A power distribution equipment operation and maintenance knowledge base construction method and system

Abstract

This application discloses a method and system for constructing a knowledge base for the operation and maintenance of power distribution equipment, relating to the field of knowledge base construction. The method includes: constructing triples based on the operation and maintenance knowledge of each power distribution device; determining the failure rate of any power distribution device based on its fault text and corresponding solutions; calculating the similarity between each pair of power distribution devices; dividing all power distribution devices into several device groups based on the similarity between each pair and the failure rate of each device; and constructing a knowledge base for the operation and maintenance of power distribution equipment based on the triples and the corresponding device groups. This application aims to solve the problem that, in traditional power distribution equipment knowledge management models, independent knowledge bases cannot meet the comprehensive diagnostic needs of multi-device faults when dealing with cross-device or complex faults.

Description

Technical Field

[0001] This application relates to the field of knowledge base construction, and in particular to a method and system for constructing a knowledge base for the operation and maintenance of power distribution equipment. Background Technology

[0002] Currently, knowledge about the operation and maintenance of power distribution equipment is mainly stored and managed independently by equipment model. For example, for different types of equipment such as transformers, circuit breakers, and distribution cabinets, the operation and maintenance knowledge base usually records their typical fault phenomena, fault causes, and corresponding maintenance measures, and then manually organizes them into structured documents or database entries.

[0003] However, with the diversification of power distribution equipment types and the increasing complexity of operating environments, traditional knowledge management models, while meeting the maintenance needs of individual devices, rely on maintenance personnel to manually compare multiple independent knowledge bases when dealing with cross-device or complex faults, lacking a systematic knowledge association mechanism. On the one hand, much knowledge and principles are interconnected, but the traditional model results in significant duplication and redundancy due to knowledge being constructed independently for each device. On the other hand, a single fault phenomenon in a power distribution system often involves multiple devices, and traditional independent knowledge bases divided by device cannot meet the needs of comprehensive diagnosis. Summary of the Invention

[0004] The main purpose of this application is to provide a method and system for constructing a knowledge base for the operation and maintenance of power distribution equipment. It aims to solve the technical problems of traditional power distribution equipment knowledge management models, which rely on maintenance personnel to manually compare multiple independent knowledge bases when dealing with cross-equipment or complex faults. These problems include the lack of a systematic knowledge association mechanism, the existence of redundant and repetitive knowledge due to independent construction of knowledge bases for each equipment, and the inability of independent knowledge bases to meet the comprehensive diagnosis needs of multiple equipment faults.

[0005] To achieve the above objectives, this application provides a method for constructing a knowledge base for the operation and maintenance of power distribution equipment, comprising: constructing triples based on the operation and maintenance knowledge of each power distribution equipment, wherein the triples include the name of the power distribution equipment, fault text, and solutions corresponding to each fault; for any power distribution equipment, determining the fault rate of the power distribution equipment based on the fault text and solutions corresponding to each fault; calculating the similarity between each pair of power distribution equipment based on the fault text and solutions corresponding to each fault, wherein the similarity is used to characterize the similarity of fault texts and solutions between each pair of power distribution equipment; dividing all power distribution equipment into several equipment groups based on the similarity between each pair of power distribution equipment and the fault rate of each power distribution equipment, wherein each equipment group includes at least one type of power distribution equipment; and constructing a knowledge base for the operation and maintenance of power distribution equipment based on the triples and the equipment groups corresponding to each power distribution equipment.

[0006] Optionally, determining the failure rate of any power distribution equipment based on the fault text of the power distribution equipment and the solutions corresponding to each fault includes: determining multiple fault types of the power distribution equipment based on the fault text of the power distribution equipment, and obtaining the occurrence frequency of each fault type within a preset period; determining the matching weight of each fault type based on the solutions corresponding to each fault in the power distribution equipment; and determining the failure rate of the power distribution equipment based on the occurrence frequency of each fault type within a preset period and the matching weight of each fault type.

[0007] Optionally, the step of calculating the similarity between each pair of power distribution devices based on the fault text of each power distribution device and the corresponding solution for each fault includes: for the target power distribution device and the power distribution device of interest among the power distribution devices, vectorizing the fault text of the target power distribution device and the corresponding solution for each fault to obtain several target fault vectors and several target solution vectors; vectorizing the fault text of the power distribution device of interest and the corresponding solution for each fault to obtain several focus fault vectors and several focus solution vectors; and determining the similarity between the target power distribution device and the power distribution device of interest based on the several target fault vectors, several target solution vectors, several focus fault vectors, and several focus solution vectors.

[0008] Optionally, determining the similarity between the target power distribution equipment and the power distribution equipment of interest based on several target fault vectors, several target solution vectors, several fault vectors of concern, and several fault vectors of concern includes: calculating a first similarity between each target fault vector and each fault vector of concern, and summing all the first similarities to obtain the fault similarity between the target power distribution equipment and the power distribution equipment of concern; for any target fault vector and any fault vector of concern whose first similarity is greater than a preset threshold, calculating a second similarity between the target solution vector corresponding to the target fault vector and the fault vector of concern corresponding to the fault vector of concern, and summing all the second similarities to obtain the solution similarity between the target power distribution equipment and the power distribution equipment of concern; and determining the similarity between the target power distribution equipment and the power distribution equipment of concern based on the fault similarity and the solution similarity.

[0009] Optionally, the step of dividing all power distribution equipment into several equipment groups based on the similarity between pairs of power distribution equipment and the failure rate of each power distribution equipment includes: forming a pair of equipment by any two power distribution equipment; constructing a similarity matrix by dividing all equipment pairs and the similarity of each equipment pair in descending order of similarity; and dividing all power distribution equipment into several equipment groups based on the similarity matrix.

