Medical precision instrument full life cycle tracing system based on digital twinning

Abstract

The invention discloses a medical precision instrument full life cycle tracing system based on digital twinning, and relates to the technical field of medical instrument management. The full-life-cycle tracing system comprises an identification recording module which is used for endowing different and unique identity identification codes for components of the medical precision instrument; the data recording module is used for recording physical working condition data, operation and maintenance event data and parameter data of the components of the medical precision instrument in the whole life cycle by taking the identity identification code as a base point; and a digital twin module. According to the method, each component is endowed with a unique identity identification code, the management degree is adjusted to the level of key parts, the deviation between a parameter prediction value output by a digital twin model and an actual measurement value is combined, and path enumeration and intersection analysis are carried out on a complex influence relationship by using a knowledge graph; according to the invention, single or multiple fault source components causing system-level performance abnormity can be automatically and quickly locked.

Description

Technical Field

[0001] This invention relates to the field of medical device management technology, specifically to a full lifecycle traceability system for precision medical devices based on digital twins. Background Technology

[0002] Surgical robots and high-end imaging equipment, among other precision medical devices, play an increasingly important role in clinical diagnosis and treatment. These devices are complex in structure and expensive, and their operational status directly affects the treatment outcome and patient safety. Therefore, establishing a traceability system that covers their entire lifecycle, from activation, use, maintenance, repair to disposal, is of great significance for ensuring medical quality, optimizing device management, and reducing operating costs.

[0003] For example, patent publication number "CN112331323A", entitled "A Traceability System and Method for the Entire Lifecycle of Medical Surgical Instruments", describes a system comprising: an application front-end, including individual surgical instruments equipped with RFID tags, multiple antennas, management equipment, and a statistics and query interface. The antennas are connected to a reader, and the statistics and query interface includes modules for managing sterilized items, project maintenance, location management, and dynamic inventory. This invention utilizes RFID tags that are significantly smaller than ordinary RFID tags, making them more suitable for instrument identification requirements; they eliminate the need for manual reading, as the tags can be directly embedded into the assets; and hundreds of tags can be read in a short time.

[0004] The aforementioned patents only manage individual surgical instruments at the overall level, lacking independent identification and refined management of core functional modules and key components within the instruments. When the instruments experience performance degradation or malfunction, the system cannot accurately trace which specific component (such as the motor, bearing, or sensor) failed. When the output performance of the instruments deviates, the existing systems can only record the result of "fault" without automatically and quickly locating the root cause of the fault. To address this, a full lifecycle traceability system for medical precision instruments based on digital twins has been invented. Summary of the Invention

[0005] The purpose of this invention is to provide a digital twin-based full lifecycle traceability system for medical precision instruments to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a full lifecycle traceability system for medical precision instruments based on digital twins, the full lifecycle traceability system comprising: Identification and Recording Module: Assigns unique identification codes to the components of precision medical devices; Data recording module: Based on the identification code, it records the physical condition data, operation and maintenance event data, and parameter data of the components of medical precision instruments throughout their entire life cycle; Digital twin module: Based on the parameter data of the components, a virtual twin model corresponding to the medical precision instrument is constructed, and the state and behavior of the virtual twin model are simulated based on physical working condition data to obtain simulation data; Knowledge Graph Module: Creates a knowledge graph with components as nodes and physical interactions between components as edges; System-level deviation set detection and tracing module: detects the measured value set of several performance parameters in medical precision instruments, extracts the predicted value set of parameters with the same attributes as the performance parameters from the simulation data, calculates the difference between the measured value set and the predicted value set to obtain the system-level deviation set, and uses the system-level deviation set to perform reverse tracing in the knowledge graph to obtain the fault node; Status assessment module: Establishes a status assessment model for evaluating the health value of faulty nodes. The status assessment model receives operation and maintenance event data and automatically updates the status of the components in the virtual twin model. Self-calibration module: When the system-level deviation set is determined to be model distortion originating from the virtual twin model, a self-calibration process is triggered to adjust the parameter data of the components in the virtual twin model in reverse. Intelligent decision-making module: When the model is not distorted, predictive maintenance instructions are generated based on the comparison between the health value of the faulty node and the preset health threshold. After executing the predictive maintenance instructions, operation and maintenance event data are obtained.

