Metering fault feature sub-mode extraction and fault diagnosis method and system

Abstract

The invention discloses a metering fault feature sub-mode extraction and fault diagnosis method and system, and relates to the technical field of power systems, and the method comprises the steps that a metering master station obtains fault category sample data of a metering low-voltage operation and maintenance work order and alarm event record feature sample data of a fault category; obtaining a metering low-voltage fault feature sub-mode of the cluster number K; determining the optimal clustering number of the fault feature sub-modes based on an elbow method, and obtaining a final fault feature sub-mode; fault diagnosis is carried out based on the fault feature sub-mode and the real-time alarm event, and generation of a metering low-voltage operation and maintenance work order is supported. According to the metering fault feature sub-mode extraction and fault diagnosis method provided by the invention, fine extraction and management of the fault feature sub-mode are realized, fault diagnosis is carried out based on the combination of the fault feature sub-mode and a real-time alarm event, the accuracy of metering low-voltage side fault diagnosis is improved, and generation of a metering low-voltage operation and maintenance work order is supported.

Description

Technical Field

[0001] The present invention relates to the technical field of electric power systems, and in particular to a method and system for extracting characteristic sub-patterns of metering faults and diagnosing faults. Background Art

[0002] The on-site operation conditions of power users are complex, and there are many types of metering abnormal events in daily operation, which makes it difficult to conduct a comprehensive and effective analysis of abnormal alarm events. At present, there is no unified and efficient analysis method for all metering alarm events that can be given by the metering automation system. Most of them use traditional methods for abnormal analysis or only analyze and process single abnormal alarm events. In the face of multiple alarm events occurring at the same time, most of them need to rely on the experience of metering personnel for judgment and analysis. The results are easily affected by subjective factors and lack theoretical support. At the same time, after the grassroots metering low-voltage fault operation and maintenance work, the on-site operation and maintenance personnel will report the fault category after the work order is completed. However, since the on-site operation and maintenance personnel will not describe the fault category in detail when filling in the fault category, there may be multiple fault causes for the same fault category. Based on this, combined with the fault category and alarm event, the characteristic sub-pattern extraction of the fault alarm event on the metering low-voltage side can realize the refined identification and extraction of the characteristic sub-patterns of the fault alarm event with different fault causes in the same fault category, which is of great significance to improving the accuracy of fault diagnosis on the metering low-voltage side.

[0003] In this regard, the method of the present invention is based on the fault categories and alarm event records of the metering automation system operation and maintenance work order, and proposes a metering fault feature sub-pattern extraction and fault diagnosis method based on the K-Medoids algorithm and the elbow method, so as to realize the refined extraction and management of the fault feature sub-pattern, and carry out fault diagnosis based on the fault feature sub-pattern and the real-time alarm event, so as to improve the accuracy of fault diagnosis on the metering low-voltage side and support the generation of metering low-voltage operation and maintenance work orders. Summary of the invention

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

[0005] Therefore, the technical problem solved by the present invention is that the existing metering alarm event analysis method lacks a unified and efficient analysis method, is unable to effectively integrate and analyze various types of metering abnormal events, and how to optimize the refined identification of different fault causes when multiple alarm events occur at the same time.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: a method for extracting characteristic sub-patterns of metrology faults and diagnosing faults, comprising:

[0007] The metering master station obtains fault category sample data of the metering low-voltage operation and maintenance work order;

[0008] The metering master station obtains the characteristic sample data of the alarm event record of the fault category;

[0009] Based on the K-Medoids clustering algorithm, set the metering low voltage fault characteristic sub-pattern clustering algorithm and clustering parameters, calculate the metering low voltage fault characteristic sub-pattern clustering standard, calculate and update the fault characteristic sub-pattern clustering center, and obtain the metering low voltage fault characteristic sub-pattern with cluster number K;

[0010] Determine the optimal clustering number of fault feature sub-patterns based on the elbow method and obtain the final fault feature sub-patterns;

[0011] Fault diagnosis is carried out based on fault characteristic sub-patterns and real-time alarm events, supporting the generation of metering low-voltage operation and maintenance work orders.

