An AI-based medical equipment operation and maintenance management system

Abstract

The application relates to the field of operation and maintenance management, and discloses an AI-based medical equipment operation and maintenance management system, which comprises an equipment monitoring module, an equipment fault prediction module, a safety detection module and an operation and maintenance execution module; the equipment monitoring module is used for monitoring the running parameters of equipment in real time and serving as equipment monitoring data; the equipment fault prediction module is based on a pre-trained recurrent neural network model, and obtains a fault prediction coefficient and a fault type after cleaning, fusing and processing the equipment monitoring data; the safety detection module analyzes the safety of equipment according to the fault prediction coefficient and the fault type of the equipment fault diagnosis module, and establishes a priority maintenance equipment list according to the analysis result; and the operation and maintenance execution module is used for executing a maintenance task according to the priority maintenance list of the safety detection module. The equipment fault prediction module and the safety detection module cooperate to realize predictive judgment on whether the equipment needs to be maintained, and the equipment needing to be maintained is prioritized, so that the purpose of efficient maintenance is achieved.

Description

Technical Field

[0001] This invention relates to the field of operation and maintenance management, and specifically to an AI-based medical equipment operation and maintenance management system. Background Technology

[0002] The operation and maintenance management of medical equipment aims to ensure its safety, reliability, and efficiency, avoid equipment failures, reduce maintenance time and costs, and improve the service level of medical institutions and patient safety. The effective operation of medical equipment directly impacts the patient experience and the timely and effective medical services they receive within medical institutions. If a large number of medical devices are not operational, it obviously poses a significant problem for patients. Existing medical equipment operation and maintenance management systems can monitor equipment performance and status in real time, predict equipment failures and errors, and prevent sudden equipment outages.

[0003] In practical application, it has been found that prioritizing equipment based on its importance and likelihood of failure is crucial to determining which equipment requires maintenance first. However, existing equipment maintenance processes do not consider the impact of different failure types. For example, some types of failures can be maintained more quickly, allowing for rapid restoration of operation even in the event of a sudden interruption. Conversely, other types of failures require longer maintenance times, leading to equipment outages, disrupting normal use, and placing excessive pressure on other similar medical equipment, ultimately causing a series of cascading effects. Summary of the Invention

[0004] The purpose of this invention is to provide an AI-based medical device operation and maintenance management system to solve the above-mentioned technical problems.

[0005] The objective of this invention can be achieved through the following technical solutions:

[0006] An AI-based medical device operation and maintenance management system includes:

[0007] The equipment monitoring module is used to monitor the operating parameters of the equipment in real time and to serve as equipment monitoring data;

[0008] The equipment fault prediction module, based on a pre-trained recurrent neural network model, cleans, fuses, and processes the equipment monitoring data to obtain fault prediction coefficients.

[0009] The safety detection module analyzes equipment safety based on the fault prediction coefficient of the equipment fault diagnosis module, and establishes a list of priority maintenance equipment based on the analysis results.

[0010] The operation and maintenance execution module is used to perform maintenance tasks based on the priority maintenance list of the security detection module.

[0011] As a further technical solution, the operating parameters include temperature, pressure, current, voltage, and usage time; the equipment fault types include mechanical faults, electrical faults, and software faults. The pressure referred to is pneumatic pressure. Each operating parameter is obtained through specialized testing instruments or sensors, which is existing technology and will not be elaborated upon here.

[0012] As a further technical solution, the process of analyzing equipment safety is as follows:

[0013] Obtain the equipment's fault prediction coefficient P;

[0014] A safety analysis model F(P, K) is established based on the fault prediction coefficient P and the maintenance difficulty coefficient K; the expression is:

[0015]

[0016] Where α and β are weighting coefficients, P represents the extreme difference of the fault prediction coefficient of the current equipment per unit time. max P min Let be the maximum and minimum values ​​of the fault prediction coefficient of the current equipment per unit time, respectively, where i represents the i-th fault type, and k is the k-th fault. i p represents the maintenance difficulty coefficient corresponding to the i-th fault type. i P0 and K0 are the preset fixed coefficients for the i-th fault type, and P0 and K0 are the threshold values ​​for the fault prediction coefficient and maintenance difficulty coefficient of the current equipment.

