Medical instrument life cycle management and tracking method and system

Abstract

The invention relates to the technical field of medical instrument information management and intelligent scheduling, in particular to a medical instrument life cycle management and tracking method and system, and the method comprises the steps: starting from the collection of original vibration signals and the generation of fractal health degrees, integrating fault risk data with equipment use plans, department allocation arrangement and maintenance resource information, and a quantum heuristic solution algorithm is utilized to complete equipment scheduling and maintenance decision making, and visualization and interactive revision are carried out through a holographic virtual reality interface. And then, the system detects a scheduling execution result, and if the health degree of the equipment cannot be recovered after multiple times of maintenance, decommissioning marking is executed in the digital twinborn body, and complete closed-loop management is formed. According to the invention, the use safety of the instrument and the resource allocation efficiency are improved, and the defects of a comparison scheme in fault identification and scheduling modes are overcome.

Description

Technical Field

[0001] The present invention relates to the technical field of medical device information management and intelligent scheduling, and particularly to a method and system for managing and tracking the life cycle of medical devices. Background Art

[0002] In modern hospitals, the usage and types of medical devices are constantly increasing. Suppliers usually need to build a comprehensive database to register device information, and hospitals also need to track the entire process of equipment from procurement, use to maintenance and retirement. However, existing solutions mostly rely on simple usage logs or static thresholds for inventory and fault management, lacking in-depth analysis of the vibration characteristics and true wear degree of the devices, resulting in the inability to detect potential faults in a timely manner. In addition, in the scheduling process, devices are often allocated manually or according to fixed rules, which is difficult to quickly respond to emergencies and cannot dynamically adapt to maintenance resource limitations, easily causing risks of duplicate purchases or serious accidents. Summary of the Invention

[0003] In view of the many problems existing in the above-mentioned prior art, the present invention provides a method and system for managing and tracking the life cycle of medical devices. The present invention collects and analyzes vibration information for medical devices, calculates the fractal health degree and fault risk, and then combines the department requirements and maintenance resources. An optimal scheduling strategy is generated by a quantum-inspired algorithm and presented on a holographic interaction interface for managers to revise. After executing the instruction, the system detects the result. If it is found that the device fails multiple times and cannot be repaired, a retirement mark is made to form a complete closed-loop management of the device life cycle. This principle of combining automation and visualization can greatly improve the utilization efficiency and usage safety of the devices.

[0004] A method for managing and tracking the life cycle of medical devices, comprising the following steps:

[0005] Obtain the unique identification data of the medical device, collect the raw vibration signal data, clean the raw vibration signal data to generate fractal dimension data, and output fault risk data according to the fractal dimension data and a preset fractal health degree threshold;

[0006] Integrate the fractal health degree, fault risk data and demand information to generate initial multi-agent scenario data, iteratively generate a device scheduling plan through a quantum-inspired solution algorithm, and form a scheduling instruction after performing an interactive operation on the device scheduling plan in a holographic virtual reality environment;

[0007] Detect the result of executing the scheduling instruction to identify abnormal event data, and make a retirement mark for the medical device that meets the retirement conditions in the digital twin to complete the closed-loop management of the entire life cycle of the medical device.

[0008] Preferably, when cleaning the original vibration signal data, the vibration waveform sequence is compared with a fixed frequency range to filter out noise signals, and by removing the peaks with numerical mutations in continuous sampling points, waveform data meeting the requirements for fractal dimension calculation is generated.

[0009] Preferably, the fractal dimension data is obtained by applying a fractal analysis algorithm to the multi-dimensional vibration trajectory of phase space reconstruction, and the complexity of the vibration mode of the medical device is represented by the comprehensive measure of the vibration amplitude and phase evolution, which is used for subsequent fractal health calculation.

[0010] Preferably, the fractal health is used to characterize the usage status of the medical device in a numerical range from zero to one hundred. When the fractal dimension data exceeds a preset threshold, the fractal health value is reduced to an interval indicating greater wear to reflect the performance degradation of the device caused by long-term vibration shock.

[0011] Preferably, when the fractal health drops below a preset fractal health threshold, failure risk data is output and the failure risk data is associated with the digital twin of the corresponding medical device, so as to record the risk status of the medical device malfunctioning in the digital twin.

[0012] Preferably, the demand information consists of the device usage plan, department transfer arrangement, and maintenance resource allocation parameters, and together with the fractal health and failure risk data, multi-agent initial scenario data is generated for subsequent iteration of the quantum-inspired solution algorithm.

[0013] Preferably, when the quantum-inspired solution algorithm performs iterative calculations on the discrete optimization model established for the multi-agent interaction process, for scenarios with resource conflicts, medical devices with failure risk data exceeding the threshold are preferentially arranged to enter the maintenance or replacement process.

[0014] Preferably, the holographic virtual reality environment is used to display the distribution location of the device scheduling plan among hospital departments, and allows managers to generate scheduling instructions after interactive modification based on the visualization results, and write the scheduling instructions to the digital twin for subsequent execution.

[0015] Preferably, meeting the retirement condition means that the fractal health is continuously lower than the preset fractal health threshold and the failure risk data cannot be restored to a usable state after being triggered for maintenance multiple times. When identifying abnormal event data, a retirement mark is executed on the medical device, and the retirement information is registered in the digital twin to form a closed-loop record.

[0016] A medical device life cycle management and tracking system for implementing the medical device life cycle management and tracking method as described above, the system includes:

[0017] An identification and fractal analysis module, which is used to obtain the unique identification data of medical devices and collect the original vibration signal data, clean the original vibration signal data to generate fractal dimension data, and output fault risk data according to the fractal dimension data and a preset fractal health threshold;

[0018] A quantum scheduling holographic interaction module, which is used to integrate the fractal health and fault risk data with demand information to construct multi-agent initial scenario data, iteratively generate a device scheduling scheme through a quantum-inspired solution algorithm, and form a scheduling instruction after performing an interaction operation on the device scheduling scheme in a holographic virtual reality environment;

[0019] An anomaly detection and retirement module, which is used to detect the result of executing the scheduling instruction to identify anomaly event data, and mark the corresponding medical device for retirement in the digital twin when the retirement condition is met, so as to complete the closed-loop management of the entire life cycle of the medical device.

[0020] Compared with the prior art, the advantages and beneficial effects of the present invention are as follows:

[0021] The present invention realizes the precise quantification of the device wear state through fractal vibration analysis means;

[0022] The present invention realizes the global optimization of complex multi-department device requirements and maintenance resources through quantum-inspired solution technology means;

[0023] The present invention realizes an intuitive and visual scheduling scheme and simplifies the manual revision process through holographic virtual reality interaction means;

[0024] The present invention realizes the closed-loop information tracking from factory to retirement through digital twin retirement marking means. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 It is a flowchart of the method of the present invention;

[0026] Figure 2 It is a flowchart of the identification and fractal analysis process in the present invention;

[0027] Figure 3 It is a flowchart of the quantum scheduling holographic interaction process in the present invention;

[0028] Figure 4 It is a structural block diagram of the system of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0029] Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present disclosure. In the following detailed description, for the sake of explanation, many specific details are set forth to provide a thorough understanding of the embodiments of the present disclosure.

[0030] As Figure 1 shown, a method for the life cycle management and tracking of medical devices includes the following steps:

[0031] Obtain the unique identification data of the medical device, collect the original vibration signal data, clean the original vibration signal data to generate fractal dimension data, and output fault risk data according to the fractal dimension data and a preset fractal health threshold;

[0032] With the help of a sensing patch or a micro-vibration monitoring component, continuously or periodically record the vibration waveform sequence generated by the medical device during operation and handling. By performing operations such as denoising, filtering, and peak removal on the waveform sequence, waveform information available for fractal calculation is obtained, and then the fractal dimension is calculated at the phase space or time series level to measure the vibration complexity. According to the numerical value of the fractal dimension, it is compared with a set fractal health threshold. If the fractal dimension exceeds the threshold, the system will output fault risk data and synchronously record it in the digital twin of the medical device, indicating the existence of potential fault hazards or high wear conditions. Through this link, abnormal wear of the device can be identified early, providing pre-information for subsequent equipment scheduling and maintenance, thus realizing closed-loop management.

[0033] In practical applications, the unique identification data is usually provided by hospitals or manufacturers and includes device models, serial numbers, production batches, and registration information. Before collecting any vibration data, it is necessary to first perform "identity registration" on the device to ensure accurate matching to the same device in subsequent digital twin management and scheduling decisions.

[0034] For example: In a certain device information system, the unique identification data of newly-instock surgical instruments is entered to form a device file. The identification data is written into the system background by scanning the code, and subsequently, the vibration patch can associate the collected original vibration signals with the corresponding device according to this identification data.

