wavelet coefficient of the jth layer, represents the wavelet coefficient after threshold shrinkage, j is the number of wavelet decomposition layers, and k represents the kth wavelet coefficient.

[0077] 1.4) inverse transform the wavelet after the shrinkage processing to obtain a signal estimation value;

[0078] 1.5) perform signal median filtering processing on the signal estimation value to obtain the denoised historical fault information.

[0079] The calculation formula used in the median filtering is:

[0080]

[0081] Further, in the embodiment, the calculation formula used in the normalization processing is:

[0082]

[0083] In the formula, x represents a parameter of the historical fault information, x min represents the minimum value of the parameter of the historical fault information, x max represents the maximum value of the parameter of the historical fault information.

[0084] S20, classifying the historical fault information by using a clustering algorithm;

[0085] After step S10 is performed, the data of the historical fault information in a unified format is obtained, and the present step aims to classify the data. Specifically, the present embodiment uses a clustering algorithm to classify the data.

[0086] In a specific embodiment, step S20 further includes the following sub-steps:

[0087] 2.1) determining an initial clustering cluster center, an iteration number threshold and a clustering number based on the historical fault information;

[0088] In the present step, the iteration number threshold is the maximum iteration number in the iteration process.

[0089] 2.2) calculating the Manhattan distance of all data points in the historical fault information to the cluster center and assigning the data points to the cluster center closest to the data points; wherein the calculation formula of the Manhattan distance is:

[0090] c = |x1-x2| + |y1-y2|;

[0091] 2.3) using a mean function to obtain a new cluster center of the clustering cluster and calculating the offset of the new cluster center;

[0092] 2.4) judging whether the difference between the offset and the iteration number threshold is less than a preset value; if yes, the clustering ends; if no, returning to perform step 2.1).

[0093] S30, constructing an original fault diagnosis model, training the original fault diagnosis model by using the classification result, and generating a target fault diagnosis model;

[0094] In this step, based on the classification result, the classified historical fault information is first divided into a training set and a test set, and the division ratio is preferably 7:3.

[0095] Further, an original fault diagnosis model is constructed, specifically, an original fault diagnosis model with a three-layer structure is constructed based on an artificial neural network. Here, a fully connected neural network is preferred, such as Figure 2 , Figure 2 which is a schematic diagram of the principle of the fully connected neural network. The first layer of the neural network is an input layer, the second layer is a fault diagnosis layer, which includes fault detection and diagnosis of known fault types that are mutually independent in terms of occurrence probability, and the third layer is an adaptive detection and diagnosis layer, which includes identification and diagnosis of new fault types and early faults in an adaptive manner.

[0096] Further, the diagnosis algorithm of the fault diagnosis layer is:

[0097] The data feature vector to be detected is defined as S = {s1, s2, s3,... s n}, and the i-th fault identifier in the fault diagnosis layer is Z = {z i1 ,z i2 ,z i3 ,...z in};

[0098] The fusion degree of the data feature vector and the fault identifier is:

[0099]

[0100] where θ t represents the weight ratio of the k-th characteristic parameter, and satisfies 0 < D(s, z i ) < 1;

[0101] The final fusion degree calculation formula after transformation is:

[0102]

[0103] where 0 < f(s, z i ) < 1.

[0104] When the output test result reaches the preset standard, the target fault diagnosis model is generated.

[0105] S40, input the power communication data to be analyzed into the target fault diagnosis model, match the output fault information with the database, and obtain the fault cause and diagnosis strategy of the output fault information.

[0106] In this embodiment, when fault analysis is needed, only the power communication data to be analyzed needs to be collected in real time, and then input into the target fault diagnosis model to obtain the diagnosis result. It should be noted that the standard for judging whether the power communication fault occurs in this embodiment is that when f(s,z i )∈(0.67, 0.89), it is judged that the power communication fault occurs.

[0107] Finally, in order to enrich the cases in the database, after each diagnosis, the fault information obtained by the fault diagnosis result is matched with the fault information in the database. If the same fault information can be matched, the fault cause and diagnosis strategy can be found correspondingly, so as to provide operation guidance for maintenance personnel. If it cannot be matched, it can be saved as a new case in the database to realize the effect of self-learning of the database.

[0108] The first embodiment of the present application analyzes the power communication fault based on the artificial neural network, avoids the tediousness and subjectivity of manual analysis, can accurately and intelligently analyze the power communication fault, and further effectively improves the reliability of the power communication.

[0109] Second embodiment:

[0110] This embodiment is another embodiment of the present application. Different from the first embodiment, this embodiment provides a verification test of the power communication fault analysis method based on the artificial neural network. In order to verify the technical effects adopted in the method, the method of the present application is compared with the traditional fault diagnosis method based on the BP algorithm for test, the test results are compared by scientific means, and the real effect of the method is verified.

[0111] The test samples are selected as normal data, fault data, suspected abnormal data and abnormal data, each of which has 40 record samples. When the BP algorithm is used for learning, the fast training method is used, 70% of the samples are used as training samples, and 30% of the samples are used as test samples. After repeated observation and learning, the measured results are compared with the results obtained by using the method of the present application. The comparison results are shown in the following table.

