Maintenance tool closed-loop tracing method and system based on double-phase identification before and after work

Abstract

The application relates to the field of industrial maintenance process management, and in particular to a maintenance tool closed-loop tracing method and system based on pre- and post-operation dual-phase identification. The method comprises the following steps: acquiring tool preposition feature data sets, performing initial attribute registration and change type classification on each tool, and generating tool initial information sets; acquiring tool postposition feature data sets, processing according to the differential collection strategies and feature disintegration logic corresponding to the deformation consumption class, the property change class and the appearance stable class according to the tool category labels, and generating tool change information sets; based on this, performing state matching analysis on each tool, extracting state change characteristics and comparing with maintenance quality correlation rules, and generating a maintenance quality tracing report. The application can realize multi-dimensional perception of the physical form, material properties and microscopic wear of the tool, discover operation abnormalities and potential risks in advance, and improve the accuracy of maintenance quality tracing and closed-loop management capability.

Description

Technical Field

[0001] This application relates to the field of industrial maintenance process management, and in particular to a closed-loop traceability method and system for maintenance tools based on dual-phase identification before and after operation. Background Technology

[0002] In industries with high safety and precision requirements, such as aero-engines, nuclear facilities, semiconductor manufacturing, and precision machining, the management of maintenance tools and the traceability of maintenance quality are crucial for ensuring production safety and product quality. Existing tool traceability systems typically employ technologies such as RFID tags, barcodes, or QR codes to record the borrowing and returning status of tools, and combine this with material coding to achieve identification and quantity verification. Some advanced systems further integrate weighing modules or simple visual inspection devices to confirm overall weight changes or macroscopic appearance integrity of the tools. These technical solutions, by establishing electronic ledgers for tool issuance and return, achieve digital recording of the tool circulation process and, to some extent, prevent tool loss.

[0003] However, current technologies lack effective means to perceive subtle changes in the physical form of tools, potential degradation of material properties, or microscopic surface wear that occur during maintenance operations. Existing traceability logic mainly focuses on the existence and quantity of tools, failing to directly correlate the specific differences in the tool's condition before and after use with the final quality of the maintenance operation. This makes it difficult to detect potential problems such as operational anomalies, environmental pollution residues, or foreign object residues from the evolution of the tool's condition in advance, resulting in a lack of objective data support for maintenance quality evaluation based on changes in the tool itself. Summary of the Invention

[0004] This application provides a closed-loop traceability method and system for maintenance tools based on dual-phase identification before and after operation, in order to solve the above problems.

[0005] In a first aspect, this application provides a closed-loop traceability method for maintenance tools based on dual-phase identification before and after operation, the method comprising: Obtain the tool pre-operation feature dataset for the pre-operation phase of the maintenance operation. Based on the tool pre-operation feature dataset, perform initial attribute registration and change type classification for each tool to generate an initial tool information set. Obtain the tool post-maintenance feature dataset after the maintenance operation. Based on the category label of each tool in the initial tool information set, process it according to the differentiated acquisition strategy and feature deconstruction logic corresponding to the deformation consumption class, property change class and appearance stability class to generate tool change information set. Based on the initial information set and the change information set of the tools, a state matching analysis is performed on each tool according to its category. The state change features of the tool before and after the operation are extracted, and the state change features are compared with the maintenance quality association rules to generate a maintenance quality traceability report.

[0006] Optionally, the process of generating the initial information set for the tool includes: Obtain the tool pre-feature dataset, which includes the shape outline, surface texture, initial quality, and component topology of each tool; Based on the shape and surface texture, and combined with the material coding rules, each tool is initially registered for attributes to generate a unique tool identifier. Based on the component topology and operating environment characteristics, each tool is automatically classified into one of the following categories: deformation consumption, property change, or appearance stability, and a unique tool identifier is bound to it to generate an initial information set for the tool.

[0007] Optionally, the process of generating the tool change information set includes: Based on the category label of each tool in the initial tool information set, differentiated acquisition strategies and feature deconstruction logic are selected for the deformation consumption class, property change class, and appearance stability class, respectively: For deformation-consuming tools, perform deformation feature capture and component topology reconstruction; For tools that involve changes in properties, perform multimodal performance parameter remeasurement and deviation calculation; For tools with stable appearance, perform high-precision registration of micro-textures; The post-feature data obtained from the deformation consumption tools, the property change tools, and the appearance stability tools are summarized to generate the tool change information set.

[0008] Optionally, the process of performing deformation feature capture and component topology reconstruction includes: The point cloud data of the deformation-consuming tool after the operation is collected by a 3D scanning device, and all macroscopic geometric deformation areas in the current outline of the tool that exceed the initial reference range are identified. The point cloud data after the operation is matched with the initial geometric dimensions and relative positions of components recorded in the initial information set of the tool to find residual feature regions and infer the source identity of the deformed components. Simultaneously, the number and shape of all the deformable components are counted, the spatial topological relationship between the components is reconstructed, and the deformation feature record of the deformation consumption tool is generated.

[0009] Optionally, the step of performing multimodal performance parameter remeasurement and deviation calculation includes: The mass data of the tool with the property change type is obtained by intelligent weighing equipment at the post-operation stage, and compared with the initial mass to calculate the mass change. AI visual recognition technology is used to collect images of the tool surface, and deep learning convolution algorithms are used to identify the surface cleanliness level, coating integrity and color characteristics to obtain surface state parameters. The composition and concentration of characteristic gases escaping from the tool surface are collected by a micro gas sensor array, and the chemical changes of the tool material are calculated by matching them with knowledge of chemical degradation gases. The mass change, surface state parameters, and chemical change are used together as performance change characteristics, and compared with preset performance thresholds to classify performance change levels.

[0010] Optionally, the high-precision registration of micro-textures includes: Microscopic surface images of the appearance-stabilized tool in the same region as the tool's initial information set in the subsequent time phase were acquired using a high-resolution imaging device. A feature point cloud registration algorithm is used to match minute scratches, wear patterns or laser-engraved edges in two images before and after the operation, and the matching score is calculated. If the matching degree is lower than the preset threshold, it is determined that the tool has been swapped or has not been used. If the matching degree is qualified, the newly emerging minor damage areas after the operation are further identified, their location and shape are recorded, and the damage feature record of the appearance-stabilized tool is generated.

[0011] Optionally, the process of generating the maintenance quality traceability report includes: Based on the initial information set and the change information set of the tools, a state matching analysis is performed on each tool according to its category to extract the state change features of the tool before and after the operation. The state change characteristics are compared with the maintenance quality association rules one by one to determine the return status, functional failure risk and potential impact on the equipment being maintained for each tool. The maintenance quality association rules include deformation association rules and performance association rules; Based on the comparison results of all tools, the maintenance quality traceability report is generated and output.

[0012] Optionally, the deformation association rules include: Obtain the area of ​​the macroscopic geometric deformation region of the deformation-consuming tool and the spatial topological deviation value between the components after deformation; The total deformation rate of the tool is calculated by superimposing the area of ​​the macroscopic geometric deformation region with the cumulative area of ​​the tool's historical deformation recorded in the tool's initial information set. If the total deformation rate of the tool exceeds the preset plastic deformation threshold of the tool, the deformation correlation result is determined to be forced scrap. If the total deformation rate of the tool is below the plastic deformation threshold, but the spatial topology deviation causes the clearance between the tool and the equipment being repaired to exceed the design tolerance range, then the deformation correlation result is determined to be that the component needs to be replaced.

[0013] Optionally, the performance association rules include: Obtain the performance change level and chemical change amount of the aforementioned property change tools; The chemical change is compared with the initial material properties and historical performance degradation curves recorded in the initial information set of this tool to calculate the predicted remaining life of the current tool material. If the performance change level is severe abnormality and the amount of chemical change exceeds the critical threshold for material degradation, then the performance correlation result is determined to be that the material is unusable due to failure. If the performance change level is slightly abnormal but the amount of chemical change shows a continuous accelerating upward trend and the predicted remaining life of the material is lower than the duration of the next maintenance cycle, then the performance correlation result is determined to be that the material needs to be replaced in advance. If the performance change level is normal but an unexpected gas type appears, the performance correlation result is determined to require in-depth chemical testing.

[0014] This application provides a closed-loop traceability method for maintenance tools based on dual-phase identification before and after maintenance operations. This method constructs a closed-loop condition monitoring system for the entire lifecycle by acquiring feature datasets of the tool in two key phases before and after maintenance operations. First, based on multi-dimensional data such as shape contour, surface texture, initial mass, and component topology obtained in the first phase, the tool's initial attributes are registered and automatically categorized according to its characteristics into deformation-consumption, property-change, or appearance-stable categories, thus establishing a high-fidelity digital profile of the tool and laying the foundation for subsequent differentiated processing. Subsequently, in the second phase, customized acquisition and deconstruction strategies are executed for different category labels: for deformation-consumption tools, 3D scanning and key point matching are used to reconstruct component topology, accurately quantifying macroscopic geometric deformation; for property-change tools, intelligent weighing, AI visual recognition, and micro gas sensing technologies are integrated to comprehensively capture subtle changes in mass, cleanliness, and chemical composition; for appearance-stable tools, high-resolution imaging and feature point cloud registration are used to achieve sub-micron-level identity verification and micro-damage identification. Furthermore, the system performs state matching analysis on the data from before and after the changes, extracts specific state change characteristics, and compares them one by one with preset deformation and performance correlation rules. This not only determines the tool's return status but also deeply assesses the risk of functional failure and its potential impact on the equipment being repaired. This solves the problem that existing technologies in tool traceability only focus on identity and quantity, failing to perceive changes in physical form, material property degradation, or microscopic wear, resulting in a lack of objective data to support repair quality. Therefore, it avoids the risks of operational anomalies, environmental pollution spread, and equipment failure caused by the failure to detect hidden tool damage, residual contamination, or performance degradation in a timely manner, improving the controllability of the repair process, the accuracy of quality traceability, and the safety and reliability of production operations.

