An intelligent temperature and humidity monitoring method and system for an IVD reagent library
By combining three layers of monitoring equipment and a deep learning model, the temperature and humidity of the IVD reagent library are dynamically controlled, solving the problems of insufficient monitoring accuracy and low response efficiency. This achieves high-precision prediction and rapid response, reduces energy consumption, and meets stringent regulatory requirements.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-04-13
- Publication Date
- 2026-07-14
AI Technical Summary
Existing temperature and humidity monitoring technologies in IVD reagent libraries suffer from insufficient monitoring accuracy and scope, making it impossible to predict abnormal trends. This leads to the failure of high-value reagents due to local temperature and humidity drift, and also results in low response efficiency.
A three-layer monitoring system is used in conjunction with a deep learning anomaly prediction model. The data collection frequency is dynamically set, and real-time prediction and control are performed through edge computing nodes. The system optimizes equipment operation by using a reagent-equipment-environment linkage model and generates audit trail reports by combining blockchain storage.
It achieves high-precision, multi-dimensional temperature and humidity monitoring, enabling early prediction of anomalies and targeted adjustments, shortening response time, reducing equipment energy consumption, and meeting stringent regulatory requirements.
Smart Images
Figure CN122384901A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of reagent storage and monitoring technology, specifically to an intelligent temperature and humidity monitoring method and system for IVD reagent libraries. Background Technology
[0002] Currently, in the management and storage of in vitro diagnostic reagents, temperature and humidity monitoring is a crucial step in ensuring the effectiveness, stability, and accuracy of test results. Existing IVD reagent storage temperature and humidity monitoring technologies mostly employ conventional environmental monitoring methods, such as evenly distributing several temperature and humidity sensors within the storage room, setting fixed upper and lower alarm thresholds, and triggering audible and visual alarms when the monitored data exceeds the preset range, requiring manual intervention by management personnel. Some technologies have introduced an Internet of Things (IoT) architecture, enabling remote viewing of monitoring data and basic historical data recording functions, and pushing alarm information via SMS or an app. In terms of control, existing systems are usually linked to the central air conditioning or independent dehumidifiers in the storage room. When temperature and humidity exceed the limits, they simply and crudely start or stop the relevant equipment, attempting to bring the environmental parameters back to the normal range by running it at full power.
[0003] Existing technologies have many insurmountable technical defects in practical applications, including the following:
[0004] 1. Insufficient monitoring accuracy and scope: Existing technologies ignore the temperature and humidity differences in different areas of IVD reagent storage (such as doorways, windowsills, and the top and bottom shelves), and cannot detect changes in the microenvironment inside reagent packaging boxes, resulting in a large number of monitoring blind spots. For specific reagents that require strict cold chain (2~8℃) or freezing (below -20℃) storage, irreversible failures can easily occur due to local instantaneous temperature and humidity drift. 2. Existing technologies lack the ability to predict abnormal trends. They often only issue alarms after the temperature and humidity have exceeded the standards and the reagents have been exposed to unqualified environments. By this time, damage has already occurred, and the response process relies on manual operation, resulting in low processing efficiency and an inability to intervene in the early stages of anomalies. This makes them severely inadequate in protecting high-value and highly sensitive IVD reagents.
[0005] Based on the above problems, there is an urgent need to design an intelligent temperature and humidity monitoring system for IVD reagent libraries to solve the technical problems mentioned above. Summary of the Invention
[0006] In view of this, the purpose of the present invention is to provide an intelligent temperature and humidity monitoring method and system for IVD reagent libraries, so as to solve the technical problems of insufficient monitoring accuracy and dimensions and lack of predictive function for abnormal trends in the prior art.
