Medical equipment state monitoring and early warning method and system based on full life cycle

Abstract

The invention discloses a medical equipment state monitoring and early warning method and system based on a full life cycle, and relates to the technical field of medical equipment state monitoring and early warning. Comprising a full-life-cycle data acquisition module, a multi-dimensional data preprocessing module, a full-life-cycle state evaluation module, a hierarchical early warning decision module, a full-life-cycle management module and an edge-cloud collaborative storage module. The full-life-cycle data acquisition module is used for acquiring multi-dimensional data of the medical equipment in the whole stage from purchase acceptance, operation and use, operation and maintenance to scrap evaluation, and outputting original data in the whole stage. Differentiation quantitative indexes are set for different life cycle stages, an evaluation model is constructed by adopting an analytic hierarchy process and dynamic weight adjustment, and an early warning threshold value and a four-level hierarchical response mechanism are dynamically adjusted in combination with a gradient lifting tree algorithm. The problems that a traditional evaluation mode is poor in adaptability, early warning misinformation lags behind, response pertinence is insufficient, and clinical safety guarantee is weak are solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of medical equipment state monitoring and early warning, and in particular to a medical equipment state monitoring and early warning method and system based on a whole life cycle. BACKGROUND

[0002] In the modern medical service system, medical equipment, as the core infrastructure for clinical diagnosis, treatment, rehabilitation and scientific research, directly determines the accuracy of diagnosis and treatment and the safety of patients' lives, and is the key support for ensuring the high-quality development of medical services. Its core work is to realize physiological parameter detection, lesion positioning, treatment intervention and other functions through precise mechanical structure, electronic control system and sensing technology, covering all categories from routine examination equipment to large-scale precision treatment equipment, and widely used in various medical scenes such as hospitals, clinics and research institutions, becoming an indispensable part of the development of the medical industry.

[0003] At present, although a preliminary monitoring and maintenance system has been formed in the field of medical equipment management, there are still significant technical shortcomings and application bottlenecks: some management modes only focus on data collection in a single running stage of the equipment, lack of coverage of the whole life cycle of procurement inspection, operation and maintenance, scrap assessment, the problem of data fragmentation is prominent, and a complete management data chain cannot be formed; the existing data processing relies on simple screening means, it is difficult to effectively eliminate explicit abnormal values and implicit interference data, and the format difference problem of multi-source heterogeneous data has not been solved, resulting in uneven data quality and affecting the accuracy of state evaluation; the state evaluation mostly uses a fixed weight system, without adjusting dynamic factors such as equipment service life and clinical dependence, and cannot adapt to the characteristic differences of different life cycle stages, and the evaluation results are not targeted enough; the early warning mechanism is mostly based on fixed threshold triggering, without dynamic optimization of historical fault data and operation rules of the same type of equipment, prone to early warning false alarm or lag, and lacks a hierarchical classification of emergency handling schemes, with low operation response efficiency; at the same time, the whole-process management closed loop is missing, the equipment maintenance plan is out of touch with the actual state, and the key data tracing capability is insufficient, combined with the fact that data storage does not take into account the real-time response and security backup needs, there is a risk of data leakage or delayed calling, which is difficult to meet the needs of modern medical equipment fine and intelligent management. SUMMARY

[0004] The present application aims to provide a medical equipment state monitoring and early warning method and system based on a whole life cycle, which solves the technical problems raised in the background art.

[0005] To achieve the above object, the present application provides the following technical solutions: a medical equipment state monitoring and early warning system based on a full life cycle, comprising a full life cycle data acquisition module, a multi-dimensional data preprocessing module, a full life cycle state evaluation module, a hierarchical early warning decision module, a full life cycle management module, and an edge-cloud collaborative storage module; The full life cycle data acquisition module is used to acquire multi-dimensional data of the medical equipment in all stages from procurement acceptance, operation and use, operation and maintenance, to scrap evaluation, and output raw data in all stages; The multi-dimensional data preprocessing module performs outlier rejection, data standardization, and multi-source data fusion processing on the raw data in all stages, eliminates format differences and noise interference of data from different sources, and outputs clean fusion data; The full life cycle state evaluation module, based on the clean fusion data, combines the characteristics of different life cycle stages of the medical equipment, adopts "analytic hierarchy process + dynamic weight adjustment" to construct an evaluation model, calculates the equipment health degree in real time through quantitative indicators, and outputs the state evaluation result; wherein, the dynamic weight adjustment takes the proportion of the actual service life of the equipment to the designed life and the clinical application dependence coefficient as the core basis, and the sum of the weights of each stage after adjustment is 1; The hierarchical early warning decision module sets a dynamic early warning threshold based on the gradient boosting tree algorithm trained historical data according to the state evaluation result, determines the early warning level and generates the corresponding emergency treatment suggestion, and outputs the hierarchical early warning information and decision scheme; The full life cycle management module formulates the equipment maintenance plan, updates the life cycle stage division based on the state evaluation result and the early warning information, records the equipment full-process operation and maintenance information, and forms a closed-loop management; The edge-cloud collaborative storage module adopts an edge node to store real-time operation data to ensure millisecond-level response, and a cloud to store full life cycle historical data and perform encrypted backup, so as to realize data security and efficient calling.

