Equipment fault early warning detection method and system for medical equipment department

Abstract

The invention relates to an equipment fault early warning detection method and system for a medical equipment department. The method comprises the following steps: acquiring operation state data and maintenance time data of each medical device; calculating a first fault probability of each medical device based on the fault period judgment rule and the maintenance time data; selecting the medical equipment with the first fault probability greater than a first fault probability threshold to obtain a first monitoring equipment set, and selecting the medical equipment with the first fault probability less than the first fault probability threshold to obtain a first to-be-monitored equipment set; quantifying the operation condition of the medical equipment based on the operation state data to obtain change trend data, and selecting the medical equipment of which the change trend data is greater than a preset trend threshold to obtain a second monitoring equipment set; and based on the first monitoring equipment set and the second monitoring equipment set, the fault probability of the medical equipment is calculated according to the operation state data, and fault early warning information is generated. By adopting the method, the pertinence and efficiency of equipment fault early warning can be improved.

Description

Technical Field

[0001] This invention belongs to the field of equipment fault diagnosis, and in particular relates to a method and system for early warning and detection of equipment faults in the medical equipment sector. Background Technology

[0002] With the development of equipment fault diagnosis technology, equipment status monitoring technology based on real-time acquisition of multi-source sensor data has emerged. This technology features continuous data, rich dimensions, and rapid response, which in turn promotes the popularization of predictive maintenance for medical equipment as a new management approach.

[0003] Existing equipment condition monitoring technologies based on real-time acquisition of multi-source sensor data primarily employ a combination of real-time monitoring with fixed thresholds and time-based preventative maintenance (TBM). Specifically, the monitoring system continuously collects equipment operating parameters (such as current, temperature, and noise) through sensors and sets a static safety threshold for each parameter. An alarm is triggered when a parameter exceeds the threshold. Simultaneously, regardless of the actual operating status of the equipment, comprehensive fault detection calculations are performed according to a preset time cycle. A unified diagnostic model is applied to historical data from all equipment to generate maintenance recommendations. This approach largely relies on manual experience to set thresholds and maintenance cycles, aiming to prevent sudden equipment failures through regular inspections and calculations.

[0004] However, the current method of using fixed thresholds and periodic calculations has obvious limitations. For example, static thresholds cannot adapt to the dynamic changes of individual equipment and different stages of use, which can easily lead to false alarms or missed alarms. Furthermore, periodically performing full fault calculations on all equipment consumes a lot of computing resources, with many stable equipment being analyzed repeatedly, resulting in serious computational redundancy. More importantly, existing technologies fail to fully utilize the periodic patterns contained in historical equipment faults and the abrupt changes in real-time data, making it impossible to achieve high-efficiency predictive maintenance while ensuring computational efficiency, resulting in insufficient targeting and efficiency of equipment fault warnings. Summary of the Invention

[0005] Therefore, it is necessary to provide a medical equipment fault early warning detection method and system that can ensure computational efficiency while achieving high-performance predictive maintenance, thereby improving the pertinence and efficiency of equipment fault early warning.

[0006] Firstly, this application provides a method for early warning detection of equipment failure in the medical equipment department, including:

[0007] Acquire operational status data and maintenance time data for each medical device; operational status data includes current data, voltage data, temperature data, and noise sound pressure level data;

[0008] Based on the preset fault cycle judgment rules and maintenance time data, the probability of each medical device being in a fault cycle is calculated to obtain the first fault probability; medical devices with the first fault probability greater than the preset first fault probability threshold are selected to obtain the first monitoring device set, and medical devices with the first fault probability less than the first fault probability threshold are selected to obtain the first monitoring device set.

[0009] For each medical device in the first set of monitoring devices, the operating status of the medical device is quantified based on the operating status data to obtain the trend data of each medical device; medical devices whose trend data is greater than the preset trend threshold are selected to obtain the second set of monitoring devices;

[0010] Based on the first and second monitoring device sets, the failure probability of medical equipment is calculated according to the operating status data, and failure early warning information is generated.

[0011] Furthermore, the fault cycle determination rule is obtained through the following method:

[0012] Obtain historical maintenance record data; historical maintenance record data includes equipment model, fault type, and fault occurrence time data for each medical device;

[0013] For each equipment model, the fault interval time for each fault type is calculated based on the fault occurrence time data, thus obtaining the fault interval time set corresponding to each equipment model.

[0014] For each equipment model, the failure rate variation trend and average failure time of each failure type are solved by constructing a likelihood function, thereby obtaining the failure rate variation characteristic data and failure occurrence time characteristic data corresponding to each equipment model.

[0015] For each equipment model, based on the failure rate variation characteristics and failure occurrence time characteristics, a failure cycle probability function for each failure type of each medical device is constructed to obtain the failure cycle judgment rule.

[0016] Furthermore, based on preset fault cycle judgment rules and maintenance time data, the probability of each medical device being in a fault cycle is calculated to obtain the first fault probability, including:

[0017] Based on the maintenance time data of each medical device, the continuous operating time of each medical device is determined, and the testing time data of each medical device is obtained.

[0018] Based on the testing time data and fault cycle judgment rules of each medical device, the probability of each medical device being in the fault cycle of each fault type is calculated, and the first fault probability of each medical device is obtained.

[0019] Furthermore, the first fault probability threshold is obtained by the following method:

[0020] Based on the equipment model and failure time data of each medical device, the historical failure rate of each equipment model of medical device is calculated;

[0021] For each medical device model, based on historical failure rates, the first failure probability threshold for each model is calculated using the following formula:

[0022]

[0023] in, It is the first The first failure probability threshold for each device model It is the basic threshold. It is an adjustment factor. It is the first Historical failure rate of each equipment model.

[0024] Furthermore, for each medical device in the first monitoring device group, the operating status of the medical device is quantified based on the operating status data to obtain the trend data of each medical device, including:

[0025] For each medical device in the first monitoring device group, the working status of the medical device is determined based on the operating status data, and the corresponding working status label of each medical device is obtained.

[0026] For each medical device in the first monitoring group, the operating status data is standardized based on the working status label to obtain standardized operating status data for each medical device;

[0027] The standardized operating status data is compared with the preset mutation threshold to obtain the mutation comparison result; medical devices whose standardized operating status data is greater than the mutation threshold are selected from the first set of devices to be monitored to obtain the first set of devices to be determined for monitoring.