[0007] Furthermore, the difference between the measured values ​​in the measured value set and the predicted values ​​in the parameter prediction value set with the same performance parameter attributes is obtained, a deviation threshold is established, and those with a difference greater than the deviation threshold are classified into the system-level deviation set. The reverse tracing includes: In a knowledge graph, several propagation paths leading to the generation of a system-level deviation set are obtained by enumerating paths based on the system-level deviation set. By analyzing the intersection nodes between several propagation paths, it can be determined whether the intersection node is a faulty node.

[0008] Furthermore, the path enumeration includes: starting from the end node in the knowledge graph that is associated with the parameters in the system-level deviation set, traversing backwards through all influence paths leading to the source to obtain the propagation path; Determining whether a junction node is a faulty node includes: The number of times each intersection node appears in the several fault propagation paths is counted, an occurrence threshold is set, and intersection nodes whose occurrence count exceeds the occurrence threshold are identified as fault nodes.

[0009] Furthermore, based on health scores and faulty nodes, predictive maintenance instructions are generated, including: Establish a health threshold; when the health value falls below the health threshold, generate a predictive maintenance instruction. The predictive maintenance instructions include maintenance instructions and replacement instructions; Obtain the actual number of times the health value of a component exceeds the health threshold, establish a count threshold, execute a replacement command on components whose actual count exceeds the count threshold, and execute a maintenance command on components whose actual count is less than or equal to the count threshold.

[0010] Furthermore, the state assessment model includes: Establish an initial health value, calculate the cumulative wear and tear of the component based on its historical workload and theoretical lifespan, and reduce the health value based on the cumulative wear and tear. When the system-level deviation set is traced back to the component, and the virtual twin model is not determined to be distorted, the reduction in the health value is controlled according to the magnitude of the specific value in the system-level deviation set.

[0011] Furthermore, the method for determining the distortion of the virtual twin model includes: Check the persistence of system-level deviations: determine whether the system-level deviations persist within the detection period; Obtain direct detection data of the components corresponding to the faulty node, and determine whether the direct detection data is within the normal range; Comprehensive judgment: When the system-level deviation persists and the direct detection data is within the normal range, the virtual twin model is judged to be distorted.

[0012] Furthermore, the reverse adjustment of parameter data for components in the virtual twin model includes: Parameter data debugging is performed on the components corresponding to the faulty node until the system-level deviation set of the faulty node is eliminated; Parameter data debugging involves adjusting the set of predicted parameter values ​​based on different parameter data for computer simulation components until the system-level deviation set is eliminated.

[0013] Furthermore, the operation and maintenance event data includes maintenance data and component replacement data; When the digital twin module receives component replacement data, it resets the health values ​​of the nodes associated with the component replacement data. When the digital twin module receives maintenance data, it restores the health value of the node associated with the maintenance data to a predefined increment.

[0014] Compared with the prior art, the beneficial effects of the present invention are: This digital twin-based medical precision instrument full lifecycle traceability system adjusts the management level to the critical component level by assigning a unique identification code to each component. It combines the deviation between the parameter prediction values ​​output by the digital twin model and the actual measurement values, and uses knowledge graphs to perform path enumeration and intersection point analysis on complex influence relationships. It can automatically and quickly identify one or more fault source components that cause system-level performance abnormalities, and generate targeted predictive maintenance instructions for one or more fault source components, making it convenient for personnel to carry out inspection and replacement operations on the fault source components.