[0006] Preferably, the full life cycle data acquisition module comprises a procurement acceptance data acquisition unit, a real-time operation data acquisition unit, an operation and maintenance data acquisition unit, and a scrap evaluation data acquisition unit; the multi-dimensional data specifically comprises equipment factory technical parameters, operation condition data, use environment data, and operation and maintenance operation records; The procurement acceptance data acquisition unit is used to acquire factory detection parameters, quality inspection reports, on-site installation and debugging records, and acceptance standard compliance data of the medical equipment, and output procurement acceptance raw data; The operation real-time data acquisition unit is connected with the device control system through a sensor, collects voltage, current, core component temperature, cumulative running time and fault feedback code data in the device operation process, and outputs operation real-time raw data; the operation and maintenance data acquisition unit is used for collecting device fault maintenance records, part replacement model and quantity information, periodic calibration reports and maintenance execution cycle data, and outputting operation and maintenance raw data; The scrap evaluation data acquisition unit is used for collecting device key component aging detection data, actual operation performance attenuation curve and residual value evaluation report data, and outputting scrap evaluation raw data; the procurement acceptance raw data, operation real-time raw data, operation and maintenance raw data and scrap evaluation raw data are summarized to form full-stage raw data and output.

[0007] Preferably, the multi-dimensional data preprocessing module comprises an abnormal value processing unit and a multi-source data fusion unit. The abnormal value processing unit first screens out explicit abnormal values deviating from the mean value by more than 3 times the standard deviation in the full-stage raw data through the 3σ principle, then performs density analysis on data distribution by using an isolated forest algorithm to capture implicit abnormal data hidden in normal data distribution, and finally marks and removes the identified explicit and implicit abnormal data, and outputs the abnormal data; The multi-source data fusion unit first formats the structured data in the abnormal data, extracts features from the unstructured data, then strengthens the key information weight through an attention mechanism, realizes feature alignment and information complementarity of different types of data, forms clean fusion data in a unified format, and outputs the clean fusion data.

[0008] Preferably, the full-life-cycle state evaluation module comprises a stage characteristic evaluation unit and a health degree calculation unit. The stage characteristic evaluation unit verifies the conformity of the collected data and the factory standard parameters for the device procurement acceptance stage, and the quantitative index is "parameter conformity rate"; For the operation stage, the performance stability index in the continuous operation process of the device is analyzed, and the quantitative index includes "operation parameter fluctuation rate" and "fault feedback code occurrence frequency"; For the operation and maintenance stage, the effectiveness of the performance recovery of the device after maintenance is evaluated, and the quantitative index is "performance recovery rate"; for the scrap stage, the reasonableness of the device aging degree and residual value evaluation is judged, and the quantitative index includes "key component aging rate" and "residual value rate deviation", and the stage evaluation sub-results corresponding to each stage are outputted; The health degree calculation unit adopts the analytic hierarchy process, combines medical equipment operation and maintenance management practical experience, determines that the basic weights of the evaluation sub-results in the procurement and acceptance, operation, operation and maintenance and scrap stages are 15%, 40%, 30% and 15% respectively; and based on the ratio of the actual service life of the equipment to the design life, denoted as K, K [0, 1], a dynamic adjustment factor is constructed according to the clinical application dependence coefficient, and the adjustment logic is: first, the total up-regulation amount = 40% * 0.2K + 30% * 0.3C is calculated, then the down-regulation amount is allocated according to the basic weight ratio of the procurement and acceptance stage and the scrap stage, 1:1, and finally the weight of each stage = basic weight ± corresponding adjustment amount, and the total weight after adjustment is 1; the comprehensive health degree of the equipment is calculated through the weighted sum formula: comprehensive health degree = Σ (each stage evaluation sub-result * adjusted weight), the state evaluation result is formed and output.

[0009] Preferably, the hierarchical early warning decision module comprises a warning level determination unit and an intelligent pushing unit. The early warning level determination unit divides the equipment comprehensive health degree into four levels: normal, mild early warning, moderate early warning and severe early warning; based on the equipment operation history health degree data in the last three months and the fault statistical results of the same type of equipment in the last five years, a gradient boosting tree algorithm is trained to generate an adaptive model to dynamically adjust the threshold interval of each early warning level: when the fault occurrence rate of the same type of equipment in the last six months increases by ≥20% compared with the historical average, the lower limit of each level threshold is lowered by a preset value; when the health degree fluctuation rate of the equipment in the last three months is ≤5%, the lower limit of each level threshold is increased by a preset value; the early warning level is determined according to the interval where the current health degree is located and output. The intelligent pushing unit generates targeted processing suggestions according to the early warning level: the mild early warning pushes a routine daily maintenance prompt to the mobile terminal of the front-line operation and maintenance personnel; The moderate early warning pushes a detailed maintenance scheme and a spare parts list to the department head, and synchronously triggers the spare parts application process; The severe early warning triggers the device emergency shutdown suggestion, when the device failure may affect the safety of clinical diagnosis and treatment, the device control system generates a shutdown instruction, which is synchronously pushed to the hospital equipment management department, the relevant clinical departments and the operation and maintenance emergency team, with the current fault data of the device, the replacement device scheduling suggestion, forming the hierarchical early warning information and decision scheme and output.

[0010] Preferably, the whole life cycle management module comprises a maintenance plan generation unit and a whole process traceability unit. The maintenance plan generation unit determines the preventive maintenance period based on the equipment comprehensive health degree, the early warning level and the actual use frequency, automatically generates a plan list containing specific maintenance items, required spare part models and specifications and recommended maintenance time window, and supports manual adjustment and optimization. The full-process traceability unit records the key data change records, the state evaluation results of each stage, and the execution situation of the early warning processing in time sequence from the purchase to the scrap of the equipment, forms an unalterable electronic file of the life cycle of the equipment, and supports the reverse tracing of the key information of any link.

[0011] Preferably, the edge-cloud collaborative storage module comprises an edge node storage unit, a cloud storage unit and a data encryption unit. The edge node storage unit is deployed locally in each clinical department of the hospital, and is specially used for storing the real-time data of the equipment running for nearly 72 hours, supporting the fast data retrieval and equipment state checking of the local terminal of the department, and guaranteeing the millisecond-level data response speed. The cloud storage unit adopts a distributed storage architecture, is used for storing the historical data of the equipment in the whole life cycle, realizes the elastic expansion and off-site disaster recovery backup of the data, supports the authorized data sharing and statistical analysis across departments and across hospital areas, and realizes the precise matching of the permissions and the data, and guarantees the safety of the medical data. The data encryption unit adopts the national secret SM4 algorithm to encrypt the transmission data between the edge node and the cloud, adopts a role permission hierarchical management mechanism for the stored data, grades the user role types and the data sensitivity, and enables different roles to only access the stored data of the corresponding permission level, so as to realize the precise matching of the permissions and the data, and guarantee the safety of the medical data.