[0010] Optionally, the step of constructing a power distribution equipment operation and maintenance knowledge base based on the triplet and the equipment group corresponding to each power distribution equipment includes: establishing a node tree based on the equipment group corresponding to each power distribution equipment, wherein the node tree includes several nodes, the connection relationship between nodes, and several branches of each node, wherein one node corresponds to one power distribution equipment, the connection relationship is used to characterize the similarity, and the several branches include the fault text of each power distribution equipment in the triplet and the solution corresponding to each fault; and constructing a power distribution equipment operation and maintenance database based on the node tree.

[0011] Optionally, determining the matching weight of each fault type based on the solution corresponding to each fault in the power distribution equipment includes: classifying all fault types into three types based on the solution corresponding to each fault in the power distribution equipment, the three types including a first type, a second type, and a third type; and determining the matching weight of each fault type based on the matching weight corresponding to each type.

[0012] Furthermore, to achieve the above objectives, this application also provides a power distribution equipment operation and maintenance knowledge base construction system, comprising: a triplet construction module, used to construct triplets based on the operation and maintenance knowledge of each power distribution equipment, wherein the triplet includes the name of the power distribution equipment, fault text, and solutions corresponding to each fault; a single power distribution equipment failure rate determination module, used to determine the failure rate of any power distribution equipment based on the fault text of the power distribution equipment and solutions corresponding to each fault; a similarity determination module, used to calculate the similarity between pairs of power distribution equipment based on the fault text of each power distribution equipment and solutions corresponding to each fault, wherein the similarity is used to characterize the similarity of fault texts and solutions between pairs of power distribution equipment; a power distribution equipment grouping module, used to divide all power distribution equipment into several equipment groups based on the similarity between pairs of power distribution equipment and the failure rate of each power distribution equipment, wherein each equipment group includes at least one power distribution equipment; and a knowledge base construction module, used to construct a power distribution equipment operation and maintenance knowledge base based on the triplets and the equipment groups corresponding to each power distribution equipment.

[0013] This application also provides a device for constructing a knowledge base for the operation and maintenance of power distribution equipment, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the methods in any of the above possible implementations.

[0014] This application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the method in any of the possible implementations described above.

[0015] This application proposes a method and system for constructing a knowledge base for the operation and maintenance of power distribution equipment. First, it achieves structured processing of operation and maintenance knowledge by constructing triples containing the name of the power distribution equipment, fault text, and solution. Then, using this triple information, it calculates the similarity between power distribution equipment and groups the equipment accordingly. Each group of equipment exhibits high consistency in fault characteristics and solutions. Based on this, a node tree is further constructed to intuitively display the similarity relationships and fault solutions between equipment. This greatly facilitates the rapid location and diagnosis of cross-equipment or complex faults by operation and maintenance personnel, solving the problems of knowledge redundancy and insufficient comprehensive diagnostic capabilities in traditional knowledge management models, and significantly improving operation and maintenance efficiency. Attached Figure Description

[0016] Figure 1 One of the flowcharts for the method of constructing a knowledge base for operation and maintenance of power distribution equipment provided in the embodiments of this application;

[0017] Figure 2 Flowchart 2 of the method for constructing a knowledge base for operation and maintenance of power distribution equipment provided in the embodiments of this application;

[0018] Figure 3 The third flowchart of the method for constructing a knowledge base for operation and maintenance of power distribution equipment provided in the embodiments of this application;

[0019] Figure 4 The fourth flowchart of the method for constructing a knowledge base for operation and maintenance of power distribution equipment provided in the embodiments of this application;

[0020] Figure 5 A structural block diagram of a power distribution equipment operation and maintenance knowledge base construction system provided in an embodiment of this application;

[0021] Figure 6 This is a schematic diagram of the structure of a power distribution equipment operation and maintenance knowledge base construction device provided in an embodiment of this application.

[0022] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0023] It should be understood that the specific embodiments described herein are merely illustrative of this application and are not intended to limit this application.

[0024] Currently, knowledge about the operation and maintenance of power distribution equipment is mainly stored and managed independently by equipment model. For example, for different types of equipment such as transformers, circuit breakers, and distribution cabinets, the operation and maintenance knowledge base usually records their typical fault phenomena, fault causes, and corresponding maintenance measures, and then manually organizes them into structured documents or database entries.

[0025] However, with the diversification of power distribution equipment types and the increasing complexity of operating environments, traditional knowledge management models, while meeting the maintenance needs of individual devices, rely on maintenance personnel to manually compare multiple independent knowledge bases when dealing with cross-device or complex faults, lacking a systematic knowledge association mechanism. On the one hand, many knowledge points and principles are interconnected, but the traditional model results in significant duplication and redundancy due to the independent construction of knowledge by device. For example, the phenomenon of "poor contact causing overheating" and the basic handling procedures (inspection, tightening, applying conductive paste) could be applied to various scenarios such as circuit breaker joints, disconnector switches, and busbar connections. However, the traditional method records maintenance and repair content separately for each device, increasing the workload of knowledge base construction and leading to redundant knowledge storage, requiring repetitive operations during subsequent maintenance and updates. On the other hand, a single fault phenomenon in a power distribution system often involves multiple devices, and traditional independent knowledge bases divided by device cannot meet the needs of comprehensive diagnosis. For example, maintenance personnel input "a certain line tripped" into the knowledge base, but the cause of the fault may be the circuit breaker itself, the transformer it protects may be faulty, or the cable on the line may be broken. Under the traditional knowledge management model, maintenance personnel need to consult the independent knowledge bases of multiple devices separately, which makes it difficult to quickly integrate information for comprehensive judgment. This not only prolongs the fault diagnosis time, but may also cause key causes to be missed due to fragmented information, resulting in low fault handling efficiency.