[0015] By setting up a condition assessment model, which comprehensively considers the historical workload of components, the cumulative wear and tear of theoretical life, and the real-time anomalies reflected by the system-level deviation set, the system calculates a quantitative health value. The system can issue maintenance or replacement instructions in advance before the performance of components deteriorates significantly or fails completely, guiding medical staff to carry out planned interventions, and realizing the transformation from "passive maintenance" to "proactive preventive maintenance".

[0016] By setting up a self-calibration module and combining persistent deviation judgment with independent direct detection data, the system can intelligently identify whether system-level deviations originate from real faults in physical components or distortions in the virtual model itself. When the model is determined to be distorted, the system can automatically trigger a calibration process to optimize the model parameters in reverse, ensuring that the digital twin maintains a high degree of fidelity consistent with the physical entity over the long term. Attached Figure Description

[0017] Figure 1 This is a schematic diagram illustrating the active maintenance mechanism of the present invention; Figure 2 This is a schematic diagram of the full lifecycle traceability system of the present invention; Figure 3 This is a schematic diagram illustrating the adjustment of the digital twin model of the present invention under whether it is distorted; Figure 4 This is a schematic diagram of the simulation data for this invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] like Figure 1 - Figure 3As shown, this invention provides a technical solution: a digital twin-based full lifecycle traceability system for precision medical devices, the full lifecycle traceability system comprising: Identification and Recording Module: Assigns unique identification codes to the components of precision medical devices; Data recording module: Based on the identification code, it records the physical condition data, operation and maintenance event data, and parameter data of the components of medical precision instruments throughout their entire life cycle; Digital twin module: Based on the parameter data of the components, a virtual twin model corresponding to the medical precision instrument is constructed, and the state and behavior of the virtual twin model are simulated based on physical condition data to obtain simulation data; Knowledge Graph Module: Creates a knowledge graph with components as nodes and physical interactions between components as edges; System-level deviation set detection and tracing module: detects the measured value set of several performance parameters in medical precision instruments, extracts the predicted value set of parameters with the same attributes as the performance parameters from the simulation data, calculates the difference between the measured value set and the predicted value set to obtain the system-level deviation set, and uses the system-level deviation set to perform reverse tracing in the knowledge graph to obtain the fault node; Status assessment module: Establishes a status assessment model for evaluating the health value of faulty nodes. The status assessment model receives operation and maintenance event data and automatically updates the status of the components in the virtual twin model. Self-calibration module: When the system-level deviation set is determined to be model distortion originating from the virtual twin model, the self-calibration process is triggered to adjust the parameter data of the components in the virtual twin model in reverse. Intelligent decision-making module: When the model is not distorted, predictive maintenance instructions are generated based on the comparison between the health value of the faulty node and the preset health threshold. After executing the predictive maintenance instructions, operation and maintenance event data are obtained.

[0020] Obtain the difference between the measured values ​​in the measured value set and the predicted values ​​in the parameter prediction value set with the same performance parameter attributes, establish a deviation threshold, and classify the differences that are greater than the deviation threshold into the system-level deviation set; Reverse tracing includes: In a knowledge graph, several propagation paths leading to the generation of a system-level deviation set are obtained by enumerating paths based on the system-level deviation set. By analyzing the intersection nodes between several propagation paths, it can be determined whether the intersection node is a faulty node.

[0021] Path enumeration includes: starting from the end node in the knowledge graph that is associated with the parameters in the system-level deviation set, traversing backwards through all influence paths leading to the source to obtain the propagation path; Determining whether a junction node is a faulty node includes: The number of times each intersection node appears in several fault propagation paths is counted. An occurrence threshold is set, and intersection nodes whose occurrence count exceeds the occurrence threshold are identified as fault nodes.

