Device wear intelligent identification method and system based on deep learning

Abstract

The invention relates to the technical field of artificial intelligence, and discloses an equipment wear intelligent identification method and system based on deep learning, and the method comprises the steps: collecting a vibration signal in an equipment operation state, analyzing the surface damage characteristics of a key part, and determining the wear stage of the key part; inputting the stage wear data into a preset deep learning network, and outputting an equipment low-wear mode adapted to different working conditions; then real-time wear parameters of the equipment in a low-wear mode are collected, and a global wear map is constructed in combination with the operation vibration signals; and finally, intelligent identification of the wear state of the target equipment is realized through self-adaptive decision-making of each wear contrast protocol in the atlas. According to the invention, the equipment wear identification efficiency and accuracy can be improved.

Description

Technical Field

[0001] This invention relates to a method and system for intelligent identification of equipment wear based on deep learning, belonging to the field of artificial intelligence technology. Background Technology

[0002] Equipment wear is one of the main causes of equipment performance degradation and failure in industrial production. Intelligent identification aims to monitor and evaluate the wear status of equipment in real time through automation technology, so as to prevent unexpected downtime and extend equipment life.

[0003] Currently, equipment wear identification mainly relies on traditional methods such as regular manual inspections, simple vibration analysis, or fixed threshold alarms. These methods suffer from problems such as scattered monitoring data and limited analysis tools, resulting in inaccurate wear status assessments, poor early warning timeliness, and an inability to adapt to the dynamic changes in equipment wear under complex operating conditions. In addition, existing systems lack the ability to deeply integrate multi-source data and make intelligent decisions, making it difficult to achieve accurate identification in the early stages of wear. Summary of the Invention

[0004] This invention provides a method and system for intelligent identification of equipment wear based on deep learning, the main purpose of which is to improve the efficiency and accuracy of equipment wear identification.

[0005] To achieve the above objectives, the present invention provides a deep learning-based intelligent device wear identification method, comprising: After collecting the vibration signals of the target equipment during operation, the surface damage characteristics of key components in the target equipment are analyzed. Based on the surface damage characteristics, the wear stage of the key components in the target equipment is determined; After inputting the stage wear data corresponding to the wear stage into a preset deep learning network, the system outputs a low-wear mode for the target device under different operating conditions. Real-time wear parameters of the target equipment are collected under the low-wear mode of the equipment. Based on the real-time wear parameters and the operating vibration signal, a global wear map corresponding to the target equipment is constructed. Adaptive decision-making is performed on each wear comparison protocol in the global wear map to achieve intelligent wear identification of the target device.

[0006] Optionally, constructing a global wear map corresponding to the target device based on the real-time wear parameters and the operating vibration signal includes: The key wear elements in the real-time wear parameters and the vibration floating elements in the operating vibration signal are extracted respectively. After aligning the key wear elements with the vibration floating elements, the wear distribution in the target equipment is determined based on the aligned element relationship pairs. Based on the wear comparison criteria in the wear distribution, a global wear map corresponding to the target equipment is constructed.

[0007] Optionally, determining the wear distribution in the target device based on the aligned feature relationship pairs includes: Query the wear relationship hotspots in the aligned feature relationship pairs; After mapping the wear relationship hotspots to specific component locations in the target device, the wear distribution corresponding to the specific component locations is determined.

[0008] Optionally, the analysis of surface damage characteristics of key components in the target device includes: After converting the operating vibration signal into a component vibration spectrum, a surface wear image corresponding to the key wear spectrum in the component vibration spectrum is generated; Locate the specific damaged area in the surface wear image; The damage features in the specific damage area are matched with a preset damage type library to obtain surface damage features.

[0009] Optionally, determining the wear stage of a key component in the target device based on the surface damage characteristics includes: Analyze the damage type and damage size corresponding to the surface damage features; The degree of wear of the key components in the target equipment is determined by the damage type and the damage size. After marking the wear status identifier corresponding to the key component, the wear stage of the key component is determined based on the wear degree quantity and the wear status identifier.

[0010] Optionally, inputting the stage wear data corresponding to the wear stage into a preset deep learning network includes: The wear data corresponding to the wear stage is organized into a wear input sequence to convert the wear input sequence into a sequence format tensor. After the sequence format tensor is input into the input layer of a preset deep learning network, the deep learning network is started to perform a backward processing of the sequence format tensor.

[0011] Optionally, the output adapts to the low-wear mode of the target equipment under different operating conditions, including: Obtain the operation adjustment parameters and load configuration parameters corresponding to the target device output by the deep learning network; After matching the operation adjustment parameters with the current operating conditions of the equipment, the optimal load power corresponding to the target equipment is set according to the matched operating conditions and the load configuration parameters. Based on the optimal load power, the low-wear mode of the target equipment under different operating conditions is dynamically adjusted.

[0012] Optionally, the adaptive decision-making for each wear comparison protocol in the global wear map includes: After reading the key wear indicators corresponding to the wear comparison protocol in the global wear map, analyze the wear change trend corresponding to the key wear indicators; By using a preset decision tree, a set of decision instructions corresponding to the wear change trend is generated; Identify the corresponding instructions in the decision instruction set to achieve adaptive decision-making for each of the wear comparison protocols.