[0026] To address the aforementioned issues, this application provides a method and system for constructing a knowledge base for the operation and maintenance of power distribution equipment. The following is a detailed description of the solution proposed in this application.

[0027] Figure 1 This is one of the flowcharts for a method of constructing a power distribution equipment operation and maintenance knowledge base according to an embodiment of this application. This method can be executed by a knowledge base construction device, which can specifically be a power distribution equipment operation and maintenance knowledge base construction device. (Refer to...) Figure 1 The method for constructing a knowledge base for the operation and maintenance of power distribution equipment may include the following steps:

[0028] S11. Construct a ternary set based on the operation and maintenance knowledge of each power distribution equipment. The ternary set includes the name of the power distribution equipment, the fault text, and the corresponding solution for each fault.

[0029] The fault text refers to the textual content that accurately describes information related to power distribution equipment faults, including the equipment status at the time of the fault, specific abnormal phenomena, key parameters (such as temperature, voltage, and operating time), and the fault occurrence scenario (such as after circuit breaker closing, during peak load periods, etc.). The fault text can be used to clearly define "what problem occurred with the equipment and under what circumstances." The corresponding solution for each fault refers to an operational and standardized handling process developed for the specific problem described in the fault text. This includes safety preparations before fault handling (such as power outage measures), fault diagnosis during fault handling, problem repair during fault handling, selection of tools and materials, and verification of the effectiveness of fault repair. It can be understood that the solution can be used to ensure that maintenance personnel handle faults efficiently according to the procedures.

[0030] For example, the name of the power distribution equipment could be "10kV vacuum circuit breaker," and the fault text could be: "After the 10kV vacuum circuit breaker was closed and operated under load for about 25 minutes, the infrared thermometer at the connection point showed a temperature of 82℃ (ambient temperature 28℃), exceeding the normal operating threshold of 65℃ specified in GB50150-2016. A slight 'buzzing' discharge sound could be heard at the connection point when approached, and the load current stabilized at 400A (rated current 630A)." The corresponding solution for this fault could be: "Safety preparation: First, disconnect the upstream power switch of the circuit breaker, hang a warning sign saying 'Do not close, people are working,' and after confirming with a voltage tester that there is no voltage on both sides of the circuit breaker, install a grounding wire; Fault investigation: Remove the insulating sleeve at the connection point, check the tightness of the connection bolts with a wrench, and find that the torque of the A-phase connection bolts is insufficient (standard torque 35N)." m, measured 20N The joint contact surface has a small amount of oxide layer; Repair: Sand the oxide layer on the joint contact surface with fine sandpaper until the metallic luster is exposed, apply conductive paste (model: Electrical Composite Grease 89D), and apply a standard torque of 35N. m. Tighten the bolts again; Verification of effect: Restore the insulation sleeve, remove the grounding wire and warning sign, close the upstream power switch, and run the circuit breaker under load for 30 minutes. The infrared thermometer shows that the joint temperature drops to 42℃ (ambient temperature 28℃), and there is no discharge sound. The fault is confirmed to be eliminated.

[0031] In the specific implementation process, the first step is to collect the operation and maintenance (O&M) documents of all power distribution equipment within the power distribution system. These O&M documents may include equipment fault handling reports, manufacturer technical manuals, O&M management system logs, and maintenance records. Then, the O&M documents are classified according to the type of power distribution equipment (e.g., transformers, circuit breakers, disconnect switches, etc.) to ensure centralized management of relevant documents for all types of power distribution equipment.

[0032] Furthermore, for paper documents in the operation and maintenance documentation, they are first converted into electronic documents by scanning, and then the text content is recognized using OCR (Optical Character Recognition) technology, combined with manual verification to correct errors. For electronic documents in the operation and maintenance documentation, NLP (Natural Language Processing) technology or keyword search function is used to extract the "distribution equipment name," "fault text," and "corresponding solution" for the fault. It should be noted that the "distribution equipment name" must clearly specify the equipment model, specifications, and system to which it belongs; the "fault text" must include the fault phenomenon, the scenario in which it occurred, and key parameters (such as temperature, voltage, and operating time); and the "corresponding solution" must include safety preparations, fault diagnosis, problem repair, tools and materials, and effectiveness verification.

[0033] Furthermore, the extracted three types of information ("distribution equipment name", "fault text", and "fault-related solutions") are standardized. Specifically, duplicate or invalid document information (such as multiple records of the same fault) can be deleted, and terminology can be standardized (such as unifying "connector overheating" and "terminal temperature too high" as "connector overheating"), and the step logic of the solution can be standardized (arranged in the order of "safety preparation, fault investigation, problem repair, and effect verification") to ensure that the information format of the same type of distribution equipment is consistent and the semantics are unified, avoiding the deviation of triple association due to differences in expression.

[0034] Finally, the cleaned "power distribution equipment name", "fault text" and "fault corresponding solution" are matched one by one to form a complete triplet. The constructed triplet is then entered into the database to establish the association between power distribution equipment, faults and solutions.

[0035] S12. For any power distribution equipment, determine the failure rate of the power distribution equipment based on the fault text of the power distribution equipment and the solutions corresponding to each fault.

