A device operation and maintenance archive retrieval method based on defect label string merging

Abstract

The application discloses a kind of based on defect label string merging equipment operation file search method, comprising: by defect label string is according to label order, continuous label section and section connection is split, and generate fragment boundary mark, complete the structured split of defect process, it is convenient to keep fragment boundary clear when cross node search;Again, link state of each search node is combined to generate label fragment call sequence, reduce the influence of link fluctuation on fragment call order;Further, according to fragment boundary mark, the sequence splicing of candidate label fragment is carried out, and through connection consistency check and merging comparison, the file that is similar locally but overall inconsistent is excluded;Therefore, the application improves the accuracy, integrity and stability of equipment operation file search under dispersed storage condition.

Description

Technical Field

[0001] This invention relates to the technical field of equipment operation and maintenance record management and distributed data retrieval, and in particular to a method for retrieving equipment operation and maintenance records based on defect tag string merging. Background Technology

[0002] With the development of the Industrial Internet, equipment lifecycle management, and distributed operation and maintenance platforms, inspection records, fault records, maintenance records, alarm records, and handling records during equipment operation are continuously accumulating, gradually forming large-scale equipment operation and maintenance archives. To improve the reuse efficiency of these archives, existing technologies typically focus on archive classification, keyword retrieval, field filtering, rule matching, knowledge graph association, and vectorized recall. By organizing information such as equipment identification, fault type, maintenance time, component name, and handling results, they support equipment fault tracing, operation and maintenance solution reuse, and historical case querying. In multi-node deployment scenarios, existing technologies also distribute operation and maintenance archives across edge nodes, regional nodes, and central nodes to reduce single-node storage pressure and improve local access efficiency. Meanwhile, to address the need to express the evolution process of equipment defects, operation and maintenance systems are gradually introducing organizational methods such as defect tag sequences, event association chains, and state transition records to enhance the retrieval capability for complex defect processes. Therefore, how to structure and accurately retrieve defect tag information with continuous relationships in a multi-node distributed storage environment has become an important research direction in the field of equipment operation and maintenance archive management.

[0003] However, existing technologies for retrieving equipment maintenance records mainly rely on keyword matching, single-tag filtering, or combined field condition retrieval methods. Their retrieval targets are mostly limited to single tags, single records, or independent fields. For defect tag strings formed by multiple defect tags connected in sequence within the same equipment maintenance record, there is a lack of structured processing paths oriented towards tag connection relationships. When equipment maintenance records are scattered across multiple retrieval nodes, existing technologies typically retrieve record content according to static addresses or fixed node order, rarely incorporating the link status of the retrieval nodes into the tag fragment retrieval path. This leads to link fluctuations, missing fragments, or node errors. When latency differences are significant, problems such as disordered tag fragment retrieval order, misaligned splicing boundaries, and incomplete retrieval results are prone to occur. On the other hand, existing technologies typically focus on whether tags appear when comparing retrieval results, with less emphasis on further verifying the connection consistency between adjacent tags, and less on performing boundary connection checks and merging comparisons on multiple candidate tag fragments returned across nodes. As a result, it is difficult to accurately reconstruct the equipment operation and maintenance file corresponding to the complete defect tag string. Therefore, it is evident that existing equipment operation and maintenance file retrieval technologies still have significant shortcomings in terms of fragmented organization for defect tag strings, fragmented retrieval involving link status, and sequential splicing based on boundary identifiers.

[0004] Therefore, existing equipment operation and maintenance record retrieval technologies suffer from problems such as unstable cross-node retrieval order, insufficient precision in defect tag string splicing, and weak completeness of retrieval results. The invention addresses the problem of accurately retrieving and merging defect tag strings according to their connection relationships under distributed storage conditions. Summary of the Invention

[0005] The purpose of this section is to outline some aspects of the embodiments of the present invention and to briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section, as well as in the abstract and title of the present application, to avoid obscuring the purpose of this section, the abstract and title of the invention. Such simplifications or omissions shall not be used to limit the scope of the present invention.

[0006] In view of the aforementioned existing problems, the present invention is proposed.

[0007] To solve the above-mentioned technical problems, the present invention provides the following technical solution: As a preferred embodiment of the equipment operation and maintenance file retrieval method based on defect tag string merging described in this invention, the defect tag strings in the equipment operation and maintenance file are segmented according to the tag order, continuous tag segments, and inter-segment connection relationships to form tag fragments, and fragment boundary identifiers are generated. Based on the link status of each retrieval node storing the tag fragments, a link status level sequence is formed, and a tag fragment retrieval sequence is generated according to the link status level sequence. Based on the defect label string to be inspected and the label fragment retrieval sequence, candidate label fragments are extracted from the corresponding retrieval node and sequentially spliced ​​according to the fragment boundary identifier to form a candidate label string set. The candidate tag string set and the defect tag string to be inspected are checked for connection consistency and merged for comparison, and the equipment operation and maintenance file matching the defect tag string to be inspected is output.

[0008] The beneficial effects of this invention are as follows: This invention segments the defect tag string according to tag order, continuous tag segments, and inter-segment connection relationships, and generates fragment boundary identifiers to complete the structured decomposition of the defect process, which facilitates clear fragment boundaries during cross-node retrieval. Furthermore, it generates a tag fragment retrieval sequence by combining the link status of each retrieval node, reducing the impact of link fluctuations on the fragment retrieval order. Additionally, it sequentially splices candidate tag fragments based on fragment boundary identifiers and eliminates locally similar but overall inconsistent files through connection consistency checks and merge comparisons. Therefore, this invention improves the accuracy, completeness, and stability of equipment operation and maintenance file retrieval under distributed storage conditions. Attached Figure Description

[0009] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein: Figure 1 This is a flowchart illustrating the equipment maintenance record retrieval method based on defect tag string merging as shown in this invention. Detailed Implementation

[0010] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0011] Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without inventive effort should fall within the scope of protection of this invention.