[0035] For the sensing hardware, the common practice is to paste a micro-vibration patch or an embedded accelerometer module on the surface of the device and periodically read the vibration waveform (which can be at a millisecond or second time step). To capture frequent or slight vibrations, the sampling frequency is generally set to ≥500Hz to extract high-frequency pulses. These original vibration signal data will enter the subsequent steps for fractal analysis, which is a key metric for determining device wear and fault risk.

[0036] In one embodiment, vibration patches are installed on a batch of reusable electrocardiogram monitors, and the waveform sequence is collected once every 30 seconds, with each collection lasting for 3 seconds, generating original vibration signal data with timestamps and transmitting it to the background.

[0037] A band-pass filter is adopted to filter out the noise that significantly exceeds the normal vibration frequency band of medical devices; for sudden extreme high-value points (which may be caused by electromagnetic interference or instantaneous sensor failures), they are eliminated by setting amplitude limits or statistical thresholds. The cleaned vibration sequence is mapped into the phase space dimension to reveal the non-linear characteristics of the waveform; after the reconstruction is completed, it is calculated through the fractal dimension formula. Taking the box dimension as an example, it can be denoted as:

[0038]

[0039] where D f is the fractal dimension, and N(r) represents the number of squares (or cubes, depending on the reconstruction dimension) with side length r required to cover the trajectory. If the fractal dimension value is larger, it means the local complexity of the waveform is higher; generally, in the present invention, a baseline test is first performed on the same type of device to determine a reasonable fractal dimension value range.

[0040] In one embodiment, the vibration sequence of the electrocardiogram monitor is band-pass filtered through a Matlab script, and the waveform points that exceed the expected frequency by more than 10% are eliminated. The phase space reconstruction is performed on the filtered waveform. Assuming the embedding dimension is 3, the box dimension algorithm is used to calculate D f = 2.35. This value is recorded and stored in the digital twin for subsequent threshold comparison.

[0041] One or more thresholds are set in the system in advance, such as Threshold = 2.20. If the fractal dimension of a certain device is higher than Threshold, it is regarded as "abnormal vibration characteristics". Multi-segment thresholds can also be set to divide different risk levels.

[0042] In the system, when the fractal dimension D f is greater than the threshold Threshold, "fault risk data" is immediately generated, and this result is bound to the digital twin of the corresponding medical device. RiskData = 1 indicates that the fault risk has been triggered; RiskData = 0 indicates that the fractal dimension is within the safe range. More auxiliary information can also be added, such as the current using department, sampling timestamp, etc., for comprehensive consideration during scheduling decisions.

[0043] In one embodiment, in the aforementioned monitor, if it is calculated that D f = 2.35, and the threshold is set to 2.20, the system determines that the vibration signal of the device is abnormal. "Fault risk data = 1" is written on the digital twin, and a timestamp "2023-07-01 14:05" is generated to assist the subsequent maintenance department in maintenance.

[0044] Preferably, when cleaning the original vibration signal data, the vibration waveform sequence is compared with a fixed frequency range to filter out noise signals, and the peaks with numerical mutations in continuous sampling points are removed to generate waveform data that meets the requirements for fractal dimension calculation.

[0045] In the present invention, medical device suppliers usually record the basic information of each device in their database and establish a digital file for it. After the medical device is put into use, the equipped micro-vibration sensing component will continuously or intermittently submit the original vibration signal data to the database or hospital information system. In order to perform subsequent fractal dimension calculation or fault prediction on these waveforms, data cleaning is required to avoid distortion caused by noise interference or extreme peaks. In this way, throughout the life cycle management of medical devices, from the supplier's database establishment to the hospital's use and then to subsequent maintenance or retirement, it is possible to uniformly rely on the same or interconnected database environment, making each device vibration data traceable and consistent.

[0046] When the device is in the use stage (such as in the operating room, ward monitor, inspection equipment, etc.), its vibration data will be continuously uploaded to the system. The medical device life cycle management and tracking solution needs to first clean the original waveform to remove noise and sudden peaks that do not match the fixed frequency range to ensure the accuracy of subsequent "fractal dimension calculation" and even fault prediction.

[0047] According to the type of device and its working principle, a reasonable vibration frequency band is usually determined in advance. If vibration components outside this frequency band are detected, it may be external impact or electromagnetic noise and needs to be filtered out. For abnormal spikes that appear within the sampling window (which can be caused by instantaneous impact or sensor failure), they can be marked and excluded according to the set threshold or statistical determination method.

[0048] The "waveform data that meets the requirements for fractal dimension calculation" obtained through this step can more truly reflect the vibration characteristics of the device under normal working conditions and avoid the deviation brought by noise to the subsequent determination results.

[0049] In practical applications, medical devices often have a specific main vibration frequency range during normal operation. For example, the main vibration peak of a certain monitor during operation is between 50Hz and 250Hz. By setting v min and v max (lower frequency limit and upper frequency limit) to construct a fixed frequency range. If a strong component of the detected waveform is outside this range, it is determined as a noise component and filtered out.

[0050] Band-pass filtering is adopted: for example, for v min = 50 and v max = 250, only the vibration energy within this frequency band is retained;

[0051] Or directly perform energy zeroing on the frequency bands exceeding [v min , v max in the spectrum analysis.

[0052] In practical applications, a captured time-domain vibration waveform is subjected to a fast Fourier transform (FFT), and the frequency components within the range of [v min , v max are retained; the high-frequency electromagnetic interference in the computer room environment and the ultra-low-frequency ground vibration components during the movement of the instrument are removed.

[0053] In one embodiment: When a certain supplier database detects that the vibration of an X-type injection pump is mainly concentrated between 30 Hz and 180 Hz, then v min = 30, v max = 180. If there are energy peaks less than 30 Hz or greater than 180 Hz, they are directly filtered out to make the subsequent waveform more stable.

[0054] Among consecutive sampling points, the peaks of some numerical mutations may not represent real mechanical vibrations but instantaneous anomalies. If not removed, it will cause the calculation of the fractal dimension to be distorted, and further misjudge the wear degree of the equipment.

[0055] Set a peak clipping threshold γ or use a moving window standard deviation statistic. When the sampling point exceeds (where is the average value within the window, σ is the standard deviation within the window, and α is a preset multiple), then this point is interpolated or deleted. Remove these unreasonable mutation peaks to ensure that the waveform sequence fluctuates within a reasonable range.

[0056] In the waveform sequence {x1, x2, …, x n}, each x k is compared with its front and rear neighbors. If |x k - x k-1 | > β or |x k - x k+1 | > β, and β is a certain specific threshold, then it is determined that this point may be a sudden peak and is removed or interpolated with the average value. Make the waveform sequence smoother and remove the extreme data points caused by electromagnetic shocks or accidental collisions.

[0057] In one embodiment: For a ventilator operating in an operating room, its sampled data sometimes records instantaneous spikes caused by collisions of medical staff. By removing such spikes, the subsequent calculated fractal dimension will not be affected by outliers.

[0058] The cleaning process of the present invention can be understood as a combination of frequency-domain filtering and time-domain outlier rejection. By comparing the waveform sequence with a fixed frequency-domain range for denoising and adding the rejection of peak mutations, "waveform data that meets the requirements for fractal dimension calculation" is finally obtained. This waveform can better represent the actual vibration behavior of the instrument and exclude non-primary factors such as external noise or accidental collisions.

[0059] If unfiltered noise or extreme peaks are directly input into the fractal analysis, it will significantly increase the fractal dimension and give false warnings to the system. When suppliers or hospital maintenance departments query the database, cleaner waveform records can be obtained, which helps to track the wear of equipment in the long term and also provides a more accurate basis for subsequent calculation of failure risks. Adapt to the batch management of the supplier database: Through a fixed cleaning algorithm, automatic filtering and rejection of abnormal peaks can be performed during batch processing, greatly reducing the burden of manual screening and being suitable for periodic detection of a large number of medical devices.

[0060] In one embodiment, a supplier database stores vibration monitoring records of multiple hemodialysis machines. These hemodialysis machines generate medium-frequency vibrations during dialysis, while the motor noise generated by other surrounding devices is in a higher frequency band.

[0061] When performing FFT analysis on the original waveforms of these hemodialysis machines, set ν min = 20 and ν max = 150, and only retain the frequency components within this range and filter out the remaining noise. During subsequent fractal dimension calculation, it is no longer affected by the high-power fans (400 Hz) in the computer room. At the same time, by monitoring the mutation situation of consecutive sampling points, the outliers caused by engineers occasionally hitting the machine shell are removed.