[0015] In summary, this application achieves a leap from single-identity recognition to multi-dimensional state perception by constructing a complete technical chain of pre-registration classification, post-differentiation deconstruction, state matching, and rule association. The scheme is logically rigorous, adaptable to the changing patterns of different tools, ensures the targetedness and accuracy of data collection, and forms a complete data loop from tool adoption to quality assessment, possessing systemic advantages and application value.

[0016] Secondly, this application provides a closed-loop traceability system for maintenance tools based on dual-phase identification before and after operation, the system comprising: The initial information module is used to acquire the tool pre-operation feature dataset before maintenance work, and based on the tool pre-operation feature dataset, to perform initial attribute registration and change type classification for each tool, and generate tool initial information set; The change information module is used to acquire the tool post-operation feature dataset in the post-maintenance phase. Based on the category label of each tool in the initial tool information set, it processes the data according to the differentiated acquisition strategy and feature deconstruction logic corresponding to the deformation consumption class, property change class, and appearance stability class to generate the tool change information set. The traceability report module is used to perform state matching analysis under its category for each tool based on the tool's initial information set and the tool's change information set, extract the state change features of the tool before and after the operation, compare the state change features with the maintenance quality association rules, and generate a maintenance quality traceability report. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a schematic diagram illustrating an application scenario provided in one embodiment of this application; Figure 2 A flowchart illustrating a closed-loop traceability method for maintenance tools based on pre- and post-operation dual-phase identification, as provided in an embodiment of this application; Figure 3 This is a schematic diagram of the structure of a closed-loop traceability system for maintenance tools based on dual-phase identification before and after operation, provided in an embodiment of this application. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0020] Furthermore, the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article, unless otherwise specified, generally indicates that the preceding and following related objects have an "or" relationship.

[0021] The embodiments of this application will now be described in further detail with reference to the accompanying drawings.

[0022] Figure 1 This application provides an illustration of an application scenario. In the closed-loop traceability process of maintenance tools, the method provided in this application can achieve multi-dimensional perception of the tool's physical form, material properties, and micro-wear, directly linking changes in tool status with maintenance quality, identifying operational anomalies and potential risks in advance, and improving the accuracy of maintenance quality traceability and closed-loop management capabilities.

[0023] Specifically, the method provided in this application can be applied to any server, where the server interacts with the visual AI platform to obtain the tool pre-feature dataset and tool post-feature dataset provided by the visual AI platform, and finally outputs the maintenance quality traceability report to relevant personnel.

[0024] The specific implementation method can be referred to in the following embodiments, wherein the data mentioned in the embodiments are only for reference and examples, so that relevant personnel can better understand them.

[0025] Figure 2 This is a flowchart illustrating a closed-loop traceability method for maintenance tools based on pre- and post-operation dual-phase identification, provided in one embodiment of this application. The method of this embodiment can be applied to the server in the above scenario. (See also...) Figure 2 The process, specifically the implementation steps, are as follows: Example 1: In high-safety-requirement maintenance scenarios such as aircraft engines, nuclear facilities, and semiconductor manufacturing, existing tool tracking systems typically only record the borrowing and returning status of tools or use RFID tags for identification. This existing technology cannot detect physical changes, material property alterations, or microscopic wear that occur during tool use. After maintenance is completed, the evaluation of maintenance quality often relies on final functional testing or manual inspection, lacking objective data support from changes in the tool's condition itself. This results in the inability to detect operational anomalies, environmental contamination, foreign object residue, and other potential hazards in advance, posing significant safety risks.

[0026] To address the aforementioned problems, this application provides a closed-loop traceability method for maintenance tools based on dual-phase identification before and after operations. The method includes: Step 1: Obtain the tool pre-operation feature dataset before the maintenance operation. Based on the tool pre-operation feature dataset, perform initial attribute registration and change type classification for each tool to generate the tool initial information set. The tool pre-feature dataset refers to the raw data collection that characterizes the physical state of a tool, acquired through multi-dimensional sensing devices before maintenance work begins. Sources include 3D scanners, intelligent weighing equipment, high-resolution imaging equipment, and gas sensor arrays. Specifically, this dataset includes the point cloud data of each tool's outline, surface texture images, initial mass values, and topological connections between components. Initial attribute registration refers to the process of binding the aforementioned physical characteristics with the tool's unique identifier according to material coding rules, establishing the tool's "birth certificate" in the digital space. Change type classification, based on the tool's component topology and expected operating environment characteristics, automatically categorizes tools into one of three types using pre-defined classification logic: deformation-consumption, property-change, or appearance-stable. Deformation-consumption tools refer to components prone to plastic deformation or partial wear during use, such as gaskets; property-change tools refer to components prone to chemical degradation, resistance changes, or surface contamination during use, such as anti-static wrist straps; appearance-stable tools refer to components that primarily maintain their macroscopic shape but may experience microscopic wear during use, such as precision wrenches.

[0027] For example, in a certain aero-engine maintenance scenario, the system acquires 3D point cloud data of an irregularly shaped sealing gasket, identifies the relative positional relationship between its two components—the sealing ring and the metal skeleton—and, combined with the material code Gasket-001, registers its initial attributes as a unique identifier ID: 2024-PAD-001. Subsequently, based on its component topology and the high-temperature, high-pressure operating environment of the high-pressure compressor, the system determines that it is highly susceptible to compression deformation during operation, thus automatically classifying it as a deformation-consumption type. This refined registration and classification at the pre-construction stage lays an accurate data foundation for subsequent differentiated post-processing, ensuring that tools with different characteristics can be covered by appropriate strategies.

[0028] Step 2: Obtain the tool post-maintenance feature dataset after the maintenance operation. Based on the category label of each tool in the initial tool information set, process it according to the differentiated acquisition strategy and feature deconstruction logic corresponding to the deformation consumption class, property change class and appearance stability class to generate the tool change information set. The tool post-feature dataset consists of tool status data collected again after maintenance work is completed. The differentiated acquisition strategy and feature deconstruction logic refer to executing specific data processing flows for the three types of labels generated above. For deformation-consumption tools, the corresponding strategy is to perform deformation feature capture and component topology reconstruction, aiming to quantify macroscopic geometric deformation and infer the source identity of deformed components. For property change tools, the corresponding strategy is to perform multimodal performance parameter remeasurement and deviation calculation, aiming to calculate the amount of material chemical change through multidimensional comparison of mass, visual characteristics, and gas composition. For appearance-stabilized tools, the corresponding strategy is to perform high-precision registration of micro-textures, aiming to identify minor damage or verify that the tool has not been tampered with through sub-pixel-level image matching. The tool change information set summarizes the structured state change data obtained after differentiated processing of the above three types of tools.

[0029] For example, for antistatic wrist straps classified as subject to property changes, the system initiates a multimodal data acquisition strategy in the later stages: First, it acquires the post-operation mass using an intelligent weighing device, finding an increase of 0.03g compared to the initial mass; second, it analyzes the surface image using AI visual recognition technology, detecting microparticle distribution but an intact coating; finally, it confirms no abnormal gas escape through a micro gas sensor array. The system deconstructs this data into mass change, surface state parameters, and chemical change, comparing it with preset thresholds to classify the performance change level as severely abnormal. For hex wrenches classified as stable in appearance, the system uses a high-resolution imaging device to acquire microscopic images of the end, calculates the matching degree between the old and new images using a feature point cloud registration algorithm, and further marks newly appearing 0.3mm scratches if the matching degree is qualified. Through this differentiated processing based on classification, the system can avoid over-collection of stable tools or under-collection of consumable tools, improving the targeting and accuracy of data acquisition.

[0030] Step 3: Based on the initial information set and the change information set of the tool, perform state matching analysis under the category of each tool, extract the state change characteristics of the tool before and after the operation, compare the state change characteristics with the maintenance quality association rules, and generate a maintenance quality traceability report.

[0031] State matching analysis refers to the process of comparing the generated initial baseline data with the generated change data one by one. This process is executed by the core processing module of the maintenance quality traceability system. The analysis aims to extract specific state change characteristics of the tool before and after operation, such as deformation rate, resistance drift value, and number of micro-scratches. Maintenance quality association rules are a pre-built knowledge base containing deformation association rules and performance association rules, used to define the mapping relationship between specific state change characteristics and maintenance quality risks. The comparison process involves substituting the extracted feature values ​​into the rules for logical judgment, thereby determining the tool's return status (e.g., complete return, partial loss), functional failure risk (e.g., material failure, accuracy deviation), and potential impact on the repaired equipment (e.g., risk of foreign object residue, excessive clearance). The maintenance quality traceability report is a document generated based on the comparison results of all tools. Its output can be in the form of an electronic work order attachment or visual dashboard data, used to guide subsequent work order closure, tool maintenance plan adjustments, or operator training.

[0032] For example, the system compares the edge tearing deformation characteristics of a sealing gasket with deformation association rules. The rules state that if an edge tear occurs and the offset exceeds 2°, an installation abnormality is identified. Based on this, the system determines that the gasket has a risk of uneven compression and, combined with the fact that the total deformation rate exceeds the plastic deformation threshold, generates a conclusion requiring mandatory scrapping. Simultaneously, the system incorporates the severely abnormal resistance and increased mass characteristics of the anti-static wrist strap into the performance association rules. The rules state that such a combination implies potential radioactive contamination, and the system generates a recommendation for isolation and testing. Finally, the system synthesizes the judgment results of all tools and outputs a maintenance quality traceability report including specific measures such as gasket scrapping, wrist strap isolation, and wrench calibration. Thus, this method achieves closed-loop management from tool status perception to maintenance quality assessment, avoiding subsequent equipment failures caused by hidden tool damage.