[0012] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0013] According to an embodiment of the present invention, in combination Figure 1 The flowchart shown illustrates a method for retrieving equipment operation and maintenance records based on defect tag string merging, which specifically includes the following steps: S1. Segment the defect tag strings in the equipment operation and maintenance records according to tag order, consecutive tag segments, and inter-segment connections to form tag fragments, and generate fragment boundary markers. Note that the following points should be noted in this step: In this embodiment, the equipment maintenance file refers to a collection of maintenance records formed around the same equipment during the processes of inspection, spot checks, repair, shutdown troubleshooting, retesting, and restoration to operation. The equipment maintenance file includes at least the equipment number, maintenance time sequence, defect record content, and defect tag arrangement result. The defect tag string is an ordered tag sequence obtained by organizing the defect records in the same equipment maintenance file according to the order of the records. Each defect tag in the defect tag string corresponds to a defect record that has been standardized and classified. The defect tags can be set according to the defect categories commonly used in the equipment maintenance industry, such as abnormal vibration, abnormal temperature rise, abnormal oil level, abnormal leakage, abnormal insulation, abnormal corrosion, abnormal loosening, abnormal noise, abnormal start-up and shutdown, and abnormal communication.

[0014] In a preferred example, for the equipment operation and maintenance files formed for wind turbine gearboxes, main bearings, pitch mechanisms, and transformer substations, the original inspection records, maintenance records, and fault retest records are first classified into defects, and then defect tag strings are formed according to the order in which the records appear. For example, the defect tag strings in a certain equipment operation and maintenance file after being sorted out are: vibration abnormality, vibration abnormality, vibration abnormality, temperature rise abnormality, temperature rise abnormality, leakage abnormality, leakage abnormality, leakage abnormality, insulation abnormality. The defect tag strings are not a direct splicing of the original text, but an ordered tag arrangement result formed after the original defect records have been uniformly named, synonyms merged, and sorted in order.

[0015] Preferably, this approach allows defect records generated by different recorders, work groups, and maintenance periods to be included in a unified tagging system. Furthermore, it ensures that subsequent segmentation, retrieval, splicing, and merging are all based on the same tag semantics, facilitating a one-to-one correspondence between subsequent retrieval results and the original equipment maintenance records.

[0016] It should be noted that the reason why the method provided in this embodiment first segments the defect tag string in step S1, instead of directly searching the entire defect tag string, is that when the number of equipment maintenance files is large, comparing the entire string line by line will result in a large search range, many duplicate comparisons, and a large amount of cross-node retrieval. By first segmenting according to continuous tag segments and the connection positions between segments, the long tag string can be split into multiple tag fragments with clear boundaries. Then, in subsequent steps, they are sequentially spliced ​​according to the boundary connection rules, thereby reducing the number of times the entire string is moved and irrelevant comparisons are made. Compared with the existing method of directly comparing the complete defect sequence, this embodiment does not simply rely on full string matching. Instead, it first splits the defect tag string in the equipment maintenance file into tag fragments that can be independently retrieved, sequentially connected, and traceably merged. Then, the fragment boundary markers correspond each tag fragment to the original equipment maintenance file. Therefore, in scenarios where files are distributed, node links are inconsistent, and defect tag strings are long, the retrieval path is clearer, the merging basis is more direct, and the result verification is more convenient.

[0017] S1.1. According to the order of the defect tags in the defect tag string, the previous defect tag and the next defect tag are arranged into adjacent tag pairs in sequence; when the next defect tag is arranged next to the previous defect tag, there are no other defect tags between the adjacent tag pairs, the defect tag categories of the previous defect tag and the next defect tag have not changed, and the previous defect tag and the next defect tag are arranged continuously in the same equipment operation and maintenance file, it is determined that the adjacent tag pair is connected, and the adjacent tag pairs that are connected are divided into continuous tag segments.

[0018] In this embodiment, switching refers to a change in the defect label category between two adjacent defect labels. Specifically, when the label names of the previous defect label and the next defect label are completely identical, it is determined that the defect label category has not switched; when the label names of the previous defect label and the next defect label are inconsistent, it is determined that the defect label category has switched. For example, there is no switching between vibration abnormality and vibration abnormality, but there is a switching between vibration abnormality and temperature rise abnormality.

[0019] It should be noted that this embodiment does not use fuzzy category similarity judgment, nor does it use the method of considering the same superordinate category as not switching. Instead, it uses the complete consistency of tag names as the only judgment rule for not switching. The reason for using this rule is that the equipment operation and maintenance file retrieval is oriented towards the arrangement relationship of defect tag strings. If synonym merging or category approximation judgment is introduced in step S1.1, it will lead to the instability of the boundary of continuous tag segments, thereby affecting the uniqueness of the subsequent segment boundary identification.

[0020] Furthermore, consecutive arrangement within the same equipment maintenance file means that the preceding and following defect labels are adjacent in the label order of the same equipment maintenance file, and there are no other defect labels between them. Taking a defect label string (vibration abnormality, vibration abnormality, vibration abnormality, temperature rise abnormality, temperature rise abnormality, leakage abnormality, leakage abnormality, insulation abnormality) in a certain equipment maintenance file as an example, the first vibration abnormality and the second vibration abnormality form an adjacent label pair, and the second vibration abnormality and the third vibration abnormality form an adjacent label pair. The label names are the same and the positions are adjacent, so these two adjacent label pairs are both determined to be connected and are classified into the same continuous label segment. The third vibration abnormality and the subsequent temperature rise abnormality are determined to be a category switch because the label names are inconsistent, so this forms the boundary of the continuous label segment.

[0021] S1.2. Determine the first defect label in each consecutive label segment as the first defect label, the last defect label in each consecutive label segment as the last defect label, and determine the position between the last defect label of the previous consecutive label segment and the first defect label of the next consecutive label segment as the inter-segment connection position; use the first and last defect labels in each consecutive label segment as the segment range, and the inter-segment connection position as the segmentation position to segment the defect label string to form label fragments.

[0022] In a preferred embodiment, firstly, according to each continuous tag segment obtained in step S1.1, the first defect tag and the last defect tag in each continuous tag segment are determined respectively. Then, the position between the last defect tag of the previous continuous tag segment and the first defect tag of the next continuous tag segment is determined as the inter-segment connection position. Subsequently, the first defect tag to the last defect tag of each continuous tag segment is taken as a complete segment range, and the inter-segment connection position is used as the cutting position to cut the entire defect tag string segment by segment to obtain multiple tag pieces. In other words, the cutting is not to split the continuous tag segment again, but to separate adjacent continuous tag segments with the inter-segment connection position as the cutting position, so that each tag piece is composed of continuous defect tags of the same type.