[0007] According to a first aspect of the present invention, an intelligent temperature and humidity monitoring method for an IVD reagent library is provided, applied to a temperature and humidity monitoring device, comprising: When a monitoring instruction is received, the initial data collection frequency is dynamically set based on the sensitivity coefficient and remaining shelf life identified before the reagents are stored. The three-layer monitoring equipment deployed in the reagent library is activated according to the initial data acquisition frequency to collect data, and the data collected by the three-layer monitoring equipment is sent to the processing layer edge computing node and cloud server of the temperature and humidity monitoring equipment. The edge computing nodes of the processing layer run a pre-trained deep learning anomaly prediction model. The model takes the real-time data collected by the three-layer monitoring equipment, the sensitivity coefficient, the historical temperature and humidity change curve, the reservoir environmental parameters, and the external meteorological data as input to obtain the probability and degree of deviation of the temperature and humidity of each monitoring point exceeding the tolerance threshold of its associated reagent within a future preset time window, and classifies the anomaly level. The prediction results of the deep learning anomaly prediction model are input into the preset reagent-equipment-environment linkage model, and the operating parameters of the corresponding regional control equipment are dynamically adjusted according to the output of the linkage model. All received data is encrypted and stored using blockchain technology on a cloud server. Based on a preset time period, an audit trail report for a specific reagent batch is automatically generated using the data stored on the cloud server.
[0008] Preferably, before receiving the monitoring instruction, it also includes: The temperature and humidity tolerance thresholds, storage periods, and sensitivity coefficients of different types of IVD reagents were obtained in advance, and an IVD reagent characteristic database was constructed. By deploying a reagent information collection unit at the entrance of the reagent warehouse, the identification mark on the reagent packaging is automatically read when the reagent is put into the warehouse to obtain the type, expiration date and temperature and humidity storage requirements of the current batch of reagents; The information collected by the reagent information acquisition unit is compared and verified with the parameters in the IVD reagent characteristic database. If there is no abnormality, the reagent, the reagent storage location, and the monitoring equipment deployed at that location are automatically logically associated.
[0009] Preferably, the initial data collection frequency is dynamically set based on the sensitivity coefficient and remaining shelf life identified before the reagents are stored, including: The reagent's unique identifier is determined by scanning a QR code. Based on the unique identifier, the sensitivity coefficient and remaining shelf life of the reagent are automatically retrieved from the IVD reagent characteristic database; Based on the preset risk assessment weights, the sensitivity coefficient and the remaining shelf life are weighted and calculated to obtain the risk level of the reagent within the current storage period, and the initial data collection frequency is dynamically matched according to the risk level.
[0010] Preferably, the three-layer monitoring device deployed within the reagent storage includes: Miniature embedded temperature and humidity sensors, either embedded in reagent packaging boxes or in direct contact with reagent packaging, are used to collect temperature and humidity data of the reagent storage microenvironment. Shelf level monitoring nodes are deployed at different levels of each shelf to collect temperature and humidity data for local areas of the shelf. Environmental monitoring terminals deployed in different physical areas within the reagent storage facility are used to collect baseline data on temperature and humidity of the overall environment of the storage facility.
[0011] Preferably, the edge computing nodes of the processing layer run a pre-trained deep learning anomaly prediction model, including: The edge computing nodes of the processing layer parse the received real-time data and restore the real-time temperature and humidity values, timestamps, and sensitivity coefficients of the corresponding reagents at each monitoring point of the three-layer monitoring equipment. The edge computing nodes in the processing layer obtain historical temperature and humidity variation curves, reservoir environmental parameters, and external meteorological data from the cloud server. The real-time data stream obtained from the analysis, historical temperature and humidity change curves, reservoir environmental parameters, and external meteorological data are fused to construct a time series feature vector in the input format; The edge computing node loads a pre-trained deep learning anomaly prediction model, and uses the deep learning anomaly prediction model to perform forward inference calculation on the time series feature vector, outputting the probability and prediction deviation of the temperature and humidity of each monitoring point exceeding the tolerance threshold of its associated reagent within a future preset time window. Based on the prediction results output by the deep learning anomaly prediction model, the anomaly level of each reagent is classified. The prediction results and anomaly levels are synchronized to the cloud server.
[0012] Preferably, after the edge computing nodes of the processing layer classify reagents into anomaly levels, the following steps are also included: When a minor warning is detected, a reminder message is pushed to the administrator's terminal; When a moderate warning is detected, a pre-adjustment command is automatically sent to the control equipment, and a warning message is simultaneously pushed to the administrator terminal. When a severe warning is detected, an audible and visual alarm is immediately triggered, backup control equipment is activated, and emergency alarm information is pushed to preset multi-level administrators and monitoring terminals to lock down the abnormal area.