[0028] For each medical device in the first monitoring device group, the slope of the standardized operating status data within the time window parameter is calculated based on the preset time window parameter to obtain the change trend data of each medical device.

[0029] Furthermore, for each medical device in the first monitoring device group, the working status of the medical device is determined based on the operating status data, and the corresponding working status label for each medical device is obtained, including:

[0030] For each medical device in the first monitoring device group, the Euclidean distance between the operating status data and the central data of each working state is calculated to obtain the Euclidean distance result corresponding to each working state; the working state corresponding to the smallest Euclidean distance result is determined as the working state label of the medical device.

[0031] The central data for each working state was obtained through the following method:

[0032] Acquire a preset amount of historical work record data; the historical work record data includes current data, voltage data, temperature data, and noise sound pressure level data of each medical device in its historical normal working state;

[0033] The K-Means++ clustering algorithm is used to select cluster center data from each historical work record data to obtain the initial cluster center data group;

[0034] For the remaining historical work record data, calculate the Euclidean distance with each cluster center data in the initial cluster center data group to obtain the minimum Euclidean distance result; update the cluster center data corresponding to the minimum Euclidean distance result using the historical work record data to obtain the complete cluster center data group;

[0035] According to the preset work status classification rules, the cluster center data in the complete cluster center data group are matched with each work status to obtain the center data of each work status.

[0036] Furthermore, based on the first and second monitoring device sets, the failure probability of the medical equipment is calculated according to the operating status data, and fault early warning information is generated, including:

[0037] For each medical device in the first and second monitoring device sets, the operating status data is input into the trained fault early warning autoencoder model to obtain the reconstructed operating data corresponding to each medical device.

[0038] Based on the mean square error calculation formula, the difference between the reconstruction operation data and the operation status data is calculated to obtain the reconstruction error data corresponding to each medical device.

[0039] An exponential decay function is used to map the reconstructed error data to obtain the fault likelihood of each medical device.

[0040] For each medical device in the first monitoring device set, the first failure probability and the failure likelihood are fused based on Bayes' theorem to obtain the posterior failure probability.

[0041] Based on the first failure probability and posterior failure probability of each medical device in the first monitoring device set, and the change trend data and failure likelihood of each medical device in the second monitoring device set, a fault warning information is generated.

[0042] Secondly, this application also provides a medical equipment fault early warning and detection system, including:

[0043] The data acquisition module is used to acquire the operating status data and maintenance time data of each medical device; the operating status data includes current data, voltage data, temperature data and noise sound pressure level data;

[0044] The fault cycle matching module is used to calculate the probability that each medical device is in a fault cycle based on preset fault cycle judgment rules and maintenance time data, and obtain a first fault probability; select medical devices with a first fault probability greater than a preset first fault probability threshold to obtain a first monitoring device set, and select medical devices with a first fault probability less than the first fault probability threshold to obtain a first monitoring device set.

[0045] The data mutation matching module is used to quantify the operating status of each medical device in the first set of monitored devices based on the operating status data, and obtain the change trend data of each medical device; select the medical devices whose change trend data is greater than the preset trend threshold to obtain the second set of monitored devices;

[0046] The early warning information generation module is used to calculate the failure probability of medical equipment based on the first monitoring device set and the second monitoring device set, and generate failure early warning information according to the operating status data.

[0047] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement any of the medical equipment fault early warning detection methods described in the first aspect of this application.

[0048] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements any of the medical equipment fault early warning detection methods described in the first aspect of this application.

[0049] The aforementioned medical equipment fault early warning detection method and system acquires the operating status data and maintenance time data of each medical device. The operating status data includes current data, voltage data, temperature data, and noise sound pressure level data. Based on preset fault cycle judgment rules and maintenance time data, the probability of each medical device being in a fault cycle is calculated to obtain a first fault probability. Medical devices with a first fault probability greater than a preset first fault probability threshold are selected to obtain a first monitoring device set, and medical devices with a first fault probability less than the first fault probability threshold are selected to obtain a first set of devices to be monitored. For each medical device in the first set of devices to be monitored, the operating status of the medical device is quantified based on the operating status data to obtain the change trend data of each medical device. Medical devices with change trend data greater than a preset trend threshold are selected to obtain a second monitoring device set. Based on the first and second monitoring device sets, the fault probability of the medical device is calculated according to the operating status data, and fault early warning information is generated. This method combines the fault cycle of the medical device and the sudden changes in the operating data, achieving high-efficiency predictive maintenance while ensuring computational efficiency, and improving the pertinence and efficiency of equipment fault early warning. Attached Figure Description

[0050] To more clearly illustrate the technical solutions in the embodiments or related technologies of this application, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0051] Figure 1 A flowchart illustrating a method for early warning and detection of equipment failure in a medical device department, provided as an embodiment of this application;

[0052] Figure 2 A schematic diagram of the structure of a medical equipment fault early warning and detection system provided in one embodiment of this application;

[0053] Figure 3 This is a schematic diagram of the structure of a computer device for a medical equipment fault early warning detection method provided in one embodiment of this application. Detailed Implementation

[0054] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0055] In one embodiment, such as Figure 1As shown, a method for early warning and detection of equipment failure in medical equipment is provided. This embodiment illustrates the application of this method to a terminal. It is understood that this method can also be applied to a server, and can also be applied to a system including a terminal and a server, and implemented through the interaction between the terminal and the server. In this embodiment, the method includes the following S101-S104, wherein:

[0056] S101, acquire the operating status data and maintenance time data of each medical device; the operating status data includes current data, voltage data, temperature data and noise sound pressure level data.

[0057] Specifically, the terminal continuously and synchronously collects operational status data from each medical device using sensors deployed on the devices. This operational status data includes continuously varying current, voltage, temperature, and noise sound pressure level data over time. The terminal also acquires maintenance time data for each medical device. Illustratively, the maintenance time data is used to characterize the time stamp of the last maintenance performed on the medical device.