[0013] Optionally, the wear comparison protocol includes: Key component identifiers are used to uniquely identify the monitoring components in the target device; Operating condition tags are used to record the environmental parameters of the target device during operation; Decision priority indicators are used to determine the order in which decisions are made based on the severity of wear and tear.

[0014] To address the above problems, the present invention also provides a deep learning-based intelligent device wear recognition system, the system comprising: The feature analysis module is used to collect the operating vibration signal of the target equipment in operation and then analyze the surface damage characteristics of key components in the target equipment. The stage determination module is used to determine the wear stage of the key components in the target device based on the surface damage characteristics. The mode output module is used to input the stage wear data corresponding to the wear stage into a preset deep learning network and then output a low-wear mode of the device that is adapted to the target device under different working conditions. The atlas construction module is used to collect real-time wear parameters of the target equipment under the low-wear mode of the equipment, and construct a global wear atlas corresponding to the target equipment based on the real-time wear parameters and the operating vibration signal. The wear identification module is used to make adaptive decisions on each wear comparison protocol in the global wear map to realize intelligent wear identification of the target device.

[0015] Compared to the problems described in the background art, the embodiments of the present invention, by collecting vibration signals during equipment operation and analyzing the surface damage characteristics of key components, can capture dynamic information related to component damage in real time. This directly correlates the actual operating state of the component with the damage manifestation, improving the targeting and accuracy of damage feature extraction. Furthermore, the embodiments of the present invention can accurately define the degree of wear, avoiding ambiguous judgments of component state. Through the direct correlation between damage features and wear stages, the progress of component wear is clearly presented, eliminating state perception bias and facilitating timely response to wear changes, ensuring the stability and continuity of equipment operation. Finally, the embodiments of the present invention, through the output of a low-wear mode, can accurately adapt to different operating conditions of the target equipment, avoiding mode adaptation bias caused by differences in operating conditions, and through deep learning, the stage... In-depth mining of wear data enables the low-wear mode to fully align with the actual wear characteristics of the equipment, improving the accuracy of the mode. Furthermore, this invention integrates multi-dimensional wear-related information during equipment operation, avoiding cognitive limitations caused by single data dimensions. Constructing a global wear map systematically presents the wear correlation status of each key component of the equipment, enhancing the overall comprehensiveness and objectivity of wear status assessment. Finally, this invention enables the wear identification process to accurately align with the wear characteristics corresponding to each protocol. This decision-making method dynamically adapts to complex changes in wear status, avoiding identification bias caused by fixed decision logic, ensuring adaptability to different wear scenarios, and achieving comprehensive control over the equipment's wear status. This makes intelligent wear identification more accurate. Therefore, this invention can improve the efficiency and accuracy of equipment wear identification. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating a deep learning-based intelligent identification method for device wear according to an embodiment of the present invention. Figure 2 A flowchart for determining the wear stage of a key component based on surface damage features, provided as an embodiment of the present invention; Figure 3 A schematic diagram of a module for implementing the deep learning-based intelligent device wear identification method according to an embodiment of the present invention; Figure 4 A schematic diagram of a computer device for a deep learning-based intelligent device wear recognition method according to an embodiment of the present invention; The objectives, features, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0017] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0018] This application provides a deep learning-based intelligent device wear identification method. The executing entity of the deep learning-based intelligent device wear identification method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application: a server, a terminal, etc. In other words, the deep learning-based intelligent device wear identification method can be executed by software or hardware installed on a terminal device or a server device. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster.

[0019] Reference Figure 1 The diagram shown is a flowchart illustrating a deep learning-based intelligent device wear identification method according to an embodiment of the present invention. In this embodiment, the deep learning-based intelligent device wear identification method includes: S1. After collecting the vibration signal of the target equipment in operation, analyze the surface damage characteristics of the key components in the target equipment.

[0020] This invention collects vibration signals from equipment operation and analyzes surface damage characteristics of key components, enabling real-time capture of dynamic information related to component damage. This information can directly correlate the actual operating state of the component with its damage manifestations, thereby improving the targeting and accuracy of damage feature extraction.

[0021] The target equipment refers to various mechanical equipment in industrial production scenarios that require wear condition monitoring. Its core characteristics are that it contains core components that play a key supporting role in the overall operating performance of the equipment, and it is in a continuous or intermittent operating state under actual working conditions. It is not idle, shut down for maintenance, or experimental equipment in a laboratory environment. It is the direct object of wear identification in this method, and its operating state is directly related to the collection and analysis process of wear monitoring data. The operating vibration signal refers to the physical signal transmitted in the form of waves caused by physical behaviors such as mechanical movement of internal components, contact friction between components, load action, or structural vibration during the operation of the target equipment. This signal contains quantifiable characteristic parameters such as vibration frequency, amplitude, and phase, which can objectively reflect the mechanical dynamic changes during the operation of the equipment. The surface damage features refer to the set of parameters that can accurately characterize the actual state of surface damage of key components after matching the image features of specific damage areas with a preset damage type library. These parameters include core attributes such as the specific type, morphology, size, and severity of the damage.