[0012] As a preferred solution of the method for extracting characteristic sub-patterns of metering faults and diagnosing faults described in the present invention, the fault categories are sorted according to the fault categories, and the fault category n is recorded as Y n .

[0013] As a preferred solution of the method for extracting characteristic sub-patterns of metering faults and diagnosing faults described in the present invention, wherein: the alarm events include metering equipment alarms, abnormal power consumption alarms, photovoltaic user alarms and other alarms;

[0014] The alarm events are sorted according to the alarm event sequence, and each alarm event is sorted into the Xnth vector.

[0015] As a preferred solution of the method for extracting characteristic sub-patterns of metering faults and diagnosing faults described in the present invention, the clustering algorithm and clustering parameters of the characteristic sub-patterns of metering low-voltage faults are set,

[0016] Set the number of clusters to K, and randomly select K data samples from the alarm event record feature vector data samples as the initial cluster centers, and record the kth cluster center as [Ok,1,Ok,2,…,Ok,j,…,O 56 ,].

[0017] As a preferred solution of the method for extracting characteristic sub-patterns of metering faults and diagnosing faults described in the present invention, the calculation of the clustering standard of the characteristic sub-patterns of metering low-voltage faults includes using the sum of squared errors to measure the distance between the qth alarm event record feature vector data sample and the kth clustering center, as the clustering standard of the characteristic sub-patterns of metering low-voltage faults, expressed as:

[0018]

[0019] Among them, Φ k,q Record the sum of squared errors between the feature vector data sample and the kth cluster center for the qth alarm event;

[0020] After calculating the sum of squares of the errors between the qth alarm event record feature vector data sample and the K cluster centers, if the sum of squares of the errors between a certain alarm event record feature vector data sample and a certain cluster center is the smallest, the sample is classified into the cluster cluster.

[0021] As a preferred solution of the method for extracting characteristic sub-patterns of metering faults and diagnosing faults described in the present invention, the calculation and updating of the cluster center of the fault characteristic sub-pattern is included in the kth cluster, and each alarm event record feature vector data sample in the cluster is taken as the cluster center in turn, and the sum of square errors between other samples and the cluster center is calculated, and the alarm event record feature vector data sample corresponding to the cluster center with the smallest sum of square errors is updated as the cluster center of the cluster cluster.

[0022] As a preferred solution of the method for extracting characteristic sub-patterns of metering faults and diagnosing faults of the present invention, the method of obtaining the final characteristic sub-pattern of faults includes connecting the end of the error sum of squares and the K value relationship curve by the elbow method to obtain a straight line, which is recorded as y=ax+b;

[0023] When the number of clusters is K, the difference between the error sum of squares and the value of the straight line is recorded as LK. The K value corresponding to the maximum LK is the optimal number of clusters.

[0024] Calculate the sum of squares of errors under different K values, calculate LK under different clustering numbers K values, and the K value corresponding to the maximum LK is taken as the optimal clustering number. The metering low voltage fault characteristic subpattern under this clustering number is the metering low voltage fault characteristic subpattern of fault category Yi. The final metering low voltage fault characteristic subpattern of fault category Yi is recorded as [Es,i,1,Es,i,2,…,Es,i,j,…,Es,i,56], s∈[1,…,Ui], Ui is the optimal clustering number of fault category Yi.

[0025] Another object of the present invention is to provide a metering fault characteristic sub-pattern extraction and fault diagnosis system, which can solve the problems of imprecise fault cause identification and subjective judgment affecting the results in existing metering automation systems by constructing a metering fault characteristic sub-pattern extraction and fault diagnosis system.