[0058] S102, based on the preset fault cycle judgment rules and maintenance time data, calculate the probability that each medical device is in a fault cycle to obtain a first fault probability; select medical devices with a first fault probability greater than a preset first fault probability threshold to obtain a first monitoring device set, and select medical devices with a first fault probability less than the first fault probability threshold to obtain a first monitoring device set.

[0059] Specifically, the terminal calculates the probability of each medical device being in a fault cycle of different fault types based on the maintenance time data of each device according to preset fault cycle judgment rules, obtaining a first fault probability. Medical devices with a first fault probability greater than a preset first fault probability threshold are then grouped together to obtain a first monitoring device set; conversely, medical devices with a first fault probability less than the first fault probability threshold are grouped together to obtain a first set of devices to be monitored. The preset fault cycle judgment rules are a set of mapping relationships between the continuous operating time and fault cycle probability of each medical device. These rules are used to map the maintenance time data of each medical device to the probability of each device being in a fault cycle of different fault types. They can be generated by mining periodic patterns of equipment failures from historical maintenance data using statistical learning methods. The preset first fault probability threshold represents the threshold for further fault analysis of the medical devices. When the first fault probability of a medical device is greater than its first fault probability threshold, that medical device is grouped into the first monitoring device set for further analysis. Different medical devices have different first fault probability thresholds, and these thresholds are related to the overall failure rate of each medical device. The threshold can be set according to the accuracy of fault cycle probability calculations in actual work.

[0060] S103, for each medical device in the first set of monitoring devices, the operating status of the medical device is quantified based on the operating status data to obtain the change trend data of each medical device; the medical devices whose change trend data is greater than the preset trend threshold are selected to obtain the second set of monitoring devices.

[0061] Specifically, for each medical device in the first set of monitored devices, the terminal analyzes the operational status data, quantifies the abrupt change trend and degree of change in the operational status data through tiered calculation, obtains the change trend data of each medical device, and selects medical devices from the first set of monitored devices whose change trend data exceeds a preset trend threshold to form a second set of monitored devices. The preset trend threshold is used to characterize the change trend data threshold for further fault analysis of medical devices, and can be set according to the changes in the operational status data of medical devices that have experienced faults in actual operation.

[0062] S104, based on the first and second monitoring device sets, calculate the failure probability of medical equipment according to the operating status data, and generate fault warning information.

[0063] Specifically, for each medical device in the first and second monitoring device sets, the terminal predicts the normal operating status data based on the operating status data, and calculates the fault probability based on this and the operating status data to obtain the fault likelihood of each medical device. For the medical devices in the first monitoring device set, the fault likelihood is fused with the aforementioned first fault probability to obtain the posterior fault probability and fault type. The terminal integrates the posterior fault probability and fault type of each medical device in the first monitoring device set with the fault likelihood of each medical device in the second monitoring device set to generate fault warning information.

[0064] This application provides a method for early warning and detection of equipment failures in the medical equipment sector. By acquiring the operating status data and maintenance time data of each medical device, and combining this with failure cycle judgment rules mined from historical maintenance data using statistical learning methods, the method selects devices in the failure cycle to form a first monitoring device set. For the remaining medical devices, it selects devices with data mutations by analyzing the operating status data to form a second monitoring device set. For the medical devices in the first and second monitoring device sets, it calculates the failure probability of the medical devices by comprehensively analyzing the operating status data, and generates failure warning information. This method combines the failure cycle of medical devices with the mutation of operating data, and narrows the target range that requires in-depth calculation through two rounds of device screening, effectively improving the pertinence and efficiency of equipment failure early warning.

[0065] In one embodiment, the fault cycle determination rule is obtained through the following method:

[0066] S201, Obtain historical maintenance record data; historical maintenance record data includes equipment model, fault type, and fault occurrence time data for each medical device.

[0067] Specifically, the terminal acquires historical maintenance record data, which includes the equipment model of each medical device, the type of failure when the failure occurred in the past, and the time of failure. Each medical device has only one unique equipment model. The failure type is used to characterize the type of failure that occurred in the medical device and can be set according to the actual work. The failure time data is used to characterize the time information of the failure that occurred in the medical device.

[0068] S202, for each equipment model, calculate the fault interval time for each fault type based on the fault occurrence time data, and obtain the fault interval time set corresponding to each equipment model.

[0069] Specifically, for each device type, the terminal classifies the fault occurrence time data under that device model according to the fault type to obtain the fault occurrence time set corresponding to each fault type. For each fault type's fault occurrence time set, the fault occurrence time data is arranged in chronological order. Starting from the second fault, the difference between the current fault occurrence time and the previous fault occurrence time is calculated to form the fault interval time set corresponding to each device model and fault type.

[0070] S203. For each equipment model, the failure interval time set is used to construct a likelihood function to solve the failure rate change trend and average failure time of each failure type of the equipment model, thereby obtaining the failure rate change characteristic data and failure occurrence time characteristic data corresponding to each equipment model.

[0071] Specifically, for each device model, the terminal constructs a likelihood function and uses maximum likelihood estimation to fit a Weibull distribution model based on the failure interval time set. By solving for the shape and scale parameters in the Weibull distribution model, it obtains failure rate variation characteristic data and failure occurrence time characteristic data for each device model, including each failure type. The failure rate variation characteristic data characterizes the trend of failure over time, while the failure occurrence time characteristic data characterizes the timing of failures. The Weibull distribution model is a continuous probability distribution widely used in reliability engineering and failure analysis. Key parameters in the model are the shape parameter k and the scale parameter λ. By fitting the Weibull distribution model to the failure interval time set, the terminal obtains the shape parameter k, which characterizes the failure rate variation trend: k < 1 indicates an early failure period with decreasing failure rate; k = 1 indicates an occasional failure period with constant failure rate; and k > 1 indicates a wear-out failure period with increasing failure rate. The obtained scale parameter λ characterizes the average time of failure. Maximum likelihood estimation is a classic method in statistics used to estimate probability distribution parameters. Its basic principle is to find a set of parameter values ​​such that the probability distribution model based on these parameter values ​​maximizes the probability of the currently observed sample data appearing.

[0072] S204. Based on the failure rate variation characteristic data and failure occurrence time characteristic data of each equipment model, construct the failure cycle probability function for each failure type of each medical device to obtain the failure cycle judgment rule.

