Equipment fault document processing method and device, equipment and medium

Abstract

The invention discloses an equipment fault document processing method and device, equipment and a medium. The method comprises the following steps: acquiring a to-be-processed equipment fault document; performing content analysis on the to-be-processed equipment fault document to obtain at least one unit title and corresponding unit content; and determining a classification label corresponding to the unit title according to the content of the unit title and the corresponding unit content. According to the technical scheme provided by the embodiment of the invention, the classification accuracy of the equipment fault documents is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of natural language processing, and in particular to a method, device, equipment and medium for processing equipment failure documents. Background Art

[0002] In order to strengthen the equipment management of hydropower stations, improve the health level of equipment and reduce equipment maintenance costs, hydropower stations usually record equipment failures and classify equipment failure documents.

[0003] At present, the existing classification methods of equipment fault documents can only achieve rough classification of equipment fault documents, and there is a problem of poor classification accuracy. Summary of the invention

[0004] The present invention provides a method, device, equipment and medium for processing equipment fault documents, which improve the classification accuracy of equipment fault documents.

[0005] According to one aspect of the present invention, a method for processing equipment failure documents is provided, the method comprising:

[0006] Obtain the equipment failure documents to be processed;

[0007] Parsing the content of the device failure document to be processed to obtain at least one unit title and corresponding unit content;

[0008] According to the content of the unit title and the corresponding unit content, the classification label corresponding to the unit title is determined.

[0009] According to another aspect of the present invention, there is provided a device for processing equipment failure documents, the device comprising:

[0010] The equipment failure document acquisition module is used to acquire the equipment failure document to be processed;

[0011] A content analysis module, used for analyzing the content of the equipment failure document to be processed to obtain content structured data; wherein the content structured data includes each unit title in the content and the corresponding unit content;

[0012] The multi-label classification module is used to determine the classification label corresponding to the unit title according to the content of the unit title and the corresponding unit content.

[0013] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0014] at least one processor; and

[0015] a memory communicatively connected to the at least one processor; wherein,

[0016] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the device fault document processing method described in any embodiment of the present invention.

[0017] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the device failure document processing method described in any embodiment of the present invention when executed.

[0018] The technical solution of the embodiment of the present invention obtains the equipment failure document to be processed, parses the content of the equipment failure document to be processed, obtains at least one unit title and corresponding unit content, and determines the classification label corresponding to the unit title according to the content of the unit title and the corresponding unit content. This solves the problem that the existing classification method of equipment failure documents can only achieve rough classification of equipment failure documents, and has the problems of low classification efficiency and poor classification accuracy, and improves the classification accuracy of equipment failure documents.

[0019] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0021] Figure 1 is a flowchart of a method for processing device failure documents provided according to Embodiment 1 of the present invention;

[0022] Figure 2 is a flowchart of a method for processing device failure documents provided according to Embodiment 2 of the present invention;

[0023] Figure 3 is a flowchart of a method for processing device failure documents provided according to Embodiment 2 of the present invention;

[0024] Figure 4 is a flowchart of a method for processing device failure documents provided according to Embodiment 2 of the present invention;

[0025] Figure 5 is a schematic diagram of the structure of a device failure document processing apparatus provided according to Embodiment 3 of the present invention;

[0026] Figure 6 It is a structural schematic diagram of an electronic device for implementing the device failure document processing method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0027] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0028] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0029] Embodiment 1

[0030] Figure 1 A flowchart of a device failure document processing method provided in Embodiment 1 of the present invention. The embodiment of the present invention is applicable to the case of processing device failure documents, and the method can be executed by a device failure document processing device, which can be implemented in the form of hardware and / or software, and can be configured in an electronic device that carries the device failure document processing function.

[0031] See also Figure 1 The device fault document processing method shown includes:

[0032] S110: Obtain device failure documents to be processed.

[0033] Equipment failure documents can be documents that record equipment failures. Equipment failure documents can be used as reference documents for equipment failures. Equipment failure documents to be processed can be equipment failure documents to be classified and allocated. In order to strengthen the management of faulty equipment in hydropower stations, improve the health level of equipment, and reduce equipment maintenance costs, hydropower stations usually record equipment failures, generate corresponding equipment failure documents, and maintain and manage equipment failure documents. For different hydropower stations, there may be common equipment failures. Through the maintenance and management of equipment failure documents, when equipment failures occur in hydropower stations, the equipment failures can be handled with reference to the managed historical equipment failure documents, thereby improving the efficiency and accuracy of equipment failure handling in hydropower stations. At present, in order to reduce the response time for handling equipment failure documents to be processed, the equipment failure documents to be processed can be classified according to the content of the historical failure documents, and classified into the category of similar historical failure documents, so as to reduce the identification time and disposal time of professional and technical personnel for the equipment failure documents to be processed.

[0034] However, since hydropower stations contain different equipment systems and involve different professional fields, the content of equipment failure documents requires professional technicians from multiple different fields to effectively clarify and effectively classify the equipment failure documents. At the same time, the causes and treatment methods of equipment failure documents of the same type of equipment may also be diverse. Professional technicians who handle faults downstream still need to spend more energy to further identify the content of equipment failure documents before they can finally be handed over to professional technicians in the corresponding field to handle the equipment failure. In this case, it is particularly important to efficiently, accurately and automatically complete the classification of equipment failure documents so that professional technicians in the corresponding field can handle the equipment failure.

[0035] Specifically, the equipment failure documents to be processed uploaded by the hydropower station can be obtained.

[0036] S120: parse the content of the device failure document to be processed to obtain at least one unit title and corresponding unit content.