[0026] In order to solve the above technical problems, the present invention provides the following technical solutions: a metering fault characteristic sub-pattern extraction and fault diagnosis system, comprising: a sample acquisition module, an alarm recording module, a clustering analysis module, an optimized clustering module and a fault diagnosis module; the sample acquisition module is used for the metering master station to obtain the fault category sample data of the metering low-voltage operation and maintenance work order; the alarm recording module is used for the metering master station to obtain the alarm event record characteristic sample data of the fault category; the clustering analysis module is used to set the metering low-voltage fault characteristic sub-pattern clustering algorithm and clustering parameters based on the K-Medoids clustering algorithm, calculate the metering low-voltage fault characteristic sub-pattern clustering standard, calculate and update the fault characteristic sub-pattern clustering center, and obtain the metering low-voltage fault characteristic sub-pattern with a cluster number K; the optimized clustering module is used to determine the optimal clustering number of the fault characteristic sub-pattern based on the elbow method, and obtain the final fault characteristic sub-pattern; the fault diagnosis module is used to carry out fault diagnosis based on the fault characteristic sub-pattern and real-time alarm events, and support the generation of metering low-voltage operation and maintenance work orders.

[0027] A computer device comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above-mentioned method for extracting characteristic sub-patterns of metering faults and diagnosing faults when executing the computer program.

[0028] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the above-mentioned method for extracting characteristic sub-patterns of metering faults and diagnosing faults.

[0029] Beneficial effects of the present invention: The metering fault characteristic sub-pattern extraction and fault diagnosis method provided by the present invention is based on the metering automation system operation and maintenance work order fault category and alarm event record, and proposes a metering fault characteristic sub-pattern extraction and fault diagnosis method based on the K-Medoids algorithm and the elbow method, so as to realize the refined extraction and management of fault characteristic sub-patterns, and carry out fault diagnosis based on this in combination with the fault characteristic sub-patterns and real-time alarm events, improve the accuracy of fault diagnosis on the metering low-voltage side, and support the generation of metering low-voltage operation and maintenance work orders. On the other hand, data samples can be further classified according to the regional attributes of the low-voltage substation, the scale of users in the substation, the type of terminal, the networking method and other requirements, so as to realize personalized metering fault characteristic sub-pattern extraction and achieve better fault characteristic sub-pattern extraction and fault diagnosis effects. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0031] Figure 1 An overall flow chart of a method for extracting characteristic sub-patterns of metering faults and diagnosing faults provided by an embodiment of the present invention.

[0032] Figure 2 A schematic diagram of alarm event sequence numbers of a method for extracting characteristic sub-patterns of metering faults and diagnosing faults provided by an embodiment of the present invention.

[0033] Figure 3 A schematic diagram of elbow clustering for a method for extracting characteristic sub-patterns of metering faults and diagnosing faults provided by an embodiment of the present invention.

[0034] Figure 4 An overall structural diagram of a metering fault characteristic sub-pattern extraction and fault diagnosis system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0035] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0036] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0037] Example 1

[0038] Reference Figure 1-Figure 3 , is an embodiment of the present invention, and provides a method for extracting characteristic sub-patterns of metering faults and diagnosing faults, comprising:

[0039] S1: The metering master station obtains sample data of fault categories of metering low-voltage operation and maintenance work orders:

[0040] The fault categories include 13 fault categories, including power meter communication module failure, power meter clock battery failure, power meter body failure, terminal communication module failure, SIM card failure, terminal power failure, terminal parameter setting error, user file error, abnormal power generation and consumption of photovoltaic users, wiring error, metering circuit failure, metering circuit failure, and human theft of electricity. The above fault categories are respectively recorded as fault categories Y1, Y2,…, Y13, and fault category n is recorded as Yn.

[0041] Taking the substation as a unit, all operation and maintenance work order records of the substation terminals and electric energy meters are obtained from the metering master station database, and the fault type of each operation and maintenance work order is obtained.

[0042] S2: The metering master station obtains the characteristic sample data of the alarm event record of the fault category:

[0043] Metering low voltage alarm events include four categories: metering equipment alarm, abnormal power consumption alarm, photovoltaic user alarm, and other alarms.

[0044] The metering equipment alarms include 15 alarm events, such as abnormal clock of the energy meter, change of programming times of the energy meter, low voltage of the clock battery, change of metering transformer ratio, change of pulse constant of the energy meter, change of rate period, clearing of maximum demand, stop of the energy meter, flight of the energy meter, decrease of indication, opening and closing of the metering device door, quantity difference alarm, opening alarm of the energy meter terminal box, opening alarm of the energy meter cover, and failure alarm of the energy meter metering chip. The above 15 alarm events are respectively regarded as the 1st to 15th vectors of the alarm event vector, and are respectively recorded as X1, X2,…, X15.