[0012] The medical equipment state monitoring and early warning method based on the whole life cycle comprises the following steps: Step S1, collecting the multi-dimensional data in the whole stage of the medical equipment procurement, acceptance, running, use, maintenance and repair, and scrap evaluation through the data acquisition module, and outputting the whole stage original data formed by the collection; Step S2, performing the outlier elimination, data standardization and multi-source data fusion processing on the whole stage original data through the multi-dimensional data preprocessing module, and outputting the clean fusion data; Step S3, combining the characteristics of different life cycle stages of the equipment, constructing an evaluation model by using the analytic hierarchy process + dynamic weight adjustment, calculating the comprehensive health degree of the equipment based on the clean fusion data, and outputting the state evaluation result; Step S4, setting a dynamic early warning threshold based on the gradient boosting tree algorithm according to the state evaluation result, determining the early warning level and generating an emergency processing suggestion, and outputting the graded early warning information and the decision scheme; Step S5, formulating the equipment maintenance plan based on the state evaluation result and the early warning information, recording the full-process operation and maintenance information of the equipment, and realizing the closed-loop management in the whole life cycle; Step S6, storing the real-time running data of the edge node storage equipment, storing the whole life cycle historical data of the cloud, and performing the encrypted backup, so as to guarantee the data response efficiency and the storage safety.

[0013] Preferably, the step S2 specifically comprises the following sub-steps: Step S21, using 3σ principle, |x-μ|>3σ, wherein x is a data value, μ is a mean value, and σ is a standard deviation, identifying explicit outliers in the full-stage original data, marking and temporarily storing; then constructing an anomaly detection model through the isolation forest algorithm, setting a sample anomaly score threshold, performing density clustering analysis on the data, and capturing implicit abnormal data hidden in the normal data distribution; after merging the explicit outliers and the implicit abnormal data, the explicit outliers and the implicit abnormal data are uniformly removed, and the de-anomaly data is output. Step S22, normalizing the numerical data in the de-anomaly data by using a Z-score standardization method, and mapping to a preset numerical interval; converting the text data into a fixed-dimensional feature vector by using a word embedding algorithm; calculating the weight coefficients of each feature by using an attention mechanism, strengthening the weight proportion of fault keywords and core parameter key features, realizing feature alignment and information complementation of numerical and text data, and outputting clean fusion data in a unified format.

[0014] Preferably, the step S4 specifically comprises the following sub-steps: Step S41, collecting the historical health degree data of the equipment in the past three months and the fault statistical data of the same type of equipment, training a model through a gradient boosting tree algorithm, determining a dynamic early warning threshold suitable for the running characteristics of the equipment, and triggering a corresponding early warning level when the current health degree of the equipment is lower than the corresponding threshold; Step S42, generating a targeted processing suggestion according to the early warning level: a mild early warning is pushed to the mobile terminal of the operation and maintenance personnel through a system message; a moderate early warning is synchronously sent to the department head through a system message and an email; a severe early warning triggers the generation of a work order of a hospital equipment management system, and when the equipment failure may affect the safety of clinical diagnosis and treatment, a device shutdown protection mechanism is linked, and the hospital equipment management department and the related clinical departments are synchronously pushed, and the graded early warning information and the decision scheme are output.

[0015] Compared with the related art, the medical equipment state monitoring and early warning method and system based on the whole life cycle provided by the application have the following beneficial effects: 1. The medical equipment state monitoring and early warning method and system based on the whole life cycle provided by the application realize full coverage integration of multi-dimensional data of each stage of the equipment through a whole life cycle data acquisition module, and solve the problem of traditional management.

[0016] 2. The application provides a full life cycle-based medical device state monitoring and early warning method and system, which sets differentiated quantitative indicators for different life cycle stages, uses "analytic hierarchy process + dynamic weight adjustment" to construct an evaluation model, dynamically adjusts the early warning threshold and four-level grading response mechanism through gradient boosting tree algorithm, and solves the problems of poor adaptability of traditional evaluation mode, early warning false alarm lag, insufficient response pertinence and weak clinical safety guarantee.

[0017] 3. The application provides a full life cycle-based medical device state monitoring and early warning method and system, which generates individualized maintenance plans through a full life cycle management module and forms a closed-loop management, builds an unalterable electronic file with a full-process traceability unit, combines edge-cloud collaborative storage and data encryption, and solves the problems of disconnection between maintenance plans and reality, lack of traceability, low operation and maintenance efficiency, and difficulty in balancing data response and security guarantee in traditional management. BRIEF DESCRIPTION OF DRAWINGS

[0018] Figure 1 is a flowchart of the application; Figure 2 is an extended flowchart of the full life cycle data acquisition module of the application; Figure 3 is an extended flowchart of the multi-dimensional data preprocessing module of the application; Figure 4 is an extended flowchart of the full life cycle state evaluation module of the application; Figure 5 is an extended flowchart of the grading early warning decision module of the application; Figure 6 is an extended flowchart of the full life cycle management module of the application; Figure 7 is an extended flowchart of the edge-cloud collaborative storage module of the application. DETAILED DESCRIPTION

[0019] The technical solutions in the embodiments of the application will be described clearly and completely below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, not all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of the application.