[0037] The content of the equipment failure document includes unit title and unit content. Among them, the unit title can be used to divide the content of the equipment failure document. The unit content can be the result of the division of the equipment failure document by the unit title. It can be understood that the same equipment failure document can contain at least one unit title. The unit titles of different equipment failure documents may be the same. Exemplarily, the unit title may include a description of the fault phenomenon, an analysis of the cause of the fault, a fault handling plan and precautions, etc. Different unit titles correspond to different unit contents in the equipment failure document. For example, the unit title may be "Description of the fault phenomenon"; the unit content may be "On July 4, 2020, the operator of the P power station made an on-site inspection and found that the screen of the 10kV II section small current grounding line selection device of the plant power system had no display and the fault light was on. The on-site inspection found that the 10kV The branch loads and bus voltages of Section II are normal, and the 10kV protection device has no alarm. The preliminary analysis shows that the small current grounding line selection device has an internal fault, and the maintenance personnel are notified to check the small current grounding line selection device. The maintenance personnel found that the power module of the small current grounding line selection device was faulty through on-site inspection, and the device power module needed to be replaced. "The unit title can be "Analysis of the Cause of the Fault"; the unit content can be "The maintenance personnel conducted a power outage inspection on the small current grounding line selection device. The device power supply stability requirements are high, and it is a single power module. The daily fluctuation of the factory power supply is large, and the power module has been running continuously for a long time, which eventually led to an internal fault. "The unit title can be "Fault Handling Plan"; the unit content can be "①Handle a work ticket; ②"Finger-pointing and daily description" of the dangerous points at the work site, propose and implement preventive measures, and all members of the work team confirm and sign; ③Replace the power module of the small current grounding line selection device; ④Power-on test; ⑤Clean up the tools and materials on site, explain the work, and complete the work. "The unit title can be "Precautions"; the unit content can be "It is forbidden to work with power on, and the power outage operation must be performed for module replacement and other operations. The power inspection process must be strictly followed before the operation. "The unit titles of different equipment fault documents can also be the same. Exemplarily, the unit title may also include the document title; the corresponding unit content is also the content of the document title. For example, the unit title may be "Factory power system 10kV VII section small current grounding line selection device failure"; the unit content may also be "Factory power system 10kV II section small current grounding line selection device failure".

[0038] Specifically, a layout analysis algorithm may be used to parse the content of the device failure document to be processed, and obtain at least one unit title and corresponding unit content of the device failure document.

[0039] Optionally, the content of the equipment failure document may include text, pictures, etc.

[0040] Exemplarily, when the content of the equipment failure document is in the form of text, a layout analysis algorithm may be directly used to parse the content of the equipment failure document to be processed to obtain at least one unit title and corresponding unit content of the equipment failure document.

[0041] Exemplarily, when the content of a device failure document is in the form of a picture, for example, when the device failure document is a scanned file or a PDF (portable document format) file, OCR (Optical Character Recognition) technology can be used to first convert the picture content into text content; then a layout analysis algorithm is used to parse the content of the device failure document to be processed, so as to obtain at least one unit title and corresponding unit content of the device failure document.

[0042] S130. Determine a classification label corresponding to the unit title according to the content of the unit title and the corresponding unit content.

[0043] The classification tag can be used to classify the equipment failure document. The equipment failure document can include at least one unit title. Accordingly, after determining the classification tag corresponding to the unit title, the equipment failure document also has at least one classification tag. According to the classification tags of each unit title, multi-tag classification of the equipment failure document can be achieved.

[0044] Specifically, the unit title and the corresponding unit content may be analyzed and processed to respectively determine the classification labels of the unit title and the unit content, and finally determine the classification label of the unit title.

[0045] In the prior art, the classification of equipment failure documents usually uses the document title or the full content of the document as the data set, and adopts the BERT+Bi-LSTM+CRF model (named entity recognition model) to roughly classify the equipment failure documents, and then classify the equipment failure documents into historical defect documents under the same category. This classification method of equipment failure documents has the following problems: (1) Only single-label classification is performed on the equipment failure documents, and the characteristic information of the equipment failure documents is not fully utilized. The classification method is too simple, and the classification result obtained is only a rough classification result, which requires downstream professional technicians to classify again; (2) The named entity recognition model needs to be trained before it can be used, but the number of equipment failure documents is actually not large. When there is a deviation in the training sample or the amount of data is insufficient, the classification effect of the model is poor.

[0046] The technical solution of the embodiment of the present invention obtains the device failure document to be processed, performs content analysis on the device failure document to be processed, obtains at least one unit title and corresponding unit content, determines the classification label corresponding to the unit title according to the content of the unit title and the corresponding unit content, and fully mines the characteristic information of the device failure document by processing the content of the unit title and the corresponding unit content, increases the amount of data referenced for the classification of the device failure document, and realizes multi-label classification of the device failure document by generating classification labels for the unit titles. Compared with the single-label classification of the device failure document, the category of the device failure document is refined, and the classification accuracy of the device failure document is improved; at the same time, it also avoids the problem that the model classification is time-consuming and low-accurate when the model is trained in the case of deviation in the training sample or insufficient data, and further improves the classification accuracy of the equipment failure document.

[0047] In an optional embodiment of the present invention, after determining the classification label corresponding to the unit title based on the content of the unit title and the corresponding unit content, it also includes: obtaining a plurality of processed equipment failure documents having the same classification label as the equipment failure document to be processed; performing similarity calculation on the equipment failure document to be processed and the processed equipment failure documents based on the classification labels corresponding to each of the unit titles, and determining a processed equipment failure document that is similar to the equipment failure document to be processed.

[0048] The processed equipment failure document may be a classified and assigned equipment failure document. The processed equipment failure document is a historical equipment failure document. The processed equipment failure document has been processed by professional technicians of downstream fault processing. Optionally, the processing of the equipment failure document by professional technicians of downstream fault processing may include reviewing the equipment failure document, or classifying and storing the equipment failure document. Exemplarily, after classifying the equipment failure document, the content of the equipment failure document may also be reviewed to detect whether the content recorded in the equipment failure document is reasonable. When unreasonable content is found, the equipment failure document may be modified. Optionally, after classifying the equipment failure document, the equipment failure document may also be classified and stored. Optionally, processed equipment failure documents with the same classification label may be stored in the same storage space.