[0033] This application constructs a complete closed-loop data system for the entire lifecycle of maintenance tools through the synergistic effect of the aforementioned technical features. By registering initial attributes and classifying change types in the pre-phase stage, precise digital archives are established for tools with different characteristics, solving the problem of blind data collection caused by the ambiguity of tool characteristics in traditional methods. Building upon this, the differentiated acquisition strategy and feature deconstruction logic in the post-phase stage, using classification labels to drive a targeted perception mechanism, enable the efficient and accurate capture of multi-dimensional states such as deformation, property changes, and micro-wear, overcoming the limitation of a single acquisition mode that cannot cover all tool types. Furthermore, by deeply comparing the extracted state change features with maintenance quality association rules, abstract physical changes in tools are directly transformed into specific maintenance quality risk warnings, achieving a dimensional upgrade from managing tool quantity to managing maintenance quality. This progressive processing logic not only ensures the authenticity of the tool's returned state but also detects potential hazards such as operational violations, environmental pollution, and tool failure in advance, providing quality traceability assurance for maintenance operations in industries with high safety requirements.

[0034] Example 2: In one possible implementation, this application embodiment provides a method for generating an initial tool information set, the method comprising the following steps: Step 1: Obtain the tool pre-feature dataset, which includes the shape outline, surface texture, initial quality, and component topology of each tool; The tool pre-feature dataset refers to the basic data set reflecting the physical state of the tool collected by multimodal sensing devices before maintenance work begins. The external profile refers to the geometric boundary shape of the tool in three-dimensional space, typically obtained by acquiring point cloud data using a 3D laser scanner or structured light camera, used to characterize the tool's macroscopic dimensions and shape. Surface texture refers to the tool's surface micro-roughness, scratch distribution, and material texture characteristics, typically obtained by capturing images with a high-resolution industrial camera and macro lens and extracting features such as the gray-level co-occurrence matrix, used to distinguish the uniqueness of different individuals of the same model. Initial mass refers to the tool's weight in its unused state, measured by high-precision intelligent weighing equipment, serving as a benchmark for subsequent monitoring of material loss or foreign object adhesion. Component topology refers to the connection methods, relative positions, and spatial constraints between components of a composite tool composed of multiple parts, such as the engagement state of bolts and nuts, and the nesting position of seals and metal frames. The external profile and component topology, used together, can completely describe the tool's spatial structure, while surface texture and initial mass supplement the tool's feature information from the dimensions of micromorphology and physical properties, respectively. By fusing multidimensional data in this way, we can ensure that subsequent analysis has a high-fidelity data foundation and effectively avoid identification errors caused by the lack of a single data source.

[0035] Step 2: Based on the shape and surface texture, and combined with the material coding rules, perform initial attribute registration for each tool to generate a unique tool identifier; The unique tool identifier refers to a globally unique code assigned to each repair tool, used to accurately track that specific individual throughout its entire lifecycle. This identifier is generated by hashing the collected geometric feature vector of the tool's outline with the fingerprint of its surface texture, and then combining this with preset material coding rules (such as ISO standard coding or internal ERP coding). Specifically, the system first extracts key geometric parameters of the outline (such as aspect ratio and hole coordinates), and then extracts unique features of the surface texture (such as the distribution pattern of random scratches), binding these two types of biometric-level physical information with a static material code. For example, for two identical hex wrenches, although their material codes are the same, the unique tool identifiers generated by the system will be completely different due to differences in surface texture left from manufacturing and unique scratches formed by usage history.

[0036] By co-verifying the shape and surface texture, not only is physical feature-based anti-counterfeiting verification achieved, but the problem of traditional RFID tags being easily detached or damaged, leading to identity loss, is also solved. This step aims to establish a strong correlation between the physical entity of the tool and its digital information, ensuring that all subsequent state change data can be accurately attributed to the specific tool, thereby providing a reliable identity anchor for closed-loop traceability.

[0037] Step 3: Based on the component topology and operating environment characteristics, automatically classify each tool into one of the following categories: deformation consumption, property change, or appearance stability, and bind a unique tool identifier to generate an initial tool information set; Tools categorized into three types are classified as follows: Deformation-consumable tools, which are prone to plastic deformation, breakage, or partial wear during operation, such as gaskets, fuses, and disposable locking clips; Property-change tools, which exhibit significant changes in their physicochemical properties (such as resistivity, mass, surface cleanliness, and chemical composition) during use, such as anti-static wrist straps, oil-absorbing cotton, and chemical reagent applicators; and Appearance-stable tools, which rely primarily on their geometric shape and maintain a largely unchanged macroscopic form during use, potentially experiencing only minor wear, such as wrenches, screwdrivers, and gauges. The classification process is automatically determined based on the complexity of the component topology and the characteristics of the operating environment (such as high temperature, high pressure, strong corrosion, and cleanliness requirements). If the component topology indicates the presence of easily separable parts or thin-walled structures, and the operating environment involves high-pressure extrusion, it is classified as deformation-consumable. If the operating environment involves chemical contact or electrostatically sensitive areas, and the tool material has adsorption or conductivity, it is classified as property-change tools. If the component topology is stable and the operation primarily involves mechanical fastening, it is classified as appearance-stable. For example, for a rubber seal with a metal skeleton, the system identifies its component topology as a soft-hard composite nested structure. Combined with the high temperature and high pressure environment characteristics of aero-engine maintenance, it automatically classifies it as a deformation-consumption type rather than a common appearance-stability type.

[0038] By jointly determining the component topology and operating environment characteristics, a predictive classification of tool failure modes is achieved. This classification mechanism enables subsequent post-phase data acquisition strategies to be more targeted: for deformation-related failures, the focus is on scanning geometric integrity; for failures due to property changes, the focus is on detecting physicochemical indicators; and for failures with stable appearance, the focus is on comparing microscopic textures. The resulting initial tool information set not only records the tool's origin but also pre-defines the monitoring dimensions of its health record, improving the targeting and accuracy of subsequent state matching analysis and avoiding data omissions or false alarms caused by using a single strategy to process all tools.

[0039] Example 3: In one optional implementation, this application provides a process for generating a tool change information set. The method further includes selecting differentiated acquisition strategies and feature deconstruction logic corresponding to deformation consumption, property change, and appearance stability categories based on the category label of each tool in the initial tool information set, and summarizing the processed post-feature data to generate the tool change information set.

[0040] Step 1: Based on the category label of each tool in the initial tool information set, select the differentiated acquisition strategy and feature deconstruction logic corresponding to the deformation consumption class, property change class, and appearance stability class respectively; The category label refers to a classification identifier automatically generated based on the tool's component topology and operating environment characteristics during the pre-maintenance phase, originating from the registration data of the tool's initial information set. This step dynamically routes data to the optimal processing branch based on the tool's main physical or chemical changes during use, avoiding resource waste or missed key features due to a single acquisition strategy. Specifically, if the category label is deformation-consumption type, the system automatically loads a point cloud reconstruction algorithm library for macroscopic geometric deformation; if the label is property change type, the system activates the multimodal sensor fusion module and configures quality, vision, and gas detection parameters; if the label is appearance-stable type, a high-resolution microscopic image registration engine is invoked. For example, when the system reads that a sealing gasket's label is deformation-consumption type, it skips the resistance measurement and gas analysis processes, directly activating only the 3D scanning equipment, thereby reducing the single detection time from 120 seconds to 45 seconds. Through this label-based strategy distribution mechanism, precise matching of detection resources and tool characteristics is achieved, improving overall traceability efficiency.

[0041] Step 2: For deformation-consuming tools, perform deformation feature capture and component topology reconstruction; Deformation feature capture refers to acquiring spatial geometric data after tool operation using 3D sensing technology and identifying macroscopic deformation areas exceeding the initial reference range. Component topology reconstruction refers to inferring the origin of deformed components and reconstructing the spatial connection relationships between components based on key point matching technology. This step aims to quantify changes in the structural integrity of tools prone to plastic deformation or fracture, such as gaskets and fuses. The execution entity is a detection terminal integrated with a high-precision laser scanner or structured light camera. In practice, the system first collects point cloud data after operation and compares it point by point with the initial geometric dimensions recorded in the tool's initial information set to screen out deformation areas. Subsequently, by analyzing the relative positions of residual features, the component topology network is reconstructed. For example, for irregularly shaped gaskets in aero-engine maintenance, the system captures that its thickness is compressed from 2mm to 1.2mm, and the metal skeleton and sealing ring undergo a 5° rotational offset. Based on this, the system reconstructs the deformed topology structure and confirms that no debris remains. By combining deformation feature capture with component topology reconstruction, the true shape of the tool after being subjected to force can be accurately restored, effectively identifying the risk of breakage or uneven compression caused by improper installation, and providing a geometric basis for subsequent judgment on whether to force scrapping.

[0042] Step 3: For tools involving changes in properties, perform multimodal performance parameter remeasurement and deviation calculation; The multimodal performance parameter remeasurement refers to the simultaneous acquisition of tool mass data, surface state parameters (such as cleanliness and coating integrity), and chemical changes (characteristic gas composition and concentration). Deviation calculation refers to comparing the measured data with the initial baseline value and preset performance threshold to calculate the degree of performance degradation and classify it. This step is used to capture changes in the intrinsic properties of tools such as antistatic wrist straps and chemical probes caused by environmental pollution, material degradation, or chemical reactions. During execution, intelligent weighing equipment is responsible for monitoring minute increases or decreases in mass, AI vision module is responsible for analyzing surface texture changes, and micro gas sensor array is responsible for capturing volatile organic compounds or reaction byproducts. For example, in a nuclear facility maintenance scenario, for an antistatic wrist strap, the system measured a 0.03g increase in mass (suspected dust adhesion), an increase in resistance from 3.8MΩ to 45MΩ, and no abnormal gas was detected. After comprehensive calculation, it was determined to be a severe anomaly. Through the collaborative calculation of multidimensional data on mass, vision, and gas, not only can the remaining lifespan of materials be quantified, but also hidden dangers invisible to the naked eye, such as radioactive contamination, can be detected early, ensuring the functional reliability of tools with changing properties.