[0017] The safety of the current equipment is determined by the equipment safety analysis model F(P,K).

[0018] As a further technical solution, the process for determining whether the current device is safe is as follows:

[0019] The fault prediction coefficient P of the j-th device j and maintenance difficulty coefficient K j Substituting this into the equipment safety analysis model, the fault estimate F(P) of the j-th equipment is calculated. j K j );

[0020] F(P) j K j The value is compared with the preset safety threshold F(P, K)0;

[0021] If F(P) j K j If F(P, K) ≥ F(P, K) 0, then the current device is determined to be unsafe.

[0022] Otherwise, determine that the current device is safe.

[0023] As a further technical solution, the process for obtaining the maintenance difficulty coefficient is as follows:

[0024] Mechanical faults, software faults, and electrical faults are labeled l1, l2, and l3, respectively.

[0025] The curves l1q(t), l2q(t), and l3q(t) of the frequency of mechanical faults, software faults, and electrical faults over time were obtained by fitting.

[0026] Based on the variation amplitude and extreme values ​​of the curve corresponding to the frequency of occurrence of each fault type, the maintenance difficulty coefficients k1, k2, and k3 corresponding to each fault type are calculated respectively.

[0027] As a further technical solution, the expressions for k1, k2, and k3 are as follows:

[0028]

[0029] Where l1q(t)0, l2q(t)0, and l3q(t)0 are reference curves, and l1q max l1q min These represent the maximum and minimum frequencies of mechanical failure occurrence, l2q, respectively. max l2q min These represent the maximum and minimum frequencies of software failure occurrence, respectively. max l3q mi n represents the maximum and minimum frequency of electrical fault occurrence, respectively; S1, S2, and S3 represent the fluctuation coefficients corresponding to each fault type, respectively; and t0 and t1 represent the start and end points of the statistical period, respectively.

[0030] As a further technical solution, the process of obtaining the fluctuation coefficient corresponding to each fault type is as follows:

[0031] The number of times n of the i-th type of fault occurs and the duration Δt of each maintenance operation. e Substitute into the formula The average maintenance processing time S corresponding to each fault type was calculated. i ;

[0032] Where e is the e-th time, Let be the average maintenance processing time for the i-th fault type.

[0033] As a further technical solution, the process of establishing a priority maintenance equipment list based on the analysis results is as follows:

[0034] The condition F(P) will be satisfied.j K j Add devices with a value ≥ F(P, K)0 to the list of maintained devices;

[0035] Sort each piece of equipment in the maintenance equipment list in descending order according to its maintenance difficulty coefficient K;

[0036] Each device is sequentially numbered 1, 2, 3, ... N, which forms the priority maintenance list.

[0037] The beneficial effects of this invention are:

[0038] (1) Through the above technical solution, the present invention utilizes the pre-trained recurrent neural network model in the equipment fault prediction module to obtain the fault prediction coefficient based on the equipment's operating parameters. Obviously, the larger the fault prediction coefficient, the greater the risk of the current equipment and the greater the probability of fault interruption during subsequent operation. Therefore, it needs to be maintained in a timely manner. Then, the safety detection module establishes a safety analysis model based on the fault prediction coefficient and the types of faults currently existing in the equipment to evaluate the equipment safety from multiple aspects and finally obtain a more accurate equipment fault valuation. This provides data support for the subsequent establishment of a priority maintenance list, thereby achieving the purpose of effectively allocating maintenance resources and establishing a key monitoring list.

[0039] (2) The present invention provides a process for obtaining the maintenance difficulty coefficient corresponding to each type of fault through the above technical solution. Since different equipment encounters different types of faults, the pressure on maintenance personnel is different. Therefore, by fitting the curve of the frequency of mechanical faults, software faults and electrical faults occurring over time in each unit time, and then by integrating the curve to obtain the deviation from the reference curve, the stability of each equipment encountering each type of fault can be determined. The worse the stability, the greater the maintenance difficulty may be. In order to reduce the error caused by only the change amplitude, the difference between the maximum and minimum values ​​of each type of fault is added, so as to achieve a more accurate judgment of the maintenance difficulty.