[0045] The abnormal power consumption alarm includes 21 alarm events, such as voltage loss (phase loss), current loss, phase failure, voltage reverse phase sequence, current reverse polarity, load overload, power consumption exceeding the contract capacity, power exceeding the set value, overvoltage, current overload, reactive power overcompensation, reactive power undercompensation, voltage distortion, current distortion, current imbalance, voltage imbalance, long-term flicker, large zero-sequence current, failure of electric energy meter opening and closing control, failure of electric energy meter reading, and insufficient remaining electricity fee. The above 21 alarm events are respectively regarded as the 16th to 36th vectors of the alarm event vector, and are respectively recorded as X16, X17,…, X36.

[0046] Photovoltaic user alarms include five alarm events: over-capacity power generation, total grid-connected power generation not equal to grid-connected power, power generation less than grid-connected power, abnormal utilization hours, and power flow reversal. The above five alarm events are respectively regarded as the 37th to 41st vectors of the alarm event vector, and are recorded as X37, X38, …, X41 respectively.

[0047] Other alarms include 15 items such as remote signal change, equipment power failure, equipment power-on, communication failure, monthly communication traffic limit exceeded, backup battery abnormality, SIM card replacement, measurement point parameter change, topology information change, terminal module status change, safety monitoring, unknown meter discovery, meter module replacement, file transfer situation alarm, and operating system abnormality alarm. The above 15 alarm events are respectively regarded as the 42nd to 56th vectors of the alarm event vector, and are recorded as X42, X43, …, X56 respectively.

[0048] Taking the substation as a unit, for each operation and maintenance work order, obtain all alarm event records of the substation terminal N hours before the time when the operation and maintenance work order is generated from the metering automation system database, including the event name, alarm event occurrence time and recovery time of each alarm event, eliminate the alarm events whose alarm event recovery time is before the time when the operation and maintenance work order is generated, and eliminate duplicate alarm event records.

[0049] Based on this, the feature vector of the alarm event record corresponding to a single work order can be recorded as [X1,X2,…,Xi,…,X56]. When the jth alarm event occurs before the work order occurs, the feature value of the jth alarm event is Xj=1, otherwise Xj=0. Since the historical operation and maintenance work order provides the fault category, the feature vector of the alarm event record corresponding to the fault category Yi can be obtained as [Xi,1,Xi,2,…,Xi,j,…,Xi,56]. Within the sample data range, the number of work orders with fault category Yi is Q, and the feature vector of the alarm event record corresponding to the qth work order can be recorded as [Xq,i,1,Xq,i,2,…,Xq,i,j,…,Xq,i,56].

[0050] Based on the alarm event record feature vector [Xq,i,1,Xq,i,2,…,Xq,i,j,…,Xq,i,56] of the qth work order of fault category Yi, a two-dimensional graph is drawn, where the horizontal axis is the sequence number of the alarm event and the vertical axis is the feature value of the alarm event (i.e., 0 or 1). The characteristic sample curve of the alarm event record data of the qth work order of fault category Yi can be obtained, as shown in Figure 2 shown.

[0051] S3: Based on the K-Medoids clustering algorithm, set the metering low voltage fault characteristic sub-pattern clustering algorithm and clustering parameters, calculate the metering low voltage fault characteristic sub-pattern clustering standard, calculate and update the fault characteristic sub-pattern clustering center, and obtain the metering low voltage fault characteristic sub-pattern with cluster number K:

[0052] S3.1: Set the clustering algorithm and clustering parameters of the metering low voltage fault characteristic sub-pattern:

[0053] The present invention selects the K-Medoids clustering algorithm as the clustering algorithm of the fault feature sub-pattern, sets the number of clusters to K, randomly selects K data samples from the feature vector data samples of the alarm event record as the initial clustering centers, and records the kth clustering center as [Ok,1,Ok,2,…,Ok,j,…,Ok,56].