[0043] Step 4: For appearance-stabilized tools, perform high-precision registration of micro-textures; High-precision registration of micro-textures refers to using a feature point cloud registration algorithm to perform sub-pixel-level matching of minute scratches, wear patterns, or laser-engraved edges in two high-resolution images before and after the operation, and calculating the matching score. This step is mainly used for the identification and damage assessment of tools such as hex wrenches and precision measuring tools whose macroscopic shape remains basically unchanged but may have micro-wear or have been switched. The execution method is to first lock the feature region in the initial information set, and then acquire microscopic images of the same region in the subsequent phase. The algorithm calculates the matching degree between the old and new images. If the matching degree is lower than a preset threshold (such as 95%), the tool is determined to be inconsistent (it may have been switched); if the matching degree is qualified, the newly emerging micro-damage features are further extracted. For example, for an hex wrench used for precision machine tool repair, the system finds a new 0.3mm long scratch at the end through registration. Although it does not affect the overall shape, it indicates that there may be slippage due to excessive torque. Through high-precision registration of micro-textures, it is possible to uniquely identify a tool and keenly perceive minor damage without relying on macroscopic deformation, effectively preventing quality risks caused by misuse of tools.

[0044] Step 5: Summarize the post-feature data obtained from deformation consumption tools, property change tools, and appearance stability tools to generate a tool change information set; The tool change information set is a structured dataset containing the state change characteristics of all tested tools in the subsequent time phase. It consists of deformation feature records, performance change levels, damage feature records, and corresponding tool unique identifiers. This step aims to standardize and integrate heterogeneous data scattered across different processing branches, forming a unified traceability data foundation and providing complete input for subsequent state matching analysis and quality report generation. Specifically, the system indexes and merges the generated deformation feature records, performance change level data, and damage feature records according to the tool's unique identifier. For example, tear deformation records of sealing gaskets, severe abnormality levels of anti-static wrist straps, and minor wear records of Allen wrenches are packaged and stored in the same data frame. By summarizing the differentiated processing results of the three types of tools, a comprehensive tool state view is constructed, ensuring that the maintenance quality traceability report can cover all potential risk points from macroscopic structural damage to microscopic material failure.

[0045] This application achieves the organic synergy of deformation feature capture, multimodal performance parameter calculation, and micro-texture registration by dynamically selecting differentiated acquisition strategies based on tool category labels. Building upon this, it precisely quantifies structural deformation using 3D point cloud reconstruction technology, comprehensively captures material property degradation through multi-sensor fusion, and rigorously monitors micro-wear and identity consistency using high-precision image registration. Finally, it aggregates the three types of heterogeneous data to generate a tool change information set. This not only solves the problem of missed detections caused by the inability of a single acquisition strategy to cover the change patterns of different tools in existing technologies, but also improves the targeting and accuracy of data acquisition, providing multi-dimensional data support for closed-loop traceability of maintenance quality.

[0046] Example 4: In another embodiment, this application provides a method for performing deformation feature capture and component topology reconstruction, the method comprising the following steps: Step 1: Collect post-operation point cloud data of deformation-consuming tools using a 3D scanning device, and identify all macroscopic geometric deformation areas in the current outline of the tool that exceed the initial reference range. Among them, 3D scanning equipment refers to a measuring device used to acquire high-precision 3D coordinate data of an object's surface. Its working principle involves converting the spatial position information of the tool surface into a dense point cloud set using laser triangulation or structured light projection technology. Post-operation point cloud data is a digital 3D model generated by mapping the actual physical form of a deformable consumable tool after maintenance and recycling. The initial reference range is an allowable tolerance interval set based on the original geometric dimensions recorded in the tool's initial information set, used to define the boundary between normal elastic deformation and abnormal macroscopic deformation. Macroscopic geometric deformation regions can refer to local areas on the tool surface that have undergone irreversible changes such as plastic deformation, fracture, tearing, or defects; the geometric characteristics of these regions significantly deviate from the initial design contour.

[0047] Specifically, the system controls a 3D scanning device to perform a comprehensive scan of the recovered deformed consumable tools, generating a post-operation point cloud dataset containing millions of spatial coordinate points. Subsequently, the data processing module spatially registers this point cloud data with the initial standard model stored in the tool's initial information set. By calculating the normal distance or curvature change corresponding to each point, it filters out point sets whose displacement exceeds a preset tolerance threshold, thereby identifying macroscopic geometric deformation areas. For example, for irregularly shaped sealing gaskets used in aero-engine maintenance, if their edges have tear gaps exceeding 2mm in length, or their overall thickness compression exceeds the material's elastic limit, the 3D scanning device will accurately capture these areas of missing point clouds or severe displacement and mark them as macroscopic geometric deformation areas. Through this high-precision spatial data acquisition and difference identification, the degree of physical damage to the tool can be quantitatively assessed, providing an objective basis for subsequent judgments on whether the tool has failed.

[0048] Step 2: Perform key point matching between the post-operation point cloud data and the initial geometric dimensions and relative positions of components recorded in the tool's initial information set, find residual feature regions, and infer the origin of the deformed components; Keypoint matching refers to the process of using feature descriptor algorithms (such as SIFT and FPFH) to find unique and stable corresponding point pairs between the post-processing point cloud and the initial model. Residual feature regions refer to fragments or parts with identifiable geometric features that remain on the main body or scattered in a recycling container after partial breakage, wear, or component separation of the tool. Inferring the origin of deformed components refers to determining which specific sub-component of the tool the deformed part originally belonged to by analyzing the geometric properties, material texture, or specific processing marks of the residual features.

[0049] Specifically, the system uses undeformed areas in the post-operation point cloud data as anchor points and performs iterative nearest-point (ICP) matching or feature point matching with the initial geometric dimensions and relative positions of each component recorded in the tool's initial information set. When a part of the point cloud cannot perfectly overlap with the main model, but highly matches the features of a sub-component in the initial model (such as the metal skeleton of a gasket, the lip of a sealing ring, etc.), the system determines that part is a residual feature and infers its origin based on the matching results. For example, in tracing a certain type of sealing gasket, if a scan reveals an independent arc-shaped point cloud cluster next to the main body, and key point matching shows that its curvature and thickness features are consistent with the metal skeleton part in the initial model, the system can infer that the residue is a detached metal skeleton fragment and confirm its origin. This process enables accurate location of the specific ownership of damaged parts even after structural damage to the tool, avoiding misjudgments caused by component confusion.

[0050] Step 3: Simultaneously count the number and shape of all deformable components, reconstruct the spatial topological relationship between components, and generate deformation feature records for deformation consumption tools.

[0051] The number of deformable components can refer to the number of independent parts identified as undergoing macroscopic deformation or separating from the main body. Morphology can refer to the physical properties of the deformable components, such as their geometry, volume, surface area, and surface roughness. Spatial topological relationships can refer to the connections, adjacencies, inclusions, or relative positions of the components within the tool in three-dimensional space, such as the engagement of bolts and nuts, or the encapsulation relationship between sealing rings and the skeleton. The deformation feature record is a structured data set that integrates information on the distribution of deformed areas, component identities, quantity statistics, and topological changes.

[0052] Specifically, after completing key point matching, the system performs cluster analysis on all components marked as deformed, counts their total number, and extracts their respective morphological parameters. Subsequently, based on the spatial distribution of point cloud data, it reconstructs the relative position network between components, compares the topology before and after the operation, and identifies topological changes such as connection breaks, relative displacements, or sequence errors. Finally, the statistical results and the reconstructed topological relationships are encapsulated to generate deformation feature records. By combining quantitative statistics with topology reconstruction, not only can the scale of deformation be quantified, but the assembly state at the time of tool failure can also be reconstructed, providing crucial evidence for analyzing the standardization of maintenance operations (such as whether there is non-uniform clamping).

[0053] This application achieves in-depth analysis of the post-operation state of deformable consumable tools through the synergistic effect of the aforementioned technical features. High-density point cloud data acquired using 3D scanning equipment provides a precise spatial data foundation for identifying macroscopic geometric deformation areas. Based on this, key point matching technology is used to compare the post-operation data with the initial geometric dimensions and relative positions of components. This not only accurately locates residual feature areas but also intelligently infers the origin of deformed components, solving the problem of traditional methods' difficulty in distinguishing mixed fragments. Furthermore, by statistically analyzing the number and shape of deformed components and reconstructing spatial topological relationships, the system transforms discrete deformation data into logically related structured information, fully reconstructing the internal assembly and change process of the tool. This progressive processing method, from microscopic geometric feature identification to macroscopic topological relationship reconstruction, ensures that the generated deformation feature records not only reflect the extent of tool damage but also explain how the damage occurred and which component failed. This improves the accuracy and depth of maintenance quality traceability and effectively avoids subsequent safety accidents caused by undetected hidden tool damage.