[0036] The failure rate is used to characterize the scale of failures in power distribution equipment. Specifically, it can characterize the frequency of failures and the adequacy of corresponding solutions. For example, if a power distribution device experiences a high number of different types of failures and a high total frequency of failures within a specific period (e.g., one year or one maintenance cycle), and a high proportion of these failures lack corresponding solutions or have incomplete solutions (e.g., missing steps or unclear tools), then the failure rate of that power distribution device is high. Conversely, if the power distribution device experiences fewer types of failures and a lower frequency of failures, and all failures have complete and operable solutions, then the failure rate of that power distribution device is low.

[0037] In one embodiment, reference is made to... Figure 2 , Figure 2 The second flowchart of the method for constructing a knowledge base for operation and maintenance of power distribution equipment provided in this application embodiment shows that, in step S12, for any power distribution equipment, determining the failure rate of the power distribution equipment based on the fault text of the power distribution equipment and the solutions corresponding to each fault may specifically include:

[0038] S121. Based on the fault text of the power distribution equipment, determine multiple fault types of the power distribution equipment and obtain the occurrence frequency of each fault type within a preset period;

[0039] S122. Determine the matching weight of each fault type based on the solutions corresponding to each fault in the power distribution equipment;

[0040] S123. Determine the failure rate of the power distribution equipment based on the frequency of occurrence of each fault type within a preset period and the matching weight of each fault type.

[0041] In this embodiment, a preset period is set as an operation and maintenance period. In other embodiments, the preset period can also be set as a specific time, such as 1 year. This embodiment does not impose specific restrictions on the setting of the preset period.

[0042] In the specific implementation process, any power distribution equipment is selected from all power distribution equipment as the target power distribution equipment. This embodiment takes the target power distribution equipment as an example for explanation. All fault texts in the target power distribution equipment are classified according to fault type. For example, "low oil level", "excessive core grounding current", and "leakage of high voltage side bushing" of transformer are different fault types. The frequency of occurrence of each fault type in one maintenance cycle is obtained (for example, "low oil level" occurs 3 times in one maintenance cycle and "excessive core grounding current" occurs 2 times in one maintenance cycle).

[0043] It should be noted that the total frequency of faults of the target power distribution equipment (i.e., the sum of the frequencies of all fault types) and the total number of fault types of the target power distribution equipment can be determined based on the frequency of occurrence of each fault type within a maintenance cycle; and the total frequency of faults and the total number of fault types can reflect the frequency of faults occurring in the target power distribution equipment.

[0044] Furthermore, the solutions corresponding to each fault type in the triplet are checked in turn. Based on the solution, all fault types are divided into three types, namely the first type, the second type and the third type. The matching weight of the first type is set to a, the matching weight of the second type is set to b, and the matching weight of the third type is set to c. In this embodiment, a is set to 1.8, b to 1.3 and c to 1.

[0045] Specifically, any fault type is selected as the target fault type from all fault types. If the target fault type has no solution in the triplet (e.g., the fault type is aging insulation of a new type of bushing, and no treatment steps are recorded in the triplet), the target fault type is marked as Type 1. If the target fault type has a missing solution in the triplet (e.g., the triplet only records the solution "inspect the bushing," without including safety preparation, tool and material selection, and effect verification), the target fault type is marked as Type 2. If the target fault type has a complete solution in the triplet, the target fault type is marked as Type 3. It can be understood that Type 1 can represent faults with no solution, Type 2 can represent faults with incomplete solutions, and Type 3 can represent faults with complete solutions. The matching weight of each fault type reflects whether the corresponding solution is sufficient for each fault.

[0046] Furthermore, the frequency of occurrence of the target fault type within a preset period is multiplied by the matching weight of the target fault type to obtain the fault severity of the target fault type. The fault severity of all fault types in the target power distribution equipment is summed to obtain the fault rate of the target power distribution equipment.

[0047] S13. Calculate the similarity between each pair of power distribution equipment based on the fault text of each power distribution equipment and the corresponding solution for each fault. The similarity is used to characterize the similarity of the fault text and the similarity of the solution between each pair of power distribution equipment.

[0048] In one embodiment, reference is made to... Figure 3 , Figure 3 The third flowchart of the method for constructing a knowledge base for operation and maintenance of power distribution equipment provided in this application embodiment includes, in step S13, calculating the similarity between pairs of power distribution equipment based on the fault text of each power distribution equipment and the corresponding solutions for each fault.

[0049] S131. For the target power distribution equipment and the power distribution equipment of interest in each power distribution equipment, the fault text of the target power distribution equipment and the corresponding solution of each fault are vectorized to obtain several target fault vectors and several target solution vectors. The fault text of the power distribution equipment of interest and the corresponding solution of each fault are vectorized to obtain several fault vectors of interest and several solution vectors of interest.

[0050] S132. Determine the similarity between target power distribution equipment and power distribution equipment of interest based on several target fault vectors, several target solution vectors, several fault vectors of concern, and several solution vectors of concern.

[0051] Among them, any power distribution device is selected from all power distribution devices as the target power distribution device, and any power distribution device other than the target power distribution device is selected from all power distribution devices as the power distribution device of interest.

[0052] In the specific implementation process, for all power distribution equipment, including target power distribution equipment and power distribution equipment of interest, the word2vec model is used to vectorize the fault text of the target power distribution equipment and the corresponding solutions for each fault, resulting in several target fault vectors and several target solution vectors. Similarly, the word2vec model is used to vectorize the fault text of the power distribution equipment of interest and the corresponding solutions for each fault, resulting in several fault vectors and several solution vectors of interest.

[0053] Furthermore, the first similarity between each target fault vector and each fault vector of interest is calculated using cosine similarity, and all first similarities are summed to obtain the fault similarity between the target power distribution equipment and the power distribution equipment of interest.