[0054] S3.2: Calculate the clustering criteria of the characteristic sub-patterns of metering low voltage faults:

[0055] The sum of squared errors is used to measure the distance between the qth alarm event record feature vector data sample and the kth cluster center as the clustering standard for the characteristic sub-pattern of the metering low-voltage fault. The calculation formula is as follows:

[0056]

[0057] In the formula, Φ k,q The sum of squared errors between the feature vector data sample and the kth cluster center is recorded for the qth alarm event.

[0058] After calculating the sum of squares of the errors between the qth alarm event record feature vector data sample and the K cluster centers, if the sum of squares of the errors between a certain alarm event record feature vector data sample and a certain cluster center is the smallest, then this sample is classified into the cluster. In this way, new K clusters are obtained.

[0059] S3.3: Calculate and update the fault feature sub-pattern clustering center:

[0060] In the kth cluster, each alarm event record feature vector data sample in the cluster is taken as the cluster center in turn, and the sum of square errors between other samples and the cluster center is calculated. The alarm event record feature vector data sample corresponding to the cluster center with the smallest sum of square errors is updated as the cluster center of the cluster.

[0061] S3.4: Obtain the metering low voltage fault characteristic sub-mode with cluster number K:

[0062] Repeat steps S3 and S4 until the cluster center remains unchanged, and the cluster center of the metering low voltage fault characteristic sub-pattern with cluster number K can be obtained. These K cluster centers are the metering low voltage fault characteristic sub-patterns of fault category Yi under cluster number K.

[0063] S4: Determine the optimal number of clusters of fault feature sub-patterns based on the elbow method and obtain the final fault feature sub-patterns:

[0064] like Figure 3 As shown, in order to ensure that the fault feature sub-pattern clustering results meet the uniqueness, the present invention proposes a method for determining the optimal number of clusters of fault feature sub-patterns based on the elbow method. The elbow method is a method of confirming the optimal K value using a relationship diagram between the sum of squared errors and the number of clusters K. In the elbow method, the first and last points of the curve of the relationship between the sum of squared errors and the K value are connected to obtain a straight line, which is recorded as y=ax+b. When the number of clusters is K, the difference between the sum of squared errors and the value of the straight line is recorded as LK. The K value corresponding to the maximum value of LK is the optimal number of clusters.

[0065] On this basis, the sum of squares of errors under different K values ​​is calculated, and LK under different clustering numbers K values ​​is calculated. The K value corresponding to the maximum LK is taken as the optimal clustering number. The metering low voltage fault characteristic subpattern under this clustering number is the metering low voltage fault characteristic subpattern of fault category Yi. The final metering low voltage fault characteristic subpattern of fault category Yi is recorded as [Es,i,1,Es,i,2,…,Es,i,j,…,Es,i,56], s∈[1,…,Ui], Ui is the optimal clustering number of fault category Yi.

[0066] S5: Fault diagnosis is performed based on fault characteristic sub-patterns and real-time alarm events to support the generation of metering low-voltage operation and maintenance work orders:

[0067] According to step S1 to step S4, fault characteristic sub-patterns of all fault types are obtained.

[0068] On the master station side of the metering automation system, taking the substation as a unit, when the master station collects alarm events and the abnormality still exists after data call detection, it continues to collect fault alarm events in the substation within W minutes, removes duplicate alarm records, and obtains the alarm event record feature vector [R1, R2, …, Ri, …, R56]. The error sum of squares between the alarm event record feature vector [R1, R2, …, Ri, …, R56] and the fault category Yi [Es,i,1,Es,i,2, …,Es,i,j, …,Es,i,56] is calculated, recorded as Vi,s, so that the alarm event record feature vector [R1, R2, …, Ri, …, R56] can be obtained. The sum of square errors between the event record feature vector [R1, R2, …, Ri, …, R56] and all fault feature sub-patterns of the fault category Yi is recorded as [Vi, 1, …, Vi, Ui], and the minimum value is taken as the minimum sum of square errors between the alarm event record feature vector [R1, R2, …, Ri, …, R56] and the fault feature pattern of the fault category Yi. The minimum sum of square errors between the alarm event record feature vector [R1, R2, …, Ri, …, R56] and the fault feature patterns of all fault categories are compared, and the minimum sum of square errors is recorded as the fault category of this fault. If the minimum sum of square errors of multiple fault categories are simultaneously the minimum, multiple fault categories are output.