[0112] Table 1 comparison table of experimental results.

[0113] Test sample BP algorithm The method of the present application Classification accuracy 97.26％ 98.87％ Fault precision 97.95％ 99.14％ Fault recall 82.87％ 98.10％

[0114] As can be seen from the above table, the method of the present application has stronger robustness compared with the traditional method, which embodies the effectiveness of the method of the present application.

[0115] The third embodiment is as follows:

[0116] Referring to Figure 3 In the third embodiment of the present application, an electric power communication fault analysis device is also provided, comprising:

[0117] An information collection unit 01 is configured to collect historical fault information of electric power communication, determine fault causes and diagnosis strategies of the historical fault information, and construct a database;

[0118] A clustering analysis unit 02 is configured to classify the historical fault information by using a clustering algorithm;

[0119] A model construction unit 03 is configured to construct an original fault diagnosis model according to the classification result, and generate a target fault diagnosis model when the original fault diagnosis model is trained until a preset condition is met;

[0120] A fault analysis unit 04 is configured to input electric power communication data to be analyzed into the target fault diagnosis model, match output fault information with the database, and obtain fault causes and diagnosis strategies of the output fault information.

[0121] It can be understood that the electric power communication fault analysis device provided in the embodiment is used to execute the electric power communication fault analysis method as described in the first embodiment. The electric power communication fault is analyzed based on the artificial neural network in the embodiment, which avoids the complexity and subjectivity of manual analysis, can accurately and intelligently analyze the electric power communication fault, and further effectively improves the reliability of electric power communication.

[0122] The fourth embodiment is as follows:

[0123] Referring to Figure 4 In the fourth embodiment of the present application, a terminal device is also provided, comprising:

[0124] One or more processors;

[0125] A memory coupled to the processor, configured to store one or more programs;

[0126] When the one or more programs are executed by the one or more processors, the one or more processors implement the electric power communication fault analysis method as described above.

[0127] The processor is configured to control overall operations of the terminal device to complete all or part of the steps of the power communication fault analysis method described above. The memory is configured to store various types of data to support the operation of the terminal device, which can include, for example, instructions for any application or method operating on the terminal device, and application-related data. The memory can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0128] In an exemplary embodiment, the terminal device can be implemented by one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic elements, for executing the power communication fault analysis method as described in any of the above embodiments and achieving the technical effects consistent with the above method.

[0129] In another exemplary embodiment, a computer-readable storage medium including a computer program is also provided, which, when executed by a processor, implements the steps of the power communication fault analysis method as described in any of the above embodiments. For example, the computer-readable storage medium can be the above-mentioned memory including the computer program, which can be executed by the processor of the terminal device to complete the power communication fault analysis method as described in any of the above embodiments and achieve the technical effects consistent with the above method.

[0130] Further, the operations of the processes described herein can be performed in any suitable order unless otherwise indicated herein or otherwise clearly contradicted by context. The processes described herein (or variations and / or combinations thereof) can be implemented under the control of one or more computer systems configured with executable instructions (e.g., computer programs, one or more computer programs, or one or more applications), by hardware, or combinations thereof. The computer programs include machine- readable instructions that can be executed by one or more processors.

[0131] Further, the methods can be implemented in any suitable type of computing platform operatively coupled to, including but not limited to, a personal computer, a mini-computer, a mainframe, a workstation, a network or distributed computing environment, a stand-alone or integrated computer platform, or in communication with a charged particle tool or other imaging device, and the like. Aspects of the present application can be implemented in machine-readable code stored on a non-transitory storage medium or device, whether removable or integrated to the computing platform, such as a hard disk, an optical read and / or write storage medium, RAM, ROM, and the like, such that it can be read by a programmable computer to configure and operate the computer to perform the processes described herein when the storage medium or device is read by the computer. Further, the machine-readable code, or portions thereof, can be transmitted over wired or wireless networks. The present application described herein includes these and other different types of non-transitory computer-readable storage media when such media include instructions or programs implementing the steps described above in conjunction with a microprocessor or other data processor. The present application also includes the computer itself when programmed according to the methods and techniques described herein. The computer program is capable of applying to input data to perform the functions described herein, thereby transforming the input data to generate output data that is stored to non-volatile memory. The output information can also be applied to one or more output devices, such as a display. In a preferred embodiment of the present application, the transformed data represents a physical and tangible object, including a particular visual depiction of a physical and tangible object produced on a display.

[0132] As used in this application, the terms "component," "module," "system" and the like are intended to refer to a computer-related entity, either hardware, firmware, a combination of hardware and software, software, or software in execution. For example, a component can be, but is not limited to being, a process running on a processor, a processor, an object, an executable, a thread of execution, a program, and / or a computer. By way of illustration, both an application running on a computing device and the computing device can be a component. One or more components can reside within a process and / or thread of execution and a component can be localized, partially and / or entirely, in one computer or distributed between two or more computers. Also, these components can execute from various computer readable media having various data structures stored thereon. The components can communicate by way of local and / or remote processes such as in accordance with a signal having one or more data packets (e.g., data from one component interacting with another component in a local system, distributed system, and / or across a network such as the Internet with other systems via the signal).