[0073] Specifically, for each device model, the terminal uses the cumulative distribution function to quantify the mathematical relationship between operating time and cumulative failure probability under each failure type, based on the failure rate variation characteristic data and failure occurrence time characteristic data. This yields the failure cycle probability function for each failure type of each medical device, forming the failure cycle judgment rule. The cumulative distribution function, in probability theory, describes the probability that a random variable takes a value less than or equal to a specific value. It fully defines a probability distribution, with its value range monotonically increasing from 0 to 1. The mathematical form of the specific failure cycle probability function is as follows: ,in It is medical equipment During continuous running time Afterwards, it is in the fault type The failure probability of the failure cycle. This refers to the continuous operating time of the medical equipment since its last maintenance. It is medical equipment In fault type Fault occurrence time characteristic data, It is medical equipment In fault type Failure rate variation characteristic data.

[0074] This application provides a method for early warning and detection of equipment failures in the medical equipment sector. It acquires data on the failure types and failure occurrence times of various medical equipment models during historical failure processes. By calculating the failure occurrence interval for each failure type, and using statistical inference methods to fit Weibull distribution parameters from the failure occurrence interval, it obtains failure rate change characteristic data representing the trend of failure rate over time and failure occurrence time characteristic data representing the average failure time. Using the aforementioned characteristic data, it establishes a mapping relationship between operating time and failure probability, including equipment type-failure type labels, to obtain a failure cycle probability function. This forms a failure cycle judgment rule, enabling a scientific summary of the failure cycle of medical equipment. It yields a precise dynamic probability function that quantifies the failure risk of equipment at different operating stages, narrowing the target range requiring in-depth failure probability calculation and effectively improving the efficiency of equipment failure early warning.

[0075] In one embodiment, based on preset fault cycle judgment rules and maintenance time data, the probability that each medical device is in a fault cycle is calculated to obtain a first fault probability, including:

[0076] S301, based on the maintenance time data of each medical device, determine the continuous operating time of each medical device, and obtain the testing time data of each medical device.

[0077] Specifically, the terminal calculates the continuous operating time of each medical device by subtracting the maintenance time data from the last maintenance and the real-time time, thus obtaining the testing time data for each medical device. Illustratively, the testing time data is used to characterize the time elapsed within the current maintenance cycle of the medical device.

[0078] S302, based on the test time data and fault cycle judgment rules of each medical device, calculate the probability of each medical device being in the fault cycle of each fault type, and obtain the first fault probability of each medical device.

[0079] Specifically, the terminal obtains the fault cycle probability function of each fault type for each device model according to the fault cycle judgment rule. For each medical device, the terminal inputs the test time data of the medical device into the fault cycle probability function of each fault type corresponding to the device model of the medical device to obtain the probability of the medical device being in the fault cycle of each fault type. The highest probability is selected as the first fault probability of the medical device, and the fault type corresponding to the first fault probability is determined as the fault type of the medical device.

[0080] This application provides a method for early warning and detection of equipment failure in the medical equipment sector. It calculates the actual continuous operating time of the equipment based on maintenance time data, and directly converts the operating time value into the corresponding failure probability value according to the failure cycle probability function. The failure probability with the highest probability is selected as the first failure probability. By utilizing the failure cycle pattern to identify equipment that meets the high failure cycle in the time dimension, it achieves efficient screening in the first round, narrows the target range that requires in-depth failure probability calculation, and effectively improves the efficiency of equipment failure early warning.

[0081] In one embodiment, the first failure probability threshold is obtained by the following method:

[0082] S401, based on the equipment model and failure time data of each medical device, calculates the historical failure rate of each equipment model.

[0083] Specifically, the terminal acquires data on the time of failure for each model of medical equipment. For each model, it calculates the total number of failures across all its medical devices and divides this total by the product of the total number of devices of that model and the observation duration, thereby calculating the historical failure rate of that model of medical equipment. Illustratively, the observation duration, i.e., the total equipment uptime for that model of medical equipment, can be obtained by subtracting the earliest failure occurrence time from the real-time and failure occurrence time data.

[0084] S402, For each medical device model, based on the historical failure rate, the first failure probability threshold for the medical device model is calculated using the following formula:

[0085]

[0086] in, It is the first The first failure probability threshold for each device model It is the basic threshold. It is an adjustment factor. It is the first Historical failure rate of each equipment model.

[0087] Specifically, for each model of medical equipment, the terminal calculates a first failure probability threshold for that model based on historical failure rates using a formula. (Illustratively, the...) First failure probability threshold for each device model A threshold used to characterize the first failure probability for further failure analysis of the medical device. Optionally, a base threshold. This is used to characterize the most basic probability threshold set for medical devices with extremely low historical failure rates or no historical data. Illustratively, the adjustment coefficient... This controls the impact of historical failure rates on the first failure probability threshold, and can be set according to actual operating conditions. Optionally, the... Historical failure rate of each device model It is the [number]th [item] among the historical failure rates of medical equipment models calculated by the aforementioned S401. Historical failure rate of each equipment model.

[0088] This application provides a method for early warning and detection of equipment failure in the medical equipment department. By aggregating and analyzing historical failure data, the historical failure rate of medical equipment representing each equipment model is calculated. Based on the historical failure rate, a dynamic threshold for real-time risk judgment is calculated to obtain the first failure probability threshold for each medical equipment. This enables the personalized setting of different failure cycle probability thresholds for each medical equipment, allowing for strict screening of equipment models with high failure rates, while maintaining a higher trigger threshold for relatively reliable equipment models, effectively improving the efficiency of equipment failure early warning.

[0089] In one embodiment, for each medical device in the first monitoring device group, the operating status of the medical device is quantified based on the operating status data to obtain the change trend data of each medical device, including:

[0090] S501, for each medical device in the first monitoring device group, determines the working status of the medical device based on the operating status data, and obtains the corresponding working status label for each medical device.

[0091] Specifically, for each medical device in the first monitoring device group, the terminal determines the working status of the medical device by analyzing its operational status data, and obtains a corresponding working status label for each medical device. Illustratively, the working status label is used to characterize the operating status of the medical device, providing a correct benchmark for subsequent standardized processing.