[0069] After the metering automation system determines the type of fault, it generates a metering low-voltage operation and maintenance work order, thereby supporting on-site operation and maintenance personnel in precise operation and maintenance, improving the efficiency of operation and maintenance work, and reducing grassroots operation and maintenance work.

[0070] The method of the invention is based on the fault categories and alarm event records of the metering automation system operation and maintenance work order, and proposes a metering fault feature sub-pattern extraction and fault diagnosis method based on the K-Medoids algorithm and the elbow method, so as to realize the refined extraction and management of the fault feature sub-pattern. Based on this, fault diagnosis is carried out in combination with the fault feature sub-pattern and the real-time alarm event, so as to improve the accuracy of fault diagnosis on the metering low-voltage side and support the generation of metering low-voltage operation and maintenance work orders.

[0071] Example 2, reference Figure 4 , is an embodiment of the present invention, and provides a metering fault characteristic sub-pattern extraction and fault diagnosis system, comprising:

[0072] Sample acquisition module 100, alarm recording module 200, cluster analysis module 300, cluster optimization module 400 and fault diagnosis module 500;

[0073] The sample acquisition module 100 is used for the metering master station to acquire fault category sample data of the metering low-voltage operation and maintenance work order;

[0074] The alarm recording module 200 is used for the metering master station to obtain alarm event record characteristic sample data of the fault category;

[0075] The cluster analysis module 300 is used to set the metering low voltage fault characteristic sub-pattern clustering algorithm and clustering parameters based on the K-Medoids clustering algorithm, calculate the metering low voltage fault characteristic sub-pattern clustering standard, calculate and update the fault characteristic sub-pattern clustering center, and obtain the metering low voltage fault characteristic sub-pattern with a cluster number K;

[0076] The optimized clustering module 400 is used to determine the optimal clustering number of the fault characteristic sub-pattern based on the elbow method to obtain the final fault characteristic sub-pattern;

[0077] The fault diagnosis module 500 is used to perform fault diagnosis based on fault characteristic sub-patterns and real-time alarm events, and support the generation of metering low-voltage operation and maintenance work orders.

[0078] Example 3

[0079] An embodiment of the present invention is different from the first two embodiments in that:

[0080] If the 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, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.

[0081] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0082] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0083] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, a plurality of 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, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0084] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for extracting characteristic sub-patterns of metrology faults and diagnosing faults, characterized in that: include: The metering master station obtains fault category sample data of the metering low-voltage operation and maintenance work order; The metering master station obtains the characteristic sample data of the alarm event record of the fault category; Based on the K-Medoids clustering algorithm, set the metering low voltage fault characteristic sub-pattern clustering algorithm and clustering parameters, calculate the metering low voltage fault characteristic sub-pattern clustering standard, calculate and update the fault characteristic sub-pattern clustering center, and obtain the metering low voltage fault characteristic sub-pattern with cluster number K; Determine the optimal clustering number of fault feature sub-patterns based on the elbow method and obtain the final fault feature sub-patterns; Fault diagnosis is carried out based on fault characteristic sub-patterns and real-time alarm events, supporting the generation of metering low-voltage operation and maintenance work orders.

2. The method for extracting characteristic sub-patterns of metering faults and diagnosing faults according to claim 1, characterized in that: Sort the fault categories by fault category, and record fault category n as Y n .

3. The method for extracting characteristic sub-patterns of metering faults and diagnosing faults according to claim 2, characterized in that: The alarm events include metering equipment alarms, abnormal power consumption alarms, photovoltaic user alarms and other alarms; The alarm events are sorted according to the alarm event sequence, and each alarm event is sorted into the Xnth vector.