[0054] Example 5: In another embodiment, this application provides a method for performing multimodal performance parameter remeasurement and deviation calculation, the method comprising the following steps: Step 1: Obtain the mass data of the tool with property changes in the post-operation phase through intelligent weighing equipment, compare it with the initial mass, and calculate the mass change; The change in mass refers to the difference in gravity value of a tool before and after maintenance due to material loss, increased surface deposits, or densification of its internal structure. This data is collected in real time by a high-precision intelligent weighing device after the maintenance work, and its function is to quantitatively assess the physical integrity and surface contamination of the tool. Specifically, the intelligent weighing device can be an electronic balance or an industrial-grade weighing sensor integrated into the tool return terminal. Its accuracy can be set according to the size of the tool; for example, for small precision fasteners, the weighing accuracy can be set to 0.001g, while for large hydraulic wrenches, the accuracy can be set to 0.1g. When the tool is placed in the weighing area, the device automatically reads the current mass value and performs a subtraction operation on the initial mass data recorded in the tool's initial information set to obtain the change in mass. If the change in mass is negative and exceeds a preset threshold, it usually indicates that the tool has experienced material wear or breakage; if the change in mass is positive, it may indicate that the tool surface has adsorbed foreign substances such as oil, metal debris, or radioactive dust. This high-precision quality comparison can effectively distinguish between tool wear and environmental adhesion, providing a basic physical basis for subsequent judgment on whether the tool has been contaminated or suffered structural damage.

[0055] Step 2: Collect images of the tool surface using AI visual recognition technology, and use deep learning convolution algorithms to identify the surface cleanliness level, coating integrity, and color characteristics to obtain surface state parameters; Among them, surface state parameters are a set of multi-dimensional vectors describing the physical and chemical properties of the tool's appearance, specifically including surface cleanliness level, coating integrity index, and color feature values. These parameters are obtained by acquiring tool surface images through industrial cameras or high-resolution scanning equipment deployed in tool recycling stations, and using pre-trained deep learning convolutional neural network (CNN) models to extract and classify features from the images. Their function is to intuitively reflect the microscopic morphological changes of the tool surface and the aging of the chemical coating. Specifically, the AI ​​visual recognition model can be trained based on architectures such as ResNet or YOLO, with high-resolution images of the tool surface taken after the operation. For example, when identifying coating integrity, the model can classify image pixels into four categories: intact, micro-cracked, peeling, and completely failed. If a peeling area exceeding 5% of the conductive coating on the surface of an anti-static wrist strap is detected, the coating integrity index is determined to be low. When identifying color features, by extracting the HSV color space histogram of the image, if the tool surface changes from silver-gray to dark brown, it can be inferred that high-temperature oxidation or severe oil buildup has occurred. Furthermore, by integrating cleanliness levels, coating integrity, and color characteristics, a comprehensive health profile of the tool surface can be constructed. This non-contact intelligent visual analysis can quickly capture subtle surface degradation that is difficult for the human eye to detect, avoiding secondary damage to the tool, and providing intuitive visual evidence for determining whether the tool needs cleaning, recoating, or direct disposal.

[0056] Step 3: Collect the composition and concentration of characteristic gases escaping from the tool surface using a micro gas sensor array, and combine this with knowledge of chemical degradation gases to calculate the amount of chemical change in the tool material; Chemical change refers to the degree of alteration in the composition of tool materials during operation due to chemical reactions, thermal decomposition, or adsorption. This data is collected by a miniature gas sensor array surrounding the tool recovery port or a dedicated detection chamber. The sensor array typically includes electrochemical sensors, semiconductor gas sensors, or photoionization detectors (PIDs) to specifically respond to volatile organic compounds (VOCs), acid gases, or specific industrial byproducts. Its function is to deeply probe the molecular-level changes within the tool materials, identifying potential chemical corrosion or hazardous substance residues. Specifically, the system has a built-in knowledge base of chemical degradation gases, storing characteristic gas fingerprints that various tool materials may produce under different operating conditions. For example, for rubber seals used in nuclear facility maintenance, if the sensor array detects an abnormally high concentration of trace amounts of sulfur dioxide or hydrogen sulfide, combined with knowledge base matching, it can be inferred that the rubber material has undergone sulfur bond breakage or radiation degradation. Similarly, in semiconductor maintenance scenarios, the detection of silicon tetrafluoride (SiF4) gas indicates abnormal fluorine-based plasma etching of silicon-containing components. The system performs differential calculations between the collected component concentrations and the initial baseline or environmental background values, and then weights the resulting chemical changes. By introducing a gas sensing dimension, this step can overcome the limitations of physical morphology observation and detect latent degradation within materials in advance, providing crucial early warning signals, especially for tools that show no obvious external changes but whose internal chemical structure has been damaged.

[0057] Step 4: Use the mass change, surface state parameters, and chemical change as performance change characteristics, compare them with preset performance thresholds, and classify the performance change levels.

[0058] The performance change level is a graded evaluation of the overall usability status of tools with changing properties, typically categorized as normal, slightly abnormal, severely abnormal, or failed. This level is generated based on the performance change characteristics obtained in the preceding steps, encompassing three dimensions: mass change, surface state parameters, and chemical change. It is generated by comprehensively comparing these characteristics with a preset performance threshold matrix using a multi-source data fusion algorithm. Its purpose is to provide quantitative decision-making basis for subsequent maintenance quality traceability reports, directly determining the tool's disposal strategy (such as continued use, calibration, isolation testing, or mandatory scrapping). Specifically, the preset performance threshold is not a single value, but rather a set of association rules. For example, the rules are set as follows: when the absolute value of the mass change is less than 0.1% and the surface condition parameters show that the coating is intact and the chemical change is zero, it is classified as normal; when the mass change is between 0.1% and 1%, or when there are slight scratches on the surface but no coating peeling, and no harmful gases are detected, it is classified as slightly abnormal, indicating that cleaning or calibration is required; when the detected chemical change exceeds the material degradation threshold, or when there is a significant increase in mass accompanied by surface radioactive contamination characteristics, it is directly classified as severely abnormal or failed, regardless of other indicators. Through this collaborative comparison of multi-dimensional features, the system can eliminate the interference of false alarms from a single sensor. For example, a mass increase may only be due to water contamination, but if it is accompanied by the detection of a specific chemical gas, it can be confirmed as chemical adsorption or reaction, thereby achieving accurate classification of tool performance degradation and ensuring the safety and reliability of maintenance operations.

[0059] This application constructs a comprehensive performance evaluation system for tools with changing properties through the synergistic collaboration of intelligent weighing equipment, AI visual recognition technology, and a micro gas sensor array. Mass change provides macroscopic physical evidence of wear or adhesion, surface state parameters reveal the microscopic morphology and optical characteristics of the coating, while chemical change delves into the molecular-level compositional evolution. These three elements complement each other; through the fusion analysis of multimodal data, the system can not only accurately calculate the performance change characteristics of the tool but also effectively distinguish between different causes such as normal wear, environmental pollution, and material aging. Based on this, by dynamically comparing with preset performance thresholds, the system can automatically classify precise performance change levels, thereby immediately identifying tools with potential safety hazards or functional failure risks after maintenance work is completed. This avoids secondary accidents caused by hidden tool defects, realizing a shift from passive post-incident inspection to proactive pre-incident warning, and improving the intelligence level and safety closed-loop capability of maintenance quality traceability.

[0060] Example 6: In one possible implementation, embodiments of this application provide a method for performing high-precision registration of micro-textures, the method comprising: Step 1: Acquire microscopic surface images of the same area in the tool's initial information set at the later time using a high-resolution imaging device; High-resolution imaging equipment can refer to optical microscopes, electron microscopes, or industrial-grade macro cameras with sub-micron resolution, used to acquire nanometer to micrometer-level texture details on tool surfaces. The same area can refer to a uniquely identified region locked during the initial tool attribute registration stage, based on laser engraving numbers, factory tool marks, or specific geometric feature points. The acquisition process ensures strict spatial consistency between the two images before and after the operation by aligning the imaging equipment's field of view center with the coordinates of the initially recorded feature points. For example, for an internal hex wrench, the system automatically positions itself to the edge area of ​​the laser engraving Hex-005 on its handle, acquiring a microscopic image including surrounding natural wear patterns at a resolution of 0.5 μm / pixel. This high-precision point-to-point reproduction acquisition reduces matching errors caused by shooting angle or positional deviations, providing a reliable raw data foundation for subsequent point-by-point comparisons.

[0061] Step 2: Use the feature point cloud registration algorithm to match the fine scratches, wear lines or laser engraving edges in the two images before and after the operation point by point, and calculate the matching degree score; The feature point cloud registration algorithm refers to the process of converting a two-dimensional microscopic image into a three-dimensional depth point cloud or a high-dimensional feature vector, and then using the Iterative Closest Point (ICP) algorithm or deep learning-based feature descriptor matching technology to spatially align key feature points in the image. Fine scratches, wear lines, or laser-engraved edges serve as natural fingerprint features for matching, possessing uniqueness and stability. The matching score is calculated by statistically analyzing the proportion of successfully matched feature points to the total number of feature points, combined with a weighted average of spatial geometric consistency error. For example, if the system registers 3000 feature point clouds of the wrench end recorded before operation with the point cloud collected after operation, and 2994 points overlap within the allowable error range, the matching score is calculated to be 99.8%. This step aims to objectively assess the identity of the tool through quantitative means, eliminating subjective interference from human visual judgment.

[0062] Step 3: If the matching degree is lower than the preset threshold, it is determined that the tool has been swapped or has not been used; The preset threshold is a critical value set based on historical data statistics and misjudgment rate analysis, typically between 90% and 95%, used to distinguish between minor feature shifts caused by normal wear and tear and structural differences between tools from different origins. When the matching degree is below this threshold, it indicates that the surface texture structure of the tool recovered after the operation is fundamentally different from the initially registered tool, or that the tool surface is completely covered, resulting in feature loss. For example, if the matching degree between a wrench image collected after a maintenance and the initial image is only 65%, the system directly determines that the tool has been swapped for another wrench of the same model but from a different individual, or that the operator did not actually use the tool and directly returned the spare part. This judgment logic effectively prevents the risk of unauthorized tools entering the maintenance process and ensures the traceability authenticity of the tools used in maintenance operations.