[0020] Embodiment one: Please refer to Figures 1-7The application provides a technical scheme: a medical equipment state monitoring and early warning system based on a whole life cycle, comprising a whole life cycle data acquisition module, a multi-dimensional data preprocessing module, a whole life cycle state evaluation module, a hierarchical early warning decision module, a whole life cycle management module and an edge-cloud collaborative storage module; The whole life cycle data acquisition module is used for acquiring multi-dimensional data of the medical equipment in the whole stage from procurement acceptance, operation and use, operation and maintenance to scrap evaluation, and outputting whole-stage original data; The whole life cycle data acquisition module comprises a procurement acceptance data acquisition unit, an operation real-time data acquisition unit, an operation and maintenance data acquisition unit and a scrap evaluation data acquisition unit; the multi-dimensional data specifically comprises equipment factory technical parameters, operation condition data, use environment data and operation and maintenance operation records; The procurement acceptance data acquisition unit is used for acquiring equipment factory detection parameters, quality inspection reports, on-site installation and debugging records and acceptance standard compliance data, and outputting procurement acceptance original data; The operation real-time data acquisition unit is connected with the equipment control system through a sensor, acquires voltage, current, core component temperature, cumulative operation time length and fault feedback code data in the equipment operation process, and outputs operation real-time original data; the operation and maintenance data acquisition unit is used for acquiring equipment fault maintenance records, part replacement model and quantity information, periodic calibration reports and maintenance execution cycle data, and outputs operation and maintenance original data; The scrap evaluation data acquisition unit is used for acquiring equipment key component aging detection data, actual operation performance attenuation curve and residual value evaluation report data, and outputs scrap evaluation original data; the procurement acceptance original data, the operation real-time original data, the operation and maintenance original data and the scrap evaluation original data are summarized to form whole-stage original data and are outputted; In the embodiment, the four special acquisition units of procurement acceptance, operation real-time, operation and maintenance and scrap evaluation are divided to realize whole coverage acquisition of the whole life cycle data of the medical equipment. In the procurement acceptance stage, the equipment access quality is focused, and it is ensured that the initial parameters are compliant; in the operation stage, the core condition data such as voltage and temperature are captured in real time through a sensor, and the equipment operation state is accurately reflected; in the operation stage, key operations such as maintenance and calibration are recorded, and a basis is provided for performance recovery evaluation; in the scrap stage, aging detection and residual value data are acquired, and equipment exit decision is supported. The comprehensive acquisition of multi-dimensional data provides a complete and reliable data source basis for subsequent state evaluation and early warning.

[0021] The multi-dimensional data preprocessing module performs outlier rejection, data standardization and multi-source data fusion processing on the whole-stage original data, eliminates the format difference and noise interference of different source data, and outputs clean fusion data; The multi-dimensional data preprocessing module comprises an outlier processing unit and a multi-source data fusion unit; The outlier processing unit first screens out explicit outliers in the original data of the whole phase that deviate from the mean by more than 3 times the standard deviation by the 3σ principle, and then uses the isolation forest algorithm to analyze the density of the data distribution and capture implicit abnormal data hidden in the normal data distribution. After marking the explicit and implicit abnormal data, the outliers are removed, and the de-outlier data is output. The multi-source data fusion unit first formats the structured data in the de-outlier data, extracts features from the unstructured data, and then strengthens the weight of key information through the attention mechanism to realize feature alignment and information complementarity of different types of data, forming clean fusion data in a unified format and outputting it. In this embodiment, the combination of "3σ principle + isolation forest algorithm" is used to process outliers, which can efficiently screen out explicit extreme abnormal data and accurately capture implicit abnormalities hidden in normal data distribution, avoiding the problem of missing outliers caused by a single algorithm. In the data fusion link, in view of the differences between structured and unstructured data, format normalization and feature extraction strategies are used respectively, and then the attention mechanism is used to strengthen the weight of key information such as fault keywords and core parameters, effectively eliminating the format barriers and information redundancy of data from different sources, forming clean fusion data in a unified format, and providing high-quality data support for subsequent health degree calculation.

[0022] The full life cycle state evaluation module is based on clean fusion data, combines the characteristics of medical devices in different life cycle stages, uses "analytic hierarchy process + dynamic weight adjustment" to build an evaluation model, calculates the device health degree in real time through quantitative indicators, and outputs the state evaluation result; the dynamic weight adjustment takes the proportion of the actual service life of the device to the designed life and the clinical application dependence coefficient as the core basis, and the sum of the weights of each stage after adjustment is 1; The full life cycle state evaluation module includes a stage characteristic evaluation unit and a health degree calculation unit. The stage characteristic evaluation unit verifies the conformity of the collected data and the factory standard parameters for the device procurement and acceptance stage, and the quantitative indicator is "parameter conformity rate"; For the running stage, the performance stability indicators in the continuous running process of the device are analyzed, and the quantitative indicators include "running parameter fluctuation rate" and "fault feedback code occurrence frequency"; For the maintenance stage, the effectiveness of the performance recovery of the device after maintenance is evaluated, and the quantitative indicator is "performance recovery rate"; for the scrap stage, the reasonableness of the device aging degree and residual value evaluation is judged, and the quantitative indicators include "key component aging rate" and "residual value rate deviation", and the corresponding stage evaluation sub-results of each stage are output. The health degree calculation unit adopts the analytic hierarchy process, and determines the basic weights of the evaluation sub-results in the procurement and acceptance, operation, maintenance and scrap stages as 15%, 40%, 30% and 15% respectively in combination with the practical experience of medical equipment operation and maintenance management; then, a dynamic adjustment factor is constructed based on the ratio of the actual service life of the equipment to the design life, denoted as K, K [0, 1], and the clinical application dependence coefficient, and the adjustment logic is: first, the total upward adjustment amount is calculated as 40% x 0.2K + 30% x 0.3C, then the downward adjustment amount is allocated in a 1:1 ratio according to the basic weight proportions of the procurement and acceptance stage and the scrap stage, and finally the weight of each stage = basic weight ± corresponding adjustment amount, and the total weight after adjustment is 1; the comprehensive health degree of the equipment is calculated by the weighted sum formula: comprehensive health degree = Σ (evaluation sub-result of each stage x weight after adjustment), the state evaluation result is formed and output; In the embodiment, differentiated quantitative indicators are set for the core characteristics of different life cycle stages of medical equipment to ensure the pertinence and effectiveness of the evaluation at each stage. In the weight distribution, the basic weights determined by the analytic hierarchy process are taken as the benchmark, and a dynamic adjustment factor is constructed in combination with the ratio of the actual service life of the equipment and the clinical application dependence coefficient, so that the weights of the operation stage and the maintenance stage are dynamically optimized with the service length and clinical importance of the equipment, and the adjustment amount is proportionally allocated to the procurement and acceptance stage and the scrap stage, which not only ensures the scientificity of the evaluation model, but also adapts to the dynamic changes of the equipment throughout its life cycle. The comprehensive health degree is accurately calculated by the weighted sum formula, and the real state of the equipment is fully reflected.