[0022] Based on health status values ​​and faulty nodes, predictive maintenance instructions are generated, including: Establish a health threshold; when the health value falls below the health threshold, generate a predictive maintenance instruction. Predictive maintenance instructions include maintenance instructions and replacement instructions; Obtain the actual number of times the health value of a component exceeds the health threshold, establish a count threshold, execute a replacement command on components whose actual count exceeds the count threshold, and execute a maintenance command on components whose actual count is less than or equal to the count threshold.

[0023] The state assessment model includes: Establish an initial health value, calculate the cumulative wear and tear of the component based on its historical workload and theoretical lifespan, and reduce the health value based on the cumulative wear and tear. When the system-level deviation set is traced back to the component, and the virtual twin model is not determined to be distorted, the reduction in the health value is controlled according to the magnitude of the specific value in the system-level deviation set.

[0024] Methods for judging the distortion of virtual twin models include: Check the persistence of system-level deviations: determine whether the system-level deviations persist within the detection period; Obtain direct detection data of the components corresponding to the faulty node, and determine whether the direct detection data is within the normal range; Comprehensive judgment: When system-level deviations persist and directly detected data are within the normal range, the virtual twin model is judged to be distorted.

[0025] The reverse adjustment of parameter data for components in the virtual twin model includes: Parameter data debugging is performed on the components corresponding to the faulty node until the system-level deviation set of the faulty node is eliminated; Parameter data debugging involves adjusting the set of predicted parameter values ​​based on different parameter data for computer simulation components until the system-level deviation set is eliminated.

[0026] Operations and maintenance event data includes maintenance data and component replacement data; When the digital twin module receives component replacement data, it resets the health values ​​of the nodes associated with the component replacement data. When the digital twin module receives maintenance data, it restores the health value of the node associated with the maintenance data to a predefined increment.

[0027] Each piece (or set) of precision medical equipment is assigned a unique identification code through instrument identification, and its components are also assigned unique identification codes. These codes are carried out using QR codes or RFID electronic tags as their physical carriers. This ID is permanently bound to the instrument file in the system database. At the data acquisition and sensing layer, RFID readers are deployed in key areas such as the instrument warehouse, operating room, and sterilization supply center to automatically collect information on the entry and exit status and location of instruments. Staff use handheld terminals to scan the identification codes to quickly complete operations such as instrument requisition, return, inventory, and maintenance record entry. Replacement of components is also recorded in the system. The core business logic layer is structured with sophisticated management sub-modules: the medical device file management sub-module manages basic information and also critically links to its operating procedures, precautions, and technical manuals; the full lifecycle process management sub-module covers all aspects of use, maintenance, repair, inventory and location, and disposal. By connecting with hospital information systems and other personnel systems, it records the operator and the associated surgery / patient for each use, and automatically generates maintenance and replacement plans based on preset rules (such as usage frequency and duration and physical condition data), realizing the transformation from "passive maintenance" to "proactive preventive maintenance," laying a solid foundation for improving medical safety and operational efficiency.

[0028] The identification code can use a QR code or RFID electronic tag as the physical carrier. The identification code of each component is permanently bound to the device file in the system database. When a component is replaced, the identification codes of the components before and after the replacement are identified, and the corresponding component in the virtual twin model is simulated for replacement. Identification code readers are deployed at access control points or key nodes in key areas such as device warehouses, operating rooms, and sterilization supply centers to automatically read the entry and exit status and location information of medical precision instruments. This enables refined and intelligent management of medical precision instruments from purchase to disposal, ensuring that instruments are in optimal working condition, guaranteeing medical safety, improving management efficiency, and reducing operating costs. The system can also track and manage replaced components of medical precision instruments. Furthermore, the system can perform maintenance and upkeep of medical precision instruments, recording the maintenance process. The system records maintenance details, replacement parts, maintenance personnel, and dates, along with fault reports, repair processes, replaced parts, repair costs, and downtime, creating a maintenance history archive. This facilitates subsequent traceability of replaced parts. Since these are precision medical devices, the system shifts from "passive maintenance" to "proactive preventative maintenance," effectively ensuring the precision of their operation. By assigning a unique identification code to each component, management is adjusted to the level of critical parts. Combining the deviation between predicted and actual measured values ​​from a digital twin model, and utilizing knowledge graphs to enumerate paths and analyze intersections of complex relationships, the system can automatically and quickly identify single or multiple fault-causing components that trigger system-level performance anomalies. Subsequently, targeted predictive maintenance instructions are generated for these fault-causing components, facilitating inspection and replacement by personnel.