[0040] (3) Through the above technical solution, the present invention comprehensively analyzes the historical situation of each type of fault to finally determine which type of fault or which two types of fault occur simultaneously for each medical device, which will lead to increased maintenance difficulty. Then, it focuses on monitoring and prioritizes predictive maintenance to reduce the adverse effects of this type of fault on the equipment. At the same time, it can also reduce the cascading impact of long-term disconnection on other devices, which will cause excessive pressure in a short period of time. Attached Figure Description

[0041] The invention will now be further described with reference to the accompanying drawings.

[0042] Figure 1 This is a system block diagram of the present invention. Detailed Implementation

[0043] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0044] Please see Figure 1 As shown, this invention is an AI-based medical device operation and maintenance management system, comprising:

[0045] The equipment monitoring module is used to monitor the operating parameters of the equipment in real time and to serve as equipment monitoring data;

[0046] The equipment fault prediction module, based on a pre-trained recurrent neural network model, cleans, fuses, and processes the equipment monitoring data to obtain fault prediction coefficients. The establishment and training process of the recurrent neural network model is a mature technology, so it will not be elaborated on in this article.

[0047] The safety detection module analyzes equipment safety based on the fault prediction coefficient of the equipment fault diagnosis module, and establishes a list of priority maintenance equipment based on the analysis results.

[0048] The operation and maintenance execution module is used to execute maintenance tasks according to the priority maintenance list of the safety detection module. The operating parameters include temperature, pressure, current, voltage, and usage time; the equipment fault types include mechanical faults, electrical faults, and software faults. The pressure refers to air pressure. Each operating parameter is obtained through specialized testing instruments or sensors, which is existing technology and will not be elaborated upon here.

[0049] In this embodiment, a pre-trained recurrent neural network model within the equipment fault prediction module is used to obtain fault prediction coefficients based on the equipment's operating parameters. Obviously, the larger the fault prediction coefficient, the greater the risk of the current equipment and the greater the probability of fault interruption during subsequent operation. Therefore, timely maintenance is required. Then, the safety detection module uses the fault prediction coefficients combined with the types of faults present in the current equipment to establish a comprehensive safety analysis model, which assesses equipment safety from multiple aspects and ultimately obtains a more accurate equipment fault estimate. This provides data support for establishing a priority maintenance list, thereby achieving the goal of effectively allocating maintenance resources and establishing a key monitoring list.

[0050] The process of analyzing equipment safety is as follows:

[0051] Obtain the equipment's fault prediction coefficient P;

[0052] A safety analysis model F(P, K) is established based on the fault prediction coefficient P and the maintenance difficulty coefficient K; the expression is:

[0053]

[0054] Wherein, α and β are weighting coefficients, which are determined based on historical experience data. P represents the extreme difference of the fault prediction coefficient of the current equipment per unit time. max P min Let be the maximum and minimum values ​​of the fault prediction coefficient of the current equipment per unit time, respectively, where i represents the i-th fault type, and k is the k-th fault. i p represents the maintenance difficulty coefficient corresponding to the i-th fault type. i The preset fixed coefficient for the i-th fault type is a constant, determined based on historical data. P0 and K0 are the threshold values ​​for the fault prediction coefficient and maintenance difficulty coefficient of the current equipment.

[0055] The safety of the current equipment is determined by the equipment safety analysis model F(P, K).

[0056] The process of determining whether the current device is safe is as follows:

[0057] The fault prediction coefficient P of the j-th device j and maintenance difficulty coefficient K j Substituting this into the equipment safety analysis model, the fault estimate F(P) of the j-th equipment is calculated. j K j );

[0058] F(P) j K j The value is compared with the preset safety threshold F(P, K)0;

[0059] If F(P) j K j If F(P, K) ≥ F(P, K) 0, then the current device is determined to be unsafe.

[0060] Otherwise, determine that the current device is safe.

[0061] This embodiment provides a specific method for performing equipment safety analysis. First, an equipment safety analysis model F(P, K) is established based on the fault prediction coefficient P and the maintenance difficulty coefficient K, expressed as follows: As can be seen from the above expressions, the larger the maintenance difficulty coefficient K and the fault prediction coefficient P, the greater their impact on the final fault estimate of the equipment. The extreme value difference method can represent the overall change range of the fault prediction coefficient of each equipment. Obviously, the larger the extreme value difference, the greater the stability of the fault prediction coefficient of the equipment, and therefore the greater its impact on the final fault estimate of the equipment. Obviously, the larger the fault estimate, the higher the probability of the current equipment failure, and therefore the greater the need to add it to the maintenance list. Furthermore, by comparing the fault estimate calculated for each equipment with the system's preset safety threshold, if the fault estimate exceeds the safety threshold, it means that the probability of the equipment failure is high, and the cascading impact of the failure is greater, so it needs more focused maintenance. Conversely, if the fault estimate is lower, it means that the cascading impact of the equipment is smaller, and although timely maintenance is still needed, its importance is obviously lower.