[0062] When implementing, deploy a data cleaning script at the backend of the database to perform the following operations on each incoming vibration sequence of the hemodialysis machine: perform band-pass filtering to retain 20 Hz and 150 Hz; calculate the average value and standard deviation σ within the sampling window with the same waveform length of 256, and regard the points exceeding as abnormal peaks and reject them; the output waveform is more suitable for subsequent calculation of the fractal dimension using algorithms such as the box dimension algorithm without interference from noise peaks. The database observed that the fractal dimension before cleaning was 2.55, while after cleaning it was 2.02. The latter significantly reduced external irrelevant interference, and the equipment was evaluated to be in a medium wear state and did not require immediate maintenance.

[0063] Preferably, as Figure 2 shown, the fractal dimension data is obtained by applying a fractal analysis algorithm on the multi-dimensional vibration trajectory of phase space reconstruction, and represents the complexity of the vibration mode of the medical device through the comprehensive measure of the evolution of vibration amplitude and phase, and is used for subsequent calculation of fractal health.

[0064] The present invention introduces the phase space reconstruction technology to map the vibration waveform sequence into a multi-dimensional state trajectory, so as to explore the hidden non-linear features in the vibration process of the device. When a supplier provides medical devices for a hospital or other user units, the vibration data of the corresponding device can be recorded in its database. By means of phase space reconstruction, a single time series is transformed into a multi-dimensional point set, which can more completely present the amplitude change and phase evolution law of the vibration.

[0065] The key parameters of phase space reconstruction include the embedding dimension and the delay step size. The actual working frequency band and the desired dynamic range of the medical device help to select an appropriate embedding dimension at the database level. If the generated vibration trajectory after reconstruction is more scattered or shows complex nesting, it may mean that the device experiences large vibration disturbances during use.

[0066] After the phase space reconstruction is completed, the present invention uses a fractal analysis algorithm to measure the complexity of the multi-dimensional trajectory, that is, the fractal dimension data. By comprehensively measuring the amplitude and phase, the non-linear degree of a certain vibration mode can be determined. This fractal dimension data will be stored in the digital file of the medical device (jointly managed by the supplier database or the hospital system), and will be used in the subsequent calculation of the fractal health degree to support the failure risk judgment and maintenance decision-making of the medical device.

[0067] In practical applications, the vibration amplitude can be used to evaluate whether there is excessive mechanical jitter during the operation of the device by detecting the peaks and valleys of the waveform. If the peak value in a certain area continues to increase, it may indicate progressive wear. For the phase evolution, after mapping the time series into the phase space, the distribution and surrounding mode of the data points in this multi-dimensional space can be observed. If the phase produces a large deviation or chaos over time, it means that the internal vibration mechanism is unstable.

[0068] In the system implementation, the above two aspects are combined into a comprehensive measurement, such as statistically analyzing the structural distribution of the point cloud in the local neighborhood of the phase space trajectory. If the neighborhood distribution is more complex, the fractal dimension is usually higher.

[0069] The more commonly used fractal analysis method is the box dimension (it is also possible to use other methods). The box dimension can be denoted as:

[0070]

[0071] where D f represents the fractal dimension, and N(r) is the number of covering units with side length r. After the phase space reconstruction, the algorithm will statistically analyze the change of the number of covering units with the scale on the multi-dimensional vibration trajectory, and calculate the value of the fractal dimension based on this. The higher the fractal dimension, the more "complex" the vibration trajectory is, and there may be large-scale irregular fluctuations in amplitude or phase.

[0072] Before delivering the equipment to the hospital, the supplier usually stores a set of vibration test baseline data. If the fractal dimension collected by the hospital subsequently is significantly higher than the baseline, it indicates that a significant shift has occurred in the vibration mode of the equipment. A "equipment ID - fractal dimension" comparison list can be established in the database to continuously receive the updated values uploaded by the hospital and monitor the health status of the equipment in real time.

[0073] In the hospital or a third - party maintenance center, when the fractal dimension rises to an abnormal level, the system will promptly give maintenance or decommissioning suggestions through health calculation. If it is found during the scheduling process that the fractal dimension continues to rise, the corresponding department will be instructed to reduce the usage frequency of the equipment or immediately transfer it to the maintenance department for inspection, forming a closed - loop management.

[0074] Through phase - space reconstruction, multi - dimensional information can be obtained from a single time series, presenting the changes in amplitude and phase in the same trajectory; the complexity of the vibration mode is measured using a fractal analysis algorithm and numerically represented by the fractal dimension; by comparing the fractal dimension with the fractal health threshold, abnormal vibrations of the equipment can be identified at an early stage and its degree can be quantified.

[0075] The present invention avoids judging equipment failures solely based on manual experience and more accurately identifies potential abnormalities with the help of the fractal dimension; the supplier's database provides the baseline vibration characteristics of the medical device at the hospital end, and the hospital end continuously updates the fractal dimension during use. Through network collaboration between the two parties, a unified life - cycle tracking can be formed; if the fractal dimension of the equipment gradually rises during operation, the hospital can intervene in advance to reduce the probability of downtime or surgical delays caused by sudden failures.

[0076] In one embodiment, the supplier records the vibration waveform sequence of the monitor during factory inspection, sets the phase - space embedding dimension to 3, and obtains D f = 2.05 using the box - counting dimension calculation method. Subsequently, this baseline record is uploaded to the database.

[0077] When the hospital uses the monitor, it periodically calculates the fractal dimension. If D f suddenly increases to 2.70 during a subsequent sampling, it indicates that the vibration of the equipment has become complex and chaotic, and there may be hardware loosening or internal component aging. The system automatically marks the monitor as being in a high - risk state and generates a maintenance instruction.

[0078] In another embodiment, a maintenance center compares the trajectory evolution of multiple hemodialysis machines under phase - space reconstruction and calculates the fractal dimension. It is found that the phase point cloud of one of them shows a significant deviation and D f is significantly higher than that of other equipment of the same type.

[0079] After the maintenance center docks this conclusion with the medical device database, it triggers a fault reminder and arranges for troubleshooting. After disassembly and inspection, it is found that some transmission mechanisms are worn, and parts need to be replaced in a timely manner, successfully avoiding sudden failures during actual patient use.

[0080] Preferably, the fractal health degree characterizes the usage state of the medical device in a numerical range from zero to one hundred. When the fractal dimension data exceeds a preset threshold, the fractal health degree value is reduced to an interval indicating greater wear to reflect the performance attenuation caused by long-term vibration and shock of the device.

[0081] In the present invention, the fractal health degree is distributed numerically between 0 and 100 to represent the usage state of the medical device. Through such a linear interval, suppliers and users (hospitals or third-party maintenance agencies) can intuitively record and compare the health levels of different devices at the database level.

[0082] When the fractal health degree of a certain device is close to 100, it indicates that its vibration state is relatively stable and the wear is small in the short term. When the fractal health degree is in a lower section, it implies that the device has experienced greater impact or continuous poor vibration, and there is a significant attenuation risk. After incorporating the fractal health degree data into the supplier database, the medical device management platform can measure the wear conditions of multiple devices with a unified standard, making subsequent scheduling, maintenance, or retirement decisions more well-founded.

[0083] In this numerical interval mode, when the fractal dimension data exceeds a certain fixed threshold (usually determined by the device type or historical statistics), the fractal health degree will be "forced" to be reduced from the original higher interval to an interval reflecting severe wear or high risk. The "preset threshold" here can be either based on the long-term statistical results of the same type of device or given by the device manufacturer according to the acceptance criteria and input into the supplier database or the hospital information system.

[0084] If the fractal dimension exceeds the threshold, it indicates that the vibration mode has become significantly more complex or chaotic, and the corresponding device performance is accelerating decay. In this case, the decrease in the health degree value can prompt the relevant management system to reduce the frequency, schedule maintenance, or replace it in advance.

[0085] Suppliers often conduct a vibration detection and fractal analysis during the factory acceptance of the device to set an initial health degree for the device, such as 95. This initial value is uploaded to the database as a benchmark. After the device is put into use in the hospital, as the fractal dimension is continuously monitored, the system will periodically calculate or correct the fractal health degree. When the fractal dimension is below the threshold, the relatively high health degree is maintained; otherwise, the health degree is quickly pulled down to a specific numerical range to indicate the deterioration of the device condition.

[0086] In an implementation scenario, the initial fractal health of a certain monitor is recorded as 92. If subsequent monitoring finds that the fractal dimension value is higher than the threshold for a long time, the health is decreased from 92 to 60, which is marked as the medium-risk range. This result is synchronously written into the digital file of the device, and the hospital or third-party maintenance can arrange maintenance based on this.