[0063] Step 4: If the matching degree is qualified, further identify the newly emerging minor damage areas after the operation, record their location and shape, and generate damage feature records for appearance-stable tools.

[0064] In this context, a satisfactory matching score means the matching score is higher than a preset threshold, confirming the tool's identity. Based on this, the system uses differential image processing technology to subtract the baseline texture of the pre-operation image from the post-operation image, extracting the residual region to identify newly generated minor damage. Minor damage areas include edge chipping caused by overload torque, new scratches from accidental impacts, or fatigue crack initiation points. The recorded information includes the pixel coordinates, physical dimensions (length, width, and depth estimates), and morphological classification of the damaged area. For example, after confirming the wrench's identity (99.8% matching score), the system detects a new linear scratch at the hexagonal edge, 0.3mm long and 0.01mm deep. The system records the scratch's location coordinates and morphological parameters in the damage feature record.

[0065] This application achieves a dual function of fingerprint-level authentication and micro-damage detection for appearance-stabilized tools through the synergy of high-resolution imaging equipment and feature point cloud registration algorithms. By acquiring microscopic surface images of the same area before and after operation and calculating the matching score using the registration algorithm, it can not only accurately determine whether a tool has been switched or unused, eliminating management loopholes, but also, after confirming its identity, further mine residual information in the image to identify newly emerging micro-damage, extending tool condition perception from macroscopic integrity to the microscopic defect level. This mechanism allows maintenance quality traceability to go beyond simply verifying the quantity of tools, delving into the inversion of tool usage behavior and early warning of potential risks. For example, identifying tiny edge scratches can indicate potential slippage or over-torque during operation, thus providing calibration suggestions for maintenance personnel and preventing subsequent maintenance quality accidents caused by hidden tool damage.

[0066] Example 7: In one possible implementation, this application embodiment provides a process for generating a maintenance quality traceability report, which further includes a state matching analysis based on the tool initial information set and the tool change information set, and a report generation step.

[0067] Step 1: Based on the initial information set and the change information set of the tool, perform state matching analysis under the category of each tool to extract the state change characteristics of the tool before and after the operation; State matching analysis refers to the process of logically comparing the tool baseline data registered in the preceding time phase with the tool change data collected in the subsequent time phase. The initial tool information set includes a unique identifier for each tool, a category label (deformation and consumption type, property change type, or appearance stability type), and baseline parameters such as initial geometric dimensions, mass, and texture. The tool change information set records the tool's deformation, performance change level, or damage characteristics after the operation. This step aims to quantify the amount of physical or chemical changes to the tool throughout its entire maintenance lifecycle by comparing the two datasets. Specifically, for deformation-related tools, the system superimposes the component topology reconstructed from post-operation point cloud data with the initial geometric dimensions to extract the area of ​​macroscopic geometric deformation regions and spatial topological deviation values ​​as state change features. For property-related tools, the system calculates the difference between the remeasured mass change, surface state parameters, and chemical change and the initial performance threshold to extract the performance degradation magnitude and material degradation degree as state change features. For appearance-stabilized tools, the system uses the matching score calculated by the micro-texture registration algorithm and the location of newly identified micro-damage areas to extract identity consistency features and newly added wear patterns as state change features. For example, when processing a certain aero-engine sealing gasket, the system extracted a 3mm tear deformation at its edge and a 5° rotational offset in the relative position of the components; these specific values ​​are the state change features of the tool. Through this multi-dimensional feature extraction, subtle anomalies during tool use can be accurately captured, providing objective data support for subsequent quality assessment.

[0068] Step 2: Compare the status change characteristics with the maintenance quality association rules one by one to determine the return status, functional failure risk, and potential impact on the equipment being maintained for each tool. The maintenance quality association rules are a pre-defined logical judgment library used to transform abstract state change characteristics into specific management decision conclusions. This rule library includes two main categories: deformation association rules and performance association rules, corresponding to the risk assessment logic of different types of tools. Deformation association rules are mainly used to assess the structural integrity of deformation-consuming tools and their fit precision with the equipment being maintained. By setting plastic deformation thresholds and fit clearance tolerance ranges, it determines whether the tool has reached the mandatory scrapping standard or whether components need to be replaced. Performance association rules are mainly used to assess the material life and chemical safety of tools with changing properties. By combining historical performance degradation curves and material degradation critical thresholds, it determines whether the tool has the risk of material failure, requires early replacement, or requires in-depth chemical testing. This step is executed by the inference engine of the maintenance quality traceability system, which automatically traverses and matches the corresponding association rule clauses based on the extracted state change characteristics. For example, when a severe abnormality in the resistance of an anti-static wrist strap is detected, accompanied by an increase in mass, the system matches the performance association rule regarding radioactive contamination, directly determining that the tool has a high risk of functional failure and inferring a potential impact on the equipment being repaired (such as nuclear fuel assemblies), thus generating a return status instruction requiring isolation. Similarly, when a new scratch appears on the edge of an Allen wrench, the system matches a variant of the deformation association rule, determining a risk of over-torque during operation, and prompting a visual inspection of the equipment being repaired (bearing housing). This item-by-item comparison mechanism achieves automated mapping from data features to risk levels, effectively avoiding subjective errors from manual interpretation.

[0069] Step 3: Maintenance quality association rules include deformation association rules and performance association rules; The deformation correlation rule is a dedicated judgment logic for deformation-consumable tools. Its core lies in comprehensively considering the area of ​​the macroscopic geometric deformation region and the component's spatial topology deviation value. This rule not only focuses on the deformation amount of a single operation but also incorporates the cumulative area of ​​historical deformation for superposition calculation to obtain the tool's total deformation rate, which serves as the basis for determining whether the tool has experienced plastic failure. Simultaneously, this rule compares the spatial topology deviation value with the design tolerance range to assess whether tool deformation will lead to assembly accuracy deviations. The performance correlation rule is a dedicated judgment logic for tools with property changes. Its core lies in dynamically assessing the remaining material life and chemical stability. This rule predicts the tool's remaining material life by fitting the current chemical change amount with the initial material properties and historical decay curves. Combined with the performance change level (mild, severe anomaly) and the detection results of unexpected gas types, it outputs a graded conclusion indicating material failure, the need for early replacement, or the need for in-depth testing. These two types of rules together constitute the decision-making brain for maintenance quality traceability, ensuring that tools with different physical characteristics can be subject to the most accurate evaluation standards. For example, deformation correlation rules can prevent sudden fracture accidents caused by undetected cumulative micro-deformation, while performance correlation rules can provide early warning of loss of protective function due to chemical corrosion.

[0070] Step 4: Based on the comparison results of all tools, generate and output a maintenance quality traceability report.

[0071] The maintenance quality traceability report is a summary document of the status assessment conclusions of all tools involved in the maintenance operation. This report integrates the return status of each tool (e.g., normal return, forced scrapping, requiring isolation, requiring calibration), functional failure risk level (high, medium, low), and suggestions for potential impact on the equipment being maintained (e.g., re-inspecting blades, sampling air, checking screw hole positions). The generation process involves structurally encapsulating the independent judgment results of each tool and linking them to the work order number, operator information, and operation timestamp, forming an immutable closed-loop data record. This report not only serves as the basis for tool warehousing or scrapping but also as a key credential for maintenance quality acceptance, directly guiding subsequent maintenance plan adjustments or process optimization. For example, if the report contains abnormal warnings for multiple tools, the system will automatically trigger a stop-work re-inspection process until the hidden danger is eliminated. This achieves a fully digital closed loop in the maintenance process, from tool requisition to quality assessment, improving maintenance reliability and management efficiency in industries with high safety requirements.

[0072] This application constructs a rigorous state-rule-decision reasoning chain through the synergistic effect of the above steps. Based on the precise matching of the initial tool information set and the tool change information set, the system can extract state change characteristics that reflect the actual wear and tear of the tool; with the help of a complete logic library covering deformation association rules and performance association rules, these characteristics are transformed into clear judgments of return status, functional failure risk, and potential impact on the equipment; finally, the comprehensive maintenance quality traceability report not only solves the problem that existing systems cannot detect the physical and chemical changes of the tool, but also makes hidden quality hazards explicit.

[0073] Example 8: In another optional embodiment, this application provides a method for calculating the total deformation rate and determining whether a component is scrapped or replaced in a deformation association rule. The method includes: Step 1: Obtain the area of ​​the macroscopic geometric deformation region of the deformation-consuming tool and the spatial topological deviation value between the components after deformation; The macroscopic geometric deformation area refers to the projected area or total surface area of ​​the tool during maintenance operations, caused by irreversible changes in surface or volume due to stress, wear, or plastic deformation. This area data is obtained by performing a differential operation between the 3D point cloud data acquired in post-concurrent phases and the reference geometric model recorded in the initial tool information set. Specifically, the system identifies all mesh cells that exceed the initial reference envelope and sums their areas. The spatial topology deviation value refers to the degree of deviation of the relative positional relationships (such as angles, distances, and coaxiality) between the components inside the tool from the initial state. This value is obtained by calculating the coordinate transformation matrix of the component feature points before and after deformation using a keypoint matching algorithm. For example, for an aero-engine sealing gasket, the macroscopic geometric deformation area may include the unfolded area of ​​the edge tear and the lateral area of ​​the overall compression deformation, while the spatial topology deviation value reflects the rotational offset angle or axial misalignment distance between the metal skeleton and the sealing ring. By simultaneously quantifying the local damage area and the overall structural misalignment, the multidimensional deformation characteristics of the tool under complex working conditions can be comprehensively captured, providing data input for subsequent comprehensive evaluation.