[0023] The graded early warning decision module sets dynamic early warning thresholds based on the gradient boosting tree algorithm trained on historical data, determines the early warning level and generates corresponding emergency handling suggestions, and outputs graded early warning information and decision schemes; The graded early warning decision module includes an early warning level determination unit and an intelligent pushing unit; The early warning level determination unit divides the comprehensive health degree of the equipment into four levels: normal, mild early warning, moderate early warning and severe early warning; an adaptive model is generated by training the gradient boosting tree algorithm based on the health degree data of the equipment in the last 3 months and the 5-year failure statistics of the same type of equipment, and the threshold interval of each early warning level is dynamically adjusted: when the failure occurrence rate of the same type of equipment in the last 6 months increases by ≥20% compared with the historical average, the lower limit of each level threshold is lowered by a preset value; when the health degree fluctuation rate of the equipment in the last 3 months is ≤5%, the lower limit of each level threshold is increased by a preset value; the early warning level is determined according to the interval in which the current health degree is located and output; The intelligent pushing unit generates targeted handling suggestions according to the early warning level: the mild early warning pushes a routine daily maintenance prompt to the mobile terminal of the front-line maintenance personnel; The moderate early warning pushes a detailed maintenance scheme and a spare parts list to the department head, and simultaneously triggers the spare parts requisition process; The severe early warning triggering device emergency shutdown suggestion is that when the device failure may affect the safety of clinical diagnosis and treatment, the linkage device control system generates a shutdown instruction, which is synchronously pushed to the hospital device management department, the relevant clinical departments and the operation and maintenance emergency team, with the current fault data of the device, the replacement device scheduling suggestion, to form a hierarchical early warning information and decision scheme and output; In the embodiment, the historical data is trained based on the gradient boosting tree algorithm, the early warning threshold is dynamically adjusted in combination with the recent running state of the device and the statistical results of the fault of the same type device, so that the threshold setting is more suitable for the actual running characteristics of the device, and the problems of early warning lag or false alarm caused by fixed threshold are avoided. Different processing schemes are formulated for different early warning levels, the light early warning focuses on daily maintenance, the moderate early warning links spare parts requisition, the severe early warning triggers emergency shutdown and multi-department collaborative response, forming a closed loop mechanism of "early warning-decision-execution", which not only guarantees the safety of clinical diagnosis and treatment, but also improves the accuracy and efficiency of operation and maintenance response.

[0024] The whole life cycle management module formulates the device maintenance plan, divides the life cycle stage, and records the whole process operation and maintenance information of the device based on the state evaluation results and early warning information, to form a closed loop management; The whole life cycle management module includes a maintenance plan generation unit and a whole process tracing unit; The maintenance plan generation unit determines the preventive maintenance period based on the comprehensive health degree of the device, the early warning level and the actual use frequency, automatically generates a plan list including specific maintenance items, required spare part models and specifications, and recommended maintenance time window, and supports manual adjustment and optimization; The whole process tracing unit records the key data change records, state evaluation results and early warning processing execution situation of the device in the whole stage from procurement to scrap in time sequence, forms an unalterable electronic file of the life cycle of the device, and supports reverse tracing of the key information of any link; In the embodiment, the maintenance plan generation unit formulates a personalized preventive maintenance scheme in combination with the health degree of the device, the early warning level and the use frequency, supports manual adjustment and optimization, avoids resource waste caused by excessive maintenance, and prevents device failure risk caused by insufficient maintenance. The whole process tracing unit records the key data and processing situation of the device in the whole life cycle in time sequence, forms an unalterable electronic file, realizes reverse tracing of any link, provides reliable data support for device quality tracing, operation and maintenance responsibility definition and subsequent device procurement decision, and perfects the closed loop management system of the whole life cycle of the device.

[0025] The edge-cloud collaborative storage module adopts the edge node to store real-time running data to guarantee millisecond-level response, and the cloud end to store the whole life cycle historical data and perform encrypted backup, to realize data security and efficient calling; The edge-cloud cooperative storage module comprises an edge node storage unit, a cloud storage unit and a data encryption unit; the edge node storage unit is deployed locally in each clinical department of the hospital, and is specially used for storing real-time data of equipment operation within 72 hours, supporting fast data retrieval and equipment state checking of the local terminal, and ensuring millisecond-level data response speed; The cloud storage unit adopts a distributed storage architecture, is used for storing historical data of the equipment in the whole life cycle, realizes data elastic expansion and off-site disaster recovery backup, supports authorized data sharing and statistical analysis across departments and across hospital areas; The data encryption unit adopts the national SM4 algorithm to encrypt the transmission data between the edge node and the cloud, adopts a role permission hierarchical management mechanism for the stored data, and based on the user role type and the data sensitivity level, different roles can only access the stored data of the corresponding permission level, realizes accurate matching of the permission and the data, and guarantees the safety of the medical data; In the embodiment, the edge node stores the real-time operation data within 72 hours, meets the needs of fast retrieval and state checking of the local clinical department, and guarantees the millisecond-level response speed; the cloud adopts a distributed storage architecture to store the historical data in the whole life cycle, realizes elastic expansion and off-site disaster recovery backup, and supports data sharing and statistical analysis across departments and across hospital areas. Through the SM4 algorithm to encrypt the transmission data, combined with the role permission hierarchical management mechanism, the accurate control of data access is realized, which not only guarantees the safety and privacy of the medical data, but also takes into account the efficiency and flexibility of data calling, and meets the data storage and use needs in different scenarios.