[0023] For example, if the defect tag string in a certain equipment maintenance file is: vibration abnormality, vibration abnormality, vibration abnormality, temperature rise abnormality, temperature rise abnormality, leakage abnormality, leakage abnormality, leakage abnormality, insulation abnormality, then step S1.1 obtains four consecutive tag segments, namely the first consecutive tag segment: vibration abnormality, vibration abnormality, vibration abnormality; the second consecutive tag segment: temperature rise abnormality, temperature rise abnormality; the third consecutive tag segment: leakage abnormality, leakage abnormality, leakage abnormality; and the fourth consecutive tag segment: insulation abnormality. Then, the position between the last defect tag of the first consecutive tag segment and the first defect tag of the second consecutive tag segment is used as the basis for... The first inter-segment connection position is defined as the position between the end defect label of the second consecutive label segment and the beginning defect label of the third consecutive label segment. The second inter-segment connection position is defined as the position between the end defect label of the third consecutive label segment and the beginning defect label of the fourth consecutive label segment. After the entire defect label string is divided according to these three inter-segment connection positions, four label segments are formed: label segment 1: vibration abnormality, vibration abnormality, vibration abnormality; label segment 2: temperature rise abnormality, temperature rise abnormality; label segment 3: leakage abnormality, leakage abnormality, leakage abnormality; and label segment 4: insulation abnormality.

[0024] Preferably, this segmentation method ensures that there is no category switching within each tag segment, and the connection positions between each tag segment are clearly located between adjacent consecutive tag segments. Therefore, the sequential splicing in the subsequent step S3 can be directly checked based on the segment termination tag, segment start tag, and segment sequence position, without having to look back at the entire original defect tag string again.

[0025] S1.3. Determine the first defect label in each label segment as the segment start label, the last defect label in each label segment as the segment end label, and determine the order of each label segment in the defect label string as the segment order position. Generate segment boundary identifiers based on the segment start label, segment end label and segment order position.

[0026] Specifically, the defect tag string is scanned sequentially from the beginning to the end. When the first tag fragment is encountered, its segment position is recorded as 1; when the second tag fragment that follows is encountered, its segment position is recorded as 2; and so on, until the last tag fragment in the entire defect tag string. The segment position only reflects the order of the tag fragments in the defect tag string of the same equipment maintenance file, and does not carry timestamps, equipment status values ​​or text description information.

[0027] Furthermore, the starting label, ending label, and segment sequence position of each tag segment are extracted sequentially, and these three items are combined in a fixed order to form the segment boundary identifier of that tag segment. For example, if tag segment 1 is: Vibration Abnormal, Vibration Abnormal, Vibration Abnormal, then its starting label is Vibration Abnormal, its ending label is Vibration Abnormal, and its segment sequence position is 1. Therefore, the segment boundary identifier of this tag segment can be recorded as Vibration Abnormal-Vibration Abnormal-1. If tag segment 2 is: Temperature Rise Abnormal, Temperature Rise Abnormal, then its segment boundary identifier can be recorded as Temperature Rise Abnormal-Temperature Rise Abnormal-2. If tag segment 3 is: Leakage Abnormal, Leakage Abnormal, Leakage Abnormal, then its segment boundary identifier can be recorded as Leakage Abnormal-Leakage Abnormal-3. If tag segment 4 is Insulation Abnormal, then its segment boundary identifier can be recorded as Insulation Abnormal-Insulation Abnormal-4. In the scenario of distributed storage of equipment operation and maintenance records, this segment boundary identifier can also be recorded together with the equipment number, thereby forming a correspondence between equipment number and segment boundary identifier.

[0028] It should be noted that this embodiment extracts continuous segments of categories from the entire defect tag string independently and fixes the category switching position as the boundary position of subsequent splicing, thereby transforming the entire string retrieval problem into a layer-by-layer retrieval problem of segmented retrieval + boundary connection + merging verification. Compared with the processing method of directly treating the complete tag string as a single retrieval object, this embodiment provides segmentation criteria and boundary criteria before retrieval. Therefore, in cases of multi-node distribution, local missing segments, and different link states, it is easier to obtain local candidate results first and then gradually merge them into complete matching results.

[0029] S2. Based on the link status of each retrieval node storing the tag fragments, a link status level sequence is formed, and a tag fragment retrieval sequence is generated according to the link status level sequence. Note that the following should be noted in this step: In this embodiment, each retrieval node can be an equipment operation and maintenance center server, a regional operation and maintenance substation server, a station-side edge storage host, or a maintenance team archiving terminal; each retrieval node stores at least one tag fragment and its corresponding fragment boundary identifier; in a distributed equipment operation and maintenance file management scenario, tag fragments of different equipment operation and maintenance files may be distributed across multiple retrieval nodes according to region, time, or equipment category; if the link status of each retrieval node is not considered and tag fragments are retrieved directly in a fixed node order, the first retrieved node may return slowly, have missing fragments, or be disconnected, thereby slowing down the acquisition of subsequent candidate tag fragments; therefore, in step S2 of this embodiment, the link status of each retrieval node is first sorted out, and then a tag fragment retrieval sequence is generated according to the link status level sequence, so that the order of acquiring candidate tag fragments in step S3 corresponds to the availability of the current node.

[0030] S2.1 Extract the link connectivity status, link transmission delay, and missing tag fragment return status from each retrieval node storing tag fragments, and form a link status group based on the link connectivity status, link transmission delay, and missing tag fragment return status of the same retrieval node.

[0031] For each retrieval node, a tag fragment retrieval request is first sent to that node. If the retrieval node returns a retrieval response within 2 seconds, the link connectivity status of that retrieval node is determined to be connected. If the retrieval node does not return a retrieval response within 2 seconds, the link connectivity status of that retrieval node is determined to be disconnected. Subsequently, for retrieval nodes with a connected link connectivity status, the time difference from sending the tag fragment retrieval request to receiving the first returned fragment start data from that retrieval node is calculated, and this time difference is determined as the link transmission delay.

[0032] For example, the link transmission delay of retrieval node N1 is 18 ms, the link transmission delay of retrieval node N2 is 42 ms, and the link transmission delay of retrieval node N3 is 95 ms.

[0033] For tag fragment return missing status, this embodiment adopts two statuses: complete return and missing fragment return. If the tag fragment returned by the retrieval node contains all the response tag fragments registered and stored by the node, it is determined to be a complete return. If the retrieval node does not return all the response tag fragments registered and stored by it, or if the returned result contains registered response fragments but does not actually include tag fragment content, it is determined to be a missing fragment return. Subsequently, the link connectivity status, link transmission delay, and tag fragment return missing fragment status of the same retrieval node are combined in a fixed order to form the link status group corresponding to that retrieval node. For example, the link status group of retrieval node N1 can be complete return, connected, 18 ms; the link status group of retrieval node N2 can be complete return, connected, 42 ms; the link status group of retrieval node N3 can be missing fragment return, connected, 95 ms; and the link status group of retrieval node N4 can be missing fragment return, disconnected, no delay value.