[0062] The process for obtaining the maintenance difficulty coefficient is as follows:

[0063] Mechanical faults, software faults, and electrical faults are labeled l1, l2, and l3, respectively.

[0064] The curves l1q(t), l2q(t), and l3q(t) of the frequency of mechanical, software, and electrical faults over time were obtained by fitting. Obviously, the larger the cumulative change in the frequency of a certain type of fault, the greater the threat that type of fault poses in the following operating time. If the maintenance difficulty of this type of fault is low, the demand for maintenance resources can be further reduced in the next stage, or the number of medical devices that can be kept online at the same time can be increased while the maintenance resources remain unchanged.

[0065] Based on the variation amplitude and extreme values ​​of the curve corresponding to the frequency of occurrence of each fault type, the maintenance difficulty coefficients k1, k2, and k3 corresponding to each fault type are calculated respectively.

[0066] The expressions for k1, k2, and k3 are:

[0067]

[0068] Where l1q(t)0, l2q(t)0, and l3q(t)0 are reference curves, and l1q max l1q min These represent the maximum and minimum frequencies of mechanical failure occurrence, l2q, respectively. max l2q min These represent the maximum and minimum frequencies of software failure occurrence, respectively. max l3q minS1, S2, and S3 are the maximum and minimum frequencies of electrical fault occurrence, respectively; S1, S2, and S3 are the fluctuation coefficients corresponding to each fault type, and t0 and t1 are the start and end points of the statistical period, respectively.

[0069] The process of obtaining the fluctuation coefficient corresponding to each fault type is as follows:

[0070] The number of times n of the i-th type of fault occurs and the duration Δt of each maintenance operation. e Substitute into the formula The average maintenance processing time S corresponding to each fault type was calculated. i ;

[0071] Where e is the e-th time, Let be the average maintenance processing time for the i-th fault type.

[0072] In this embodiment, a process for obtaining the maintenance difficulty coefficient corresponding to each fault type is provided. Since the pressure on maintenance personnel varies depending on the type of fault encountered by different equipment, the frequency of mechanical faults, software faults, and electrical faults over time is obtained by fitting the curves l1q(t), l2q(t), and l3q(t) of each unit time. The deviation between these curves and the reference curve is then obtained by integration. This allows us to determine the stability of each equipment when encountering each type of fault. The worse the stability, the greater the maintenance difficulty. To reduce the error caused by relying solely on the magnitude of change, the difference between the maximum and minimum values ​​of each fault type is added to achieve a more accurate assessment of maintenance difficulty. Furthermore, to further reduce the error, a fluctuation coefficient corresponding to each fault type is introduced. By calculating the standard deviation, the handling of each fault type can be accurately determined. Obviously, the smaller the fluctuation coefficient, the greater the maintenance difficulty of that fault type.

[0073] By comprehensively analyzing the historical data of each type of fault through the above technical solutions, it is possible to determine which type of fault or which two types of faults occurring simultaneously will lead to increased maintenance difficulty for each medical device. This allows for focused monitoring and priority predictive maintenance to reduce the adverse effects of this type of fault on the equipment. It also reduces the cascading impact of prolonged offline status on other devices, which could lead to excessive short-term usage pressure.

[0074] The process of establishing a priority maintenance equipment list based on the analysis results is as follows:

[0075] The condition F(P) will be satisfied. j K j Add devices with a value ≥ F(P, K)0 to the list of maintained devices;

[0076] Sort each piece of equipment in the maintenance equipment list in descending order according to its maintenance difficulty coefficient K;

[0077] Each device is numbered sequentially as 1, 2, 3, ... N, which forms the priority maintenance list.