[0022] In this embodiment of the invention, the analysis of surface damage features of key components in the target device includes: converting the operating vibration signal into a component vibration spectrum, generating a surface wear image corresponding to the key wear spectrum in the component vibration spectrum; locating specific damage areas in the surface wear image; and matching the damage features in the specific damage areas with a preset damage type library to obtain surface damage features.

[0023] The component vibration spectrum refers to the frequency domain representation obtained by converting the original vibration signal into frequency domain data through signal processing techniques such as Fourier transform and wavelet transform. It includes quantitative parameters such as frequency distribution, peak amplitude, and bandwidth range, reflecting the mechanical vibration characteristics of the critical component during operation. Its spectral structure is directly related to the component's motion state and structural integrity. The critical wear spectrum refers to specific spectral intervals or components in the component vibration spectrum that are directly related to the surface wear state of the critical component. Its frequency characteristics and amplitude variation trends fluctuate regularly with the degree of component wear. It is the core spectral information directly related to the wear state, selected from the overall vibration spectrum, excluding redundant spectral components unrelated to wear. The surface wear image refers to the visualization of abstract spectral information transformed into visual representations using image generation algorithms, with the quantitative data of the critical wear spectrum as input. The image, with its grayscale distribution, pixel density, and contour morphology, corresponds one-to-one with the wear-related state of key components, presenting the wear-related information contained in the spectrum in an intuitive image form. The specific damage area refers to a specific image region in the surface wear image that corresponds to the actual surface damage location of the key component. The image features of this region, such as grayscale differences, contour boundaries, and pixel aggregation states, directly match the damaged parts on the component surface. It can be accurately separated from the overall image through image segmentation, edge detection, and other algorithms, and is the core image region for extracting damage features. The preset damage type library refers to a database built in advance through experimental testing, actual working condition data collection, and industry technical specification compilation. It contains common surface damage types of key components of industrial equipment, such as scratches, corrosion, peeling, and cracks, and their corresponding feature parameters, such as morphological features, size range, and feature identifiers.

[0024] S2. Based on the surface damage characteristics, determine the wear stage of the key components in the target equipment.

[0025] The embodiments of the present invention can accurately define the degree of wear, avoid the ambiguity of the component status, and clearly present the progress of component wear by directly linking damage characteristics with wear stages, eliminate the bias in state perception, help respond to wear changes in a timely manner, and ensure the stability and continuity of equipment operation.

[0026] The critical components refer to the core components of the target equipment that play a core supporting role in the overall operational performance, structural stability, and operational safety. Their proper functioning directly affects the equipment's operating efficiency and operational stability; they are not auxiliary, substitute, or non-core functional components. These components are prone to wear during equipment operation due to friction, load, environmental factors, etc., making them a key focus for equipment wear monitoring and identification. Their wear status is directly related to the overall operational status assessment of the equipment. The wear stages refer to the wear progression states with clear stage attributes, defined based on the type, scope, severity, and development trend of surface damage characteristics of critical components. This represents a process definition of component wear from its initial occurrence to its gradual development. Different stages correspond to different damage characteristics, and the stage division is based on the damage evolution law and the actual operating conditions of the equipment.

[0027] In this embodiment of the invention, determining the wear stage of a key component in the target device based on the surface damage characteristics includes: analyzing the damage type and damage size corresponding to the surface damage characteristics; determining the wear degree corresponding to the key component in the target device through the damage type and the damage size; marking the wear state identifier corresponding to the key component; and determining the wear stage of the key component based on the wear degree and the wear state identifier.

[0028] The damage type refers to the classification of damage categories with specific morphology and causes, derived from the surface damage characteristics of key components. This classification is based on the appearance of the damage, its generation mechanism, and industry-standard classification criteria, covering various forms of surface damage caused by factors such as friction, load, corrosion, and fatigue. It is a qualitative definition of the essential attributes of the damage. The damage size refers to quantifiable data obtained by measuring the geometric parameters of the surface damage characteristics of key components, including core geometric indicators such as length, width, depth, area, and volume. It objectively reflects the specific size and distribution range of the damage in the spatial dimension. The wear degree measurement... Wear status refers to a set of values ​​or parameters that characterize the severity of wear on key components, obtained through a pre-defined quantitative calculation model or evaluation algorithm based on the analyzed damage type and damage size. The magnitude of these values ​​is directly related to the degree of influence of the damage type and the quantitative results of the damage size, and is a quantitative description of the wear state. Wear status identifiers refer to standardized marks that are pre-set according to the equipment operating conditions, key component characteristics, and damage evolution laws to distinguish different wear-related states. Their form may include codes, symbols, or parameter ranges. Each identifier corresponds to a specific combination of damage-related features and is standardized reference information set to help determine the wear stage.