[0054] It should be noted that in practical applications, implementers can first divide the fault text and its corresponding solutions into several descriptive terms based on the semantics of the text. Taking any fault text as an example, the fault text can first be divided into several descriptive terms, then each descriptive term can be vectorized, and finally, the multiple vectorized single vectors can be combined to form the fault vector corresponding to the fault text. When calculating the first similarity, each single vector can be multiplied by a preset coefficient. The preset coefficient ranges from 0 to 1, and the preset coefficient of each single vector can be set by the implementer. For example, if a fault text includes the descriptive terms "joint overheating" and "of", the preset coefficient of the single vector corresponding to "joint overheating" is greater than the preset coefficient of the single vector corresponding to "of". It can be understood that multiplying the single vector by the preset coefficient can effectively improve the accuracy of fault similarity calculation.

[0055] Furthermore, for any target fault vector and any concerned fault vector with a first similarity greater than a preset threshold, the second similarity between the target solution vector corresponding to the target fault vector and the concerned solution vector corresponding to the concerned fault vector is calculated using cosine similarity, and all second similarities are added together to obtain the solution similarity between the target power distribution equipment and the concerned power distribution equipment.

[0056] Furthermore, the similarity between the fault and the solution is weighted and summed to obtain the degree of similarity between the target power distribution equipment and the power distribution equipment of interest.

[0057] It should be noted that in other embodiments, other similarity calculation methods can be used to replace the cosine similarity in this embodiment. This embodiment does not impose specific limitations on the similarity calculation method.

[0058] S14. Based on the similarity between pairs of power distribution equipment and the failure rate of each power distribution equipment, all power distribution equipment is divided into several equipment groups, and each equipment group includes at least one power distribution equipment.

[0059] If a device group contains multiple power distribution devices, then the similarity between these multiple power distribution devices is greater than the preset merging threshold.

[0060] In one embodiment, reference is made to... Figure 4 , Figure 4 The fourth flowchart of the method for constructing a knowledge base for operation and maintenance of power distribution equipment provided in this application embodiment shows that, in step S14, all power distribution equipment is divided into several equipment groups based on the similarity between pairs of power distribution equipment and the failure rate of each power distribution equipment. Specifically, this may include:

[0061] S141. Form a pair of any two power distribution devices. Based on the order of similarity from large to small, construct a similarity matrix for all the pairs of devices and the similarity of each pair of devices.

[0062] S142. Based on the similarity matrix, all power distribution equipment is divided into several equipment groups.

[0063] In the specific implementation process, firstly, two power distribution devices are paired up to form a device pair until all power distribution devices are divided. Then, based on the order of similarity from largest to smallest, a similarity matrix is ​​constructed for all device pairs and the corresponding similarity of each device pair. It can be understood that the first row of the similarity matrix includes the maximum similarity value and the device pair with the largest similarity. The first column of the similarity matrix is ​​the device pair, and the second column is the similarity of the device pair. The similarity in the similarity matrix decreases as the number of rows increases. The first row and first column of the similarity matrix is ​​the device pair with the largest similarity, and the second column of the first row is the maximum similarity value.

[0064] Furthermore, in the first row of the similarity matrix, the power distribution equipment with a higher failure rate is selected as the starting center. According to the order of the similarity matrix from top to bottom, power distribution equipment with a similarity greater than the merging threshold is selected and merged with the starting center into a group until a preset number is reached to obtain the first equipment group.

[0065] For example, the similarity can be sorted in descending order, and the top 30% of similarity in this sequence can be used as a sub-similar sequence. The first similarity after the sub-similar sequence can be set as the merging threshold, and the preset number can be set to 10% of the total number of power distribution devices. Of course, in other embodiments, the merging threshold and the preset number can also be set by the implementer.

[0066] Furthermore, referring to the construction process of the first equipment group, a new starting center is selected first according to the order of the similarity matrix from top to bottom, and then according to the order of the failure rate from small to large, to obtain the second equipment group. This process is repeated until all power distribution equipment is divided, that is, each power distribution equipment has a corresponding equipment group.

[0067] It should be noted that the similarity levels can be sorted in descending order. The top 60% of similarity levels in this sequence are considered the second sub-similar sequence, and the first similarity level after the second sub-similar sequence is set as the merging threshold for the second equipment group. The preset number can still be set to 10% of the total number of power distribution devices. It is understood that as the number of equipment groups increases, the merging threshold gradually decreases. This embodiment will not elaborate further on the merging thresholds for other equipment groups. The number of power distribution devices in the last equipment group may be less than the preset number; that is, the stopping condition for equipment group division in this embodiment is that all power distribution devices have been divided.

[0068] S15. Construct a knowledge base for the operation and maintenance of power distribution equipment based on the ternary group and the corresponding equipment group of each power distribution equipment.

[0069] In the specific implementation process, a node tree is first established based on the equipment group corresponding to each power distribution equipment. The node tree includes several nodes, the connection relationship between nodes, and several branches of each node. Among them, one node corresponds to one power distribution equipment, the connection relationship can characterize the similarity between power distribution equipment, and the several branches of each node include the fault text of each power distribution equipment in the triplet and the solution corresponding to each fault.

[0070] Furthermore, a power distribution equipment operation and maintenance database is constructed based on this node tree.

[0071] The following example illustrates the power distribution equipment operation and maintenance database in this embodiment.

[0072] For example, when maintenance personnel input "poor contact causing overheating" into the power distribution equipment maintenance database, the database first preprocesses the input query text "poor contact causing overheating", that is, it breaks down the text into the core keywords "poor contact, overheating", and then calls the word2vec word vector model trained earlier to convert each keyword into a semantic vector.