[0049] Specifically, a plurality of processed device fault documents having the same classification label as the device fault document to be processed may be obtained.

[0050] Optionally, multiple processed device fault documents having a preset number of tags identical to or greater than the device fault document to be processed may be obtained. The preset number of tags identical to each other may be 2. By setting the preset number of tags identical to each other, the corresponding processed device fault documents are obtained, and in the query process of the processed device fault documents, the processed device fault documents are preliminarily screened, thereby improving the screening efficiency of the processed device fault documents similar to the device fault document to be processed.

[0051] Specifically, the classification labels corresponding to the unit titles contained in the device failure document to be processed and the classification labels corresponding to the unit titles contained in the processed device failure documents can be calculated, and similarity calculation can be performed to determine the processed device failure documents that are similar to the device failure document to be processed.

[0052] Optionally, the classification labels corresponding to the unit titles may be taken as a whole to perform similarity calculation to determine processed equipment fault documents that are similar to the equipment fault document to be processed.

[0053] Optionally, similarity calculations may be performed on the classification labels corresponding to the individual unit titles, and then the processed device fault documents similar to the device fault document to be processed may be determined as a whole.

[0054] This scheme obtains multiple processed equipment failure documents with the same classification label as the equipment failure document to be processed after determining the classification label corresponding to the unit title according to the content of the unit title and the corresponding unit content, calculates the similarity between the equipment failure document to be processed and the processed equipment failure document according to the classification label corresponding to each unit title, determines the processed equipment failure document similar to the equipment failure document to be processed, and realizes the similarity judgment of equipment failure documents of the same category, thereby realizing the precise segmentation of the equipment failure document to be processed on the basis of determining the unit title, and further improves the accuracy of the classification of equipment failure documents. At the same time, by determining the processed equipment failure document similar to the equipment failure document to be processed, it is convenient to review the equipment failure document to be processed based on the processed equipment failure document, further text mining is performed on similar processed equipment failure documents, and the equipment failure documents to be processed are classified and stored, which is convenient for the subsequent rapid processing of equipment failures and improves the efficiency and accuracy of equipment failure processing in hydropower stations.

[0055] In an optional embodiment of the present invention, the similarity calculation is performed on the device failure document to be processed and the processed device failure document according to the classification label corresponding to each of the unit titles to determine the processed device failure document similar to the device failure document to be processed, including: for the unit title, the similarity calculation is performed on the classification label of the device failure document to be processed and the classification label of the processed device failure document to obtain the label similarity of the unit title; weighted summation of the label similarities of each of the unit titles to obtain the document similarity between the device failure document to be processed and the processed device failure document; and a preset number of the processed device failure documents with larger text similarity are selected to determine them as processed device failure documents similar to the device failure document to be processed.

[0056] The label similarity can be the similarity between the classification labels of the same unit title of different equipment fault documents. By using the classification labels of the unit titles, the unit contents corresponding to each unit title are calculated similarly, which can reflect the unit contents corresponding to each unit title and improve the calculation efficiency of the similarity of the unit contents corresponding to the same unit title of different equipment fault documents. Optionally, the weights of the label similarities of different unit titles can be different. The weights of the label similarities of the unit titles are used to characterize the importance of the unit content corresponding to the unit title in the equipment fault handling. Exemplarily, the unit titles of the equipment fault documents may include the document title, fault phenomenon description, fault cause analysis, fault handling plan and precautions. Among them, the weight of the label similarity of the document title can be 0.4; the weight of the label similarity of the fault phenomenon description can be 0.2; the weight of the label similarity of the fault cause analysis can be 0.2; the weight of the label similarity of the fault handling plan can be 0.1; the weight of the label similarity of the precautions can be 0.1. In the unit contents corresponding to different unit titles, the document title can be a comprehensive description of the equipment failure; the fault phenomenon description and the fault cause analysis can be the detection and analysis content of the equipment failure; the fault handling plan and precautions can be the actual handling process and safety precautions of the equipment failure. Among the unit titles, the unit content corresponding to the document title has the highest importance. Among the unit contents corresponding to other unit titles, the unit content corresponding to the fault phenomenon description and the unit content corresponding to the fault cause analysis are of higher importance. Optionally, different weights are assigned to the label similarities of different unit titles, and the document similarities of different equipment failure documents are determined taking into account the importance of the unit contents corresponding to different unit titles, thereby further improving the accuracy of the processed equipment failure documents that are similar to the equipment failure documents to be processed. The document similarity can be the similarity between different equipment failure documents. The document similarity can be used to characterize the similarity between the processed equipment failure documents and the equipment failure documents to be processed. By calculating the document similarity between different equipment failure documents, on the one hand, it is possible to obtain processed equipment failure documents that are similar to the equipment failure document to be processed, and further text mining can be performed on the processed equipment failure documents to obtain more equipment failure processing information, providing a reference for the subsequent equipment failure processing; on the other hand, it is also possible to further classify the equipment failure documents to be processed through the processed equipment failure documents that are similar to the equipment failure document to be processed. The preset number can be used to screen the processed equipment failure documents. Exemplarily, the preset number can be 3, which is specifically set and adjusted according to the experience of the technician. By setting the preselected number, it is used to screen an appropriate number of processed equipment failure documents that are similar to the equipment failure document to be processed.

[0057] Specifically, for the same unit title, the classification labels of the device failure document to be processed and the classification labels of the processed device failure documents can be calculated to obtain the similarity of each unit title. The label similarities of each unit title can be weighted summed to obtain the document similarity between the device failure document to be processed and the processed device failure document. The similarities of each document are compared, and a preset number of processed device failure documents with greater text similarity are selected to determine as processed device failure documents similar to the device failure document to be processed.