[0029] See Figure 2The diagram shown is a flowchart illustrating a method for determining the wear stage of a key component based on surface damage characteristics, according to an embodiment of the present invention. Figure 2 In this process, the workflow begins with "input: surface damage features." Through "analyzing surface damage features," it completes feature analysis for "extracting damage type (qualitative definition: morphology, cause of formation, industry classification standards)" and "extracting damage dimensions (quantitative data: length, width, depth, and other geometric indicators)." Then, based on the "core logic of wear stage determination," it simultaneously advances two parallel stages: "determining the degree of wear through a preset quantitative calculation model / evaluation algorithm" and "marking the wear state of key components (preset standardized markings: codes / symbols / parameter ranges)." These stages respectively achieve numerical representation of wear severity and standardized marking of wear-related states. Finally, through the convergence of the outputs from these two stages, it accurately defines the wear stage of key components, forming a hierarchical derivation logic from feature analysis to stage determination.

[0030] S3. After inputting the stage wear data corresponding to the wear stage into a preset deep learning network, output a low-wear mode for the target device under different working conditions.

[0031] The embodiments of the present invention can accurately adapt to different working conditions of the target equipment by outputting a low-wear mode, avoiding mode adaptation deviation caused by differences in working conditions. Furthermore, through deep learning to deeply mine stage wear data, the low-wear mode can fully fit the actual wear characteristics of the equipment, thereby improving the accuracy of the mode.

[0032] The stage wear data refers to the collection of various wear-related data corresponding to key components when they are in a specific wear stage. This includes core information such as damage type, damage size, wear degree, and wear status identifier obtained from the analysis of that stage, as well as related data on the basic operating parameters of the equipment during that stage. These data are all obtained based on the actual monitoring, analysis, and calculation results of that wear stage, and constitute a comprehensive quantitative and qualitative data set that fully reflects the essential characteristics of the corresponding wear stage. The pre-set deep learning network refers to an artificial intelligence network model whose network structure and parameter configuration are determined after model training and parameter optimization based on a large amount of equipment wear-related sample data, including stage wear data for different operating conditions and different wear stages, as well as corresponding low-wear mode samples. It includes functional modules such as feature extraction, deep data fusion, and pattern adaptation generation, and can automatically process the input stage wear data. Furthermore, the network structure and core parameters remain fixed during the wear identification process of this method.

[0033] In this embodiment of the invention, the step of inputting the stage wear data corresponding to the wear stage into a preset deep learning network includes: organizing the stage wear data corresponding to the wear stage into a wear input sequence to convert the wear input sequence into a sequence format tensor; after inputting the sequence format tensor into the input layer of the preset deep learning network, starting the deep learning network to perform a backward processing process on the sequence format tensor.

[0034] The wear input sequence refers to a one-dimensional or multi-dimensional data sequence formed by organizing core parameters such as damage type, damage size, and wear degree according to preset rules, such as feature dimension priority or data acquisition sequence, based on stage wear data of a specific wear stage of a key component. Its data arrangement logic is compatible with the input data receiving format of the deep learning network, fully preserving the core information of the stage wear data and the correlation between parameters. It serves as an intermediate data form for converting stage wear data into a network-processable format. The sequence format tensor refers to tensor-type data obtained by standardizing the wear input sequence according to the input data structure requirements of the deep learning network. It contains key data dimension information such as sample quantity, feature dimension, and time step, and can structurally carry the ordered data and parameter correlation attributes in the wear input sequence, eliminating format differences between different data types. It is a standard data carrier that the input layer of the deep learning network can directly receive and process.

[0035] Optionally, the backward processing of the sequence format tensor by the deep learning network can be achieved by constructing a feature enhancement module using a residual network, performing multi-scale feature fusion on the sequence format tensor, and combining batch normalization to reduce the risk of gradient vanishing.

[0036] In this embodiment of the invention, the output adapting the target device to a low-wear mode under different operating conditions includes: obtaining the operation adjustment parameters and load configuration parameters corresponding to the target device output by the deep learning network; matching the operation adjustment parameters with the current operating conditions of the device, and setting the optimal load power corresponding to the target device based on the matched operating conditions and the load configuration parameters; and dynamically adjusting the low-wear mode of the target device under different operating conditions based on the optimal load power.

[0037] The operation adjustment parameters refer to the core parameter set output by a preset deep learning network for adjusting the operating state of the target equipment. These parameters cover adjustable indicators directly related to equipment operation, such as operating speed, start / stop frequency, and operating rhythm control. Their values ​​are determined based on stage wear data and equipment wear characteristic analysis results, enabling them to respond to equipment operating needs under different working conditions. The load configuration parameters refer to parameter information related to load allocation and operating load output by the deep learning network for the target equipment. These include core content such as load allocation ratio, load upper limit threshold, and load dynamic adjustment range, providing data support for load optimization under different working conditions. The current operating condition of the equipment refers to the actual working state and environmental conditions of the target equipment at a specific point in time, including key influencing factors such as real-time workload, operating environment temperature, medium pressure, and work intensity. These factors dynamically affect the equipment wear rate. The efficiency and component operating status are the direct objects of matching operating adjustment parameters, objectively reflecting the current operating background and work requirements of the equipment. The optimal load power refers to the equipment load power value that is adapted to the current operating conditions by matching the operating adjustment parameters with the current operating conditions of the equipment and combining the load configuration parameters through quantitative calculation. Its value needs to balance the equipment operating efficiency and component wear control requirements, satisfying the power requirements of the current work task of the equipment while avoiding abnormal wear caused by excessive or insufficient load. It is the core quantitative indicator for dynamically adjusting the equipment operating mode. The low-wear mode of the equipment refers to the equipment operating mode dynamically formed by taking the optimal load power as the core and combining the results of the matching of operating adjustment parameters with the current operating conditions. This mode, by standardizing the equipment operating status and optimizing the load distribution method, clarifies the optimal combination of operating parameters for the equipment under specific operating conditions, enabling the equipment to complete the work task while minimizing the risk of component wear.