[0029] This application targets medical precision devices constructed from multiple components. Each component refers to a physical entity that constitutes the medical precision device and can be independently identified and managed. Components include the complete medical precision device, functional modules, and key parts. The parameter data of each component maintains its external dimensions and material information. A medical precision device can be constructed using these components. Subsequently, a virtual twin model corresponding to the medical precision device is established in a computer using the parameter data of the components. The system models the device based on the detailed parameter data of the components. This parameter data includes, but is not limited to, the geometric dimensions (such as shape and structure), material properties (such as density and elastic modulus), physical characteristics (such as thermal conductivity and electrical conductivity), and performance specifications (such as rated power and accuracy indicators) of the components. In this way, the virtual twin model can reflect the characteristics of the physical entity with high fidelity. The identification code provides a reference point for subsequent physical condition data and operation and maintenance event data.

[0030] The recording module comprehensively collects various types of data generated by medical precision instruments throughout their entire lifecycle, mainly including physical operating condition data (such as power, load, and operating time), management data (such as operators and associated surgeries), and operation and maintenance event data (such as maintenance, repair, and component replacement records). Its role is to build a complete and continuous data chain, providing realistic and rich input data for digital twin simulation, system status assessment, and decision-making, which is a prerequisite for the system to achieve dynamic perception and intelligent analysis.

[0031] Identification code readers are deployed at key locations such as warehouses and operating rooms to automatically collect the location and status of medical devices. Personnel use handheld terminals to quickly scan and record operations such as requisitioning, returning, and maintaining the devices. Since medical precision instruments are already assembled, their internal components are inconvenient to scan. Therefore, for assembled medical precision instruments, only the identification code on the outer shell is scanned to track the entire instrument. For assembled complete devices, tracking is mainly achieved by scanning the identification code on the outer shell. At the same time, when parts are replaced, the identification codes of the old and new parts are identified. The identification code data before and after the replacement are also recorded. These are two types of data with the same parameters, but they will differ due to the different identification codes. The component replacement data includes the identification code of the old part being replaced and the identification code of the newly installed part, so as to achieve accurate traceability of the component replacement history.