[0058] This scheme calculates the similarity between the classification labels of the equipment failure document to be processed and the classification labels of the processed equipment failure documents for the unit titles to obtain the label similarity of the unit titles, performs weighted summation on the label similarities of each unit title to obtain the document similarity between the equipment failure document to be processed and the processed equipment failure documents, selects a preset number of processed equipment failure documents with larger text similarities, and determines them as the processed equipment failure documents similar to the equipment failure document to be processed, calculates the label similarity of the unit titles in advance, performs weighted summation on the similarities of each label to obtain the document similarity, determines the document similarity of different equipment failure documents taking into account the importance of the unit contents corresponding to different unit titles, and further improves the accuracy of the determined processed equipment failure documents that are similar to the equipment failure document to be processed.

[0059] In an optional embodiment of the present invention, for the unit title, similarity calculation is performed on the classification label of the device failure document to be processed and the classification label of the processed device failure document to obtain the label similarity of the unit title, including: vectorizing the classification label corresponding to each unit title of the device failure document to be processed to obtain a first classification label vector corresponding to each unit title; obtaining a second classification label vector corresponding to each unit title of the processed device failure document; for the unit title, similarity calculation is performed on the first classification label vector and the second classification label vector to obtain the label similarity of the unit title.

[0060] The first classification label vector may be a vector corresponding to the classification label of the unit title of the device failure document to be processed. The second classification label vector may be a vector corresponding to the classification label of the unit title of the processed device failure document. The first classification label vector and the second classification label vector are both vectors corresponding to the classification label of the unit title. However, the first classification label vector and the second classification label vector are respectively vectors corresponding to the classification labels of the unit titles of different device failure documents. The unit title of the device failure document may correspond to at least one classification label. Accordingly, the number of the first classification label vector and the second classification label vector is at least one.

[0061] Specifically, the classification labels corresponding to the unit titles of the device failure document to be processed can be input into the vector model to obtain the first classification label vector corresponding to the unit title. The classification labels corresponding to the unit titles of the processed device failure document can be input into the vector model to obtain the second classification label vector corresponding to the unit title. For the same unit title, the first classification label vector and the second classification label vector can be calculated to perform cosine similarity calculation to obtain the label similarity of the unit title.

[0062] This scheme obtains the first classification label vector corresponding to each unit title of the equipment failure document to be processed by vectorizing the classification label corresponding to each unit title, obtains the second classification label vector corresponding to each unit title of the processed equipment failure document, and calculates the similarity between the first classification label vector and the second classification label vector for the unit title to obtain the label similarity of the unit title. By vectorizing the classification labels of the unit title, the similarity calculation of the classification labels of the unit title of the equipment failure document to be processed and the classification labels of the unit title of the processed equipment failure document is realized, thereby improving the calculation efficiency and accuracy of the similarity of the classification labels of the unit titles.

[0063] In an optional embodiment of the present invention, the unit title includes a description of the fault phenomenon, an analysis of the cause of the fault, a fault handling solution and precautions.

[0064] The unit content corresponding to the fault phenomenon description can provide a detailed description of the equipment failure. Optionally, the fault phenomenon description may include time, fault location, operation and maintenance personnel, fault description, equipment parameters, preliminary inspection description, preliminary analysis description, maintenance personnel, preliminary maintenance description and maintenance conclusion, etc. Exemplarily, the fault phenomenon description may be "On May 27, 2015, the Q power station operator discovered that the oil level of the No. 1 reactor (1DK) C phase was lower than that of the A and B phases when patrolling the screen on the upper computer. On-site inspection found that the 1DK C phase oil level gauge indicated 4, A and B phases were 6, and the C phase oil level indication was low. Inspection of the 1DK C phase exterior and the accident oil drain showed no obvious oil leakage, and no obvious oil accumulation was found in the reactor accident oil pool. Preliminary analysis showed that there was leakage in the capsule, and the maintenance personnel were notified to check the oil level of the oil pillow. The maintenance personnel found through on-site inspection that the oil pillow capsule had sunk and there was oil accumulation on the capsule. It was determined that the capsule was damaged and needed to be replaced."

[0065] The unit content corresponding to the fault cause analysis can be used to analyze the cause of the equipment failure. Optionally, the fault cause analysis may include further maintenance description and fault generation principle analysis. For example, the fault cause analysis may be "The maintenance personnel drained the 1DK oil pillow. After opening the side cover of the oil pillow, they found a crack of about 15m on the lower right side of the capsule. Because the crack is located near the oil level sensor support, when the oil pillow is drained, filled with oil, and in operation, the oil level changes will cause friction with the oil level sensor, causing the capsule to wear and eventually break."

[0066] The unit content corresponding to the fault handling plan is used to characterize the actual fault handling process. Optionally, the fault handling plan may include various fault handling steps. Exemplarily, the fault handling plan may be "① Processing a work ticket; ② "Finger and verbalize" the dangerous points at the work site, propose and implement preventive measures, and all members of the work team confirm and sign; ③ Connect the oil drain valve at the bottom of the reactor body to the oil drain pipe, open the oil drain valve to drain the oil until the oil is no longer visible in the observation window of the gas relay; ④ Lay a plastic film and an oil collecting box under the side cover of the reactor oil pillow, remove the side cover of the reactor oil pillow, remove the flange bolts connecting the capsule and the oil pillow, and remove the damaged capsule. ⑤ Check the internal condition of the oil pillow, clean the residual oil and debris in the oil pillow, and replace the new model of oil level sensor ⑥ Unfold the new capsule and carefully place it in the oil pillow, connect the capsule and the oil pillow flange bolts, and install the oil pillow side cover. ⑦ Connect the oil filling pipe to the oil filling and draining pipe under the oil pillow, close the butterfly valve between the gas relay and the main body, and slowly inject qualified insulating oil into the oil pillow through the vacuum oil filter. Stop filling the oil when the oil level of the oil pillow is slightly higher than the operating oil level. ⑧ Open the butterfly valve between the gas relay and the main body, open the exhaust hole of the gas relay and the exhaust hole of the oil pillow to exhaust several times until no gas is discharged, and adjust the oil level of the oil pillow to the normal position. ⑨ Clean up the tools and materials on site, explain the work, and complete the work. ”

[0067] The unit content corresponding to the precautions is used to describe the safety precautions during the fault handling process. For example, the precautions may be "When filling oil, be sure to close the butterfly valve between the gas relay and the body to prevent the pressure relief valve of the reactor body from operating."