[0092] S502, for each medical device in the first monitoring device group, standardizes the operating status data based on the working status label to obtain standardized operating status data for each medical device.

[0093] Specifically, for each medical device in the first monitoring device set, the terminal selects the mean vector and standard deviation vector of the same working status label from the preset historical operating status data based on the medical device's working status label. This standardizes the operating status data, resulting in standardized operating status data for each medical device. The preset historical operating status data includes the mean vector and standard deviation vector for each medical device's historical operation. The mean vector includes the mean values ​​of current, voltage, temperature, and noise sound pressure level; the standard deviation vector includes the standard deviations of current, voltage, temperature, and noise sound pressure level. Both the mean vector and standard deviation vector carry medical device and working status labels. For any given medical device and its corresponding working status label, a unique set of mean and standard deviation vectors can be obtained. The preset historical operating status data can be collected based on actual work needs.

[0094] S503, compare the standardized operating status data with the preset mutation threshold to obtain the mutation comparison result; select medical devices from the first set of devices to be monitored whose mutation comparison result is greater than the mutation threshold, and obtain the first set of devices to be determined for monitoring.

[0095] Specifically, the terminal compares the standardized operating status data of each medical device with a preset mutation threshold to obtain the mutation comparison result for each medical device. It then selects medical devices from the first set of devices to be monitored whose standardized operating status data exceeds the mutation threshold, thus obtaining the first set of devices to be determined for monitoring. Indicatively, the mutation threshold can be set based on changes in the operating status data of medical devices that malfunction during actual operation.

[0096] S504: For each medical device in the first monitoring device group, based on the preset time window parameters, the slope of the standardized operating status data within the time window parameters is fitted and calculated to obtain the change trend data of each medical device.

[0097] Specifically, for each medical device in the first monitoring device group, the terminal selects a data segment from the standardized operating status data that is closest to the real-time time according to a preset time window parameter. The terminal then uses the least squares method to perform linear regression fitting on this data segment. The calculated slope value represents the trend of the medical device's change within the current window. This slope quantifies the rate and direction of change in the standardized operating status data; a positive value indicates an upward trend, a negative value indicates a downward trend, and the absolute value indicates the strength of the trend. The least squares method is a mathematical optimization method that solves a system of equations by minimizing the sum of squared errors, used to solve the mapping relationship of a linear system. Indicatively, the time window parameter can be set according to actual work requirements; this application does not further limit it.

[0098] This application provides a method for early warning and detection of equipment failure in the medical equipment department. By calculating the sudden change trend and trend threshold of the operating status data of each medical device in the first monitoring device set, the method selects medical devices whose operating status data meet the abnormal sudden change trend to form a second monitoring device set. This method enables the capture of early signs of sudden and gradual failures, effectively improving the pertinence and efficiency of equipment failure early warning.

[0099] In one embodiment, for each medical device in the first monitoring device set, the working status of the medical device is determined based on the operating status data, and a working status label corresponding to each medical device is obtained, including:

[0100] S601, for each medical device in the first monitoring device group, calculate the Euclidean distance between the operating status data and the center data of each working state to obtain the Euclidean distance result corresponding to each working state; determine the working state corresponding to the smallest Euclidean distance result as the working state label of the medical device.

[0101] Specifically, for each medical device in the first monitoring device set, the terminal calculates the Euclidean distance between its operating status data and the central data of each working state of the same medical device, obtaining the Euclidean distance result corresponding to the central data of each working state. The working state with the minimum Euclidean distance result is identified as the working state label of the medical device. Illustratively, the central data of each working state includes current data, voltage data, temperature data, and noise sound pressure level data characterizing the medical device in that working state.

[0102] The central data for each working state was obtained through the following method:

[0103] S6021, acquire a preset number of historical work record data; the historical work record data includes current data, voltage data, temperature data and noise sound pressure level data of each medical device in its historical normal working state.

[0104] Specifically, the terminal acquires a preset amount of historical work record data, which includes current, voltage, temperature, and noise pressure level data for each medical device during its historical normal operating conditions. The terminal then groups the historical work record data according to different medical devices, obtaining a historical work record dataset for each medical device. This is illustrative; the number of historical work record data points can be set based on the data collected during actual work.

[0105] S6022 uses the K-Means++ clustering algorithm to select cluster center data from historical work record data to obtain the initial cluster center data group.

[0106] Specifically, for each medical device's historical work record dataset, the terminal employs the K-Means++ clustering algorithm. It randomly selects a historical work record as the first initial cluster center and calculates the nearest distance between the remaining historical work records and this first initial cluster center. Based on a distance probability distribution proportional to the square of the distance, subsequent initial cluster centers are selected sequentially using a roulette wheel selection method until K centers are selected to form the initial cluster center data for that medical device. The initial cluster center data for each medical device then constitutes the initial cluster center data group. The K-Means++ clustering algorithm is an intelligent initialization algorithm for clustering. Its principle is to optimize the selection of initial cluster centers, making them as far apart as possible, thus serving as the basis for data classification.

[0107] S6023, For the remaining historical work record data, calculate the Euclidean distance with each cluster center data in the initial cluster center data group to obtain the minimum Euclidean distance result; update the cluster center data corresponding to the minimum Euclidean distance result using the historical work record data to obtain the complete cluster center data group.

[0108] Specifically, for the remaining historical work record data in the historical work record dataset of each medical device, the Euclidean distance between each historical work record data and the initial cluster center of the same medical device obtained above is calculated, and the historical work record data is assigned to the category represented by the nearest center. By adding historical work record data to the new category and the historical work record data of the original initial cluster center, the mean vector of all data points in each category is recalculated until all historical work record data is assigned, and the cluster center data of each medical device is obtained. The cluster center data of each medical device forms a complete cluster center data group.

[0109] S6024, according to the preset working status classification rules, match each cluster center data in the complete cluster center data group with each working status to obtain the center data of each working status.

[0110] Specifically, the terminal matches the cluster center data of each medical device in the complete cluster center data group with each working state according to preset working state classification rules, thereby obtaining the center data for each working state. The center data for each working state includes the medical device label. Illustratively, the preset working state classification rules can be set based on the operational data of each medical device in different working states during actual operation.