[0026] The medical equipment state monitoring and early warning method based on the whole life cycle comprises the following steps: Step S1, collecting multi-dimensional data of medical equipment procurement acceptance, operation and use, maintenance and repair, and scrap evaluation in the whole stage through a data acquisition module, and outputting the whole-stage original data formed by the collection; Step S2, performing outlier elimination, data standardization and multi-source data fusion processing on the whole-stage original data through a multi-dimensional data preprocessing module, and outputting clean fusion data; comprises the following steps: Step S1, collecting multi-dimensional data of medical equipment procurement acceptance, operation and use, maintenance and repair, and scrap evaluation in the whole stage through a data acquisition module, and outputting the whole-stage original data formed by the collection; Step S2, performing outlier elimination, data standardization and multi-source data fusion processing on the whole-stage original data through a multi-dimensional data preprocessing module, and outputting clean fusion data; Step S3, combining the characteristics of different life cycle stages of the equipment, using the analytic hierarchy process + dynamic weight adjustment to construct an evaluation model, calculating the comprehensive health degree of the equipment based on the clean fusion data, and outputting the state evaluation result; Step S4, according to the state evaluation result, set dynamic early warning threshold based on gradient boosting tree algorithm training historical data, determine the early warning level and generate emergency treatment suggestions, output graded early warning information and decision scheme; Step S5, based on the state evaluation result and early warning information, formulate equipment maintenance plan, record equipment whole process operation and maintenance information, realize whole life cycle closed loop management; Step S6, through edge node storage device real-time running data, cloud storage whole life cycle historical data and encryption backup, guarantee data response efficiency and storage security; In this embodiment, the edge node stores nearly 72 hours of real-time running data, meeting the needs of local quick retrieval and state checking of clinical departments, and guaranteeing millisecond-level response speed; the cloud uses a distributed storage architecture to store whole life cycle historical data, realizes elastic expansion and off-site disaster recovery backup, supports data sharing and statistical analysis across departments and across hospital areas. Through the encryption transmission of data by the national encryption SM4 algorithm, combined with the role permission hierarchical management mechanism, the precise control of data access is realized, which not only guarantees the security and privacy of medical data, but also takes into account the efficiency and flexibility of data calling, meeting the data storage and use needs in different scenarios.

[0027] Step S3, combined with the characteristics of different life cycle stages of the equipment, an evaluation model is constructed by using the analytic hierarchy process + dynamic weight adjustment, the comprehensive health degree of the equipment is calculated based on the clean fusion data, and the state evaluation result is output; Step S4, according to the state evaluation result, set dynamic early warning threshold based on gradient boosting tree algorithm training historical data, determine the early warning level and generate emergency treatment suggestions, output graded early warning information and decision scheme; Step S5, based on the state evaluation result and early warning information, formulate equipment maintenance plan, record equipment whole process operation and maintenance information, realize whole life cycle closed loop management; Step S6, through edge node storage device real-time running data, cloud storage whole life cycle historical data and encryption backup, guarantee data response efficiency and storage security; In this embodiment, through the six-step process of "data collection-preprocessing-state evaluation-early warning decision-life cycle management-data storage", an integrated monitoring and early warning system covering the whole life cycle of medical equipment is constructed. Each step is closely linked, data collection is the basis, preprocessing guarantees data quality, state evaluation accurately reflects the health status of the equipment, early warning decision provides targeted solutions, life cycle management realizes closed loop optimization, edge-cloud collaborative storage guarantees data security and efficient calling, the whole process not only meets the actual needs of medical equipment operation and maintenance management, but also improves the accuracy, timeliness and reliability of monitoring and early warning through algorithm model and technical means, effectively solves the problems of data fragmentation, inaccurate evaluation, early warning lag and lack of closed loop management in traditional medical equipment management.

[0028] It has to be noted that, in the present document, the terms "first", "second", etc. merely serve to identify different entities or actions from each other, without necessarily requiring or implying any actual relationship or order between these entities or actions. Moreover, the terms "comprises", "comprising", or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus. In other words, without further restriction, reference to elements will not, without more limitations, exclude additional, unrecited elements of a process, method, article, or apparatus.

[0029] While embodiments of the application have been shown and described, it is to be understood that the application is not limited to the details of the embodiments described, since numerous further modifications and changes can be apparent to one skilled in the art without departing from the spirit and scope of the application, which is defined by the claims and their equivalents.