[0038] S4. Collect the real-time wear parameters of the target equipment under the low-wear mode of the equipment, and construct the global wear map corresponding to the target equipment based on the real-time wear parameters and the operating vibration signal.

[0039] The embodiments of the present invention can integrate multi-dimensional wear-related information during equipment operation, avoid the cognitive limitations caused by a single data dimension, and construct a global wear map that can systematically present the wear correlation status of each key link of the equipment, thereby improving the overall and objective nature of wear status assessment.

[0040] The real-time wear parameters refer to the set of dynamic parameters directly related to component wear that are collected in real time and continuously through monitoring means such as sensors when the target equipment is operating in low-wear mode. They include real-time change data of surface damage of key components, operating status related parameters related to wear evolution, and physical quantity data related to component friction contact, which can objectively reflect the real-time changes in the wear state of the equipment in low-wear mode.

[0041] In this embodiment of the invention, the step of constructing a global wear map corresponding to the target device based on the real-time wear parameters and the operating vibration signal includes: extracting key wear elements from the real-time wear parameters and vibration floating elements from the operating vibration signal; aligning the key wear elements and the vibration floating elements, and determining the wear distribution in the target device based on the aligned element relationship pairs; and constructing a global wear map corresponding to the target device based on the wear comparison rules in the wear distribution.

[0042] The key wear elements refer to a subset of core parameters directly related to the wear state of components, extracted from real-time wear parameters collected in the low-wear mode of the equipment. These parameters encompass key information such as the real-time rate of damage change, frictional contact strength, and dynamic surface condition indicators. They are not all data in the real-time wear parameters, but rather a selection of parameters that provide core support for wear analysis and accurately reflect the key changing characteristics of the wear state. The vibration fluctuation elements refer to characteristic parameters extracted from the vibration signals of the target equipment that fluctuate with the equipment's operating state and wear changes. These include quantifiable indicators such as the vibration frequency fluctuation range, amplitude variation, and phase shift. Their fluctuation patterns are directly related to the equipment's wear state and represent a set of core vibration features that reflect dynamic changes after removing stable vibration components. The element relationship pairs refer to one-to-one or many-to-many data combinations formed by aligning the extracted key wear elements and vibration fluctuation elements according to preset rules. Each relationship pair contains corresponding key wear element data and vibration fluctuation data. The element data can intuitively present the inherent logical relationship between the two types of elements; the wear distribution refers to the overall description of the wear state of each key component of the target equipment, derived from the logical analysis of the element relationship pairs, including the specific location of wear on each component, the distribution of wear degree differences, and the correlation influence between different components. It is not an isolated wear state of a single component, but rather covers the overall wear spatial distribution and correlation characteristics of the equipment; the wear comparison rules refer to a set of standardized rules formulated in advance based on a large amount of equipment wear test data, industry technical standards, and actual working conditions, used to standardize wear state analysis and map construction. It includes the correspondence standards between key wear elements, vibration and floating elements and wear states, distribution classification criteria, etc.; the global wear map refers to a visualized or structured data carrier formed by integrating the correlation information of key wear elements and vibration and floating elements with the wear distribution as the core and the wear comparison rules. It can comprehensively and systematically present the wear state, distribution characteristics, and correlation relationships of each key component of the target equipment.

[0043] Furthermore, in this embodiment of the invention, determining the wear distribution in the target device based on the aligned element relationship pairs includes: querying wear relationship hotspots in the aligned element relationship pairs; mapping the wear relationship hotspots to specific component locations in the target device, and then determining the wear distribution corresponding to the specific component locations.

[0044] The wear relationship hotspot refers to the core data set selected from the aligned element relationship pairs through data analysis, where the correlation strength between key wear elements and vibration floating elements is significantly higher than other correlation combinations. Its core characteristic is that the correlation degree between the two types of elements exhibits an abnormal or significantly prominent distribution feature, which can directly point to key correlation information related to equipment wear. The specific component location refers to the actual spatial location in the physical structure of the target equipment corresponding to the wear relationship hotspot. This location clearly points to a specific key component of the equipment and its specific area, rather than a general component category or a vague equipment area. Its positioning is based on the structural design parameters of the equipment, component assembly drawings, and spatial layout during operation. It can accurately correspond to specific identifiable and monitorable component parts and is the direct landing point of the wear distribution on the physical equipment.