[0111] This application provides a method for early warning and detection of equipment failures in the medical equipment department. Based on a training process using historical work record data, it employs the K-Means++ clustering algorithm to automatically learn and define all operating modes of the equipment and their digitized central benchmark data from a complex multi-dimensional parameter space. This yields central data for each operating state, which is then matched with real-time operational status data to determine the operating status of each medical device. This reduces the reliance on manual experience-based rules for judging equipment operating status and effectively improves the targeting and efficiency of equipment failure early warning.

[0112] In one embodiment, based on a first set of monitoring devices and a second set of monitoring devices, the failure probability of the medical device is calculated according to the operating status data, and failure warning information is generated, including:

[0113] S701: For each medical device in the first and second monitoring device sets, the operating status data is input into the trained fault early warning autoencoder model to obtain the reconstructed operating data corresponding to each medical device.

[0114] Specifically, for each medical device in the first and second monitoring device sets, the terminal inputs its operational status data into a trained fault early warning autoencoder model to obtain the reconstructed operational data corresponding to each medical device. The fault early warning autoencoder model can be trained based on the historical normal operation data of each medical device and includes an encoder and a decoder. The encoder compresses the input operational status data into a low-dimensional hidden space to capture its core features, and the decoder reconstructs the original input data from these core features. The output is the reconstructed operational data corresponding to each medical device.

[0115] S702 calculates the difference between the reconstructed running data and the running status data based on the mean square error calculation formula, and obtains the reconstructed error data corresponding to each medical device.

[0116] Specifically, for each medical device in the first and second monitoring device sets, the terminal calculates the difference between its reconstructed operating data and its operating status data using the mean squared error (MSE) formula. This yields the reconstruction error data for each medical device, quantifying the degree of deviation between the current operating status and the normal mode. A larger reconstruction error value indicates a greater deviation from all learned normal modes, suggesting a higher probability of failure. The MSE formula is a statistical measure of the difference between estimated and actual values. Its principle is to calculate the average of the squares of all prediction errors, amplifying the impact of larger errors through squaring, thus more sensitively capturing significant deviations.

[0117] S703 uses an exponential decay function to map the reconstructed error data to obtain the fault likelihood of each medical device.

[0118] Specifically, for each medical device in the first and second monitoring device sets, the terminal uses an exponential decay function to map its reconstruction error data, monotonically increasing the reconstruction error to a fault likelihood, thus obtaining the fault likelihood for each medical device. A higher fault likelihood value indicates a higher conditional probability of a fault occurring under the current data. The exponential decay function is a monotonically decreasing nonlinear function; its principle is that as the independent variable increases, the function value approaches zero exponentially.

[0119] S704, for each medical device in the first monitoring device set, the first fault probability and the fault likelihood are fused based on Bayes' theorem to obtain the posterior fault probability.

[0120] Specifically, for each medical device in the first monitoring device set, the terminal integrates its first failure probability with its failure likelihood based on Bayes' theorem to obtain the posterior failure probability of that medical device. Bayes' theorem is a fundamental probability principle describing how to calculate the posterior probability given prior probabilities and likelihoods. Its core idea is to use new evidence—the failure likelihood—to update our original belief about a hypothesis (such as device failure) regarding the first failure probability.

[0121] S705 generates fault warning information based on the first fault probability and posterior fault probability of each medical device in the first monitoring device set, and the change trend data and fault likelihood of each medical device in the second monitoring device set.

[0122] Specifically, the terminal integrates the fault type and posterior fault probability in the first fault probability of each medical device in the first monitoring device set, and integrates the change trend data and fault likelihood of each medical device in the second monitoring device set to generate fault warning information.

[0123] This application provides a method for early warning and detection of equipment failures in the medical equipment sector. It utilizes a deep autoencoder to deeply reconstruct real-time data from high-risk and abnormal equipment, obtaining reconstructed operational data characterizing the normal operation of each medical device. The method calculates the error between this reconstructed data and the operational status data, obtaining reconstruction error data. This reconstruction error data is then transformed into a probabilistic fault likelihood through exponential mapping. For medical devices determined to be in a failure cycle, Bayesian updates are used to deeply fuse long-term patterns with real-time evidence, yielding a highly reliable posterior fault probability. Finally, fault warning information is generated. This method leverages the representation learning capabilities of deep learning and the reasoning capabilities of Bayesian statistics to combine the failure cycle of medical devices with abrupt changes in operational data. This achieves high-efficiency predictive maintenance while ensuring computational efficiency, thereby improving the targeting and efficiency of equipment failure warnings.

[0124] In the aforementioned medical equipment fault early warning and detection method, by acquiring the operating status data and maintenance time data of each medical device, and combining this with the fault cycle judgment rules mined from historical maintenance data using statistical learning methods, devices in the fault cycle are selected to form a first monitoring device set. For the remaining medical devices, devices with data mutations are selected through operating status data analysis to form a second monitoring device set. For the medical devices in the first and second monitoring device sets, the fault probability of the medical devices is calculated through comprehensive analysis of the operating status data, and fault early warning information is generated. This method combines the fault cycle of medical devices with the mutation of operating data, and narrows the target range that requires in-depth calculation through two rounds of device screening, effectively improving the pertinence and efficiency of equipment fault early warning.

[0125] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0126] Based on the same inventive concept, this application also provides a medical equipment fault early warning detection system for implementing the aforementioned medical equipment fault early warning detection method. The solution provided by this device is similar to the implementation scheme described in the above method; therefore, the specific limitations of one or more embodiments of the medical equipment fault early warning detection system provided below can be found in the limitations of the medical equipment fault early warning detection method described above, and will not be repeated here.

[0127] In one exemplary embodiment, such as Figure 2 As shown, a medical equipment fault early warning and detection system 200 is provided, comprising:

[0128] The data acquisition module 201 is used to acquire the operating status data and maintenance time data of each medical device; the operating status data includes current data, voltage data, temperature data and noise sound pressure level data;

[0129] The fault cycle matching module 202 is used to calculate the probability that each medical device is in a fault cycle based on the preset fault cycle judgment rules and maintenance time data, and obtain a first fault probability; select medical devices with a first fault probability greater than a preset first fault probability threshold to obtain a first monitoring device set, and select medical devices with a first fault probability less than the first fault probability threshold to obtain a first monitoring device set.