[0133] The preferred embodiments of the present application have been described above with the aid of numerous reference to drawings of specific embodiments of the application. These specific embodiments are indicative, however, of only a few of the various ways in which the principles of the application can be employed. Other objects and advantages of the application will become apparent to those skilled in the art from the subsequent detailed description, which, taken in conjunction with the drawings, discloses the embodiments. This description concludes with the Claim set forth below.

Claims

1. A method for analyzing power communication faults, characterized in that, include: Collect historical fault information of power communication, determine the causes of the faults and diagnostic strategies, and build a database; Based on the historical fault information, the initial cluster centroid, iteration number threshold, and number of clusters are determined; Calculate the Manhattan distance from all data points in the historical fault information to the centroid, and assign the data points to the centroids closest to the data points; The new centroids of the clusters are determined using the mean function, and the offset of the new centroids is calculated. Determine whether the difference between the offset and the iteration number threshold is less than a preset value; If so, then the clustering ends; If not, return to the steps of determining the initial cluster centroids, iteration threshold, and number of clusters; Construct an original fault diagnosis model, and train the original fault diagnosis model using the classification results to generate a target fault diagnosis model; The power communication data to be analyzed is input into the target fault diagnosis model, and the output fault information is matched with the database to obtain the fault cause and diagnosis strategy related to the output fault information. The construction of the original fault diagnosis model includes: The original fault diagnosis model is constructed based on an artificial neural network with a three-layer structure. The first layer is the input layer, the second layer is the fault diagnosis layer, which is used to detect and diagnose faults of known fault types, and the third layer is the adaptive detection and diagnosis layer, which includes the adaptive identification and diagnosis of unknown fault types that the second layer failed to identify. The diagnostic algorithm for the fault diagnosis layer is as follows: Define the feature vector of the data to be detected as: The i-th fault identifier in the fault diagnosis layer is ; The degree of fusion between the data feature vector and the fault identifier is: ； In the formula, Let represent the weight ratio of the k-th feature parameter, and satisfy . ; After conversion, we get: ； In the formula, ; The criteria for determining whether a power communication failure has occurred are as follows: when At that time, it was determined that a power communication failure had occurred.

2. The power communication fault analysis method according to claim 1, characterized in that, After collecting historical fault information from power communication, the method further includes data preprocessing of the historical fault information, including: Determine whether the historical fault information conforms to the database case standard; If so, then the historical fault information that conforms to the database case standard will be cleaned, denoised, and normalized. If not, then remove historical fault information that does not meet the database case standards.

3. The power communication fault analysis method according to claim 2, characterized in that, The historical fault information conforming to the database case standard is denoised, including: The historical fault information is converted into electrical signals; Perform multi-level wavelet transform on the electrical signal to obtain the low-frequency and high-frequency coefficients of the wavelets; The high-frequency coefficients are then subjected to shrinkage processing; The signal estimate is obtained by performing an inverse transform on the contracted wavelet; The estimated signal value is subjected to median filtering to obtain the denoised historical fault information.

4. A power communication fault analysis device, characterized in that, include: The information acquisition unit is used to collect historical fault information of power communication, determine the fault causes and diagnostic strategies of the historical fault information, and build a database; The clustering analysis unit is used to determine the initial cluster centroid, iteration number threshold, and number of clusters based on the historical fault information. Calculate the Manhattan distance from all data points in the historical fault information to the centroid, and assign the data points to the centroids closest to the data points; use the mean function to find the new centroids of the clusters, and calculate the offset of the new centroids; determine whether the difference between the offset and the iteration number threshold is less than a preset value; if yes, the clustering ends; if no, return to the steps of determining the initial cluster centroids, iteration number threshold, and number of clusters. The model building unit is used to build the original fault diagnosis model, and to train the original fault diagnosis model using the classification results to generate the target fault diagnosis model. The construction of the original fault diagnosis model includes: constructing an original fault diagnosis model with a three-layer structure based on an artificial neural network, wherein the first layer is the input layer, the second layer is the fault diagnosis layer, which is used to detect and diagnose faults of known fault types, and the third layer is the adaptive detection and diagnosis layer, which includes identifying and diagnosing unknown fault types that the second layer failed to identify in an adaptive manner. The diagnostic algorithm of the fault diagnosis layer is as follows: Define the data feature vector to be detected as... The i-th fault identifier in the fault diagnosis layer is ; The degree of fusion between the data feature vector and the fault identifier is: ； In the formula, Let represent the weight ratio of the k-th feature parameter, and satisfy . ; After conversion, we get: ； In the formula, ; The criteria for determining whether a power communication failure has occurred are as follows: when At that time, it is determined that a power communication failure has occurred; The fault analysis unit is used to input the power communication data to be analyzed into the target fault diagnosis model, match the output fault information with the database, and obtain the fault cause and diagnosis strategy corresponding to the output fault information.

5. A terminal device, characterized in that, include: One or more processors; A memory, coupled to the processor, for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the power communication fault analysis method as described in any one of claims 1-3.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the power communication fault analysis method as described in any one of claims 1-3.

Images