[0045] S5. Perform adaptive decision-making on each wear comparison protocol in the global wear map to achieve intelligent wear identification of the target device.

[0046] The embodiments of the present invention enable the wear identification process to accurately match the wear characteristics corresponding to each protocol. This decision-making method can dynamically adapt to the complex changes in wear status, avoid identification deviations caused by fixed decision logic, ensure adaptability to different wear scenarios, achieve comprehensive control over the wear status of equipment, and make intelligent wear identification more accurate.

[0047] The wear comparison protocol refers to a set of structured rules in the global wear map that carries the logic for determining wear status. It is constructed based on wear comparison rules, element relationships, related features, and wear distribution. It clearly defines the correspondence between specific wear-related element combinations and wear status determination results. It includes core contents such as wear feature threshold standards, determination logic process, and status classification rules. It is not a scattered determination basis, but a standardized and executable decision reference formed for different wear scenarios.

[0048] In this embodiment of the invention, the adaptive decision-making for each wear comparison protocol in the global wear map includes: reading the key wear indicators corresponding to the wear comparison protocols in the global wear map, and then analyzing the wear change trends corresponding to the key wear indicators; generating a set of decision instructions corresponding to the wear change trends through a preset decision tree; and identifying the corresponding instructions in the set of decision instructions to achieve adaptive decision-making for each wear comparison protocol.

[0049] The key wear indicators refer to quantitative parameters extracted from the wear comparison protocol of the global wear map that can directly characterize the core features of the wear state. These parameters cover core data indicators related to wear degree and damage evolution, and are not all parameters in the protocol, but rather key quantitative factors that play a decisive role in determining the wear state after screening. Their numerical changes directly reflect the core dynamics of the wear state. The wear change trend refers to the evolution law of the indicator over time or under changing operating conditions, obtained through data analysis based on real-time collected data and historical accumulated data of the key wear indicators. This includes characteristics such as the rise, fall, stable fluctuation, or sudden change of the indicator value, which can objectively present the development direction and rate of change of the wear state. The preset decision tree refers to a pre-set decision tree based on... A tree-shaped decision model, determined through model training and structural optimization, is based on a large amount of wear scenario sample data, wear comparison protocol rules, and decision logic. This model includes a root node, internal decision nodes, and leaf nodes. Each node corresponds to specific wear index judgment conditions. The decision logic corresponding to different wear change trends is clarified through branch paths, forming the core algorithm model for generating decision instruction sets. The corresponding instructions refer to specific execution instructions that perfectly match the currently analyzed wear change trend within the decision instruction set generated by the preset decision tree for a specific wear change trend. These instructions have clear operation directions and judgment logic. They are not all instructions in the instruction set, but rather targeted instructions adapted to the current wear state after filtering, directly supporting the adaptive decision-making of the wear comparison protocol.

[0050] Furthermore, in this embodiment of the invention, the wear comparison protocol includes: a key component identifier, used to uniquely identify the monitoring component in the target device; an operating condition label, used to record the environmental parameters during the operation of the target device; and a decision priority identifier, used to determine the order of decision processing based on the severity of wear.

[0051] The key component identifier refers to standardized identification information set to uniquely distinguish the key components to be monitored in the target equipment. Its form may include character codes, numerical sequences, or combined identifiers. Determined based on equipment structural design drawings, component assembly lists, and monitoring requirements, it can accurately correspond to a specific single or group of key components, avoiding identification confusion between different monitored components during wear identification. It is the core information carrier of the associated monitoring objects in the wear comparison protocol. The operating condition label refers to a standardized label attached to the wear comparison protocol used to record parameters related to the environment and working conditions of the target equipment during operation. It includes key operating condition parameters such as operating temperature, medium pressure, workload, and operating time. Its data comes directly from real-time monitoring results during equipment operation, not preset simulated data, and can objectively reflect the equipment operating background conditions corresponding to the protocol. The decision priority identifier refers to a standardized mark pre-set to clarify the order of decision-making based on the wear severity of the key components corresponding to the wear comparison protocol. Its form may include priority levels, numerical sorting, or code identifiers. The setting is based on the weight of the impact of wear degree on equipment operating safety and stability, without involving subjective judgment, and can provide a clear execution order reference when multiple protocols are processed in parallel.

[0052] Specifically, the embodiments of the present invention can realize intelligent wear identification of the target equipment, thereby enabling real-time monitoring of the wear dynamics and damage evolution of key components of the equipment, accurately capturing the full-cycle change information from the initial wear stage to the development stage, breaking the limitations of traditional monitoring methods. Its core lies in integrating multi-dimensional wear-related data and intelligent decision-making logic to comprehensively present the distribution characteristics and correlation patterns of equipment wear, providing objective and comprehensive status basis for equipment maintenance strategy formulation and operating parameter optimization, and helping to achieve forward-looking management of equipment wear.