[0074] Step 2: Overlay the area of ​​the macroscopic geometric deformation region with the cumulative area of ​​the tool's historical deformation recorded in the initial tool information set, and calculate the tool's total deformation rate; The total tool deformation rate is a core indicator characterizing the cumulative damage over the tool's entire lifecycle. Its calculation logic involves linearly or non-linearly superimposing the area of ​​the macroscopic geometric deformation region generated by the current operation with the cumulative deformation area already formed in the tool's historical maintenance records. The historical cumulative deformation area comes from the tool's initial information set, which is dynamically updated after each maintenance operation, recording the total permanent deformation generated by the tool in all operations since it was put into use. Specifically, the system reads the historical cumulative value corresponding to the tool's unique identifier from the database. In addition to the newly detected deformation area The current total deformation area is obtained. The total deformation rate of the tool is then calculated as a percentage by dividing the total deformation area by the standard reference area or the initial effective area of ​​the tool. Taking a certain type of high-pressure sealing ring as an example, if its historical cumulative deformation area is 15mm², the newly added tearing and compression deformation area in this operation is 5mm², and the standard reference area is 100mm², then the calculated total deformation rate of the tool is 20%. This dynamic cumulative assessment mechanism effectively avoids the limitations of judging based solely on the deformation of a single operation, and can keenly capture potential failure risks caused by long-term fatigue effects. Even if the single deformation is small, it can be identified in time when the cumulative effect has reached a critical state.

[0075] Step 3: If the total deformation rate of the tool exceeds the preset plastic deformation threshold of the tool, the deformation correlation result is determined to be forced scrap. The preset plastic deformation threshold is a critical value pre-set based on the mechanical properties of the tool material (such as yield strength and fracture toughness) and a safety factor. It is used to determine whether a tool has lost its structural integrity for continued use. This threshold can be differentiated for different tool models and materials, and is usually stored in a maintenance quality-related rule base. When the calculated total deformation rate of the tool exceeds the preset threshold, it indicates that the tool material has entered the deep plastic deformation stage or is close to the fracture limit. Continued use will greatly increase the probability of tool breakage, functional failure, or safety accidents. Therefore, the system automatically triggers a mandatory scrapping command. For example, for titanium alloy fasteners, if the preset plastic deformation threshold is 25%, and the total deformation rate calculated in a certain test reaches 26%, the system will directly determine that the fastener must be forcibly scrapped and strictly prohibit it from re-entering the maintenance site. This judgment logic ensures the safety of high-stress critical components and eliminates potential quality hazards caused by defective tools from the source.

[0076] Step 4: If the total deformation rate of the tool is below the plastic deformation threshold, but the spatial topology deviation causes the clearance between the tool and the equipment being repaired to exceed the design tolerance range, then the deformation correlation result is determined to be that the component needs to be replaced.

[0077] The design tolerance range refers to the allowable dimensional error range of the assembly interface of the equipment being repaired, and is a key parameter to ensure the operational accuracy and sealing performance of the equipment. Even if the total deformation rate of the tool does not exceed the scrap threshold of the material itself, indicating that the main material of the tool has not failed, if the spatial topological deviation value of its internal components (such as bending angle, torsion) is too large, causing the actual mating clearance between the tool and the equipment being repaired during installation or being installed to exceed the upper or lower limit of the tolerance allowed by the design specifications, it means that the tool can no longer meet the requirements of precision assembly. At this time, the system determines that the deformation correlation result requires component replacement, that is, although the tool has not reached the material scrapping standard, it cannot perform the current maintenance task due to loss of geometric accuracy and must be replaced or corrected. For example, the total deformation rate of a precision locating pin is only 10% (below the scrap threshold of 15%), but its axis has undergone a bending offset of 0.05mm (spatial topological deviation), causing the mating clearance when it is inserted into the bearing seat hole to increase from the designed 0.01mm to 0.06mm, exceeding the design tolerance range of 0.03mm. The system will determine that the locating pin needs component replacement. This dual-judgment mechanism takes into account both material lifespan and assembly precision, preventing secondary damage to equipment or a decline in maintenance quality caused by tool geometric deformation.

[0078] This application, through the construction of the aforementioned deformation correlation rules, achieves a leap from single instantaneous detection to cumulative assessment throughout the entire lifecycle. By superimposing the area of ​​the macroscopic geometric deformation region with the cumulative area of ​​historical deformation to calculate the tool's total deformation rate, it considers not only the impact damage of the current operation but also the cumulative effect of long-term fatigue. This allows for the accurate identification of tools with small individual deformations but already on the verge of fatigue failure, avoiding missed detections. Simultaneously, the introduction of a comparison logic between spatial topological deviation values ​​and fit clearance design tolerances compensates for the shortcomings of focusing solely on material deformation rate, ensuring the tool's applicability at the geometric accuracy level. The synergistic effect of these two aspects enables the system to both force scrapping based on material limits and determine the need to replace components based on assembly accuracy, forming a multi-layered, high-precision quality risk control system.

[0079] Example 9: In another optional embodiment, this application provides a method for determining performance association rules, the method comprising the following steps: Step 1: Obtain the performance change level and chemical change amount of the property change tool; The performance change level refers to a quantitative classification based on the comparison of multimodal detection data (such as mass change, surface state parameters, etc.) of the tool in the subsequent time phase with a preset threshold. Specifically, it can include three levels: normal, slightly abnormal, and severely abnormal, used to characterize the degree of deviation of the tool's current macroscopic performance. The chemical change amount refers to the material chemical degradation index calculated by matching the characteristic gas composition and concentration emanating from the tool surface collected by a micro gas sensor array with a chemical degradation gas knowledge base, reflecting the changes in the molecular structure inside the tool material or the degree of surface chemical reaction. For example, for an antistatic wristband, if its resistance value is detected to increase from the initial 3.8MΩ to 45MΩ, and a trace amount of radioactive dust is adsorbed on the surface causing a slight increase in mass, the system classifies its performance change level as severely abnormal; at the same time, if the gas sensor does not detect abnormal gas, the chemical change amount is recorded as zero or the baseline value. This step aims to provide a core input data foundation for subsequent life cycle prediction and disposal decisions, ensuring that the judgment criteria cover both macroscopic performance and microscopic chemical evolution.

[0080] Step 2: Compare the amount of chemical change with the initial material properties and historical performance degradation curves recorded in the initial information set of this tool, and calculate the predicted value of the remaining life of the current tool material. The initial material properties include the tool's material composition, density, hardness, and chemical stability parameters at the time of manufacture. These data are registered and uniquely identified during the tool's initial information set generation phase. The historical performance degradation curve is a time-performance relationship model fitted based on the performance degradation data of similar tools in multiple past maintenance operations, describing the natural aging law of materials under specific working conditions. The process of calculating the predicted remaining life of the current tool material involves substituting the real-time collected chemical change amount into the historical performance degradation curve, combining it with the initial material properties for deviation correction, thereby estimating the estimated time or remaining number of operations required for the tool to degrade from its current state to the failure threshold. For example, for a nozzle of an ion etching machine, the system reads that its initial material is silicon carbide, and the historical curve shows that the thickness decreases by 0.01 mm per 100 hours of operation in a fluorine-based plasma environment. This test found that the chemical erosion corresponding to the silicon tetrafluoride gas concentration is equivalent to the loss of 150 hours of normal operation. Based on this, the system dynamically adjusts the degradation slope and calculates that the predicted remaining life of the nozzle is only 60% of the time of the next maintenance cycle. This algorithm, which combines static attributes with dynamic trends, can accurately quantify the remaining usable value of tools and avoid misjudgments caused by relying solely on a single detection result.

[0081] Step 3: If the performance change level is severe and the amount of chemical change exceeds the critical threshold for material degradation, the performance correlation result is determined to be that the material is unusable due to failure. The critical threshold for material degradation is an upper limit of chemical change set according to the safety baseline of the tool. Once this threshold is exceeded, it means that the internal structure of the material has been irreversibly damaged or there are serious safety hazards (such as radioactive contamination or residues of highly toxic substances). The judgment logic adopts an AND gate mechanism, requiring that severe macroscopic performance deterioration and deep microscopic chemical degradation occur simultaneously to eliminate false alarms caused by a single sensor failure. For example, when a protective glove used for the maintenance of a nuclear facility is judged to have severely abnormal performance (such as a hole or resistor failure), and the gas sensor detects that the concentration of characteristic gas escaping from its surface exceeds the critical threshold for material decomposition and release, the system will immediately lock the tool's status, determine that the material is unusable due to failure, and trigger a mandatory scrapping process, prohibiting it from entering the maintenance site again, thereby completely preventing the risk of personal injury or environmental pollution caused by tool failure.

[0082] Step 4: If the performance change level is slightly abnormal but the amount of chemical change shows a continuous accelerating upward trend and the predicted remaining life of the material is lower than the duration of the next maintenance cycle, then the performance correlation result is determined to be that the material needs to be replaced in advance. The continuous accelerating upward trend can refer to a gradual increase in the increment of chemical change over multiple consecutive work cycles, indicating that the tool is in a rapid deterioration phase rather than a stable linear wear. This judgment logic focuses on the match between the rate of tool performance degradation and its remaining life. Even if the current macroscopic performance only shows a slight abnormality (not yet completely failed), if it is predicted to fail before the next maintenance cycle, a preventative replacement mechanism must be initiated. For example, in the monitoring of a precision bearing grease, although its viscosity decrease is still within the range of slight abnormality, the release rate of volatile organic compounds (VOCs) shows an exponential increase, and the calculated remaining lubrication life is only 20 hours, while the next planned maintenance is scheduled for 50 hours later. In this case, the system determines that replacement is necessary in advance and automatically generates a spare parts requisition instruction, arranging for immediate replacement after the current work order ends, effectively avoiding catastrophic failure due to lubrication failure in the next equipment cycle.

[0083] Step 5: If the performance change level is normal but an unexpected gas type appears, the performance correlation result is determined to require in-depth chemical testing.