[0032] Management data includes information on operators, surgeries, and projects; physical condition data such as power, load, running time, and cycle count; and state and behavior simulations of the virtual twin model based on this physical condition data. This involves linking usage time and duration to the medical precision instruments input into the virtual twin model for state and behavior simulation. The simulation obtains simulation data of the virtual twin model under the given physical condition data. The state and behavior simulation of the virtual twin model includes using finite element analysis to analyze the equipment components. Mechanical modeling simulates the structural response under conditions such as mechanical stress and vibration; fluid dynamics analysis models the flow characteristics of fluids inside the equipment, simulating temperature and pressure field distributions. Simulation data is obtained through finite element analysis and fluid dynamics analysis. Through the above simulation process, the corresponding parameter prediction set is finally calculated at the output of the virtual model. The digital twin module is the core computing engine of this system. Based on the detailed parameter data (geometric dimensions, material properties, etc.) of the components, it constructs a virtual twin model that highly matches the physical instrument. Its significance lies in creating a dynamic mirror image of the physical entity in digital space. By receiving physical condition data from the recording module, this module uses finite element analysis, fluid dynamics analysis, and other techniques to perform multi-physics state and behavior simulation on the virtual model, thereby obtaining simulation data reflecting the internal state of the instrument (such as parameter prediction values, internal temperature, and forces). Figure 4 As shown, simulation data is obtained through a virtual twin model. Since various sensors exist in actual precision medical devices, and their presence can impact these devices, the virtual twin model also simulates these sensors. When maintenance event data is received, it automatically updates the state of the corresponding components in the model, achieving synchronous evolution between the digital model and the physical entity. This also provides a basis for shifting from "passive maintenance" to "proactive preventative maintenance." Based on the actual situation, the measured values ​​of key performance parameters and the predicted values ​​of parameters from the digital twin module are collected to obtain a system-level deviation set. The difference between the measured values ​​and the predicted values ​​of the corresponding key performance parameters is then used to establish a deviation... If the difference threshold is greater than the deviation threshold, it will be classified into the system-level deviation set; otherwise, it will not. Based on the system-level deviation set, the abnormal state of components in medical precision instruments can be clearly identified in actual situations, thus facilitating early maintenance and replacement of components in abnormal states. Through the setting of the condition assessment model, the condition assessment model comprehensively considers the historical workload of the component, the cumulative wear and tear of the theoretical lifespan, and the real-time anomalies reflected by the system-level deviation set to calculate a quantitative health value. The system can issue maintenance or replacement instructions in advance before the component performance deteriorates significantly or fails completely, guiding medical staff to carry out planned interventions and realizing the transformation from "passive maintenance" to "proactive preventive maintenance".

[0033] The measured values ​​of performance parameters include direct sensing measurements: for precision instruments with built-in sensors (such as surgical robots and high-end anesthesia machines), key parameters such as torque, speed, pressure, temperature, flow rate, and image grayscale values ​​at the output end are directly acquired by the sensors. The virtual twin model receives the same input commands and operating condition data (such as set power, operating mode, and ambient temperature) as the physical instrument. The same applies to directly detected data. Note that in reality, the measured value set cannot comprehensively detect and cover all operating parameters of precision medical instruments, but the simulation data can comprehensively cover all operating parameters of the virtual twin model. The virtual twin model performs operational simulation based on physical operating condition data during calculation, thereby obtaining the state and behavior simulation of the virtual twin model based on physical operating condition data. At the same time, it extracts parameter predictions with the same attributes as the measured values ​​of performance parameters from the simulation data. The virtual twin model, based on its current part model parameters, performs real-time simulation calculations using built-in physical laws (such as the mechanical equations of FEA and the Navier-Stokes equations of CFD). Finally, it simulates and calculates the corresponding parameter prediction set at the output of the virtual model. The difference between the parameter prediction set and the measured value set yields a system-level deviation set. This system-level deviation set reflects the comprehensive deviation between the physical device and the virtual model across multiple performance dimensions. During source tracing, the system starts from nodes in the knowledge graph closest to the output and traces all possible influence paths in reverse, forming a network of possible fault paths. Within this network, by analyzing the interactions and mapping relationships between multiple propagation paths, the contribution of path intersection nodes as fault origins is calculated, thereby locating possible fault nodes. Note that there may not be a single fault node; multiple fault nodes may exist.

[0034] The operation and maintenance event data includes maintenance data and component replacement data. Maintenance data includes the maintenance of components, and component replacement data includes the replacement of identification codes before and after replacement. When a maintenance event involves replacing a new part, the health value of the corresponding node is reset to its initial health state. When a maintenance event involves maintenance or repair, the real-time health value of the corresponding node is restored by a predefined increment based on the maintenance or repair content, decreasing the health value from a large to a small value. For example, if the health value of a component is set to 100, and the component is detected as a faulty node (i.e., the source of the fault), the health value of the component will drop to 80. When a component is replaced, the health value of the corresponding component in the virtual twin model changes... If the health value is 100, and maintenance is required, the corresponding component's health value in the virtual twin model will change from 40 to 80. When the model is not distorted, based on the comparison between the health value of the fault node and the preset health threshold, if the health value of the fault node (i.e., the component) is less than the preset health threshold, a predictive maintenance instruction will be triggered. The status assessment module establishes a status assessment model and assesses a health value for each component (e.g., the initial value is 100). This value is mainly deducted based on the cumulative wear calculated from the component's historical workload and theoretical lifespan. When a component is traced back to a fault node and the system determines that the virtual twin model is not distorted, its health value will be further deducted based on the specific value of the system-level deviation set.