[0068] This solution simplifies the content of equipment failure documents by concretizing unit titles into fault phenomenon description, fault cause analysis, fault handling plan and precautions, while recording equipment failures, ensuring the classification accuracy of equipment failure documents, and at the same time improving the classification efficiency of equipment failure documents.

[0069] Embodiment 2

[0070] Figure 2A flowchart of a device fault document processing method provided in the second embodiment of the present invention. Based on the above embodiments, the embodiment of the present invention concretizes "determining the classification label corresponding to the unit title according to the content of the unit title and the corresponding unit content" as "generating unit structured data corresponding to the unit title according to the unit title and the corresponding unit content; combining the unit structured data corresponding to each unit title to obtain content structured data; inputting the content structured data into a multi-label classification model, and outputting the classification label corresponding to each unit title; wherein the multi-label classification model generates the classification label corresponding to the unit title by semantically understanding the unit title and the corresponding unit content", thereby realizing multi-label classification of device fault documents, fully mining the content of the device fault documents, and at the same time, improving the efficiency of equipment fault document classification. It should be noted that for the parts not described in detail in the embodiments of the present invention, please refer to the description of other embodiments.

[0071] See also Figure 2 The device fault document processing method shown includes:

[0072] S210: Obtain device failure documents to be processed.

[0073] S220: parse the content of the device failure document to be processed to obtain at least one unit title and corresponding unit content.

[0074] S230. Generate unit structured data corresponding to the unit title according to the unit title and the corresponding unit content.

[0075] The unit structured data may be structured data generated by the unit title and the corresponding unit content. The unit structured data may be used to characterize the corresponding unit content under the unit title. For example, the unit structured data may be "Document Title: Transformer (Reactor) Oil Conservator Oil Level Abnormality".

[0076] Specifically, the unit title and the corresponding unit content may be combined to generate unit structured data corresponding to the unit title.

[0077] S240: Combine the unit structured data corresponding to each of the unit titles to obtain content structured data.

[0078] The content structured data may be structured data corresponding to the content of the equipment failure document. The content structured data may be used to characterize the content of the equipment failure document. For example, the content structured data may be {"document title": AAA, "fault phenomenon description": BBB, "fault cause analysis": CCC, "handling solution": DDD, "precautions": EEE}.

[0079] S250. Input the content structured data into a multi-label classification model, and output the classification label corresponding to each unit title; wherein the multi-label classification model generates the classification label corresponding to the unit title by semantically understanding the unit title and the corresponding unit content.

[0080] The multi-label classification model can be used to perform multi-label classification on the equipment failure document. The input of the multi-label classification model can be content structured data. The output of the multi-label classification model can be the classification label corresponding to each unit title of the equipment failure document. Exemplarily, the multi-label classification model can be a large language model. The multi-label classification model can generate the classification label corresponding to the unit title by semantically understanding the unit title and the corresponding unit content.

[0081] Specifically, the content structured data can be input into a multi-label classification model to output the classification label corresponding to each unit title.

[0082] The technical solution of the embodiment of the present invention obtains the equipment failure document to be processed, performs content analysis on the equipment failure document to be processed, obtains at least one unit title and corresponding unit content, generates unit structured data corresponding to the unit title according to the unit title and the corresponding unit content, combines the unit structured data corresponding to each unit title to obtain content structured data, inputs the content structured data into a multi-label classification model, and outputs classification labels corresponding to each unit title; wherein the multi-label classification model generates classification labels corresponding to the unit titles by semantically understanding the unit titles and the corresponding unit contents, and utilizes the multi-label classification model to realize multi-label classification of equipment failure documents, fully mines the content of equipment failure documents, improves the efficiency of equipment failure document classification, and reduces the classification cost of equipment failure documents while ensuring the classification effect of equipment failure documents.

[0083] In an optional embodiment of the present invention, the step of inputting the content structured data into a multi-label classification model and outputting classification labels corresponding to each unit title comprises: inputting the content structured data and a preset prompt template into the multi-label classification model; and performing multi-label classification on the content structured data to obtain classification labels of at least one category corresponding to each unit title.

[0084] The preset prompt template can be used to instruct the multi-label classification model to output the corresponding content according to the preset prompt template. Optionally, the preset prompt template may include the unit content corresponding to the unit title, the function of the prompt template and the corresponding classification label. Optionally, the same unit title may correspond to at least one classification label of a category. In this way, the content of the classification label is enriched. Exemplarily, the preset prompt template can be a description of a known phenomenon BBB, and from the above, multi-label classification is performed, including classification labels such as "category A", "category B" and "category C", and the content in Json format is output.

[0085] Specifically, the content structured data and the preset prompt template can be input into the multi-label classification model. The multi-label classification model performs multi-label classification on the content structured data according to the preset prompt template to obtain at least one category classification label of each unit title.