[0084] Unexpected gas types refer to chemical substances that should theoretically not be produced under the current tool materials and standard operating conditions. Their presence often indicates abnormal chemical reactions, the introduction of foreign contaminants, or uncontrolled process parameters. This judgment logic reflects a keen ability to detect potential hidden risks, that is, identifying potential crises through the abnormal fingerprints of trace gases when routine performance indicators appear to be qualified. For example, in the maintenance work of a semiconductor cleanroom, the performance parameters (such as size, weight, and surface cleanliness) of a certain stainless steel tool are all within the normal range, but the gas sensor detects trace amounts of chlorine (Cl2), even though the tool material does not contain chlorine and there is no chlorine source in the working environment; the system immediately determines that in-depth chemical testing is required, automatically triggering a secondary analysis process using a mass spectrometer to check for hidden cleaning fluid residues, cross-contamination, or unknown corrosion reactions, thereby eliminating potential quality hazards in their infancy.

[0085] This application achieves a shift in maintenance quality management from passive response to proactive prediction through a progressive judgment based on the aforementioned performance correlation rules. By acquiring the performance change level and chemical change amount of tools with property changes, and combining this with initial material properties and historical performance degradation curves to calculate the predicted remaining life, the system can not only identify the current failure state but also foresee future degradation trends. Based on this, differentiated handling strategies are set for different scenarios: for extreme cases of severe anomalies and supercritical thresholds, the material is directly deemed unusable, ensuring zero remaining risk; for gradual changes of mild anomalies, accelerated trends, and insufficient life, early replacement is required for preventative maintenance, avoiding sudden downtime; for cases of normal performance but concealed unexpected gases, in-depth chemical testing is required to uncover potential process defects. This application collectively constitutes a highly sensitive and reliable tool health assessment system, improving the safety and economy of maintenance operations.

[0086] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0087] Figure 3 This is a schematic diagram of the structure of a closed-loop traceability system for maintenance tools based on pre- and post-operation dual-phase recognition, as provided in an embodiment of this application. Figure 3 As shown, the maintenance tool closed-loop traceability system 300 based on dual-phase recognition before and after operation in this embodiment includes: an initial information module 301, a change information module 302, and a traceability report module 303.

[0088] The initial information module 301 is used to acquire the tool pre-operation feature dataset before the maintenance operation, and based on the tool pre-operation feature dataset, to perform initial attribute registration and change type classification for each tool, and generate a tool initial information set; The change information module 302 is used to acquire the tool post-operation feature dataset in the post-maintenance phase. Based on the category label of each tool in the tool initial information set, it processes the data according to the differentiated acquisition strategy and feature deconstruction logic corresponding to the deformation consumption class, property change class and appearance stability class to generate the tool change information set. The traceability report module 303 is used to perform state matching analysis under its category for each tool based on the tool initial information set and the tool change information set, extract the state change features of the tool before and after the operation, compare the state change features with the maintenance quality association rules, and generate a maintenance quality traceability report.

[0089] The system in this embodiment can be used to execute the methods of any of the above embodiments, and its implementation principle and technical effect are similar, so they will not be described again here.

Claims

1. A closed-loop traceability method for maintenance tools based on dual-phase identification before and after operation, characterized in that, include: Obtain the tool pre-operation feature dataset for the pre-operation phase of the maintenance operation. Based on the tool pre-operation feature dataset, perform initial attribute registration and change type classification for each tool to generate an initial tool information set. Obtain the tool post-maintenance feature dataset after the maintenance operation. Based on the category label of each tool in the initial tool information set, process it according to the differentiated acquisition strategy and feature deconstruction logic corresponding to the deformation consumption class, property change class and appearance stability class to generate tool change information set. Based on the initial information set and the change information set of the tools, a state matching analysis is performed on each tool according to its category. The state change features of the tool before and after the operation are extracted, and the state change features are compared with the maintenance quality association rules to generate a maintenance quality traceability report.

2. The method according to claim 1, characterized in that, The process of generating the initial information set for the tool includes: Obtain the tool pre-feature dataset, which includes the shape outline, surface texture, initial quality, and component topology of each tool; Based on the shape and surface texture, and combined with the material coding rules, each tool is initially registered for attributes to generate a unique tool identifier. Based on the component topology and operating environment characteristics, each tool is automatically classified into one of the following categories: deformation consumption, property change, or appearance stability, and a unique tool identifier is bound to it to generate an initial information set for the tool.

3. The method according to claim 2, characterized in that, The process of generating the tool change information set includes: Based on the category label of each tool in the initial tool information set, differentiated acquisition strategies and feature deconstruction logic are selected for the deformation consumption class, property change class, and appearance stability class, respectively: For deformation-consuming tools, perform deformation feature capture and component topology reconstruction; For tools that involve changes in properties, perform multimodal performance parameter remeasurement and deviation calculation; For tools with stable appearance, perform high-precision registration of micro-textures; The post-feature data obtained from the deformation consumption tools, the property change tools, and the appearance stability tools are summarized to generate the tool change information set.

4. The method according to claim 3, characterized in that, The process of performing deformation feature capture and component topology reconstruction includes: The point cloud data of the deformation-consuming tool after the operation is collected by a 3D scanning device, and all macroscopic geometric deformation areas in the current outline of the tool that exceed the initial reference range are identified. The point cloud data after the operation is matched with the initial geometric dimensions and relative positions of components recorded in the initial information set of the tool to find residual feature regions and infer the source identity of the deformed components. Simultaneously, the number and shape of all the deformable components are counted, the spatial topological relationship between the components is reconstructed, and the deformation feature record of the deformation consumption tool is generated.

5. The method according to claim 3, characterized in that, The process of performing multimodal performance parameter remeasurement and deviation calculation includes: The mass data of the tool with the property change type is obtained by intelligent weighing equipment at the post-operation stage, and compared with the initial mass to calculate the mass change. AI visual recognition technology is used to collect images of the tool surface, and deep learning convolution algorithms are used to identify the surface cleanliness level, coating integrity and color characteristics to obtain surface state parameters. The composition and concentration of characteristic gases escaping from the tool surface are collected by a micro gas sensor array, and the chemical changes of the tool material are calculated by matching them with knowledge of chemical degradation gases. The mass change, surface state parameters, and chemical change are used together as performance change characteristics, and compared with preset performance thresholds to classify performance change levels.

6. The method according to claim 3, characterized in that, The high-precision registration of micro-textures includes: Microscopic surface images of the appearance-stabilized tool in the same region as the tool's initial information set in the subsequent time phase were acquired using a high-resolution imaging device. A feature point cloud registration algorithm is used to match minute scratches, wear patterns or laser-engraved edges in two images before and after the operation, and the matching score is calculated. If the matching degree is lower than the preset threshold, it is determined that the tool has been swapped or has not been used. If the matching degree is qualified, the newly emerging minor damage areas after the operation are further identified, their location and shape are recorded, and the damage feature record of the appearance-stabilized tool is generated.

7. The method according to claim 3, characterized in that, The process of generating the maintenance quality traceability report includes: Based on the initial information set and the change information set of the tools, a state matching analysis is performed on each tool according to its category to extract the state change features of the tool before and after the operation. The state change characteristics are compared with the maintenance quality association rules one by one to determine the return status, functional failure risk and potential impact on the equipment being maintained for each tool. The maintenance quality association rules include deformation association rules and performance association rules; Based on the comparison results of all tools, the maintenance quality traceability report is generated and output.

8. The method according to claim 7, characterized in that, The deformation association rules include: Obtain the area of ​​the macroscopic geometric deformation region of the deformation-consuming tool and the spatial topological deviation value between the components after deformation; The total deformation rate of the tool is calculated by superimposing the area of ​​the macroscopic geometric deformation region with the cumulative area of ​​the tool's historical deformation recorded in the tool's initial information set. If the total deformation rate of the tool exceeds the preset plastic deformation threshold of the tool, the deformation correlation result is determined to be forced scrap. If the total deformation rate of the tool is below the plastic deformation threshold, but the spatial topology deviation causes the clearance between the tool and the equipment being repaired to exceed the design tolerance range, then the deformation correlation result is determined to be that the component needs to be replaced.

9. The method according to claim 7, characterized in that, The performance association rules include: Obtain the performance change level and chemical change amount of the aforementioned property change tools; The chemical change is compared with the initial material properties and historical performance degradation curves recorded in the initial information set of this tool to calculate the predicted remaining life of the current tool material. If the performance change level is severe abnormality and the amount of chemical change exceeds the critical threshold for material degradation, then the performance correlation result is determined to be that the material is unusable due to failure. If the performance change level is slightly abnormal but the amount of chemical change shows a continuous accelerating upward trend and the predicted remaining life of the material is lower than the duration of the next maintenance cycle, then the performance correlation result is determined to be that the material needs to be replaced in advance. If the performance change level is normal but an unexpected gas type appears, the performance correlation result is determined to require in-depth chemical testing.

10. A closed-loop traceability system for maintenance tools based on dual-phase recognition before and after operation, characterized in that: The method applied to any one of claims 1-9 includes: The initial information module is used to acquire the tool pre-operation feature dataset before maintenance work, and based on the tool pre-operation feature dataset, to perform initial attribute registration and change type classification for each tool, and generate tool initial information set; The change information module is used to acquire the tool post-operation feature dataset in the post-maintenance phase. Based on the category label of each tool in the initial tool information set, it processes the data according to the differentiated acquisition strategy and feature deconstruction logic corresponding to the deformation consumption class, property change class, and appearance stability class to generate the tool change information set. The traceability report module is used to perform state matching analysis under its category for each tool based on the tool's initial information set and the tool's change information set, extract the state change features of the tool before and after the operation, compare the state change features with the maintenance quality association rules, and generate a maintenance quality traceability report.

Images