[0035] The self-calibration module is crucial for maintaining the long-term accuracy of the system. The method for determining virtual twin model distortion includes: first, checking whether system-level deviations persist within the detection period (a preset time period); second, acquiring direct detection data (e.g., through offline manual inspection or independent sensor measurement) of the physical components corresponding to the faulty node; and third, determining whether this data is within the normal range, i.e., whether the component is in normal working condition. If the deviation persists and the direct detection data is normal, the virtual twin model is determined to be distorted. The self-calibration module will then adjust the parameter data of the components corresponding to the faulty node, simulating their output under different parameters using a computer until the system-level deviation set is eliminated, completing the model's self-calibration. Finally, the virtual twin model is adjusted in reverse. Because the virtual twin model is calculated too ideally by computers, the output parameters of the components in the model may differ significantly from the actual parameters. Therefore, a self-calibration module adjusts the parameter data of the components in the virtual twin model. Although the adjusted parameter data differs from the actual parameter data, it makes the output parameters of the two closer. Through the settings of the self-calibration module, by combining the judgment of persistent deviation with independent direct detection data, the system can intelligently identify whether the system-level deviation is due to a real fault of the physical component or the distortion of the virtual model itself. When it is determined that the model is distorted, the system can automatically trigger the calibration process to optimize the model parameters in reverse, ensuring that the digital twin and the physical entity maintain a high degree of fidelity in the long term.

[0036] After personnel execute predictive maintenance instructions, maintenance event data is generated. This data is used to update the status of components in the virtual twin model. Predictive maintenance instructions, including repair quality and replacement instructions, correspond to maintenance data and component replacement data. The intelligent decision-making module generates predictive maintenance instructions based on health values ​​and fault node information. The system sets a health threshold; when the health value of a component falls below this threshold, a maintenance instruction is triggered. Instruction types include maintenance and replacement instructions. The system also counts the actual number of times a component's health value falls below the threshold. If the number exceeds the set threshold, a replacement instruction is generated; otherwise, a maintenance instruction is generated. When personnel execute these instructions, new maintenance event data (such as component replacement and maintenance records) is generated. This data is fed back to the digital twin module and the status assessment module to update the status of the corresponding component in the virtual twin model (e.g., resetting health after replacement, restoring a health increment after maintenance), thus forming a complete closed loop from perception, diagnosis, decision-making to execution.

[0037] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended embodiments and their equivalents.

Claims

1. A digital twin-based full lifecycle traceability system for precision medical devices, characterized in that, The full lifecycle traceability system includes: Identification and Recording Module: Assigns unique identification codes to the components of precision medical devices; Data recording module: Based on the identification code, it records the physical condition data, operation and maintenance event data, and parameter data of the components of medical precision instruments throughout their entire life cycle; Digital twin module: Based on the parameter data of the components, a virtual twin model corresponding to the medical precision instrument is constructed, and the state and behavior of the virtual twin model are simulated based on physical working condition data to obtain simulation data; Knowledge Graph Module: Creates a knowledge graph with components as nodes and physical interactions between components as edges; System-level deviation set detection and tracing module: detects the measured value set of several performance parameters in medical precision instruments, extracts the predicted value set of parameters with the same attributes as the performance parameters from the simulation data, calculates the difference between the measured value set and the predicted value set to obtain the system-level deviation set, and uses the system-level deviation set to perform reverse tracing in the knowledge graph to obtain the fault node; Status assessment module: Establishes a status assessment model for evaluating the health value of faulty nodes. The status assessment model receives operation and maintenance event data and automatically updates the status of the components in the virtual twin model. Self-calibration module: When the system-level deviation set is determined to be model distortion originating from the virtual twin model, a self-calibration process is triggered to adjust the parameter data of the components in the virtual twin model in reverse. Intelligent decision-making module: When the model is not distorted, predictive maintenance instructions are generated based on the comparison between the health value of the faulty node and the preset health threshold. After executing the predictive maintenance instructions, operation and maintenance event data are obtained.