[0130] The data mutation matching module 203 is used to quantify the operating status of each medical device in the first set of monitored devices based on the operating status data, and obtain the change trend data of each medical device; and select the medical devices whose change trend data is greater than the preset trend threshold to obtain the second set of monitored devices.

[0131] The early warning information generation module 204 is used to calculate the failure probability of medical equipment based on the first monitoring device set and the second monitoring device set, and generate failure early warning information according to the operating status data.

[0132] Furthermore, the fault cycle determination rule is obtained through the following method:

[0133] Obtain historical maintenance record data; historical maintenance record data includes equipment model, fault type, and fault occurrence time data for each medical device;

[0134] For each equipment model, the fault interval time for each fault type is calculated based on the fault occurrence time data, thus obtaining the fault interval time set corresponding to each equipment model.

[0135] For each equipment model, the failure rate variation trend and average failure time of each failure type are solved by constructing a likelihood function, thereby obtaining the failure rate variation characteristic data and failure occurrence time characteristic data corresponding to each equipment model.

[0136] For each equipment model, based on the failure rate variation characteristics and failure occurrence time characteristics, a failure cycle probability function for each failure type of each medical device is constructed to obtain the failure cycle judgment rule.

[0137] Furthermore, the fault cycle matching module is also used for:

[0138] Based on the maintenance time data of each medical device, the continuous operating time of each medical device is determined, and the testing time data of each medical device is obtained.

[0139] Based on the testing time data and fault cycle judgment rules of each medical device, the probability of each medical device being in the fault cycle of each fault type is calculated, and the first fault probability of each medical device is obtained.

[0140] Furthermore, the first fault probability threshold is obtained by the following method:

[0141] Based on the equipment model and failure time data of each medical device, the historical failure rate of each equipment model of medical device is calculated;

[0142] For each medical device model, based on historical failure rates, the first failure probability threshold for each model is calculated using the following formula:

[0143]

[0144] in, It is the first The first failure probability threshold for each device model It is the basic threshold. It is an adjustment factor. It is the first Historical failure rate of each equipment model.

[0145] Furthermore, the data mutation matching module includes:

[0146] The working label matching unit is used to determine the working status of each medical device in the first monitoring device group based on the operating status data, and obtain the working status label corresponding to each medical device.

[0147] The data standardization unit is used to standardize the operating status data of each medical device in the first monitoring device group based on the working status label, so as to obtain the standardized operating status data of each medical device.

[0148] The mutation determination unit is used to compare standardized operating status data with a preset mutation threshold to obtain a mutation comparison result; and to select medical devices from the first set of devices to be monitored whose standardized operating status data is greater than the mutation threshold, thereby obtaining the first set of devices to be determined for monitoring.

[0149] The mutation degree calculation unit is used to calculate the slope of the standardized operating status data of each medical device in the first monitoring device group according to the preset time window parameters, so as to obtain the change trend data of each medical device.

[0150] Furthermore, the work tag matching unit is also used for:

[0151] For each medical device in the first monitoring device group, the Euclidean distance between the operating status data and the central data of each working state is calculated to obtain the Euclidean distance result corresponding to each working state; the working state corresponding to the smallest Euclidean distance result is determined as the working state label of the medical device.

[0152] The central data for each working state was obtained through the following method:

[0153] Acquire a preset amount of historical work record data; the historical work record data includes current data, voltage data, temperature data, and noise sound pressure level data of each medical device in its historical normal working state;

[0154] The K-Means++ clustering algorithm is used to select cluster center data from each historical work record data to obtain the initial cluster center data group;

[0155] For the remaining historical work record data, calculate the Euclidean distance with each cluster center data in the initial cluster center data group to obtain the minimum Euclidean distance result; update the cluster center data corresponding to the minimum Euclidean distance result using the historical work record data to obtain the complete cluster center data group;

[0156] According to the preset work status classification rules, the cluster center data in the complete cluster center data group are matched with each work status to obtain the center data of each work status.

[0157] Furthermore, the early warning information generation module is also used for:

[0158] For each medical device in the first and second monitoring device sets, the operating status data is input into the trained fault early warning autoencoder model to obtain the reconstructed operating data corresponding to each medical device.

[0159] Based on the mean square error calculation formula, the difference between the reconstruction operation data and the operation status data is calculated to obtain the reconstruction error data corresponding to each medical device.

[0160] An exponential decay function is used to map the reconstructed error data to obtain the fault likelihood of each medical device.

[0161] For each medical device in the first monitoring device set, the first failure probability and the failure likelihood are fused based on Bayes' theorem to obtain the posterior failure probability.

[0162] Based on the first failure probability and posterior failure probability of each medical device in the first monitoring device set, and the change trend data and failure likelihood of each medical device in the second monitoring device set, a fault warning information is generated.

[0163] In one embodiment, such as Figure 3 A computer device is provided, comprising:

[0164] At least one processor 301, and a memory 302 communicatively connected to at least one of the processors 301: the memory stores application code that can be executed by at least one of the processors, the application code being executed by at least one of the processors to enable at least one of the processors to perform a medical equipment fault early warning detection method as described above.

[0165] Computer equipment may also include: sensor 303.

[0166] The processor 301, memory 302 and sensor 303 can be connected via a bus or other means, with the bus being an example in the figure.

[0167] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0168] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative. The components described as separate parts may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this disclosure according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0169] The above-described embodiments are merely illustrative of several implementation methods of the embodiments of this application, and their descriptions are relatively specific and detailed. However, they should not be construed as limiting the scope of the patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the embodiments of this application, and these modifications and improvements all fall within the protection scope of the embodiments of this application.

Claims

1. A method for detecting a failure of a medical equipment department device, characterized by, The method includes: Acquire operational status data and maintenance time data for each medical device; the operational status data includes current data, voltage data, temperature data, and noise sound pressure level data; Based on the preset fault cycle judgment rules and the maintenance time data, the probability of each medical device being in a fault cycle is calculated to obtain a first fault probability; the medical devices with the first fault probability greater than the preset first fault probability threshold are selected to obtain a first monitoring device set, and the medical devices with the first fault probability less than the first fault probability threshold are selected to obtain a first monitoring device set. For each medical device in the first set of monitored devices, the operating status of the medical device is quantified based on the operating status data to obtain the change trend data of each medical device; the medical devices whose change trend data is greater than a preset trend threshold are selected to obtain the second set of monitored devices; Based on the first monitoring device set and the second monitoring device set, the failure probability of the medical device is calculated according to the operating status data, and failure warning information is generated.