4. The method for extracting characteristic sub-patterns of metering faults and diagnosing faults according to claim 3, characterized in that: The clustering algorithm and clustering parameters of the characteristic sub-pattern of metering low-voltage fault are set as follows: Set the number of clusters to K, and randomly select K data samples from the alarm event record feature vector data samples as the initial cluster centers, and record the kth cluster center as [Ok,1,Ok,2,…,Ok,j,…,O 56 ,].

5. The method for extracting characteristic sub-patterns of metering faults and diagnosing faults according to claim 4, characterized in that: The calculation of the clustering standard of the characteristic sub-pattern of the metering low-voltage fault includes using the sum of squared errors to measure the distance between the qth alarm event record characteristic vector data sample and the kth cluster center as the clustering standard of the characteristic sub-pattern of the metering low-voltage fault, which is expressed as: Among them, Φ k,q Record the sum of squared errors between the feature vector data sample and the kth cluster center for the qth alarm event; After calculating the sum of squares of the errors between the qth alarm event record feature vector data sample and the K cluster centers, if the sum of squares of the errors between a certain alarm event record feature vector data sample and a certain cluster center is the smallest, the sample is classified into the cluster cluster.

6. The method for extracting characteristic sub-patterns of metering faults and diagnosing faults according to claim 5, characterized in that: The calculation and update of the fault feature sub-pattern cluster center is included in the kth cluster cluster, and each alarm event record feature vector data sample in the cluster cluster is taken as the cluster center in turn, and the sum of square errors between other samples and the cluster center is calculated. The alarm event record feature vector data sample corresponding to the cluster center with the smallest sum of square errors is updated as the cluster center of the cluster cluster.

7. The method for extracting characteristic sub-patterns of metering faults and diagnosing faults according to claim 6, characterized in that: The method of obtaining the final fault characteristic sub-mode includes connecting the beginning and the end of the error sum of squares and K value relationship curves by using the elbow method to obtain a straight line, which is recorded as y=ax+b; When the number of clusters is K, the difference between the error sum of squares and the value of the straight line is recorded as LK. The K value corresponding to the maximum LK is the optimal number of clusters. Calculate the sum of squares of errors under different K values, calculate LK under different clustering numbers K values, and the K value corresponding to the maximum LK is taken as the optimal clustering number. The metering low voltage fault characteristic subpattern under this clustering number is the metering low voltage fault characteristic subpattern of fault category Yi. The final metering low voltage fault characteristic subpattern of fault category Yi is recorded as [Es,i,1,Es,i,2,…,Es,i,j,…,Es,i,56], s∈[1,…,Ui], Ui is the optimal clustering number of fault category Yi.

8. A system using the method for extracting characteristic sub-patterns of metering faults and diagnosing faults as claimed in any one of claims 1 to 7, characterized in that: include: A sample acquisition module (100), an alarm recording module (200), a cluster analysis module (300), an optimized clustering module (400), and a fault diagnosis module (500); The sample acquisition module (100) is used by the metering master station to acquire fault category sample data of the metering low-voltage operation and maintenance work order; The alarm recording module (200) is used for the metering master station to obtain alarm event recording characteristic sample data of the fault category; The cluster analysis module (300) is used to set the metering low voltage fault characteristic sub-pattern clustering algorithm and clustering parameters based on the K-Medoids clustering algorithm, calculate the metering low voltage fault characteristic sub-pattern clustering standard, calculate and update the fault characteristic sub-pattern clustering center, and obtain the metering low voltage fault characteristic sub-pattern with a cluster number K; The optimized clustering module (400) is used to determine the optimal clustering number of the fault characteristic sub-pattern based on the elbow method to obtain the final fault characteristic sub-pattern; The fault diagnosis module (500) is used to carry out fault diagnosis based on fault characteristic sub-modes and real-time alarm events, and to support the generation of metering low-voltage operation and maintenance work orders.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method for extracting characteristic sub-patterns of metering faults and diagnosing faults 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 a processor, the steps of the method for extracting characteristic sub-patterns of metering faults and diagnosing faults according to any one of claims 1 to 7 are implemented.

Images