2. The medical precision instrument full life cycle traceability system based on digital twin as described in claim 1, characterized in that: Obtain the difference between the measured values ​​in the measured value set and the predicted values ​​in the parameter prediction value set with the same performance parameter attributes, establish a deviation threshold, and classify the differences that are greater than the deviation threshold into the system-level deviation set; The reverse tracing includes: In a knowledge graph, several propagation paths leading to the generation of a system-level deviation set are obtained by enumerating paths based on the system-level deviation set. By analyzing the intersection nodes between several propagation paths, it can be determined whether the intersection node is a faulty node.

3. The medical precision instrument full life cycle traceability system based on digital twin as described in claim 2, characterized in that: The path enumeration includes: starting from the end node in the knowledge graph that is associated with the parameters in the system-level deviation set, traversing backwards through all influence paths leading to the source to obtain the propagation path; Determining whether a junction node is a faulty node includes: The number of times each intersection node appears in the several fault propagation paths is counted, an occurrence threshold is set, and intersection nodes whose occurrence count exceeds the occurrence threshold are identified as fault nodes.

4. The medical precision instrument full life cycle traceability system based on digital twin as described in claim 1, characterized in that: Based on health status values ​​and faulty nodes, predictive maintenance instructions are generated, including: Establish a health threshold; when the health value falls below the health threshold, generate a predictive maintenance instruction. The predictive maintenance instructions include maintenance instructions and replacement instructions; Obtain the actual number of times the health value of a component exceeds the health threshold, establish a count threshold, execute a replacement command on components whose actual count exceeds the count threshold, and execute a maintenance command on components whose actual count is less than or equal to the count threshold.

5. A medical precision instrument full lifecycle traceability system based on digital twins according to claim 4, characterized in that: The state assessment model includes: Establish an initial health value, calculate the cumulative wear and tear of the component based on its historical workload and theoretical lifespan, and reduce the health value based on the cumulative wear and tear. When the system-level deviation set is traced back to the component, and the virtual twin model is not determined to be distorted, the reduction in the health value is controlled according to the magnitude of the specific value in the system-level deviation set.

6. A medical precision instrument full lifecycle traceability system based on digital twins according to claim 5, characterized in that: The method for determining the distortion of the virtual twin model includes: Check the persistence of system-level deviations: determine whether the system-level deviations persist within the detection period; Obtain direct detection data of the components corresponding to the faulty node, and determine whether the direct detection data is within the normal range; Comprehensive judgment: When the system-level deviation persists and the direct detection data is within the normal range, the virtual twin model is judged to be distorted.

7. A medical precision instrument full lifecycle traceability system based on digital twins according to claim 1, characterized in that: The reverse adjustment of parameter data for components in the virtual twin model includes: Parameter data debugging is performed on the components corresponding to the faulty node until the system-level deviation set of the faulty node is eliminated; Parameter data debugging involves adjusting the set of predicted parameter values ​​based on different parameter data for computer simulation components until the system-level deviation set is eliminated.

8. A medical precision instrument full lifecycle traceability system based on digital twins according to claim 1, characterized in that: The operation and maintenance event data includes maintenance data and component replacement data; When the digital twin module receives component replacement data, it resets the health values ​​of the nodes associated with the component replacement data. When the digital twin module receives maintenance data, it restores the health value of the node associated with the maintenance data to a predefined increment.

Images