[0034] S2.2 First, arrange the search nodes in order of their missing fragment status according to the tag fragmentation return status. Then, arrange the search nodes with the same missing fragment status according to the link connectivity status. Finally, arrange the search nodes with the same missing fragment status and link connectivity status according to the link transmission delay to form a link status level sequence.

[0035] In this embodiment, the link status level is divided into four levels: Level 1 is complete return and connectivity; Level 2 is missing fragment return and connectivity; Level 3 is a delayed node that returns completely but only resumes connectivity after a response time exceeding 2 seconds; and Level 4 is disconnected. Specifically, the nodes are first arranged according to the missing fragment return status, with complete returns preceding missing fragment returns; then, they are arranged according to the connectivity status, with connectivity preceding disconnected nodes if both return completely or with missing fragments; finally, they are arranged according to the link transmission delay, with nodes having shorter transmission delays preceding nodes with longer transmission delays if both return missing fragments and connectivity are the same. The resulting link status level sequence is a sequence of search nodes arranged from highest to lowest priority.

[0036] For example, among the four retrieval nodes N1, N2, N3, and N4, N1 is a complete return, connected, and 18 ms; N2 is a complete return, connected, and 42 ms; N3 is a missing fragment return, connected, and 95 ms; and N4 is a missing fragment return, disconnected, and has no delay value. The resulting link status level sequence is N1, N2, N3, and N4.

[0037] The link status level sequence is the result of sequential arrangement of retrieval node identifier + link status group + level order. In actual storage, the link status level sequence can also be based solely on the order of retrieval node identifiers, with the corresponding link status group as an auxiliary record of the arrangement result. The reason for adopting this sorting method is that the subsequent step S3 of this invention requires sequential retrieval of tag fragments from each retrieval node. If nodes that return complete results and have low latency are retrieved first, it is easier to obtain tag fragments with complete boundaries and stable retrieval, thereby accelerating the formation of the first candidate tag fragment and subsequent candidate tag fragments.

[0038] S2.3. Arrange the tag fragments stored in each retrieval node in sequence according to the order in the link status level sequence to generate a tag fragment retrieval sequence.

[0039] Specifically, the retrieval order of the search nodes is first determined according to the link status level sequence. Then, within each search node, the tag fragments stored in that search node are arranged in ascending order of their fragment position. Subsequently, the arrangement results of all tag fragments in the previous search node are concatenated with the arrangement results of all tag fragments in the next search node to form the overall tag fragment retrieval sequence. For example, if search node N1 stores tag fragments Vibration Anomaly-Vibration Anomaly-1 and Leakage Anomaly-Leakage Anomaly-3, search node N2 stores tag fragments Temperature Rise Anomaly-Temperature Rise Anomaly-2, and search node N3 stores tag fragments Insulation Anomaly-Insulation Anomaly-4, and the link status level sequence is N1, N2, N3, then the tag fragment retrieval sequence can be Vibration Anomaly-Vibration Anomaly-1, Leakage Anomaly-Leakage Anomaly-3, Temperature Rise Anomaly-Temperature Rise Anomaly-2, and Insulation Anomaly-Insulation Anomaly-4. If multiple fragments of the same type exist within a node, they are still arranged in ascending order of their fragment position.

[0040] It should be noted that by incorporating the network-side condition of node link status into the retrieval process in advance, the order in which subsequent candidate tag fragments are obtained is constrained by both the splicing capability of fragment boundaries and the retrieval capability of nodes. Compared to existing fixed-node polling or average retrieval methods, this embodiment does not initiate retrieval in the same order for each retrieval node. Instead, it first completes the link status classification and then prioritizes retrieving tag fragments from nodes that have returned complete information, are connected, and have low latency. This reduces the failure of the first round of retrieval due to disconnected nodes, missing fragment nodes, and high-latency nodes. It also reduces the number of times fragments are repeatedly spliced ​​due to missing fragments during the formation of subsequent candidate tag strings. In an environment where device operation and maintenance files are stored in multiple nodes, the source of tag fragments that are more likely to constitute a complete candidate tag string can be identified more quickly.

[0041] S3. Based on the defect label string and label fragment retrieval sequence, extract candidate label fragments from the corresponding retrieval nodes, and sequentially concatenate them according to the fragment boundary markers to form a candidate label string set. Note that the following should be noted in this step: S3.1. Retrieve tag fragments sequentially from each retrieval node according to the order of tag fragment retrieval. Compare the starting tag of the fragment with the first tag in the defect tag string to be inspected. The tag fragment with the same starting tag is determined as the first candidate tag fragment.

[0042] Specifically, starting from the first item in the tag segment retrieval sequence, tag segments are retrieved sequentially. For each tag segment retrieved, the starting tag of that segment is extracted and compared with the first tag in the defect tag string to be inspected. When the tag names are completely identical, the tag segment is determined to meet the starting condition and is identified as the first candidate tag segment. When the tag names are inconsistent, the next tag segment is retrieved according to the tag segment retrieval sequence until a tag segment that meets the starting condition is found.

[0043] In this embodiment, a completely consistent comparison means that the character names are completely identical, and no fuzzy similarity, alias substitution, or higher-level class merging is used for judgment. For example, when the first label of the defect label string to be inspected is temperature rise abnormality, only the label segments whose starting labels are also temperature rise abnormality can be determined as the first candidate label segments. Label segments whose starting labels are bearing temperature rise abnormality or high temperature abnormality are not considered as the first candidate label segments.

[0044] For example, if the defect label string to be inspected is: abnormal temperature rise, abnormal temperature rise, abnormal leakage, abnormal leakage, abnormal leakage, then when the label segment abnormal temperature rise - abnormal temperature rise - 2 is encountered in the label segment retrieval sequence, since its segment start label abnormal temperature rise is consistent with the first label abnormal temperature rise of the defect label string to be inspected, the label segment is determined as the first candidate label segment.

[0045] It should be noted that the defect tag string to be inspected in this embodiment is a sequence of target defect tags given by the retrieval initiator based on the current retrieval conditions, and the defect tag string to be inspected adopts the same tag naming rule as the defect tags in the equipment operation and maintenance file.