Claims

1. A full life cycle based medical device condition monitoring and early warning system, characterized in that: The system comprises a full life cycle data collection module, a multi-dimensional data preprocessing module, a full life cycle state evaluation module, a hierarchical early warning decision module, a full life cycle management module, and an edge-cloud collaborative storage module. The full life cycle data collection module is used for collecting multi-dimensional data of medical equipment in the whole stage from procurement acceptance, operation use, operation and maintenance to scrap evaluation, and outputting whole-stage original data. The multi-dimensional data preprocessing module performs outlier rejection, data standardization and multi-source data fusion processing on the whole-stage original data, eliminates the format difference and noise interference of data from different sources, and outputs clean fusion data. The full life cycle state evaluation module, based on the clean fusion data, combines the characteristics of medical equipment in different life cycle stages, adopts "analytic hierarchy process + dynamic weight adjustment" to construct an evaluation model, calculates the equipment health degree in real time through quantitative indexes, and outputs the state evaluation result; wherein the dynamic weight adjustment takes the proportion of the actual service life of the equipment to the designed life and the clinical application dependence coefficient as the core basis, and the sum of the weights of each stage after adjustment is 1. The hierarchical early warning decision module sets a dynamic early warning threshold based on the gradient boosting tree algorithm training historical data according to the state evaluation result, determines the early warning level and generates the corresponding emergency treatment suggestion, and outputs the hierarchical early warning information and decision scheme. The full life cycle management module formulates equipment maintenance plan, updates life cycle stage division based on the state evaluation result and early warning information, records equipment whole-process operation and maintenance information, and forms a closed-loop management. The edge-cloud collaborative storage module adopts edge node to store real-time operation data to ensure millisecond-level response, and cloud storage to store full life cycle historical data and perform encrypted backup, so as to realize data security and efficient calling.

2. The full life cycle based medical device condition monitoring and alerting system as claimed in claim 1, wherein: The full life cycle data collection module comprises a procurement acceptance data collection unit, a running real-time data collection unit, an operation and maintenance data collection unit and a scrap evaluation data collection unit; the multi-dimensional data specifically comprises equipment factory technical parameters, running condition data, use environment data and operation and maintenance record; The procurement acceptance data collection unit is used for collecting medical equipment factory detection parameters, quality inspection report, on-site installation and debugging record and acceptance standard compliance data, and outputting procurement acceptance original data. The running real-time data collection unit collects voltage, current, core component temperature, cumulative running time and fault feedback code data in the equipment running process through sensor and equipment control system, and outputs running real-time original data; the operation and maintenance data collection unit is used for collecting equipment fault maintenance record, part replacement model and quantity information, periodic calibration report and maintenance execution cycle data, and outputting operation and maintenance original data; The scrap evaluation data collection unit is used for collecting equipment key component aging detection data, actual running performance attenuation curve and residual value evaluation report data, and outputting scrap evaluation original data; The procurement acceptance original data, running real-time original data, operation and maintenance original data and scrap evaluation original data are summarized to form whole-stage original data and output.

3. The full life cycle based medical device condition monitoring and alerting system as claimed in claim 1, wherein: The multi-dimensional data preprocessing module comprises an outlier processing unit and a multi-source data fusion unit; The outlier processing unit first uses the 3σ principle to filter out explicit outliers in the original data of the entire stage that deviate from the mean by more than 3 times the standard deviation. Then, it uses the isolated forest algorithm to perform density analysis on the data distribution, captures the hidden outliers in the normal data distribution, marks the identified explicit and hidden outliers, removes them uniformly, and outputs the outlier data. The multi-source data fusion unit first formats the structured data in the anomaly removal data, extracts features from the unstructured data, and then strengthens the weight of key information through an attention mechanism to achieve feature alignment and information complementarity of different types of data, forming clean fused data in a unified format and outputting it.

4. The full life cycle based medical device condition monitoring and alerting system as claimed in claim 1, wherein: The full life cycle status assessment module includes a stage characteristic assessment unit and a health calculation unit; The aforementioned stage characteristic evaluation unit is designed for the equipment procurement and acceptance stage, verifying the conformity of the collected data with the factory standard parameters, with the quantitative indicator being "parameter conformity rate". For the operation phase, analyze the performance stability indicators of the equipment during continuous operation. The quantitative indicators include "operational parameter volatility" and "frequency of fault feedback codes". For the operation and maintenance phase, the effectiveness of equipment performance recovery after maintenance is evaluated, and the quantitative indicator is "performance recovery rate"; For the scrapping stage, the rationality of the equipment aging degree and residual value assessment is judged, and quantitative indicators include "aging rate of key components" and "residual value rate deviation", and the corresponding stage assessment sub-results are output. The health status calculation unit adopts the analytic hierarchy process (AHP) and combines practical experience in medical equipment operation and maintenance management to determine the basic weights of the evaluation sub-results for the procurement and acceptance, operation, maintenance, and scrapping stages as 15%, 40%, 30%, and 15%, respectively. Then, based on the ratio of the actual service life of the equipment to its design life (denoted as K, K∈[0,1]) and the clinical application dependence coefficient, a dynamic adjustment factor is constructed. The adjustment logic is as follows: first, calculate the total upward adjustment amount = 40% × 0.2K + 30% × 0.3C; then, distribute the downward adjustment amount according to the 1:1 ratio of the basic weights for the procurement and acceptance and scrapping stages. Finally, the weight of each stage = basic weight ± corresponding adjustment amount, and the total weight after adjustment is 1. The overall health status of the equipment is calculated using the weighted summation formula: Overall Health Status = Σ (Evaluation sub-results of each stage × adjusted weights), forming the status assessment result and outputting it.