[0073] Then, the database traverses all branches of the node tree and extracts the fault text vectors under each node branch one by one (for example, the fault text of node A is "the joint temperature is 82°C after 25 minutes of closing and carrying the load, accompanied by a slight discharge sound", and the fault text of node B is "the contact temperature is 68°C after opening, there is no discharge but the contact gap of the finger exceeds the standard"). The similarity between each semantic vector and each fault text vector is calculated by cosine similarity. Fault texts and corresponding nodes with similarity greater than the preset semantic matching threshold are selected and these fault texts are used as matching fault texts.

[0074] Secondly, the database locates and matches the device nodes and device groups to which the fault texts belong. For example, the matching degree of 2 fault texts of node A (vacuum circuit breaker), 1 fault text of node D (SF6 circuit breaker), and 1 fault text of node B (disconnecting switch) is greater than the preset semantic matching threshold, and these nodes all belong to "first device group (10kV switch type)", that is, the number of matching fault texts is the largest in the first device group.

[0075] Finally, the database extracts the solution branches corresponding to the matching fault text. That is, based on the triplet association relationship, it retrieves the solutions corresponding to the matching fault text in nodes A, D, and B respectively, and feeds back the search results to the maintenance personnel according to the structure of "device group, device node, fault text, solution". For example, the left side of the display interface can display the node tree topology diagram of the first device group (marking the similarity between nodes), and the right column can display the matching fault text of each node (highlighted in red with related expressions such as "overheating" and "poor contact") and the corresponding solutions. In addition, in this embodiment, the database can also provide cross-device solution comparison function (for example, comparing the solutions of nodes A, D, and B, highlighting common steps such as "polishing oxide layer, applying conductive paste, and torque tightening"), thereby helping maintenance personnel to quickly obtain cross-device fault handling experience and solving the problem of traditional knowledge bases requiring manual comparison of documents from multiple devices.

[0076] Based on the above embodiments, Figure 5 Here is a structural block diagram of a power distribution equipment operation and maintenance knowledge base construction system according to one embodiment of this application, such as... Figure 5 As shown, the power distribution equipment operation and maintenance knowledge base construction system 200 may include: a ternary group construction module 210, a single power distribution equipment failure rate determination module 220, a similarity determination module 230, a power distribution equipment grouping module 240, and a knowledge base construction module 250.

[0077] Among them, the ternary group construction module 210 is used to construct ternary groups based on the operation and maintenance knowledge of each power distribution equipment. The ternary group includes the name of the power distribution equipment, the fault text, and the solution corresponding to each fault.

[0078] The single power distribution equipment failure rate determination module 220 is used to determine the failure rate of any power distribution equipment based on the fault text of the power distribution equipment and the solutions corresponding to each fault.

[0079] The similarity determination module 230 is used to calculate the similarity between each pair of power distribution equipment based on the fault text of each power distribution equipment and the solution corresponding to each fault. The similarity is used to characterize the similarity of the fault text and the similarity of the solution between each pair of power distribution equipment.

[0080] The power distribution equipment grouping module 240 is used to divide all power distribution equipment into several equipment groups based on the similarity between pairs of power distribution equipment and the failure rate of each power distribution equipment, and each equipment group includes at least one power distribution equipment.

[0081] The knowledge base construction module 250 is used to construct a knowledge base for the operation and maintenance of power distribution equipment based on the triplet and the equipment group corresponding to each power distribution equipment.

[0082] In an exemplary embodiment, the single power distribution equipment failure rate determination module 220 is further configured to, for any power distribution equipment, determine multiple failure types of the power distribution equipment based on the failure text of the power distribution equipment, and obtain the occurrence frequency of each failure type within a preset period; determine the matching weight of each failure type based on the solution corresponding to each failure in the power distribution equipment; and determine the failure rate of the power distribution equipment based on the occurrence frequency of each failure type within a preset period and the matching weight of each failure type.

[0083] In an exemplary embodiment, the similarity determination module 230 can also be used to vectorize the fault text of the target power distribution equipment and the corresponding solutions for each fault for the target power distribution equipment and the power distribution equipment of interest in each power distribution equipment, to obtain a number of target fault vectors and a number of target solution vectors; and to vectorize the fault text of the power distribution equipment of interest and the corresponding solutions for each fault for each fault, to obtain a number of attention fault vectors and a number of attention solution vectors; and to determine the similarity between the target power distribution equipment and the power distribution equipment of interest based on the number of target fault vectors, the number of target solution vectors, the number of attention fault vectors and the number of attention solution vectors.

[0084] In an exemplary embodiment, the similarity determination module 230 can also be used to calculate a first similarity between each target fault vector and each fault vector of interest, and add all the first similarities to obtain the fault similarity between the target power distribution equipment and the power distribution equipment of interest; for any target fault vector and any fault vector of interest whose first similarity is greater than a preset threshold, calculate a second similarity between the target solution vector corresponding to the target fault vector and the solution vector of interest corresponding to the fault vector of interest, and add all the second similarities to obtain the solution similarity between the target power distribution equipment and the power distribution equipment of interest; and determine the similarity between the target power distribution equipment and the power distribution equipment of interest based on the fault similarity and the solution similarity.

[0085] In an exemplary embodiment, the power distribution equipment grouping module 240 can also be used to group any two power distribution equipment into a pair, and construct a similarity matrix by forming all the pairs of equipment and the similarity of each pair of equipment based on the order of similarity from large to small; and divide all the power distribution equipment into several equipment groups based on the similarity matrix.