[0046] S3.2. Based on the segment termination label and segment order position of the first candidate tag segment, perform connection checks on the subsequently retrieved tag segments; when the segment start label of the next tag segment is adjacent to the segment termination label of the previous tag segment, and the segment order position of the next tag segment is immediately following the segment order position of the previous tag segment, the next tag segment is determined as the subsequent candidate tag segment.

[0047] In this embodiment, the tag fragments retrieved subsequently refer to the remaining tag fragments retrieved following the tag fragment retrieval sequence after the first candidate tag fragment.

[0048] Furthermore, the method for performing connectivity checks on subsequently retrieved tag segments specifically includes: using the segment termination label and segment sequence position of the previous candidate tag segment as a benchmark, checking each subsequently retrieved tag segment to see if it simultaneously meets the following two conditions: The first condition is that the segment start label of the subsequent tag segment is immediately after the termination position of the previous candidate tag segment in the defective tag string; the second condition is that the segment sequence position of the subsequent tag segment immediately follows the segment sequence position of the previous candidate tag segment; for example, the segment sequence position of the previous candidate tag segment... If the value is 2, then the segment order position of the subsequent candidate label segment should be 3. If the segment order position of the subsequently retrieved label segment is 4, then the segment order continuity condition is not met. Specifically, first determine the termination position of the previous candidate label segment based on its coverage in the defect label string to be inspected, and then check whether the starting label of the subsequent label segment is consistent with the next position label in the defect label string to be inspected. If they are consistent, then check whether its segment order position is the segment order position of the previous candidate label segment plus 1. Only when both conditions are met simultaneously will the subsequent label segment be determined as the subsequent candidate label segment.

[0049] Taking the defect label string to be inspected (temperature rise abnormality, temperature rise abnormality, leakage abnormality, leakage abnormality, leakage abnormality) as an example, if the first candidate label segment is temperature rise abnormality-temperature rise abnormality-2, which covers the first two labels of the defect label string to be inspected, then the subsequent segment should be the label segment whose starting label is leakage abnormality and whose segment sequence position is 3. When leakage abnormality-leakage abnormality-3 is retrieved, since its starting label is consistent with the third label of the defect label string to be inspected, and its segment sequence position is immediately after 2, it can be determined as the subsequent candidate label segment.

[0050] S3.3. According to the determination order of the first candidate label segment and each subsequent candidate label segment, the adjacent candidate label segments are spliced ​​together in sequence, and the boundary connection check is performed on the segment termination label and the segment start label at the splicing point to obtain the candidate label string.

[0051] In a preferred embodiment, candidate segment sequences are first formed by arranging the first candidate tag segment and each subsequent candidate tag segment in the order of their determination. Then, adjacent two tag segments in the candidate segment sequence are spliced ​​together end to end. During splicing, all the tag arrangements of the previous candidate tag segment are retained, and all the tag arrangements of the next candidate tag segment are appended after the previous candidate tag segment to form the spliced ​​tag arrangement.

[0052] Specifically, the method for boundary connection checking is as follows: check whether the termination label of the previous candidate label segment and the starting label of the next candidate label segment are in an adjacent position relationship in the defect label string to be inspected; if they are in an adjacent position relationship, the boundary connection is determined to be successful; if they are not in an adjacent position relationship, the splicing path is terminated; for example, the defect label string to be inspected is: abnormal temperature rise, abnormal temperature rise, abnormal leakage, abnormal leakage, abnormal leakage, abnormal leakage, abnormal leakage, abnormal the previous candidate label segment is: abnormal temperature rise, abnormal temperature rise, abnormal leakage, abnormal leakage, abnormal leakage, abnormal leakage, abnormal leakage, abnormal temperature rise and abnormal leakage are adjacent in the defect label string to be inspected, so the boundary connection check at this splicing point is passed, and the candidate label string obtained after splicing is: abnormal temperature rise, abnormal temperature rise, abnormal leakage, abnormal leakage, abnormal leakage, abnormal leakage.

[0053] S3.4. Perform tag count, first tag, and last tag checks on each candidate tag string and the defect tag string to be inspected. When the candidate tag string and the defect tag string to be inspected have the same number of tags, the same first tag, and the same last tag, the candidate tag string is included in the candidate tag string set.

[0054] Specifically, the tag count check involves counting the total number of tags in the candidate tag string and comparing it with the total number of tags in the defective tag string to be inspected. If the two counts are the same, the tag count check is considered passed. The first tag check involves extracting the first tag in the candidate tag string and comparing it with the first tag in the defective tag string to be inspected. If the two counts are the same, the first tag check is considered passed. The last tag check involves extracting the last tag in the candidate tag string and comparing it with the last tag in the defective tag string to be inspected. If the two counts are the same, the last tag check is considered passed. Only when a candidate tag string passes the tag count check, the first tag check, and the last tag check simultaneously is the candidate tag string included in the candidate tag string set.

[0055] It should be noted that this embodiment moves the work that originally required comparing the complete equipment operation and maintenance records line by line to a layer-by-layer screening of the tag segment boundaries and order relationships. It first eliminates irrelevant paths with mismatched starting tags, discontinuous segment order, unconnected boundaries, and mismatched lengths, and only retains candidate tag strings that meet the complete candidate conditions. Compared with the existing method of performing a full bit-by-bit comparison of the entire defect sequence, this embodiment has already completed multiple rounds of shrinkage through the first tag, segment order position, boundary connection, and number of tags before the complete connection consistency check. Therefore, the number of candidate results entering step S4 is less, and the subsequent merging and verification is more direct; thus reducing irrelevant files entering the final comparison stage.