[0053] Compared to the problems described in the background art, the embodiments of the present invention, by collecting vibration signals during equipment operation and analyzing the surface damage characteristics of key components, can capture dynamic information related to component damage in real time. This directly correlates the actual operating state of the component with the damage manifestation, improving the targeting and accuracy of damage feature extraction. Furthermore, the embodiments of the present invention can accurately define the degree of wear, avoiding ambiguous judgments of component state. Through the direct correlation between damage features and wear stages, the progress of component wear is clearly presented, eliminating state perception bias and facilitating timely response to wear changes, ensuring the stability and continuity of equipment operation. Finally, the embodiments of the present invention, through the output of a low-wear mode, can accurately adapt to different operating conditions of the target equipment, avoiding mode adaptation bias caused by differences in operating conditions, and through deep learning, the stage... In-depth mining of wear data enables the low-wear mode to fully align with the actual wear characteristics of the equipment, improving the accuracy of the mode. Furthermore, this invention integrates multi-dimensional wear-related information during equipment operation, avoiding cognitive limitations caused by single data dimensions. Constructing a global wear map systematically presents the wear correlation status of each key component of the equipment, enhancing the overall comprehensiveness and objectivity of wear status assessment. Finally, this invention enables the wear identification process to accurately align with the wear characteristics corresponding to each protocol. This decision-making method dynamically adapts to complex changes in wear status, avoiding identification bias caused by fixed decision logic, ensuring adaptability to different wear scenarios, and achieving comprehensive control over the equipment's wear status. This makes intelligent wear identification more accurate. Therefore, this invention can improve the efficiency and accuracy of equipment wear identification.

[0054] like Figure 3 The diagram shown is a functional block diagram of the device wear intelligent identification system based on deep learning according to the present invention.

[0055] The deep learning-based intelligent device wear identification system 300 described in this invention can be installed in electronic devices. Depending on the functions implemented, the deep learning-based intelligent device wear identification system may include a feature analysis module 301, a stage determination module 302, a pattern output module 303, a map construction module 304, and a wear identification module 305. The module described in this invention can also be called a unit, referring to a series of computer program segments that can be executed by the processor of an electronic device and perform a fixed function, stored in the memory of the electronic device.

[0056] In this embodiment of the invention, the functions of each module / unit are as follows: The feature analysis module 301 is used to collect the operating vibration signal of the target device in operation and then analyze the surface damage characteristics of key components in the target device. The stage determination module 302 is used to determine the wear stage of the key components in the target device based on the surface damage characteristics. The mode output module 303 is used to input the stage wear data corresponding to the wear stage into a preset deep learning network, and then output a low-wear mode of the device that is adapted to the target device under different working conditions. The atlas construction module 304 is used to collect the real-time wear parameters of the target equipment under the low-wear mode of the equipment, and construct a global wear atlas corresponding to the target equipment based on the real-time wear parameters and the operating vibration signal. The wear identification module 305 is used to make adaptive decisions on each wear comparison protocol in the global wear map to realize intelligent wear identification of the target device.

[0057] In detail, the modules in the deep learning-based intelligent device wear recognition system 300 described in this embodiment of the invention employ the same methods as described above during use. Figure 1 The method uses the same deep learning-based intelligent identification method for device wear as described above and can produce the same technical effect, so it will not be repeated here.

[0058] In one embodiment, a computer device is provided, which may be a server or a client, and its internal structure diagram may be as follows: Figure 4 As shown, the computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile and / or volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The network interface is used to communicate with external clients via a network connection. When the computer program is executed by the processor, it implements the functions or steps of a deep learning-based intelligent device wear identification method on the server or client side.

[0059] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to perform the following steps: After collecting the vibration signals of the target equipment during operation, the surface damage characteristics of key components in the target equipment are analyzed. Based on the surface damage characteristics, the wear stage of the key components in the target equipment is determined; After inputting the stage wear data corresponding to the wear stage into a preset deep learning network, the system outputs a low-wear mode for the target device under different operating conditions. Real-time wear parameters of the target equipment are collected under the low-wear mode of the equipment. Based on the real-time wear parameters and the operating vibration signal, a global wear map corresponding to the target equipment is constructed. Adaptive decision-making is performed on each wear comparison protocol in the global wear map to achieve intelligent wear identification of the target device.

[0060] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor: After collecting the vibration signals of the target equipment during operation, the surface damage characteristics of key components in the target equipment are analyzed. Based on the surface damage characteristics, the wear stage of the key components in the target equipment is determined; After inputting the stage wear data corresponding to the wear stage into a preset deep learning network, the system outputs a low-wear mode for the target device under different operating conditions. Real-time wear parameters of the target equipment are collected under the low-wear mode of the equipment. Based on the real-time wear parameters and the operating vibration signal, a global wear map corresponding to the target equipment is constructed. Adaptive decision-making is performed on each wear comparison protocol in the global wear map to achieve intelligent wear identification of the target device.

[0061] It should be noted that the functions or steps that can be implemented by the computer-readable storage medium or computer device described above can be referred to the relevant descriptions on the server side and client side in the foregoing method embodiments. To avoid repetition, they will not be described one by one here.