[0087] The threshold setting principle is based on the statistical history vibration characteristics of the same type of device or the mechanical test results of the manufacturer. For example:

[0088] For endoscope devices, the fractal dimension threshold is set to 2.3;

[0089] For ventilators or monitors, it is set to 2.5;

[0090] The supplier registers the threshold value for each model device in the database for subsequent determination of health.

[0091] If the vibration data corresponds to the fractal dimension D f exceeds the threshold value Threshold, the system will reduce the fractal health to a low range (such as 30 - 50), or decrease it to an exact value according to a certain functional relationship.

[0092] In practical applications, the calculated fractal dimension D f is compared with the threshold Threshold:

[0093] If D f ≤Threshold, the health will only be adjusted downwards or slightly on the original basis;

[0094] If D f >Threshold, the health will directly drop significantly to the range indicating greater wear.

[0095] At this time, the device ID and the updated health are written back to the medical device database together, and the timestamp and department location are marked for subsequent retrieval and linkage.

[0096] The continuous attenuation model converts the complexity of the cumulative vibration of the device (measured by the fractal dimension) into an easily understandable health. Through the threshold mechanism, a non-linear attenuation method can be formed: once exceeding the threshold, the health suddenly drops to the warning level, indicating that "the performance has significantly decayed due to long-term vibration shock".

[0097] This mapping can be in the form of a linear or piecewise function. For example, when the fractal dimension exceeds the threshold by less than 0.1, the health drops to 50; if it exceeds the threshold by more than 0.2, the health drops to 30. The specific values vary depending on the supplier or hospital strategy.

[0098] Map the complex fractal dimension changes into intuitive values between 0 and 100. Without in-depth understanding of non-linear theory, hospital departments or maintenance departments can quickly grasp the "health status" of equipment. In scenarios with a large number of devices, the system can quickly screen out devices with low health status for allocation or maintenance; if the health status is below 20, it can even be directly included in the retirement assessment process. Suppliers can set the basic health status and thresholds in the product database, and users can update the health status based on daily monitoring. If it is observed that the health status of a device drops below the threshold multiple times, the supplier can recommend replacing parts or recycling.

[0099] In one embodiment, the supplier database sets the fractal dimension threshold for this type of monitor to 2.40 and the initial fractal health status to 95. After half a year, a hospital collects a calculated fractal dimension value of 2.35, which does not exceed the threshold, so the health status is slightly adjusted to 90; if the fractal dimension rises to 2.45 (exceeding the threshold by 0.05) three months later, the health status immediately drops to 70. In the hospital's digital twin and the supplier database, the device is marked as moderately worn, and it is recommended to arrange maintenance within one month to avoid subsequent failure risks.

[0100] In another embodiment, the supplier designates a fractal dimension threshold of 2.10 for hemodialysis equipment. If it is exceeded, the health status directly drops to 60. The hemodialysis center observes that the health status of a certain device drops from an initial 96 to 60, and further sampling three days later finds that the fractal dimension continues to rise, and the health status drops to 50. The management of the hemodialysis center lists it as a priority for maintenance and suspends its use during this period to ensure patient safety; the supplier can also obtain this data remotely and incorporate it into product improvement or recall plans.

[0101] Preferably, when the fractal health status drops below a preset fractal health status threshold, output failure risk data and associate the failure risk data with the digital twin of the corresponding medical device, so as to record the risk status of the medical device malfunctioning in the digital twin.

[0102] In the present invention, each medical device has a fractal health status (ranging from 0 to 100) to quantify its wear degree or vibration state. When this health status value drops below a preset fractal health status threshold (such as 50 or 30, etc.), the system will automatically identify it as "increased failure risk". At this time, "failure risk data" is output and associated with the digital twin of the device, providing a basis for subsequent maintenance and scheduling decisions.

[0103] A digital twin refers to the state simulation and information recording of a device in a virtual space in a one-to-one correspondence. Suppliers can establish a basic file (serial number, model, initial fractal health level, etc.) for each piece of equipment in advance in their database. After the equipment enters the hospital for use, the hospital information system will continuously update the information of this twin, such as changes in the fractal health level, actual usage records, etc. Once the health level drops below the threshold and fault risk data is output, this digital twin is marked as having a fault risk, thus realizing the mirror tracking of the real device.

[0104] In practical applications, suppliers can set one or more thresholds for each type of medical device based on factory inspection or historical statistics of similar devices. For example, for an electrocardiogram monitor, the threshold can be set to 40; for a hemodialysis machine, it can be set to 50. A higher threshold means that even minor wear will trigger an alarm, while a lower threshold indicates a more lenient risk tolerance.

[0105] The supplier's database reserves a column "fractal health level threshold" for device models M1, M2, etc. The hospital system reads this threshold for comparison when updating the fractal health level. The comparison logic is:

[0106] If H d <T h ,then output RiskData

[0107] H d represents the current fractal health level value, and T h represents the fractal health level threshold. When H d is less than T h , it is determined that the device enters the fault risk state.

[0108] The fault risk data can include: the unique ID of the device; the fault risk level (such as "medium risk" or "high risk"); the timestamp; the associated reasons (what the health level value is and what the threshold is). The system can automatically perform the above comparison daily or after each sampling; if it is triggered repeatedly for a period of time, it means that the device is continuously deteriorating, and it can be recorded in the hospital maintenance schedule and the maintenance department can be notified.

[0109] After the system identifies the fault risk, it writes the "fault risk data" into the "dynamic information area" of the digital twin of this device, adding a new entry such as "RiskData = 1, trigger time = 2026-03-15 10:22". The supplier's database or the hospital system can synchronize the data based on the unique ID of the device, enabling the remote platform to also see that the device is in a risk state.

[0110] This means that from this point on, the digital twin will carry a fault risk flag, such as "isRisk = true". In the medical device life cycle management platform, the system will highlight or mark it in red in the device list to prompt that it needs to be scheduled or repaired first.

[0111] In one embodiment, the fractal health of a certain infusion pump has continuously dropped from 65 to 45 within the recent week, and the threshold is 50. The system detects that "45 < 50" and immediately generates "fault risk data" with a marked level of "medium risk" and writes it into the digital twin. The maintenance staff receives the warning and then checks and discovers that the internal peristaltic tube is slightly damaged, and it is recommended to replace the parts to prevent infusion interruption.

[0112] This process takes the "fractal health threshold" as the automatic recognition threshold for fault risks. As long as the health of the device no longer meets the safe range, risk data will be immediately output. At the same time, the digital twin is regarded as the "mirror image" of the device in the information space, and the risk status is recorded in its dynamic fields, enabling the system or administrator to promptly grasp the health status and potential problems of the device when viewing.

[0113] Once the health drops below the threshold, the system will let the fault risk data drive the maintenance process, demand scheduling, or department warning, significantly reducing medical accidents or clinical interruptions. The time points and causes of the occurrence of fault risks are all retained in the digital twin, which is valuable for subsequent statistics and improvement. Suppliers can also collect large sample data based on this to optimize the device design or adjust the threshold setting. When device risks occur, information not only circulates within the hospital but also leaves a mark in the supplier's database. If a device triggers fault risks multiple times, the system can upgrade it to "high risk", and then, in combination with other indicators, finally determine whether to retire or send it back to the factory for repair to achieve full life cycle tracking.

[0114] In another embodiment, a large hospital uses 100 hemodialysis devices, each with a fractal health threshold = 55. The fractal health of each device is regularly monitored. If the data of a certain device this month is 52, which is less than 55, it immediately triggers RiskData: "risk occurred" and writes it into the digital twin. The serial number of this device and the trigger time are recorded in the supplier's database, and the field "health = 52" is added.

[0115] The system includes this device in the list to be checked; the department reduces the high - load use of it according to the prompt, or directly arranges a replacement plan; if the health recovers after subsequent maintenance, both the large hospital database and the supplier can synchronously see the updated status.

[0116] Integrate the fractal health, fault risk data, and demand information to generate initial multi - agent scenario data, iteratively generate a device scheduling plan through a quantum - inspired solution algorithm, and form a scheduling instruction after performing interactive operations on the device scheduling plan in a holographic virtual reality environment;

[0117] The present invention uses a numerical range from 0 to 100 to describe the current usage status of the device. A higher value indicates a lower fractal dimension and stable vibration, while a lower value indicates a higher fractal dimension and complex vibration characteristics.

[0118] Assume there is a fractal health parameter H, where H = 100 represents an ideal vibration state and H = 0 indicates extremely unstable vibration. In the medical device database constructed by the supplier, the latest value of H can be stored in the digital file of the device and associated with the device identification information.