[0086] This solution performs multi-label classification on the content structured data by inputting the content structured data and the preset prompt template into the multi-label classification model, and obtains at least one category of classification labels corresponding to each unit title. Based on the preset prompt template, the content structured data is subjected to multi-label classification, thereby improving the efficiency of multi-label classification. At the same time, by obtaining at least one category of classification labels corresponding to each unit title, the number of classification labels is further enriched, thereby improving the classification accuracy of equipment fault documents.

[0087] Figure 3 The present invention provides a flowchart of a method for processing device failure documents in accordance with a second embodiment of the present invention. Figure 3 This is a preferred embodiment of a method for processing equipment failure documents provided in Embodiment 2 of the present invention.

[0088] See also Figure 3 The device fault document processing method shown includes:

[0089] S301. Obtain historical equipment failure documents.

[0090] Specifically, historical equipment failure documents may be obtained.

[0091] S302: Parse the historical equipment failure document to obtain the text content of the historical equipment failure document.

[0092] Specifically, if the content of the historical equipment failure document is in the form of a PDF or scanned file, OCR technology can be used to convert the image content into text content.

[0093] S303: Perform layout analysis on the text content of the historical equipment failure document to generate content structured data.

[0094] Specifically, a layout analysis algorithm may be used to perform layout analysis on the text content of the historical equipment failure document to obtain content structured data of the historical equipment failure document.

[0095] Exemplarily, a layout analysis algorithm may be used to structure the data of historical equipment failure documents to generate content structured data such as {"document title": AAA, "failure phenomenon description": BBB, "failure cause analysis": CCC, "failure handling solution": DDD, "precautions": EEE}.

[0096] S304: Perform significant label analysis on the content structured data to obtain content structured classification labels, and perform similar text annotation on similar historical equipment failure documents.

[0097] Specifically, significant label analysis can be performed on the content structured data of historical equipment failure documents in advance, and professional technicians can classify the historical equipment failure documents and set classification labels. Similarity can be set for each historical equipment failure document.

[0098] For example, significant label analysis can be performed on the content structured data of historical equipment failure documents, and classification labels can be set by professional technicians. Among them, classification labels can include generators, transformers, lifting equipment, and turbines. Similarity annotation can be performed on each historical equipment failure document, for example, historical equipment failure document X is similar to historical equipment failure document Y.

[0099] S305: Obtain a large language model, and obtain a Prompt prompt based on the classification label of the content structured data.

[0100] Specifically, the Prompt words (ie, preset prompt templates) of the large language model (ie, multi-label classification model) can be repeatedly designed according to the classification labels of the content structured data set by professional technicians.

[0101] Exemplarily, the preset prompt template can be a description of a known phenomenon BBB, and from the above, multi-label classification is performed, including classification labels such as "category A", "category B" and "category C", and the content in Json format is output.

[0102] S306: Input the content structured data into the large language model to obtain content structured data with multiple labels.

[0103] Specifically, the content can be structured and input into a large language model, and multi-label classification can be performed based on the large language model to obtain the classification label corresponding to each unit title.

[0104] Exemplarily, the classification result output by the large model may be {"unit title": ["category A", "category B"]}, where category A and category B are classification labels corresponding to the unit title.

[0105] Optionally, the equipment fault documents can be classified according to the classification labels corresponding to the unit titles of the equipment fault documents, ensuring the consistency of at least two classification labels. There are different equipment fault documents under each classification label.

[0106] S307 , dividing the content structured data with multiple labels into a training data set and a verification data set.

[0107] Specifically, the content structured data with multiple labels can be used for subsequent analysis and training of the classification of equipment failure documents. The content structured data can be divided into a verification data set and a training data set. Optionally, the verification data set and the training data set also have similarity annotations.

[0108] S308. Input the classification label corresponding to the unit title of the training data set into the vector model to obtain the unit label vector corresponding to the unit title.

[0109] Specifically, under the training data set, the classification label corresponding to the unit title of the training data set can be input into the vector model to obtain the unit label vector corresponding to the unit title.

[0110] Exemplarily, the input of the vector model may be “abnormal oil level of transformer (reactor) oil pillow”; the output of the vector model may be [0.00002, 0.34252, 0.012012, ....].

[0111] S309 , storing the unit label vectors corresponding to the unit titles of the content structured data in a vector database according to the unit titles.

[0112] The vector database may be a Faiss vector database. Compared with other databases, the query efficiency of the vector database is higher.

[0113] Specifically, according to each unit title, the unit label vectors corresponding to the unit titles of the content structured data may be stored in the vector database respectively.

[0114] Exemplarily, the output results of the vector model, such as {Title A: "0.00002, 0.34252, 0.012012, ...."}, and the unit label vectors corresponding to the unit titles of the content structured data can be stored separately in the vector database according to the title, phenomenon description, cause analysis, treatment plan and precautions.

[0115] S310 . Perform weighted summation on the label similarities corresponding to the unit titles of different historical equipment failure documents to obtain the document similarities between the different historical equipment failure documents.

[0116] Specifically, a large language model can be used for multi-label classification based on a validation data set, and unit label vectors corresponding to different unit titles can be obtained. The historical equipment failure documents in the vector database can be recalled based on the classification results and auxiliary rules. The label similarities between different unit label vectors can be calculated for different unit titles based on cosine similarity. The label similarities corresponding to the unit titles of different historical equipment failure documents can be weighted and summed to obtain the similarities between different historical failure documents. Among them, the weights of the label similarities of different unit titles can be different to make the similarity calculation more accurate.

[0117] S311. Based on the similarity annotations of different historical equipment failure documents in the verification data set, the weights of the similarities of each unit label are adjusted and updated to make the order of the document similarities between different historical equipment failure documents consistent with the similarity annotations.

[0118] Specifically, based on the pre-annotated similar historical equipment failure documents, in the validation data set, based on the cosine theorem, the label similarity between the unit label vectors of different unit titles can be calculated respectively. Different weights can be set so that the order of document similarity between different historical equipment failure documents is consistent with the similarity annotation.