2. The method of claim 1, wherein, The fault cycle determination rule is obtained through the following method: Acquire historical maintenance record data; the historical maintenance record data includes the equipment model, fault type, and fault occurrence time of each of the medical devices; For each of the aforementioned equipment models, the fault interval time for each of the aforementioned equipment models is calculated based on the fault occurrence time data, thereby obtaining a fault interval time set corresponding to each of the aforementioned equipment models; For each set of fault interval times for each of the aforementioned equipment models, the failure rate variation trend and average time of failure occurrence for each fault type of the aforementioned equipment model are solved by constructing a likelihood function, thereby obtaining the fault rate variation characteristic data and fault occurrence time characteristic data corresponding to each of the aforementioned equipment models; For each of the device models, the failure rate variation characteristic data and the failure occurrence time characteristic data are used to construct the failure cycle probability function for each of the failure types of each medical device, and the failure cycle judgment rule is obtained.

3. The method of claim 2, wherein, The method of calculating the probability that each medical device is in a fault cycle based on the preset fault cycle judgment rules and the maintenance time data, to obtain the first fault probability, includes: Based on the maintenance time data of each medical device, the continuous operating time of each medical device is determined, and the testing time data of each medical device is obtained. Based on the detection time data of each medical device and the fault cycle judgment rule, the probability of each medical device being in the fault cycle of each fault type is calculated to obtain the first fault probability of each medical device.

4. The method of claim 3, wherein, The first fault probability threshold is obtained by the following method: Based on the equipment model and the time of failure of each of the medical devices, the historical failure rate of each of the equipment models of the medical devices is calculated; For each of the aforementioned device models of medical devices, based on the historical failure rate, the first failure probability threshold for the medical device of that device model is calculated using the following formula: in, It is the first The first failure probability threshold for each device model It is the basic threshold. It is an adjustment factor. It is the first Historical failure rate of each equipment model.

5. The method of claim 1, wherein, For each of the medical devices in the first set of monitored devices, the operational status of the medical device is quantified based on the operational status data to obtain the trend data of each medical device, including: For each of the medical devices in the first set of monitored devices, the working status of the medical device is determined based on the operating status data, and a working status label corresponding to each medical device is obtained. For each of the medical devices in the first set of monitored devices, the operating status data is standardized based on the working status label to obtain standardized operating status data for each of the medical devices. The standardized operating status data is compared with a preset mutation threshold to obtain a mutation comparison result; from the first set of devices to be monitored, the medical devices whose mutation comparison result is greater than the mutation threshold are selected to obtain the first set of devices to be determined for monitoring. For each of the medical devices in the first monitoring device set, the slope of the standardized operating status data in the time window parameter is calculated based on the preset time window parameter to obtain the change trend data of each medical device.

6. The method of claim 5, wherein, For each of the medical devices in the first set of monitored devices, the operating status of the medical device is determined based on the operating status data, and an operating status label corresponding to each medical device is obtained, including: For each of the medical devices in the first monitoring device set, the Euclidean distance between the operating status data and the center data of each working state is calculated to obtain the Euclidean distance result corresponding to each working state; the working state corresponding to the smallest Euclidean distance result is determined as the working state label of the medical device. The central data for each of the aforementioned working states are obtained through the following method: Acquire a preset number of historical work record data; the historical work record data includes the current data, voltage data, temperature data, and noise sound pressure level data of each of the medical devices in their historical normal working state; The K-Means++ clustering algorithm is used to select cluster center data from the historical work record data to obtain the initial cluster center data group; For the remaining historical work record data, calculate the Euclidean distance with each of the cluster center data in the initial cluster center data group to obtain the minimum Euclidean distance result; update the cluster center data corresponding to the minimum Euclidean distance result using the historical work record data to obtain the complete cluster center data group; According to the preset work status classification rules, each cluster center data in the complete cluster center data group is matched with each work status to obtain the center data of each work status.

7. The method of claim 1, wherein, The step of calculating the failure probability of the medical equipment based on the first monitoring device set and the second monitoring device set, and generating failure early warning information according to the operating status data, includes: For each of the medical devices in the first monitoring device set and the second monitoring device set, the operating status data is input into the trained fault early warning autoencoder model to obtain the reconstructed operating data corresponding to each medical device; Based on the mean square error calculation formula, the difference between the reconstructed running data and the running status data is calculated to obtain the reconstructed error data corresponding to each medical device. The reconstruction error data is mapped using an exponential decay function to obtain the fault likelihood corresponding to each medical device; For each of the medical devices in the first monitoring device set, the first fault probability and the fault likelihood are fused based on Bayes' theorem to obtain the posterior fault probability. The fault warning information is generated based on the first fault probability and the posterior fault probability of each medical device in the first monitoring device set, and the trend data and fault likelihood of each medical device in the second monitoring device set.

8. A medical equipment department device failure early warning detection system, characterized by, The system includes: The data acquisition module is used to acquire the operating status data and maintenance time data of each medical device; the operating status data includes current data, voltage data, temperature data and noise sound pressure level data; The fault cycle matching module is used to calculate the probability that each medical device is in a fault cycle based on the preset fault cycle judgment rules and the maintenance time data, and obtain a first fault probability; select the medical devices whose first fault probability is greater than a preset first fault probability threshold to obtain a first monitoring device set; and select the medical devices whose first fault probability is less than the first fault probability threshold to obtain a first monitoring device set. The data mutation matching module is used to quantify the operating status of each medical device in the first set of monitored devices based on the operating status data, and obtain the change trend data of each medical device; and select the medical devices whose change trend data is greater than a preset trend threshold to obtain a second set of monitored devices. The early warning information generation module is used to calculate the failure probability of the medical equipment based on the first monitoring device set and the second monitoring device set, and generate failure early warning information according to the operating status data.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images