[0119] In the present invention, each device stores a "fractal dimension threshold" to determine whether the vibration complexity enters the high-risk area. If the fractal dimension exceeds this threshold, the system regards the fractal health H as experiencing a phased decline, manifested as a shift in the value towards the "greater wear" direction.

[0120] The threshold can be regarded as δ. When the fractal dimension D f is greater than δ, H undergoes a significant attenuation and is updated to a lower range. For example, for a device with an original H = 80, after detecting that D f substantially exceeds δ, H may drop to between 50 and 60, and this change process is recorded in the database.

[0121] The supplier retains the basic vibration data and the initial health value of each device in the database for factory acceptance and delivery. During the use and maintenance phase, when the hospital or a third-party maintenance center detects that the fractal dimension exceeds the threshold, the fractal health decreases accordingly and is written into the digital file of the device. The system can then determine whether the device should be immediately maintained or the usage frequency reduced. During the retirement or refurbishment phase, if the health has been in an overly low range for a long time, the management platform will include it in the retirement list or send it back to the factory for refurbishment in subsequent steps, completing the full life cycle tracking.

[0122] The fractal health can be controlled using a piecewise function. If the fractal dimension D f is within the safe range, H does not change significantly; if D f > δ, then H significantly decreases in one or more iterations. The following logic can be constructed:

[0123]

[0124] where α is the attenuation coefficient, which determines the degree of decline in health.

[0125] Each time a new H new is calculated, it is located and overwrites the previous record in the database using the unique device ID, so that each organization can obtain the latest value when querying.

[0126] The system calculates the fractal dimension of the device vibration in a regular sampling or event-triggered manner. If a continuous over-threshold situation occurs, it indicates that the device has been subjected to vibration shocks for a long time and its performance has significantly decayed, and the health level can directly enter the "severely worn" range (low value).

[0127] Threshold control is achieved by pre-setting the fractal dimension threshold δ at the supplier or hospital management end. Once it is detected that the device vibration exceeds δ, attenuation operations are performed at the health level. Numerical range representation: A linear or segmented range from 0 to 100 is easy for management personnel to understand, and they can see the general state of the device without delving into the details of the fractal algorithm.

[0128] If the management platform of the present invention finds that the health level of a certain device has dropped from 80 to 45, it can directly determine that its wear degree has increased significantly. Through the same health index, different devices such as electrocardiogram monitors, infusion pumps, hemodialysis machines, etc. can be incorporated into the same management system, facilitating cross-category scheduling and comparison by hospitals or maintenance companies. When the health level is low to a certain lower range, subsequent processes can automatically trigger maintenance or retirement evaluation to complete the closed-loop control of the entire life cycle.

[0129] Preferably, as Figure 3 shown, the demand information consists of the device usage plan, department transfer arrangements, and maintenance resource allocation parameters, and jointly generates multi-agent initial scenario data with the fractal health level and fault risk data for subsequent iteration of the quantum-inspired solution algorithm.

[0130] In the present invention, "demand information" mainly includes elements such as device usage plans, department transfer arrangements, and maintenance resource allocation parameters, and is one of the important inputs for the quantum-inspired solution algorithm to obtain multi-agent initial scenario data. By combining the fractal health level and fault risk data, a comprehensive multi-agent scenario can be constructed to simulate the scheduling, use, and maintenance process of medical devices in the whole hospital (or multiple hospital areas).

[0131] The device usage plan is usually given by the department or the operating room scheduling system, indicating the quantity and type of certain instruments required within a specific time period. Department transfer arrangements refer to the mechanism for departments to transfer or share devices, such as the temporary borrowing of a monitor by the emergency department. Maintenance resource allocation parameters include available time slots for maintenance personnel, capacity of equipment to be repaired, spare parts inventory, etc. When a device has a fault risk, the system needs to evaluate whether it can be repaired or replaced in a timely manner.

[0132] Suppliers usually have access to the basic information of the equipment, maintenance manuals, and factory parameters. This information can be combined with the hospital's usage plans and transfer processes to form a more comprehensive database system. In the management and tracking of the medical device life cycle, hospitals not only need to consider the fractal health of the equipment itself but also factors such as scheduling, transfer, and maintenance resources to make optimal allocation and maintenance decisions.

[0133] The hospital's surgical scheduling or department daily usage statistics system automatically generates a list of required instruments. For example, if a department has a planned cardiothoracic surgery in the next week, the system will package the specific dates, required instrument models, and quantities into demand information records.

[0134] When there is an equipment sharing or emergency call mechanism among multiple departments, it is necessary to record the priority and approval process for how departments borrow equipment from each other. Example: The Cardiovascular Medicine Department can give priority to temporarily borrowing a monitor from the General Internal Medicine Department during a sudden rescue or temporarily borrowing a replacement device when there is a risk of equipment failure.

[0135] How many spare workstations are available in the maintenance center, whether the spare parts inventory is sufficient, and how many devices a maintenance staff can repair in a certain week all affect whether high-risk equipment can be processed in a timely manner. If a certain consumable or core component is in short supply, even if there is a risk of equipment failure, it cannot be repaired immediately. The scheduling algorithm needs to consider preferentially allocating other relatively healthy devices to key departments.

[0136] Fractal health: A value obtained through vibration analysis (ranging from 0 to 100). If the health is too low, it means the equipment is severely worn. Fault risk data: Generated when the fractal dimension exceeds a specific threshold, indicating potential equipment failure risks. Combining the above two with demand information can create multiple "agents" in the initial scenario data of multi-agent systems:

[0137] Department agent: Its goal is to meet surgical or monitoring needs;

[0138] Maintenance agent: Hopes to preferentially repair equipment with low health and high fault risks under resource constraints;

[0139] Equipment agent: Carries its own fractal health and fault risk status, competing for or applying for maintenance.

[0140] When implementing the system, demand information, fractal health, and fault risk data can be encapsulated into one or more data tables and then combined through a program into an "initial scenario". The specific structure includes:

[0141] DeviceState: Identifies the device ID, fractal health H, and fault risk data R;

[0142] DemandState: The quantity and type of equipment required by each department at different time periods;

[0143] MaintenanceState: The available time periods of maintenance personnel and the inventory of spare parts.

[0144] When the system invokes the quantum-inspired solution algorithm (or other multi-agent scheduling modules), it will load the above data all at once and initiate the subsequent iterative process.

[0145] Multi-factor integration: Originally, hospital scheduling only considered "demand information" or judged availability solely based on "health status". This invention combines the two and adds a third dimension of "maintenance resources" to construct a more comprehensive initial scenario; Self-consistent evolution: In multi-agent interaction, the department agent submits a usage request, the maintenance agent repairs equipment with high failure risk based on the spare parts inventory, and the quantum-inspired algorithm outputs a "scheduling plan" after comprehensive consideration.

[0146] When the failure risk is high and the demand is urgent, repair or replacement can be prioritized; when maintenance resources are sufficient, more equipment with relatively low fractal health can be repaired. Department transfer information can help reduce the waste of duplicate equipment purchases or idle equipment and improve the utilization rate of overall medical resources. If the supplier database stores factory maintenance parameters and common failure modes, the system can incorporate these parameters into the multi-agent scenario to schedule maintenance or allocate spare parts more reasonably.

[0147] In one embodiment, for the batch usage plan and maintenance scheduling of monitors, both the cardiology department and the surgery department need a large number of monitors next week, but the maintenance center has limited repair slots. Therefore, the system writes the following into the multi-agent initial scenario data: the equipment quantity requirements of the two departments; the maintenance center can repair at most 3 monitors simultaneously; the fractal health and failure risk data of the monitors.

[0148] After the quantum-inspired solution algorithm iterates, some monitors with low health and high failure risk are prioritized for rapid repair to ensure they are repaired before the surgery peak to meet the urgent needs of the departments.

[0149] In another embodiment, for the cross-department transfer of hemodialysis machines and the constraints of spare parts inventory, when both the nephrology department and the general internal medicine department have hemodialysis requirements, the system states in the initial scenario data "the fractal health of hemodialysis machines, the scheduling requirements of the internal medicine department, and the spare parts surplus of the maintenance department". The multi-agent automatically determines that the internal medicine department can temporarily lend several hemodialysis machines with good health to the nephrology department. If any hemodialysis machine has a high-risk failure, it will be repaired using maintenance resources. This achieves the optimal allocation and timely repair of equipment.

[0150] Preferably, when the quantum-inspired solution algorithm performs iterative calculations on the discrete optimization model established for the multi-agent interaction process, in the scenario of resource conflict, medical devices with failure risk data exceeding the threshold are preferentially arranged to enter the maintenance or replacement process.