[0062] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0063] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0064] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0065] Finally, it should be noted that in the above embodiments, each embodiment can be combined with each other or independent. Deleting any one of them will not affect the technical implementation of other embodiments. The above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A deep learning-based intelligent identification method for equipment wear, characterized in that, The method includes: After collecting the vibration signals of the target equipment during operation, the surface damage characteristics of key components in the target equipment are analyzed. Based on the surface damage characteristics, the wear stage of the key components in the target equipment is determined; After inputting the stage wear data corresponding to the wear stage into a preset deep learning network, the system outputs a low-wear mode for the target device under different operating conditions. Real-time wear parameters of the target equipment are collected under the low-wear mode of the equipment. Based on the real-time wear parameters and the operating vibration signal, a global wear map corresponding to the target equipment is constructed. Adaptive decision-making is performed on each wear comparison protocol in the global wear map to achieve intelligent wear identification of the target device.

2. The intelligent device wear identification method based on deep learning as described in claim 1, characterized in that, The process of constructing a global wear map of the target equipment based on the real-time wear parameters and the operating vibration signal includes: The key wear elements in the real-time wear parameters and the vibration floating elements in the operating vibration signal are extracted respectively. After aligning the key wear elements with the vibration floating elements, the wear distribution in the target equipment is determined based on the aligned element relationship pairs. Based on the wear comparison criteria in the wear distribution, a global wear map corresponding to the target equipment is constructed.

3. The intelligent device wear identification method based on deep learning as described in claim 2, characterized in that, The step of determining the wear distribution in the target device based on the aligned element relationship pairs includes: Query the wear relationship hotspots in the aligned feature relationship pairs; After mapping the wear relationship hotspots to specific component locations in the target device, the wear distribution corresponding to the specific component locations is determined.

4. The intelligent device wear identification method based on deep learning as described in claim 1, characterized in that, The analysis of the surface damage characteristics of key components in the target equipment includes: After converting the operating vibration signal into a component vibration spectrum, a surface wear image corresponding to the key wear spectrum in the component vibration spectrum is generated; Locate the specific damaged area in the surface wear image; The damage features in the specific damage area are matched with a preset damage type library to obtain surface damage features.

5. The intelligent device wear identification method based on deep learning as described in claim 1, characterized in that, The determination of the wear stage of key components in the target equipment based on the surface damage characteristics includes: Analyze the damage type and damage size corresponding to the surface damage features; The degree of wear of the key components in the target equipment is determined by the damage type and the damage size. After marking the wear status identifier corresponding to the key component, the wear stage of the key component is determined based on the wear degree quantity and the wear status identifier.

6. The intelligent device wear identification method based on deep learning as described in claim 1, characterized in that, The step of inputting the stage wear data corresponding to the wear stage into a preset deep learning network includes: The wear data corresponding to the wear stage is organized into a wear input sequence to convert the wear input sequence into a sequence format tensor. After the sequence format tensor is input into the input layer of a preset deep learning network, the deep learning network is started to perform a backward processing of the sequence format tensor.

7. The intelligent device wear identification method based on deep learning as described in claim 1, characterized in that, The output adapts to the low-wear mode of the target equipment under different operating conditions, including: Obtain the operation adjustment parameters and load configuration parameters corresponding to the target device output by the deep learning network; After matching the operation adjustment parameters with the current operating conditions of the equipment, the optimal load power corresponding to the target equipment is set according to the matched operating conditions and the load configuration parameters. Based on the optimal load power, the low-wear mode of the target equipment under different operating conditions is dynamically adjusted.

8. The intelligent device wear identification method based on deep learning as described in claim 1, characterized in that, The adaptive decision-making process for each wear comparison protocol in the global wear map includes: After reading the key wear indicators corresponding to the wear comparison protocol in the global wear map, analyze the wear change trend corresponding to the key wear indicators; By using a preset decision tree, a set of decision instructions corresponding to the wear change trend is generated; Identify the corresponding instructions in the decision instruction set to achieve adaptive decision-making for each of the wear comparison protocols.

9. The intelligent device wear identification method based on deep learning as described in claim 8, characterized in that, The wear comparison protocol includes: Key component identifiers are used to uniquely identify the monitoring components in the target device; Operating condition tags are used to record the environmental parameters of the target device during operation; Decision priority indicators are used to determine the order in which decisions are made based on the severity of wear and tear.

10. A deep learning-based intelligent equipment wear recognition system, characterized in that, The system includes: The feature analysis module is used to collect the operating vibration signal of the target equipment in operation and then analyze the surface damage characteristics of key components in the target equipment. The stage determination module is used to determine the wear stage of the key components in the target device based on the surface damage characteristics. The mode output module is used to input the stage wear data corresponding to the wear stage into a preset deep learning network and then output a low-wear mode of the device that is adapted to the target device under different working conditions. The atlas construction module is used to collect real-time wear parameters of the target equipment under the low-wear mode of the equipment, and construct a global wear atlas corresponding to the target equipment based on the real-time wear parameters and the operating vibration signal. The wear identification module is used to make adaptive decisions on each wear comparison protocol in the global wear map to realize intelligent wear identification of the target device.

Images