[0086] In an exemplary embodiment, the knowledge base construction module 250 can also be used to build a node tree based on the equipment group corresponding to each power distribution equipment. The node tree includes several nodes, the connection relationship between nodes, and several branches of each node. Each node corresponds to one power distribution equipment, the connection relationship is used to characterize the degree of similarity, and the several branches include the fault text of each power distribution equipment in the triplet and the solution corresponding to each fault. Based on the node tree, a power distribution equipment operation and maintenance database is constructed.

[0087] In an exemplary embodiment, the single power distribution equipment failure rate determination module 220 can also be used to classify all failure types into three types based on the solutions corresponding to each failure in the power distribution equipment, including a first type, a second type, and a third type; and to determine the matching weight of each failure type based on the matching weight corresponding to each type.

[0088] Those skilled in the art should understand that the division of the various modules in the embodiments is merely a logical functional division. In actual applications, they can be fully or partially integrated onto one or more actual carriers. These modules can be implemented entirely in software through processing unit calls, entirely in hardware, or a combination of software and hardware. It should be noted that each module in the power distribution equipment operation and maintenance knowledge base construction system in this embodiment corresponds one-to-one with each step in the power distribution equipment operation and maintenance knowledge base construction method in the aforementioned embodiments. Therefore, the specific implementation of this embodiment can refer to the implementation of the aforementioned power distribution equipment operation and maintenance knowledge base construction method, which will not be repeated here.

[0089] Based on the above embodiments, Figure 6This is a schematic diagram of the structure of a power distribution equipment operation and maintenance knowledge base according to one embodiment of this application, such as... Figure 6 As shown, the electronic device may include: a processor 310, a communication interface 320, a memory 330, and a communication bus 340, wherein the processor 310, the communication interface 320, and the memory 330 communicate with each other through the communication bus 340. The processor 310 can call logical instructions in the memory 330 to execute a method for constructing a power distribution equipment operation and maintenance knowledge base. This method includes: constructing triples based on the operation and maintenance knowledge of each power distribution equipment, where each triple includes the power distribution equipment name, fault text, and corresponding solutions for each fault; for any given power distribution equipment, determining its failure rate based on the fault text and corresponding solutions; calculating the similarity between each pair of power distribution equipment based on the fault text and corresponding solutions, where the similarity represents the similarity of fault texts and solutions between each pair of power distribution equipment; dividing all power distribution equipment into several equipment groups based on the similarity between each pair and the failure rate of each power distribution equipment, where each equipment group includes at least one power distribution equipment; and constructing a power distribution equipment operation and maintenance knowledge base based on the triples and the corresponding equipment groups.

[0090] Furthermore, the logical instructions in the aforementioned memory 330 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the 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 to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0091] Based on the above embodiments, in another aspect, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute a method for constructing a power distribution equipment operation and maintenance knowledge base provided by the above methods. The method includes: constructing triples based on the operation and maintenance knowledge of each power distribution equipment, wherein the triples include the name of the power distribution equipment, fault text, and solutions corresponding to each fault; for any power distribution equipment, determining the fault rate of the power distribution equipment based on the fault text of the power distribution equipment and the solutions corresponding to each fault; calculating the similarity between each pair of power distribution equipment based on the fault text of each power distribution equipment and the solutions corresponding to each fault, wherein the similarity is used to characterize the similarity of fault texts and solutions between each pair of power distribution equipment; dividing all power distribution equipment into several equipment groups based on the similarity between each pair of power distribution equipment and the fault rate of each power distribution equipment, wherein each equipment group includes at least one power distribution equipment; and constructing a power distribution equipment operation and maintenance knowledge base based on the triples and the equipment groups corresponding to each power distribution equipment.

[0092] Based on the above embodiments, in another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements a method for constructing a power distribution equipment operation and maintenance knowledge base provided by the methods described above. This method includes: constructing triples based on the operation and maintenance knowledge of each power distribution equipment, the triples including the name of the power distribution equipment, fault text, and solutions corresponding to each fault; for any power distribution equipment, determining the fault rate of the power distribution equipment based on the fault text and solutions corresponding to each fault; calculating the similarity between pairs of power distribution equipment based on the fault text and solutions corresponding to each fault, the similarity being used to characterize the similarity of fault texts and solutions between pairs of power distribution equipment; dividing all power distribution equipment into several equipment groups based on the similarity between pairs of power distribution equipment and the fault rate of each power distribution equipment, each equipment group including at least one power distribution equipment; and constructing a power distribution equipment operation and maintenance knowledge base based on the triples and the equipment groups corresponding to each power distribution equipment.

[0093] The above are merely preferred embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A method for constructing a power distribution equipment operation and maintenance knowledge base, characterized in that, The method includes: Based on the operation and maintenance knowledge of each power distribution equipment, a ternary set is constructed. The ternary set includes the name of the power distribution equipment, the fault text, and the corresponding solution for each fault. For any power distribution equipment, the failure rate of the power distribution equipment is determined based on the fault text of the power distribution equipment and the solutions corresponding to each fault. The similarity between each pair of power distribution devices is calculated based on the fault text of each power distribution device and the corresponding solution for each fault. The similarity is used to characterize the similarity of the fault text and the similarity of the solution between each pair of power distribution devices. Based on the similarity between pairs of power distribution equipment and the failure rate of each power distribution equipment, all power distribution equipment is divided into several equipment groups, and each equipment group includes at least one power distribution equipment. A knowledge base for the operation, maintenance and repair of power distribution equipment is constructed based on the aforementioned ternary set and the corresponding equipment group of each power distribution equipment. Specifically, determining the failure rate of any power distribution equipment based on its fault text and corresponding solutions includes: For any power distribution equipment, multiple fault types of the power distribution equipment are determined based on the fault text of the power distribution equipment, and the occurrence frequency of each fault type within a preset period is obtained; The matching weight of each fault type is determined based on the solutions corresponding to each fault in the power distribution equipment. The failure rate of the power distribution equipment is determined based on the frequency of occurrence of each fault type within a preset period and the matching weight of each fault type.