[0151] In the present invention, different departments, maintenance centers, and even the supplier side can be abstracted as "agents", which compete or cooperate with each other during the use, scheduling, and maintenance of medical devices. For example, the department agent needs the equipment to meet the surgical needs, the maintenance agent needs to repair high-failure-risk equipment, and the supplier has spare parts inventory or replacement equipment available for allocation. When resources conflict (for example, multiple departments compete for the same available device, or the capacity of the maintenance center is limited), the system requires a discrete optimization model to find the optimal or sub-optimal allocation plan.

[0152] Discrete optimization includes: transforming the decision variables of multiple agents (such as which department a device is allocated to, which devices need to be repaired or replaced immediately) into variables in Boolean or integer form, and establishing an objective function (such as minimizing the superposition of failure risks, maximizing the satisfaction rate of departmental requirements, etc.) and constraint conditions (such as the upper limit of maintenance resources).

[0153] The quantum-inspired solution algorithm uses mechanisms such as quantum annealing or similar to simulated annealing to perform multiple rounds of iteration on the discrete optimization model. Each round of iteration is looking for a feasible solution for resource allocation, trying to balance the requirements of each department and maintenance resources, and finally forming a "device scheduling plan". During this process, if a medical device with failure risk data exceeding the threshold is in a contention state (resource conflict), then the present invention preferentially incorporates it into the repair or replacement process to avoid potential risks caused by its continued use.

[0154] Failure risk data: Obtained through the previous fractal dimension calculation or other diagnostic means. If it exceeds the pre-set threshold (RiskThreshold), it means that the device has a high potential failure rate or has suffered serious wear.

[0155] Supplier database: At the time of device factory shipment, there are basic maintenance indicators, lists of replaceable parts, etc. When the hospital side writes the latest failure risk data of the device into the database, the supplier can also know whether it is necessary to supplement spare parts or dispatch professional maintenance personnel.

[0156] Common situations of resource conflict include:

[0157] Multi-department contention: When the quantity of a certain type of medical device required by two departments is greater than the available inventory, the system determines it as a resource conflict. Insufficient capacity of the maintenance center: If the maintenance center can only repair k devices in the short term, but the actual high-risk devices have reached m (m > k).

[0158] Limited supplier backup equipment: Conflicts can also occur if a hospital needs to immediately replace multiple faulty devices, but the supplier can only provide a small amount of replacement inventory in the short term.

[0159] When the quantum-inspired solution algorithm performs iterative calculations on the above conflicts, the core principle of the present invention is to prioritize high risks. If the risk data RiskData of a certain device exceeds RiskThreshold, then a weight of "must be repaired or replaced first" is assigned to this device in the objective function or constraints of the model, prompting the algorithm to tend to quickly remove this device from the usage queue under resource conflicts and arrange for repair or apply for a backup replacement machine.

[0160] In practical applications, in the present invention, the quantum-inspired solution algorithm can use a type of QUBO (Quadratic Unconstrained Binary Optimization) or Ising model. The following only shows the key points. This content is publicly available technology, so all formulas do not need to be written here:

[0161] Boolean variable x i,j , indicating "whether device i is assigned to department j".

[0162] Boolean variable y i , indicating "whether device i is sent to the maintenance process".

[0163] Boolean variable z i , indicating "whether device i applies for a replacement device".

[0164] The objective function includes:

[0165] Minimizing the failure risk: If RiskData(i) is too high, then force y i = 1 or z i = 1.

[0166] Maximizing the satisfaction of department requirements: Expect x i,j to be able to cover the requirements of department j.

[0167] Resource usage constraints: The number of repairs that the maintenance center can handle in a day cannot exceed M, and the total amount of replacement machines that the supplier can provide is limited, etc.

[0168] Make decisions on "which devices are used by whom and which high-risk devices are sent for repair" randomly or based on a greedy strategy to form a preliminary solution. In each round of annealing or iteration, the algorithm randomly flips some variables and evaluates the change in the objective function. If it can reduce the overall conflict or improve the demand satisfaction, the change is retained. When the algorithm detects a shortage of maintenance resources and there are multiple high-risk devices, and it is necessary to select which ones to repair first, devices with a fault risk RiskData exceeding the threshold are given higher priority. Form a device scheduling plan, including information such as "which devices continue to be used", "which are sent for maintenance and replacement", and "which departments' demands cannot be met temporarily", for the hospital to execute.

[0169] In one embodiment, in a multi-department conflict scenario, the monitor gives priority to repairing high-fault-risk devices. Both the ICU and the Cardiac Surgery Department need additional monitors, and the current upper limit of devices that the maintenance center can repair is 2. The hospital detects that 3 monitors have fault risk data exceeding the threshold.

[0170] During quantum-inspired iteration, it is found that the ICU must be given priority, and the demand of the Cardiac Surgery Department is equally urgent, but the maintenance center can only repair 2 immediately. Therefore, the 2 with the "highest fault risk" are arranged to be sent for repair first, and when available, they will be used by the ICU first. The 3rd one also has a relatively high fault risk, so a replacement machine from the supplier is directly applied for. If there is 1 replacement machine in stock, it will be allocated immediately; if not, wait for the next round of release of maintenance resources.

[0171] The intensive care needs of patients are optimally met, high-risk devices are quickly disposed of or replaced, and the continued use of potential faulty products is avoided.

[0172] In another embodiment, there is a conflict in the replacement resources of hemodialysis machines. Suddenly, the fractal health of multiple hemodialysis machines in the hemodialysis center has declined and triggered a high fault risk, and there are only 2 replacement machines that the supplier can allocate.

[0173] During the quantum-inspired process, it is found that 5 hemodialysis machines all have risk values exceeding the threshold, but only 2 can be replaced, and the other 3 may queue for repair. The system calculates and determines that the replacement is given priority to the 2 most severely worn and with the most urgent usage plan (such as the need for overtime dialysis at night), and the remaining devices are gradually repaired by the maintenance center.

[0174] The quantum-inspired solution ensures that critical patients get normally operating hemodialysis equipment, and after the management accepts the plan, the scheduling execution instruction can be issued.

[0175] In quantum-inspired iteration, a higher penalty or weight is set for devices with fault risk data exceeding the threshold, so that when allocating conflicts, such devices are preferentially arranged out of the usage queue and enter the repair or replacement process. Multiple agents correspond to different objectives or restrictions such as departments, maintenance, and suppliers. The quantum-inspired algorithm continuously tries to locally flip the solution in high-dimensional search, thus approaching the globally optimal solution.

[0176] Prioritizing the processing of high-risk equipment can prevent serious failures from occurring during clinical use; if the hospital's maintenance capacity is insufficient, it can automatically request the supplier to replace the inventory or dispatch more engineers to shorten the downtime due to failures; in conflict scenarios, the system can dynamically allocate limited resources to maximize equipment availability and meet demand.

[0177] Preferably, the holographic virtual reality environment is used to display the distribution locations of the equipment scheduling plan among hospital departments, and allows managers to generate scheduling instructions after making interactive modifications based on the visualization results, and write the scheduling instructions to the digital twin for subsequent execution.

[0178] In the present invention, the quantum-inspired solution algorithm outputs an equipment scheduling plan, that is, under given conditions such as departmental requirements, equipment health status, and maintenance resources, it determines which department or maintenance process each piece of equipment should be assigned to. The holographic virtual reality environment can present this allocation information to managers in a three-dimensional and interactive manner, giving them a visual understanding of the equipment allocation situation in the entire hospital.

[0179] Instead of only viewing the scheduling results in text or flat charts, information such as "3 pieces of equipment are currently allocated to the ICU department" and "2 pieces of equipment in the cardiology department are at risk of failure and are ready for repair" can be viewed in an interactive holographic scenario.

[0180] The supplier database stores the basic information of each piece of equipment, such as factory attributes, replaceable parts, warranty terms, etc. The holographic virtual reality environment can call this information when presenting the scheduling plan (for example, when a certain piece of equipment is severely worn, one can view at a glance whether the supplier has a standby machine or a warranty channel). Each piece of equipment has a digital twin record between the hospital information system and the supplier database, including health status, failure risk, location information, etc. When managers modify the scheduling plan in the holographic environment, the new scheduling instructions need to be updated to the digital twin data of the equipment for subsequent execution and tracking.

[0181] In the holographic environment, managers can see the layout of a hospital (or multiple hospital campuses), and each department is marked in three-dimensional space. According to the scheduling plan, the system places the allocation locations of each piece of equipment (such as the ICU, operating room, maintenance center) in the corresponding areas in the form of icons or cursors for easy visual identification.