[0078] This embodiment provides a method for establishing a priority maintenance list, specifically, based on F(P) j K j The condition F(P, K) ≥ F(P, K) 0 is used to initially establish a list of maintenance equipment. Then, the equipment is sorted in descending order according to the maintenance difficulty coefficient of each equipment to obtain a priority maintenance equipment list. This aims to prioritize the maintenance of equipment with high maintenance difficulty, so as to avoid excessively long maintenance interruptions due to equipment deterioration. Through the above technical solution, the equipment with high maintenance difficulty can be monitored and maintained in advance in a targeted manner. This can avoid the problem of continuous offline of a certain type of equipment during the same period when maintenance resources are insufficient, and can maintain the maximum effective operation of medical equipment even in more severe situations.

[0079] It should be noted that the calculation formulas and all parameters involved in the calculations in this invention have been dimensionless beforehand. The process of dimensionless processing is well known in the industry and will not be described here.

[0080] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. An AI-based medical equipment operation and maintenance management system, characterized by, The application relates to a device safety maintenance method and device. The device monitoring module is used for monitoring the running parameters of the device in real time and serving as device monitoring data; The device fault prediction module is used for obtaining fault prediction coefficients by cleaning, fusing and processing the device monitoring data based on a pre-trained recurrent neural network model; The safety detection module is used for analyzing the safety of the device according to the fault prediction coefficients of the device fault diagnosis module and establishing a priority maintenance device list according to the analysis result; The operation and maintenance execution module is used for executing the maintenance task according to the priority maintenance list of the safety detection module; The process of analyzing the safety of the device is as follows: Obtaining a failure prediction coefficient of a device ; According to the fault prediction coefficient And the maintenance difficulty coefficient Establish a device safety analysis model The expression is: ； in, , These are the weighting coefficients. This represents the extreme value difference of the fault prediction coefficient of the current equipment per unit time. , These represent the maximum and minimum values ​​of the fault prediction coefficient for the current equipment per unit time, respectively. For the first Types of faults For the first The maintenance difficulty coefficient corresponding to each type of fault. For the first Preset fixed coefficients for each type of fault. , The threshold values ​​for the current equipment's fault prediction coefficient and maintenance difficulty coefficient; Through a device security analysis model Determining whether the current device is secure; The process of obtaining the maintenance difficulty coefficient is as follows: mechanical failure, software failure, electrical failure are marked as , , ; Obtaining the change curve of the frequency of mechanical failure, software failure, and electrical failure per unit time over time by fitting , , ; According to the change range and extreme value of the change curve corresponding to the frequency of each fault type, the maintenance difficulty coefficient corresponding to each fault type is calculated , , ; The , , Expression is: ； ； ； wherein, , , is a reference curve, , are maximum and minimum values of the frequency of occurrence of mechanical faults, respectively, are maximum and minimum values of the frequency of occurrence of software faults, respectively, are maximum and minimum values of the frequency of occurrence of electrical faults, respectively; , , are the coefficients of variation corresponding to each type of fault, respectively, , are the start and end points of the statistical period, respectively. 2.The AI-based medical device operation and maintenance management system of claim 1, wherein The running parameters include temperature, pressure, current, voltage and service time length; the device fault types include mechanical faults, electrical faults and software faults. 3.The AI-based medical device operation and maintenance management system of claim 1, wherein The process of judging whether the current device is safe is as follows: The failure prediction coefficient and the maintenance difficulty coefficient of the first device are substituted into a device safety analysis model to calculate a failure estimate of the first device. ;​​​​ comparing with a preset safety threshold value; If then the current device is judged as unsafe; Otherwise, it is judged that the current device is safe. 4.The AI-based medical device operation and maintenance management system of claim 1, wherein The process of obtaining the fluctuation coefficient corresponding to each fault type is as follows: The first Number of times each type of failure occurs and the duration of each maintenance procedure. Substitute into the formula ; wherein, is the time, is the mean of the maintenance processing duration for the fault type. 5.The AI-based medical device operation and maintenance management system of claim 3, wherein The process of establishing the priority maintenance device list according to the analysis result is as follows: adding a device that meets the criteria to the maintenance device list; arranging each of the devices in the list of maintenance devices in order of a maintenance difficulty coefficient in descending order; Each device is sequentially numbered as 1, 2, 3,..., N, that is, as a priority maintenance device list.

Images