2. The power distribution equipment operation and maintenance knowledge base construction method of claim 1, wherein The similarity between each pair of power distribution devices is calculated based on the fault text of each power distribution device and the corresponding solution for each fault, including: For the target power distribution equipment and the power distribution equipment of interest among the power distribution equipment, the fault text of the target power distribution equipment and the corresponding solution of each fault are vectorized to obtain several target fault vectors and several target solution vectors. The fault text of the power distribution equipment of interest and the corresponding solution of each fault are vectorized to obtain several fault vectors of interest and several solution vectors of interest. The similarity between the target power distribution equipment and the power distribution equipment of interest is determined based on several target fault vectors, several target solution vectors, several fault vectors of concern, and several solution vectors of concern.

3. The power distribution equipment operation and maintenance knowledge base construction method of claim 2, wherein, The determination of the similarity between the target power distribution equipment and the power distribution equipment of interest based on several target fault vectors, several target solution vectors, several fault vectors of concern, and several solution vectors of concern includes: Calculate the first similarity between each target fault vector and each fault vector of interest, and sum all the first similarities to obtain the fault similarity between the target power distribution equipment and the power distribution equipment of interest; For any target fault vector and any concerned fault vector with a first similarity greater than a preset threshold, calculate the second similarity between the target solution vector corresponding to the target fault vector and the concerned solution vector corresponding to the concerned fault vector, and add all the second similarities to obtain the solution similarity between the target power distribution equipment and the concerned power distribution equipment; The degree of similarity between the target power distribution equipment and the power distribution equipment of interest is determined based on the fault similarity and the solution similarity.

4. The power distribution equipment operation and maintenance knowledge base construction method of claim 1, wherein, Based on the similarity between pairs of power distribution devices and the failure rate of each power distribution device, all power distribution devices are divided into several equipment groups, including: Any two power distribution devices are grouped into a device pair. Based on the order of similarity from largest to smallest, a similarity matrix is ​​constructed for all device pairs and the similarity of each device pair. Based on the similarity matrix, all power distribution equipment is divided into several equipment groups.

5. The power distribution equipment operation and maintenance knowledge base construction method of claim 1, wherein, The construction of a power distribution equipment operation and maintenance knowledge base based on the three-element combination and the corresponding equipment group of each power distribution device includes: A node tree is established based on the equipment group corresponding to each power distribution equipment. The node tree includes several nodes, the connection relationship between nodes, and several branches of each node. Each node corresponds to one power distribution equipment. The connection relationship is used to characterize the similarity. The several branches include the fault text of each power distribution equipment in the triplet and the solution corresponding to each fault. A power distribution equipment operation and maintenance database is constructed based on the node tree.

6. The power distribution device operation and maintenance knowledge base construction method of claim 1, wherein The determination of the matching weight for each fault type based on the solutions corresponding to each fault in the power distribution equipment includes: Based on the solutions corresponding to each fault in the power distribution equipment, all fault types are divided into three types, namely, the first type, the second type, and the third type. The matching weights for each fault type are determined based on the matching weights corresponding to each type.

7. A knowledge base construction system for the operation and maintenance of power distribution equipment, characterized in that, include: The ternary group construction module is used to construct ternary groups based on the operation and maintenance knowledge of each power distribution equipment. The ternary group includes the name of the power distribution equipment, the fault text, and the solution corresponding to each fault. A single power distribution equipment failure rate determination module is used to determine the failure rate of any power distribution equipment based on the fault text of the power distribution equipment and the corresponding solutions for each fault. The similarity determination module is used to calculate the similarity between each pair of power distribution equipment based on the fault text of each power distribution equipment and the corresponding solution for each fault. The similarity is used to characterize the similarity of the fault text and the similarity of the solution between each pair of power distribution equipment. The power distribution equipment grouping module is used to divide all power distribution equipment into several equipment groups based on the similarity between pairs of power distribution equipment and the failure rate of each power distribution equipment. Each equipment group includes at least one power distribution equipment. The knowledge base construction module is used to construct a knowledge base for the operation and maintenance of power distribution equipment based on the triplet and the equipment group corresponding to each power distribution equipment. The single power distribution equipment failure rate determination module is further configured to: for any power distribution equipment, determine multiple failure types of the power distribution equipment based on the failure text of the power distribution equipment, and obtain the occurrence frequency of each failure type within a preset period; determine the matching weight of each failure type based on the solution corresponding to each failure in the power distribution equipment; and determine the failure rate of the power distribution equipment based on the occurrence frequency of each failure type within a preset period and the matching weight of each failure type.

8. The power distribution device operation and maintenance knowledge base construction system according to claim 7, characterized by, The knowledge base construction module is further configured to establish a node tree based on the device groups to which the power distribution devices correspond, the node tree comprising a plurality of nodes, a connection relationship between the nodes, and a plurality of branches of each node, wherein a node corresponds to a power distribution device, the connection relationship is used to represent the similarity degree, and the plurality of branches comprise fault texts of each power distribution device in the triplets and solutions corresponding to each fault; A power distribution device operation and maintenance database is constructed based on the node tree.

Images