[0182] Display information such as the unique identifier of the equipment, fractal health status, and whether it has a failure risk. For example, "safe equipment" and "high-risk equipment" can be distinguished by color. If the equipment is in the maintenance or replacement process, it can be displayed with a specific mark in the holographic environment.

[0183] Managers can select device icons in the holographic interface, drag and drop them into a certain department space, or click the "Maintenance" button to send them to the maintenance center. If they are not satisfied with the solution given by the quantum-inspired solution algorithm (for example, if they want to prioritize ensuring the number of devices in the emergency room), they can manually override the solution and reassign the target department for the device.

[0184] Once the interaction is completed, the system will form one or more new scheduling instructions, describing which devices are to be moved where or how to handle the failure risk. Through the digital twin interface, these scheduling instructions (including the target department, operation timestamp, execution priority) are written into the digital twin records of the corresponding devices, enabling the backend system or logistics link to perform actual operations (such as robotic transportation, manual allocation) accordingly.

[0185] After receiving the scheduling instructions, hospital warehouse or department staff complete the device movement or send it for maintenance inspection, and update the "execution result" in the digital twin. If it is necessary to transfer spare machines from the supplier or apply for specific components, "Request supplier support" can also be marked in the holographic environment, and the system will automatically generate a request entry and record it in the digital twin or the supplier database.

[0186] In practical applications, project the multi-dimensional scheduling results onto the holographic interface to reduce the understanding burden brought by text or table methods; the quantum-inspired solution algorithm gives an initial solution, and the holographic environment allows managers to make "final fine-tuning", while automatically writing the modified results back to the digital twin to form a closed loop.

[0187] Managers can clearly see the current device allocation plan at a glance, especially in the case of a large number of devices, where it is difficult to intuitively feel with traditional reports. Adjust the allocation through VR / AR control pens or gesture operations to avoid cumbersome operations or secondary input; at the same time, ensure that the system automatically records and synchronizes the execution after the change. If a certain device is of high risk and there are no spare maintenance resources in the hospital, it can be directly switched to the "Supplier replacement" state in the holographic scene, and the system will request the supplier database to retrieve available replacement machines and automatically generate shipping instructions.

[0188] In one embodiment, the ICU, cardiac surgery department, and internal medicine department altogether need a number of monitors. After the quantum-inspired algorithm allocates the devices to each department, the number of monitors owned by each department and their health status are displayed on the holographic screen in a three-dimensional floor plan. The manager discovers that the number of cardiac surgery schedules has increased and one more device is needed. Directly select a monitor with a health status of 80 in the holographic interface, "drag and drop" it from the internal medicine department to the cardiac surgery area, and submit the change. The system automatically generates a new "device scheduling instruction", and the digital twin updates the target department of this monitor to "cardiac surgery department". The warehouse robot or manual handling then executes the allocation.

[0189] In another embodiment, there is a risk warning for 3 hemodialysis machines in the hemodialysis center. The quantum-inspired algorithm recommends sending 2 of them to the maintenance center. However, the management believes that 1 of them is more urgent than expected, so they manually adjust and mark "this one also needs immediate repair" in the holographic environment. The scheduling instruction is written into the digital twin and reported to the maintenance department or the supplier to provide a replacement machine. The real-time visualization is updated to show that the status of the maintenance center becomes "number of machines in queue for repair = 3", and the supplier receives a "request for replacement" message for standby machine scheduling.

[0190] Detect the result of executing the scheduling instruction to identify abnormal event data, and mark the medical device that meets the retirement conditions as retired in the digital twin, so as to complete the closed-loop management of the entire life cycle of the medical device.

[0191] In the present invention, when the scheduling instruction (for example, a certain device is assigned to a certain department or sent to the maintenance center) is actually executed, the system will detect the execution process and result, compare with the expected goal, and judge whether there is abnormal event data. The so-called "abnormal events" include:

[0192] Scheduling failure: The device is lost during transportation or not delivered to the destination on time; Abnormal use: The department reports that the device cannot work properly or a fault alarm occurs; Maintenance exception: Maintenance timeout, lack of key spare parts resulting in the device being unable to be repaired; Life cycle termination condition: Multiple maintenance attempts are ineffective or the device health cannot be restored, meeting the retirement conditions.

[0193] Once it is determined that the abnormal event belongs to the situation of "unable to continue using" or "exceeding the retirement threshold", etc., the system will add a "retirement mark" to the device in the digital twin to complete the closed-loop management of the entire life cycle, that is, all stages from "normal use to final retirement" are recorded and traceable.

[0194] The supplier database can store the complete records of each device from factory to retirement, including factory parameters, fractal health evolution, fault and maintenance history, etc. When the hospital decides to retire a device, this status will be updated in the database to ensure that the supplier understands the reason for the device's decommissioning, whether it can be recycled and refurbished, etc.

[0195] The digital twin is mainly deployed on the hospital side, but it is necessary to synchronize with the supplier whether the device has been officially taken out of use, so that the supplier can provide recycling or remanufacturing services later, or allocate new devices from the inventory for replacement.

[0196] For each "device transfer" or "maintenance instruction", the system will record the planned completion time, the person in charge, and the execution path (such as the logistics route). If it is not completed within the time limit or the department reports that the device cannot work, "abnormal event data" will be automatically generated, marking information such as "execution failed" or "faulty use".

[0197] When the device is actually operating in the department, if the fractal health continuously decreases or the failure risk is triggered multiple times, it can also be recognized as an abnormal event, such as "still malfunctioning after maintenance" or "too high failure frequency". These events are also written into the digital twin to provide a basis for subsequent retirement decisions.

[0198] Examples of retirement conditions: When the cumulative number of times the device has a failure risk ≥ N times and maintenance fails to restore the health to a certain value, it is considered to meet the retirement conditions; or when the service life exceeds the established threshold and the health cannot be restored to the safe range through maintenance. A "retirement determination logic table" can be stored in the digital twin, and a joint judgment is made based on the health, maintenance records, failure risk data, etc.

[0199] Once it is confirmed that the device no longer has safety and economic feasibility, the system makes a terminating modification to the status field of the device in the digital twin (for example, updates "status" from "in-use" to "retired"). At the same time, a "retirement completion timestamp" can be generated, and the reason for retirement is noted (such as "multiple failures cannot be repaired").

[0200] Subsequent recycling or scrapping processing: After the retirement mark is generated, if the supplier supports the recycling and refurbishment process, the supplier database automatically shows that the device is in the "waiting for recycling" state; if there is no such option, a formal scrapping record is made. The hospital can view the statistics of retired devices in the life cycle management system, compare the evolution of the fractal health and the maintenance cost; the supplier can also audit which devices are mainly retired due to component aging or design defects, which is convenient for improving the design.

[0201] Through continuous monitoring of scheduling execution and device usage, once it is found that the device status does not match the expectation, abnormal event data is generated; if the abnormal reason cannot be eliminated or the health has dropped extremely, the system retires the device based on a set of preset rules and ends its information tracking in the hospital.

[0202] In the present invention, from factory - use - maintenance - transfer - final retirement, it is all connected in series by the same digital twin and database; abnormal events automatically trigger retirement or maintenance reminders to prevent the device from being used continuously in a high-risk state; after retirement, the digital twin still retains historical records, which is beneficial for the supplier or the hospital to conduct statistical analysis on the failure reasons, service life, etc.

[0203] Preferably, meeting the retirement conditions means that the fractal health continuously falls below the preset fractal health threshold and the failure risk data cannot be restored to the usable state after being triggered for maintenance multiple times. When identifying abnormal event data, a retirement mark is executed on the medical device, and the retirement information is registered in the digital twin to form a closed-loop record.

[0204] In the present invention, the usage status of medical devices is mainly determined by two major indicators: "fractal health degree" and "fault risk data". If the fractal health degree of a certain device is continuously lower than the set threshold for a long time, and the fault risk is triggered multiple times (and the health degree still cannot be restored to the safe range after maintenance), it is determined that the device "cannot be restored to the available state", which means it enters an irreversible aging or fault stage.

[0205] The fractal health degree continuously being lower than the threshold: This indicates that the vibration characteristics or wear degree of the device has been in an abnormal range for a long time; the fault risk is triggered multiple times and the maintenance fails: This represents that the system has given the device a maintenance opportunity, but still cannot improve the fractal health degree to the minimum available standard. When the above conditions are met, the system will regard the device as a "retired device" and will no longer continue to be put into clinical or daily use to avoid potential safety hazards or additional costs.