[0056] S4. Perform connection consistency verification and merging comparison on the candidate tag string set and the defect tag string to be inspected, and output the equipment operation and maintenance file that matches the defect tag string to be inspected. Note that the following should be noted in this step: S4.1 Extract adjacent tag pairs from each candidate tag string in sequence, and extract the corresponding adjacent tag pairs according to the order of arrangement in the defect tag string to be inspected; S4.2. Verify each pair of adjacent labels in each candidate label string against the corresponding pair of adjacent labels in the defect label string to be inspected. When the preceding label of an adjacent label pair is consistent, the following label is consistent, and the order of the labels is consistent, the adjacent label pair is determined to be connected. S4.3 Count the adjacent tag pairs that are connected in each candidate tag string; when all adjacent tag pairs in the same candidate tag string are connected, the candidate tag string is determined as the merged candidate tag string. S4.4 Perform a merge comparison on each candidate tag string; when the first tag, the last tag, and the order of each adjacent tag pair are consistent among multiple candidate tag strings, merge the multiple candidate tag strings into the target tag string and determine the target tag string as the merged tag string corresponding to the defect tag string to be inspected. S4.5. Based on the fragment boundary identifiers of each candidate tag fragment that makes up the target tag string to be merged, determine the equipment operation and maintenance file to which each candidate tag fragment belongs. Specifically, first, the fragment boundary identifier corresponding to each candidate tag fragment that makes up the target tag string to be merged is extracted, and then a backtracking search is performed based on the equipment maintenance file fragment registration table. The equipment maintenance file fragment registration table records the corresponding relationship between the equipment number, the equipment maintenance file number, the fragment boundary identifier, and the storage and retrieval node. Therefore, when the fragment boundary identifier of a certain candidate tag fragment is obtained, the equipment maintenance file number corresponding to the fragment boundary identifier can be located in the equipment maintenance file fragment registration table. For example, the fragment boundary identifier Temperature Rise Anomaly - Temperature Rise Anomaly - 2 corresponds to equipment maintenance file A-2026-017 in the equipment maintenance file fragment registration table, and the fragment boundary identifier Leakage Anomaly - Leakage Anomaly - 3 also corresponds to equipment maintenance file A-2026-017. Then it can be determined that these two candidate tag fragments belong to the same equipment maintenance file A-2026-017. If the boundary identifier of another candidate tag fragment corresponds to equipment maintenance file B-2026-041, it indicates that the candidate tag fragment does not belong to the aforementioned equipment maintenance file. Preferably, by using the segment boundary marker to trace back the equipment operation and maintenance records, the candidate tag strings can be re-corresponded to the actual record entities, avoiding the problem of only completing the matching at the tag level but failing to fall into the specific operation and maintenance records; S4.6 Merge and verify the candidate tag segments belonging to the same equipment operation and maintenance file, and determine the equipment operation and maintenance file containing the complete target tag string as the equipment operation and maintenance file that matches the defect tag string to be inspected; Specifically, the candidate tag segments belonging to the same equipment maintenance file are merged and verified, including: S4.6.1 Arrange the candidate tag fragments in order of their segment positions and check whether the segment positions of adjacent candidate tag fragments are continuous. S4.6.2 For adjacent candidate tag segments with consecutive segment order positions, check whether the segment termination tag of the previous candidate tag segment is connected to the segment start tag of the next candidate tag segment. S4.6.3 Merge candidate tags that are consecutive in sequence and whose segment termination tag is connected to the segment start tag. Then, check the number of tags, the first tag, the last tag, and the order of adjacent tag pairs against the merged tag string. When the number of tags, the first tag, the last tag, and the order of adjacent tag pairs are consistent with the merged tag string, the equipment maintenance file to which the merged tag file belongs is determined to be the matching equipment maintenance file. S4.7 Output the equipment maintenance file that matches the defect label string to be inspected, and the target label string to be merged in the equipment maintenance file.

[0057] It should be noted that the candidate tag string set formed in step S3 only completes the screening of the first and last conditions, length conditions, and fragment boundary connection conditions. However, there may still be cases where tag fragments in different equipment maintenance files happen to have the same first tag, last tag, and number of tags. Therefore, in step S4, it is necessary to further check whether the order of adjacent tag pairs is completely consistent, and to merge and verify the candidate tag fragments from the same equipment maintenance file to exclude cases of cross-file splicing or local sequence consistency but inconsistent overall source. Compared with the method of directly giving the file result based solely on the consistency between the candidate tag string and the defect tag string to be inspected, this embodiment adds two layers of verification: tag connection consistency verification and same file merging verification. Therefore, the correspondence between the output equipment maintenance file and the merged target tag string is more stable, thereby eliminating false matching results formed by mixing fragments from different equipment maintenance files.

[0058] The following is a preferred example of steps S3 to S4, based on a specific application scenario. Assume a wind farm's equipment maintenance archive contains three equipment maintenance files: A-2026-017, B-2026-041, and C-2026-062. The defect tag string corresponding to equipment maintenance file A-2026-017 is: vibration abnormality, vibration abnormality, temperature rise abnormality, temperature rise abnormality, leakage abnormality, leakage abnormality, leakage abnormality, insulation abnormality. After segmentation in step S1, four tag segments are formed: tag segment A1 (vibration abnormality, vibration abnormality), with the segment boundary marked as vibration abnormality-vibration abnormality-1; tag segment A2 (temperature rise abnormality, temperature rise abnormality), with the segment boundary marked as temperature rise abnormality-temperature rise abnormality-2; tag segment A3 (leakage abnormality, leakage abnormality, leakage abnormality), with the segment boundary marked as leakage abnormality-leakage abnormality-3; tag segment A4... The insulation is abnormal, and the segment boundary is marked as insulation abnormality-insulation abnormality-4; the defect tag string corresponding to equipment maintenance file B-2026-041 is temperature rise abnormality, temperature rise abnormality, leakage abnormality, leakage abnormality, noise abnormality; after segmentation, it forms tag segment B1 temperature rise abnormality, temperature rise abnormality, segment boundary marked as temperature rise abnormality-temperature rise abnormality-1; tag segment B2 leakage abnormality, leakage abnormality, segment boundary marked as leakage abnormality-leakage abnormality-2; tag segment B3 noise abnormality, segment boundary marked as noise abnormality-noise abnormality-3; the defect tag string corresponding to equipment maintenance file C-2026-062 is temperature rise abnormality, temperature rise abnormality, leakage abnormality, leakage abnormality, leakage abnormality; after segmentation, it forms tag segment C1 temperature rise abnormality, temperature rise abnormality, segment boundary marked as temperature rise abnormality-temperature rise abnormality-1; tag segment C2 leakage abnormality, leakage abnormality, leakage abnormality, segment boundary marked as leakage abnormality-leakage abnormality-2.

[0059] Assume that retrieval node N1 stores tag fragments A2 and A3, retrieval node N2 stores tag fragments B1, B2, and B3, and retrieval node N3 stores tag fragments C1 and C2. Step S2.1 yields the following link status groups: N1 is complete return, connected, 21 ms; N2 is complete return, connected, 37 ms; and N3 is missing fragment return, connected, 82 ms. Following the sorting rules in step S2.2, the resulting link status level sequence is N1, N2, N3. Therefore, the generated tag fragment retrieval sequence can be A2, A3, B1, B2, B3, C1, C2. At this point, the defect tag string provided by the retrieval initiator is abnormal temperature rise, abnormal temperature rise, abnormal leakage, abnormal leakage, abnormal leakage.