5. The full life cycle based medical device condition monitoring and alerting system as claimed in claim 1, wherein: The hierarchical early warning decision-making module includes an early warning level determination unit and an intelligent push unit; The warning level determination unit divides the overall health of the equipment into four levels: normal, mild warning, moderate warning, and severe warning. Based on the equipment's historical health data over the past 3 months and the 5-year failure statistics of the same model of equipment, an adaptive model is generated through gradient boosting tree algorithm training. The threshold ranges for each warning level are dynamically adjusted: when the failure rate of the same model of equipment in the past 6 months increases by ≥20% compared to the historical average, the lower limit of each level threshold is lowered by the preset score; when the equipment's health fluctuation rate in the past 3 months is ≤5%, the lower limit of each level threshold is raised by the preset score; the warning level is determined and output according to the current health range. The intelligent push unit generates targeted handling suggestions based on the warning level: for mild warnings, routine maintenance reminders are pushed to the mobile devices of front-line maintenance personnel. A moderate warning pushes a detailed maintenance plan and spare parts list to the department head, simultaneously triggering the spare parts requisition process; Severe warnings trigger emergency equipment shutdown recommendations. When equipment failure may affect the safety of clinical diagnosis and treatment, the linkage equipment control system generates a shutdown command and pushes it to the hospital equipment management department, relevant clinical departments and maintenance emergency team simultaneously. It also includes the current equipment failure data and alternative equipment scheduling suggestions, forming a graded warning information and decision-making plan and outputting it.

6. The full-life cycle based medical device condition monitoring and alerting system as claimed in claim 1, wherein: The full lifecycle management module includes a maintenance plan generation unit and a full-process traceability unit; The maintenance plan generation unit determines the preventive maintenance cycle based on the overall health of the equipment, the warning level, and the actual usage frequency. It automatically generates a plan list that includes specific maintenance items, required spare parts models and specifications, and suggested maintenance time windows, and supports manual adjustment and optimization. The full-process traceability unit records key data changes, status assessment results, and early warning processing implementation status of the equipment from procurement to scrapping in chronological order, forming an unalterable electronic lifecycle archive of the equipment, supporting reverse tracing of key information at any stage.

7. The full-life cycle based medical device condition monitoring and alerting system as claimed in claim 1, wherein: The edge-cloud collaborative storage module includes an edge node storage unit, a cloud storage unit, and a data encryption unit. The edge node storage unit is deployed locally in various clinical departments of the hospital and is specifically used to store nearly 72 hours of real-time equipment operation data. It supports local terminals in the departments to quickly retrieve data and view equipment status, ensuring millisecond-level data response speed. The cloud storage unit adopts a distributed storage architecture to store historical data throughout the entire lifecycle of the device, enabling elastic data expansion and off-site disaster recovery backup, and supporting authorized data sharing and statistical analysis across departments and hospitals; The data encryption unit uses the national cryptographic SM4 algorithm to encrypt the data transmitted between the edge node and the cloud. It adopts a role-based hierarchical management mechanism for stored data, which is based on user role type and data sensitivity level. Different roles can only access the stored data with corresponding permission levels, so as to achieve precise matching of permissions and data and ensure the security of medical data.

8. The method for medical device status monitoring and early warning based on whole life cycle, which is applied to the system for medical device status monitoring and early warning based on whole life cycle as claimed in any one of claims 1-7, characterized in that, Includes the following steps: Step S1: Collect multi-dimensional data from the entire process of medical equipment procurement and acceptance, operation and use, maintenance and repair, and scrapping assessment through the data acquisition module, summarize the data to form the raw data for the entire process, and output it. Step S2 involves using a multi-dimensional data preprocessing module to remove outliers, standardize data, and fuse multi-source data from the raw data across all stages, outputting clean fused data. Step S3: Combining the characteristics of different life cycle stages of the equipment, an assessment model is constructed using the analytic hierarchy process (AHP) and dynamic weight adjustment. The overall health of the equipment is calculated based on cleanroom fusion data, and the status assessment results are output. Step S4: Based on the state assessment results, set a dynamic early warning threshold using historical data trained by the gradient boosting tree algorithm, determine the early warning level and generate emergency handling suggestions, and output graded early warning information and decision-making schemes. Step S5: Based on the status assessment results and early warning information, formulate an equipment maintenance plan, record the equipment's full-process operation and maintenance information, and realize closed-loop management of the entire life cycle; Step S6: Real-time data is processed through edge node storage devices, and historical data throughout the entire lifecycle is stored in the cloud and backed up with encryption to ensure data response efficiency and storage security.

9. The full life cycle based medical device condition monitoring and alerting method as claimed in claim 8, wherein: Step S2 specifically includes the following sub-steps: Step S21: Using the 3σ principle, |x-μ|>3σ, where x is the data value, μ is the mean, and σ is the standard deviation, identify explicit outliers in the original data across all stages, mark them, and temporarily store them. Then, construct an anomaly detection model using the Isolation Forest algorithm, set an anomaly score threshold for the samples, perform density clustering analysis on the data, and capture latent anomalies hidden in the normal data distribution. After merging explicit and latent anomalies, remove them uniformly and output the de-anomaly data. Step S22: Normalize the numerical data in the outlier data using the Z-score standardization method and map it to a preset numerical range. Textual data is converted into fixed-dimensional feature vectors using a word embedding algorithm; the weight coefficients of each feature are calculated through an attention mechanism to strengthen the weight ratio of fault keywords and key features of core parameters, thereby achieving feature alignment and information complementarity between numerical and textual data, and outputting clean fused data in a unified format.

10. The full life cycle based medical device condition monitoring and alerting method as claimed in claim 8, wherein: Step S4 specifically includes the following sub-steps: Step S41: Collect the device's historical health data for the past 3 months and the fault statistics of the same model of device. Train the model using the gradient boosting tree algorithm to determine the dynamic early warning threshold that is adapted to the device's operating characteristics. When the device's current health is lower than the corresponding threshold, trigger the corresponding early warning level. Step S42: Generate targeted handling suggestions based on the warning level: mild warnings are pushed to the mobile devices of maintenance personnel via system messages; The moderate warning was simultaneously sent to the department head via system message and email; Severe warnings trigger the generation of work orders in the hospital equipment management system. When equipment failure may affect the safety of clinical diagnosis and treatment, the system will activate the equipment shutdown protection mechanism and simultaneously push the information to the hospital equipment management department and relevant clinical departments, outputting tiered warning information and decision-making solutions.

Images