[0119] For example, the document similarity can be calculated using "title 0.4*similarity 0.8+phenomenon description 0.2*similarity 0.9+cause analysis 0.2*similarity 0.87+treatment plan 0.1*similarity 0.9+notes 0.1*similarity 0.9". Among them, the weight of the title is 0.4; the label similarity of the title is 0.8; the weight of the phenomenon description is 0.2; the label similarity of the phenomenon description is 0.9; the weight of the cause analysis is 0.2; the label similarity of the cause analysis is 0.87; the weight of the treatment plan is 0.1; the label similarity of the treatment plan is 0.9; the weight of the notes is 0.1; the label similarity of the notes is 0.9. According to the order of document similarity, the weight that is consistent with the similarity annotation can be selected as the weight corresponding to the label similarity of each unit title.

[0120] Figure 4 The present invention provides a flowchart of a method for processing device failure documents in accordance with a second embodiment of the present invention. Figure 4 This is a preferred embodiment of a method for processing equipment failure documents provided in Embodiment 2 of the present invention.

[0121] See also Figure 4 The device fault document processing method shown includes:

[0122] S401. Obtain device failure documents to be processed.

[0123] S402: Perform layout analysis on the equipment failure document to be processed to obtain content structured data.

[0124] S403: Input the content structured data into the large language model to obtain the unit label corresponding to the unit title of the equipment failure document.

[0125] S404: Input the unit label corresponding to the unit title into the vector model to obtain a first unit label vector.

[0126] S405 , obtaining a processed device failure document having the same classification label as the device failure document to be processed, and querying a vector database for a second unit label vector of the processed device failure document.

[0127] S406 , performing cosine similarity calculation on the first unit label vector and the second unit label vector to obtain label similarity of the unit title.

[0128] S407 . Perform weighted summation on the label similarities of each unit title according to the updated weights of the label similarities of each unit, and obtain the document similarity between the device fault document to be processed and the processed device fault document.

[0129] S408 , sorting the similarities of the documents, and selecting a preset number of processed device failure documents with relatively high document similarities as processed device failure documents similar to the device failure document to be processed.

[0130] This solution makes full use of various features of equipment failure documents, namely, the significant label features and hidden semantic features of equipment failure documents, adopts a large language model, and efficiently performs text classification by selecting appropriate preset prompt templates. While ensuring the classification effect, it reduces the cost of multi-label classification and also realizes further similar text mining of equipment failure documents.

[0131] Embodiment 3

[0132] Figure 5 A schematic diagram of the structure of a device failure document processing apparatus provided in Embodiment 3 of the present invention. The embodiment of the present invention can be applied to the case of processing device failure documents, the apparatus can execute a device failure document processing method, the apparatus can be implemented in the form of hardware and / or software, and the apparatus can be configured in an electronic device that carries a device failure document processing function.

[0133] See also Figure 5 The device for processing equipment failure documents shown in the figure comprises: an equipment failure document acquisition module 510, a content analysis module 520 and a multi-label classification module 530. The equipment failure document acquisition module 510 is used to acquire the equipment failure document to be processed; the content analysis module 520 is used to perform content analysis on the equipment failure document to be processed to obtain at least one unit title and corresponding unit content; the multi-label classification module 530 is used to determine the classification label corresponding to the unit title according to the content of the unit title and the corresponding unit content.

[0134] The technical solution of the embodiment of the present invention obtains the device failure document to be processed, performs content analysis on the device failure document to be processed, obtains at least one unit title and corresponding unit content, determines the classification label corresponding to the unit title according to the content of the unit title and the corresponding unit content, and fully mines the characteristic information of the device failure document by processing the content of the unit title and the corresponding unit content, increases the amount of data referenced for the classification of the device failure document, and realizes multi-label classification of the device failure document by generating classification labels for the unit titles. Compared with the single-label classification of the device failure document, the category of the device failure document is refined, and the classification accuracy of the device failure document is improved; at the same time, it also avoids the problem that the model classification is time-consuming and low-accurate when the model is trained in the case of deviation in the training sample or insufficient data, and further improves the classification accuracy of the equipment failure document.

[0135] In an optional embodiment of the present invention, the multi-label classification module 530 includes: a unit structured data generation unit, which is used to generate unit structured data corresponding to the unit title according to the unit title and the corresponding unit content; a content structured data generation unit, which is used to combine the unit structured data corresponding to each of the unit titles to obtain content structured data; a classification label output unit, which is used to input the content structured data into a multi-label classification model and output the classification label corresponding to each unit title; wherein the multi-label classification model generates the classification label corresponding to the unit title by performing semantic understanding of the unit title and the corresponding unit content.

[0136] In an optional embodiment of the present invention, the classification label output unit includes: a model input subunit, used to input the content structured data and a preset prompt template into the multi-label classification model; a classification label output subunit, used to perform multi-label classification on the content structured data to obtain a classification label of at least one category corresponding to each unit title.

[0137] In an optional embodiment of the present invention, the device also includes: a processed equipment failure document acquisition module, which is used to obtain a plurality of processed equipment failure documents having the same classification label as the equipment failure document to be processed after determining the classification label corresponding to the unit title based on the content of the unit title and the corresponding unit content; a similar equipment failure document determination module, which is used to calculate the similarity between the equipment failure document to be processed and the processed equipment failure document based on the classification labels corresponding to each of the unit titles, and determine a processed equipment failure document that is similar to the equipment failure document to be processed.

[0138] In an optional embodiment of the present invention, a similar equipment failure document determination module includes: a label similarity determination unit, which is used to calculate the similarity between the classification label of the equipment failure document to be processed and the classification label of the processed equipment failure document with respect to the unit title, so as to obtain the label similarity of the unit title; a document similarity determination unit, which is used to perform weighted summation on the label similarities of each of the unit titles, so as to obtain the document similarity between the equipment failure document to be processed and the processed equipment failure document; and a similar equipment failure document determination unit, which is used to select a preset number of the processed equipment failure documents with a larger text similarity, and determine them as processed equipment failure documents similar to the equipment failure document to be processed.