[0060] First, A2 is retrieved according to the tag segmentation sequence. Since the starting tag of A2 is "Temperature Rise Anomaly," which matches the first tag "Temperature Rise Anomaly" in the defect tag string to be inspected, A2 is determined as the first candidate tag segment. Based on the ending tag "Temperature Rise Anomaly" and the segment sequence position 2 of A2, the subsequent retrieved tag segments are checked for connectivity. When A3 is retrieved, the starting tag of A3 is "Leakage Anomaly," which matches the next tag in the defect tag string to be inspected that immediately follows "Temperature Rise Anomaly." Furthermore, the segment sequence position of A3 is 3, which immediately follows the segment sequence position 2 of A2. Therefore, A3 is determined as a subsequent candidate tag segment. The candidate tag string obtained by concatenating A2 and A3 is "Temperature Rise Anomaly, Temperature Rise Anomaly, Leakage Anomaly, Leakage Anomaly, Leakage Anomaly." The tag count, first tag, and last tag checks are performed on this candidate tag string. The results are as follows: the tag count is 5 for all segments, the first tag is "Temperature Rise Anomaly" for all segments, and the last tag is "Leakage Anomaly" for all segments. Therefore, this candidate tag string is included in the candidate tag string set.

[0061] Continuing with the tag segment retrieval sequence, B1 is retrieved. The starting tag for B1 is also "temperature rise anomaly," therefore B1 is identified as another candidate tag segment. Subsequently, when retrieving B2, its starting tag is "leakage anomaly," and its segment sequence position 2 immediately follows B1's segment sequence position 1. Therefore, B2 can be used as a subsequent candidate tag segment for B1. The candidate tag string obtained by concatenating B1 and B2 is "temperature rise anomaly, temperature rise anomaly, leakage anomaly, leakage anomaly." However, in step S3.4, the number of tags in this candidate tag string is 4, which is different from the number of tags 5 in the defect tag string to be inspected. Since they are consistent, they are not included in the candidate tag string set. When C1 is retrieved again, C1 can also be identified as the first candidate tag fragment. When C2 is retrieved, the starting tag of C2's fragment is leakage anomaly, and its fragment sequence position 2 immediately follows C1. Therefore, the candidate tag string obtained by splicing C1 and C2 is temperature rise anomaly, temperature rise anomaly, leakage anomaly, leakage anomaly, leakage anomaly. This candidate tag string also passes the three checks in step S3.4 and is included in the candidate tag string set. So far, there are a total of 2 candidate tag strings in the candidate tag string set, which come from A2+A3 and C1+C2 respectively.

[0062] The adjacent label pairs of the two candidate label strings and the defect label string to be inspected were checked for consistency. The adjacent label pairs in the defect label string to be inspected were, in order, temperature rise anomaly-temperature rise anomaly, temperature rise anomaly-leakage anomaly, leakage anomaly-leakage anomaly, and leakage anomaly-leakage anomaly. The adjacent label pairs in the candidate label string formed by A2+A3 were completely consistent with this, so all adjacent label pairs were determined to be consistent. The adjacent label pairs in the candidate label string formed by C1+C2 were also completely consistent with this, so all of them were also determined to be consistent. Thus, the two candidate label strings corresponding to A2+A3 and C1+C2 were both determined to be merged candidate label strings.

[0063] The two candidate tag strings to be merged are compared and merged. Since the first tag of both is abnormal temperature rise and the last tag is abnormal leakage, and the order of each adjacent tag pair is consistent, they can be regarded as candidate paths to be merged corresponding to the defect tag string to be inspected. However, steps S4.5 and S4.6 need to further check whether the equipment operation and maintenance file to which they belong can form a complete target tag string to be merged within the file. According to the equipment operation and maintenance file segment registration table, A2 and A3 both belong to equipment operation and maintenance file A-2026-017, and C1 and C2 both belong to equipment operation and maintenance file C-2026-062. The candidate tag segments A2 and A3 in equipment operation and maintenance file A-2026-017 are arranged in the order of segment position to obtain segment position 2 and 3, which are consecutive. The candidate tag segments C1 and C2 in equipment operation and maintenance file C-2026-062 are arranged in the order of segment position to obtain segment position 1 and 2, which are consecutive. Check if there is a connection between the segment termination label "Temperature Rise Anomaly" in A2 and the segment start label "Leakage Anomaly" in A3; the result is that they are connected. Check if there is a connection between the segment termination label "Temperature Rise Anomaly" in C1 and the segment start label "Leakage Anomaly" in C2; the result is also that they are connected. Compare the merged label arrangement of A2 and A3 with the target label string. Both have 5 labels, the first label is "Temperature Rise Anomaly" in both cases, the last label is "Leakage Anomaly" in both cases, and the adjacent label pairs are arranged in the same order. Therefore, equipment maintenance file A-2026-017 is identified as a matching equipment maintenance file. Similarly, equipment maintenance file C-2026-062 is also identified as a matching equipment maintenance file. Finally, output the equipment maintenance files A-2026-017 and C-2026-062 that match the defect label string to be inspected, and simultaneously output the target label string for merging in their respective equipment maintenance files as "Temperature Rise Anomaly, Temperature Rise Anomaly, Leakage Anomaly, Leakage Anomaly, Leakage Anomaly".

[0064] As can be seen from the above example, steps S3 to S4 are not simply a matching process at the label level. Instead, they first constrain the fragment retrieval order with the link status level sequence, then constrain the candidate path formation with the fragment start label, fragment end label, and fragment order position, and finally check the connection consistency and file merging through adjacent labels to output the final matched device maintenance file. This processing method makes the device maintenance file retrieval results not only have similarity at the label level, but also a complete correspondence at the source file level. Therefore, it is more suitable for file retrieval of historical defect evolution paths in device maintenance scenarios.

[0065] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for retrieving equipment operation and maintenance records based on defect tag string merging, characterized in that, include: The defect tag strings in the equipment operation and maintenance records are segmented according to the tag order, continuous tag segments, and inter-segment connection relationships to form tag fragments, and fragment boundary identifiers are generated. Based on the link status of each retrieval node storing the tag fragments, a link status level sequence is formed, and a tag fragment retrieval sequence is generated according to the link status level sequence. Based on the defect label string to be inspected and the label fragment retrieval sequence, candidate label fragments are extracted from the corresponding retrieval node and sequentially spliced ​​according to the fragment boundary identifier to form a candidate label string set. The candidate tag string set and the defect tag string to be inspected are checked for connection consistency and merged for comparison, and the equipment operation and maintenance file matching the defect tag string to be inspected is output.