[0206] Abnormal event data. When the device repeatedly experiences high-risk faults or the maintenance center records that its "health degree is still low after repair", abnormal event data will be formed; retirement mark. Update the status of the device in the digital twin (such as changing from "in use" to "retired") and record the reason for retirement (multiple faults, unable to repair, etc.). This process in the closed-loop management of the present invention helps hospitals and suppliers accurately locate the end point of the device's life cycle, and thus facilitates subsequent recycling or scrapping.

[0207] In practical applications, one or more thresholds can be set. For example, θ represents the minimum available standard (such as 60 points). If the device health degree H is continuously lower than θ for more than a specified duration or number of maintenance times, it is determined that the device "cannot maintain the available state". The system regularly (such as after each device use or daily data summary) calculates the fractal health degree and compares it with θ; if there is no improvement after consecutive days or multiple maintenance operations, it meets the condition of "continuously lower than the fractal health degree threshold".

[0208] When the vibration characteristics of the device exceed the threshold, the system outputs the fault risk data R and arranges maintenance; if R is still detected to be over the limit or the fractal health degree cannot be raised back to θ after maintenance, it is regarded as a maintenance failure; repeated similar results indicate that the device is difficult to recover. Example: The device triggers the fault risk three times, and each time it is sent for maintenance but the health degree is lower than θ. The system determines that it has lost its continuous use value.

[0209] When the system determines that the device meets the retirement conditions, it automatically updates the "status" field of the device in the digital twin to "retired", and at the same time records the retirement time "t_retire" and the reason "maintenance fail after N attempts", etc. After the recording is completed, the hospital system or the supplier database can uniformly regard it as "the end of the device's life cycle" and remove it from the allocation list; if necessary, it can be included in the disassembly or recycling plan.

[0210] The present invention does not simply rely on one failure or one low health to eliminate equipment, but uses the dual standard of "persistently low health + multiple failure risk triggers and ineffective maintenance" to avoid premature retirement of repairable equipment. Each maintenance failure or failure in use is an abnormal event data. After accumulation, the system determines that the equipment is indeed irreversible and then executes the retirement mark to ensure that the process is reasonable and well-founded.

[0211] The present invention does not allow high-failure risk equipment to continue to be used, thereby reducing the probability of medical accidents; if the equipment recovers well after one maintenance, it can be used again, otherwise try again multiple times; if it is always ineffective, it will be retired decisively to avoid wasting maintenance costs; suppliers can learn the reasons for retirement in the database and conduct subsequent recycling or technical improvements.

[0212] like Figure 4 As shown, a medical device lifecycle management and tracking system is used to implement the medical device lifecycle management and tracking method, and the system includes:

[0213] The identification and fractal analysis module is used to obtain the unique identification data of the medical device and collect the original vibration signal data, generate fractal dimension data after cleaning the original vibration signal data, and output the fault risk data according to the fractal dimension data and the preset fractal health threshold; the identification and fractal analysis module mainly includes RFID / scanning equipment for reading the unique identification of the device, and cooperates with the micro-vibration sensor patch or accelerometer to obtain the vibration signal. The local or back-end server can perform filtering, peak removal and fractal algorithms, process the vibration waveform and output the fault risk data.

[0214] The quantum scheduling holographic interaction module is used to integrate the fractal health and fault risk data with the demand information to construct the multi-agent initial scenario data, iteratively generate the equipment scheduling plan through the quantum-inspired solution algorithm, and form the scheduling instructions after performing interactive operations on the equipment scheduling plan in the holographic virtual reality environment; the quantum scheduling holographic interaction module uses high-performance servers or quantum annealing hardware to complete the multi-agent discrete optimization operation. Through the holographic / VR display device and gesture detection device, the management personnel can visualize and modify the scheduling plan, and write the final scheduling instructions back to the system.

[0215] The abnormal detection and retirement module is used to detect the results of executing the scheduling instructions to identify abnormal event data, and to mark the corresponding medical devices for retirement in the digital twin when the retirement conditions are met, thereby completing the closed-loop management of the medical devices throughout their life cycle; the abnormal detection and retirement module uses logistics robots, access control scanning devices, etc. to monitor the scheduling execution process to determine whether the equipment is in place or undergoing maintenance. Equipment that cannot be repaired is marked for retirement in the digital twin server to form a closed-loop record and complete the full life cycle management.

[0216] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects.

[0217] The above are only the embodiments of the present application and are not intended to limit the present application. For those skilled in the art, various changes and modifications can be made to the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included within the scope of the claims of the present application.

Claims

1. A method for medical device life cycle management and tracking, characterized in that, It includes the following steps: Obtain the unique identification data of the medical device, collect the original vibration signal data, clean the original vibration signal data to generate fractal dimension data, and output fault risk data according to the fractal dimension data and a preset fractal health threshold; Integrate the fractal health, fault risk data, and demand information to generate multi-agent initial scenario data, iteratively generate a device scheduling plan through a quantum-inspired solution algorithm, and perform interactive operations on the device scheduling plan in a holographic virtual reality environment to form a scheduling instruction; Detect the result of executing the scheduling instruction to identify abnormal event data, and mark the retirement of the medical device that meets the retirement conditions in the digital twin, thereby completing the closed-loop management of the entire life cycle of the medical device.

2. The method for medical device life cycle management and tracking according to claim 1, wherein, When cleaning the original vibration signal data, compare the vibration waveform sequence with a fixed frequency domain range to filter out noise signals, and eliminate the peaks of numerical mutations in continuous sampling points to generate waveform data that meets the requirements for fractal dimension calculation.

3. The method for medical device life cycle management and tracking according to claim 1, wherein The fractal dimension data is obtained by applying a fractal analysis algorithm to the multi-dimensional vibration trajectory of phase space reconstruction, and represents the complexity of the vibration mode of the medical device through a comprehensive measure of the evolution of vibration amplitude and phase, which is used for subsequent fractal health calculation.

4. The method for medical device life cycle management and tracking according to claim 1, wherein The fractal health characterizes the usage state of the medical device in a numerical range from zero to one hundred. When the fractal dimension data exceeds a preset threshold, the fractal health value is reduced to an interval indicating greater wear to reflect the performance degradation caused by long-term vibration impact on the device.

5. The method for medical device life cycle management and tracking according to claim 1, wherein, When the fractal health drops below a preset fractal health threshold, output the fault risk data and associate the fault risk data with the digital twin of the corresponding medical device, thereby recording the risk state of the medical device malfunctioning in the digital twin.

6. The method for managing and tracking the life cycle of a medical device according to claim 1, wherein The demand information consists of the device usage plan, department transfer arrangements, and maintenance resource allocation parameters, and jointly generates multi-agent initial scenario data with the fractal health and fault risk data for subsequent iteration of the quantum-inspired solution algorithm.

7. The method for managing and tracking the life cycle of a medical device according to claim 1, wherein When the quantum-inspired solution algorithm iteratively calculates the discrete optimization model established for the multi-agent interaction process, it preferentially arranges the medical devices with fault risk data exceeding the threshold to enter the maintenance or replacement process for scenarios with resource conflicts.

8. The method for medical device life cycle management and tracking according to claim 1, characterized in that, The holographic virtual reality environment is used to display the distribution location of the device scheduling plan among hospital departments, allows managers to generate a scheduling instruction after making interactive modifications based on the visualization results, and writes the scheduling instruction to the digital twin for subsequent execution.

9. The method for medical device life cycle management and tracking according to claim 1, characterized in that, Meeting the retirement conditions means that the fractal health is continuously lower than the preset fractal health threshold and the fault risk data cannot be restored to a usable state after being triggered for maintenance multiple times. When identifying abnormal event data, perform a retirement mark on the medical device and register the retirement information in the digital twin to form a closed-loop record.

10. A medical device lifecycle management and tracking system for implementing the medical device lifecycle management and tracking method according to any one of claims 1 to 9, characterized in that, The system includes: An identification and fractal analysis module, which is used to obtain the unique identification data of medical devices and collect the original vibration signal data, clean the original vibration signal data to generate fractal dimension data, and output fault risk data according to the fractal dimension data and a preset fractal health threshold; A quantum scheduling holographic interaction module, which is used to integrate the fractal health and fault risk data with demand information to construct multi-agent initial scenario data, iteratively generate a device scheduling scheme through a quantum-inspired solution algorithm, and perform interactive operations on the device scheduling scheme in a holographic virtual reality environment to form a scheduling instruction; An anomaly detection and retirement module, which is used to detect the result of executing the scheduling instruction to identify anomaly event data, and mark the corresponding medical device for retirement in the digital twin when the retirement condition is met, so as to complete the closed-loop management of the entire life cycle of medical devices.

Images