[0139] In an optional embodiment of the present invention, the label similarity determination unit includes: a first classification label vector determination subunit, used to vectorize the classification label corresponding to each unit title of the device fault document to be processed, and obtain a first classification label vector corresponding to each unit title; a second classification label vector acquisition subunit, used to acquire a second classification label vector corresponding to each unit title of the processed device fault document;

[0140] The label similarity determination subunit is used to calculate the similarity of the first classification label vector and the second classification label vector for the unit title to obtain the label similarity of the unit title.

[0141] In an optional embodiment of the present invention, the unit title includes a description of the fault phenomenon, an analysis of the cause of the fault, a fault handling solution and precautions.

[0142] The device failure document processing apparatus provided in the embodiment of the present invention can execute the device failure document processing method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0143] In the technical solution of the embodiment of the present invention, the acquisition, storage and application of the device failure documents to be processed, the processed device failure documents and the second classification label vector, etc., all comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0144] Embodiment 4

[0145] Figure 6 A schematic diagram of the structure of an electronic device 600 that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0146] like Figure 6 As shown, the electronic device 600 includes at least one processor 601, and a memory connected to the at least one processor 601, such as a read-only memory (ROM) 602, a random access memory (RAM) 603, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 601 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 602 or the computer program loaded from the storage unit 608 to the random access memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the electronic device 600 can also be stored. The processor 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0147] Multiple components in the electronic device 600 are connected to the I / O interface 605, including: an input unit 606, such as a keyboard, a mouse, etc.; an output unit 607, such as various types of displays, speakers, etc.; a storage unit 608, such as a disk, an optical disk, etc.; and a communication unit 609, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 609 allows the electronic device 600 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0148] The processor 601 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 601 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The processor 601 executes the various methods and processes described above, such as a device failure document processing method.

[0149] In some embodiments, the device failure document processing method may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 608. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 600 via the ROM 602 and / or the communication unit 609. When the computer program is loaded into the RAM 603 and executed by the processor 601, one or more steps of the device failure document processing method described above may be performed. Alternatively, in other embodiments, the processor 601 may be configured to execute the device failure document processing method in any other appropriate manner (e.g., by means of firmware).

[0150] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0151] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0152] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0153] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0154] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0155] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship between the client and the server is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or a cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS (Virtual Private Server) services.

[0156] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0157] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for processing equipment failure documents, characterized in that: The method comprises: Obtain the equipment failure documents to be processed; Parsing the content of the device failure document to be processed to obtain at least one unit title and corresponding unit content; According to the content of the unit title and the corresponding unit content, the classification label corresponding to the unit title is determined.

2. The method according to claim 1, characterized in that The step of determining the classification label corresponding to the unit title according to the content of the unit title and the corresponding unit content includes: Generating unit structured data corresponding to the unit title according to the unit title and the corresponding unit content; Combining the unit structured data corresponding to each of the unit titles to obtain content structured data; The content structured data is input into a multi-label classification model, and the classification label corresponding to each unit title is output; wherein the multi-label classification model generates the classification label corresponding to the unit title by semantically understanding the unit title and the corresponding unit content.

3. The method according to claim 2, characterized in that The step of inputting the content structured data into a multi-label classification model and outputting classification labels corresponding to each unit title includes: Inputting the content structured data and the preset prompt template into the multi-label classification model; The content structured data is subjected to multi-label classification to obtain at least one category of classification labels corresponding to each unit title.

4. The method according to claim 1, characterized in that: After determining the classification label corresponding to the unit title according to the content of the unit title and the corresponding unit content, the method further includes: Acquire a plurality of processed device failure documents having the same classification label as the device failure document to be processed; According to the classification labels corresponding to the unit titles, similarity calculation is performed on the device failure document to be processed and the processed device failure document to determine the processed device failure document that is similar to the device failure document to be processed.

5. The method according to claim 4, characterized in that The calculating the similarity between the device failure document to be processed and the processed device failure document according to the classification label corresponding to each unit title, and determining the processed device failure document similar to the device failure document to be processed, comprises: For the unit title, similarity calculation is performed between the classification label of the to-be-processed equipment fault document and the classification label of the processed equipment fault document to obtain the label similarity of the unit title; Performing weighted summation on the label similarities of each of the unit titles to obtain the document similarity between the to-be-processed equipment failure document and the processed equipment failure document; A preset number of the processed equipment failure documents having greater text similarity are selected and determined as processed equipment failure documents similar to the equipment failure document to be processed.

6. The method according to claim 5, characterized in that The step of calculating the similarity between the classification label of the device failure document to be processed and the classification label of the processed device failure document for the unit title to obtain the label similarity of the unit title includes: Vectorizing the classification labels corresponding to the unit titles of the device fault document to be processed to obtain first classification label vectors corresponding to the unit titles; Obtaining a second classification label vector corresponding to each of the unit titles of the processed equipment failure document; For the unit title, similarity calculation is performed on the first classification label vector and the second classification label vector to obtain label similarity of the unit title.

7. The method according to claim 1, characterized in that The unit title includes fault phenomenon description, fault cause analysis, fault handling plan and precautions.

8. A device for processing equipment failure documents, characterized in that: The device comprises: The equipment failure document acquisition module is used to acquire the equipment failure document to be processed; A content analysis module, used for performing content analysis on the device failure document to be processed to obtain at least one unit title and corresponding unit content; The multi-label classification module is used to determine the classification label corresponding to the unit title according to the content of the unit title and the corresponding unit content.

9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the device fault document processing method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the device failure document processing method according to any one of claims 1 to 7 when executed.

Images