2. The equipment operation and maintenance record retrieval method based on defect tag serialization according to claim 1, characterized in that, Forming the tag fragments includes: According to the order of the defect tags in the defect tag string, the previous defect tag and the next defect tag are sequentially formed into adjacent tag pairs; when the next defect tag is arranged immediately next to the previous defect tag, there are no other defect tags between the adjacent tag pairs, the defect tag categories of the previous defect tag and the next defect tag have not changed, and the previous defect tag and the next defect tag are arranged continuously in the same equipment operation and maintenance file, it is determined that the adjacent tag pair is connected, and the connected adjacent tag pairs are divided into continuous tag segments; The first defect label in each consecutive label segment is determined as the first defect label, the last defect label in each consecutive label segment is determined as the last defect label, and the position between the last defect label of the previous consecutive label segment and the first defect label of the next consecutive label segment is determined as the inter-segment connection position. The defect label string is segmented using the first and last defect labels in each consecutive label segment as the segment range and the inter-segment connection position as the segmentation position to form the label fragment.

3. The equipment operation and maintenance record retrieval method based on defect tag serialization according to claim 2, characterized in that, Generating the fragment boundary identifier includes: The first defect label in each tag segment is determined as the segment start label, the last defect label in each tag segment is determined as the segment end label, and the order of each tag segment in the defect label string is determined as the segment order position. The segment boundary identifier is generated based on the segment start label, the segment end label and the segment order position.

4. The equipment operation and maintenance record retrieval method based on defect tag serialization according to claim 1, characterized in that, Forming the link state level sequence includes: For each retrieval node storing the tag fragments, extract the link connectivity status, link transmission delay, and tag fragment return missing status, and form a link status group based on the link connectivity status, link transmission delay, and tag fragment return missing status of the same retrieval node; First, the search nodes are arranged in order of their missing fragment status according to the tag fragmentation. Then, the search nodes with the same missing fragment status according to the link connectivity status are arranged in order of their missing fragment status. Finally, the search nodes with the same missing fragment status and link connectivity status according to the link transmission delay are arranged in order of their missing fragment status and link connectivity status, thus forming the link status level sequence.

5. The equipment operation and maintenance record retrieval method based on defect tag serialization according to claim 4, characterized in that, Generate a tag fragment retrieval sequence according to the link state level sequence, including: According to the order in the link status level sequence, the tag fragments stored in each retrieval node are arranged sequentially to generate the tag fragment retrieval sequence.

6. The equipment operation and maintenance record retrieval method based on defect tag serialization according to claim 1, characterized in that, Extracting the candidate tag fragments includes: According to the order of the tag fragment retrieval sequence, the tag fragments are retrieved sequentially in each retrieval node, and the starting tag of the fragment is compared with the first tag in the defect tag string to be inspected. The tag fragments with the same starting tag are determined as the first candidate tag fragments. Based on the segment termination label and segment sequence position of the first candidate tag segment, a connection check is performed on the subsequently retrieved tag segments; when the segment start label of the next tag segment is adjacent to the segment termination label of the previous tag segment, and the segment sequence position of the next tag segment is immediately following the segment sequence position of the previous tag segment, the next tag segment is determined as the subsequent candidate tag segment.

7. The equipment operation and maintenance record retrieval method based on defect tag serialization according to claim 6, characterized in that, Forming the candidate tag string set includes: According to the determination order of the first candidate label segment and each subsequent candidate label segment, adjacent candidate label segments are spliced ​​together in sequence, and the boundary connection check is performed on the segment termination label and segment start label at the splicing point to obtain the candidate label string. Each candidate tag string is compared with the defect tag string to be inspected by checking the number of tags, the first tag, and the last tag. When the number of tags, the first tag, and the last tag are the same as those of the defect tag string to be inspected, the candidate tag string is included in the candidate tag string set.

8. The equipment operation and maintenance record retrieval method based on defect tag serialization according to claim 1, characterized in that, The connection consistency check and merging comparison between the candidate tag string set and the defect tag string to be inspected include: The adjacent tag pairs in each candidate tag string are extracted sequentially, and the corresponding adjacent tag pairs are extracted according to the order of the arrangement in the defect tag string to be inspected. Each adjacent tag pair in the candidate tag string is checked against the corresponding adjacent tag pair in the defect tag string to be inspected; when the preceding tag of the adjacent tag pair is the same, the following tag is the same, and the order of the tags is the same, the adjacent tag pair is determined to be connected. Count the adjacent tag pairs that are connected in each candidate tag string; when all adjacent tag pairs in the same candidate tag string are connected, the candidate tag string is determined as the merge candidate tag string. Merge and compare each candidate tag string; when the first tag, the last tag, and the order of each adjacent tag pair are consistent among multiple candidate tag strings, merge the multiple candidate tag strings into a target tag string, and determine the target tag string as the merged tag string corresponding to the defect tag string to be inspected.

9. The equipment operation and maintenance record retrieval method based on defect tag serialization according to claim 8, characterized in that, Output the equipment maintenance file that matches the defect label string to be inspected, including: Based on the fragment boundary identifier of each candidate tag fragment that makes up the merged target tag string, determine the device operation and maintenance file to which each candidate tag fragment belongs; The candidate tag segments belonging to the same equipment operation and maintenance file are merged and verified, and the equipment operation and maintenance file containing the complete target tag string is determined as the equipment operation and maintenance file that matches the defect tag string to be inspected; Output the equipment maintenance file that matches the defect tag string to be inspected and the target tag string to be merged in the equipment maintenance file.

10. The equipment operation and maintenance record retrieval method based on defect tag serialization according to claim 9, characterized in that, The process of merging and verifying candidate tag fragments belonging to the same equipment maintenance file includes: Arrange the candidate tag segments in the order of their segments, and check whether the segments of adjacent candidate tag segments are consecutive. For adjacent candidate tag segments with consecutive segment order positions, check whether the segment termination tag of the previous candidate tag segment is connected to the segment start tag of the next candidate tag segment. Candidate tags whose segment order is consecutive and whose segment termination tag is connected to the segment start tag are merged. The merged tag arrangement is then compared with the target tag string for tag count, first tag, last tag, and adjacent tag pair order. When the merged tag arrangement matches the target tag string in terms of tag count, first tag, last tag, and adjacent tag pair order, the corresponding device maintenance file is determined to be the matching device maintenance file.

Images