[0013] Preferably, the method further includes: The reagent-equipment-environment linkage model integrates the real-time temperature and humidity requirements of reagents stored in the comprehensive area, the deviation between the current temperature and humidity and the target values, changes in the external environment, and the energy efficiency curve of the equipment itself. Based on the reagent-equipment-environment linkage model, the following dynamic control algorithm formula is constructed:
[0014] Among them, P For optimal equipment operating power; T set The target temperature is set based on the characteristics of the reagents stored in the area; H set The target humidity is set based on the characteristics of the reagents stored in the area; T act This is the current measured temperature; H act The current measured humidity; ΔT tol The allowable temperature deviation range for the reagent; ΔH tol λ1 represents the allowable humidity deviation range of the reagent; E(P) represents the energy consumption function of the equipment under power P; λ1, λ2, and λ3 are the weighting coefficients between control accuracy, humidity stability, and energy consumption, respectively.
[0015] Preferably, using data stored on a cloud server, an audit trail report for a specific reagent batch is automatically generated, including: The system obtains the query conditions input by the user and extracts all related data for a specified time period or a specific reagent batch based on the query conditions. The extracted data range includes the original temperature and humidity monitoring values, equipment operation logs, abnormal alarm records, control command issuance records, and on-site handling operation records of the administrator. The extracted data is processed and converted into formats automatically to generate an audit trail report that meets the preset standards. The audit trail report includes temperature and humidity change curves, closed-loop processing records of abnormal events from occurrence to completion, detailed information on reagent batches, and verification status certificates of all equipment related to the monitoring period.
[0016] According to a second aspect of the present invention, an intelligent temperature and humidity monitoring system for an IVD reagent library is provided, comprising: The control module is used to dynamically set the initial data acquisition frequency based on the sensitivity coefficient and remaining shelf life identified before the reagents are put into storage when a monitoring command is received. The sensing module is used to start the three-layer monitoring device deployed in the reagent library to collect data according to the initial data collection frequency, and send the data collected by the three-layer monitoring device to the processing layer edge computing node and cloud server of the temperature and humidity monitoring device. The transmission module is used to encrypt and send the collected data to the processing module and the cloud server through a redundant transmission network; The processing module is used to run a pre-trained deep learning anomaly prediction model using the edge computing nodes of the processing layer. It takes the real-time data collected by the three-layer monitoring equipment, the sensitivity coefficient, the historical temperature and humidity change curve, the reservoir environmental parameters and external meteorological data as the model input to obtain the probability and prediction deviation of the temperature and humidity of each monitoring point exceeding the tolerance threshold of its associated reagent within a future preset time window, and classifies the anomaly level. The control module is used to input the prediction results of the deep learning anomaly prediction model into the preset reagent-equipment-environment linkage model, and dynamically adjust the operating parameters of the corresponding area control equipment according to the output of the linkage model. The storage and traceability module is used to encrypt and store all received data using blockchain technology on a cloud server. Based on a preset time period, it automatically generates an audit trail report for a specific reagent batch using the data stored on the cloud server.
[0017] Preferably, the sensing module includes: Miniature embedded temperature and humidity sensor, embedded in reagent packaging box or in direct contact with reagent packaging, is used to collect temperature and humidity data of the microenvironment in which the reagent is stored; Shelf layer monitoring nodes are deployed at different levels of each shelf to collect temperature and humidity data for localized areas of the shelf. The environmental monitoring terminal is deployed in different physical areas within the reagent storage facility to collect baseline data on temperature and humidity of the overall environment.
[0018] The technical solution provided by this invention may include the following beneficial effects: It is understood that the technical solution presented in this invention, upon receiving a monitoring instruction, can determine the data acquisition frequency based on the reagent's sensitivity coefficient and remaining shelf life, and activate the three-layer monitoring equipment within the reagent storage facility to collect data. Edge computing nodes, through a deep learning anomaly prediction model, combined with real-time acquired data and relevant data, predict the probability and degree of deviation of temperature and humidity exceeding the reagent's tolerance threshold at monitoring points, and classify the anomaly level. Then, relying on a reagent-equipment-environment linkage model, the operating parameters of the corresponding area's control equipment are dynamically adjusted. This technical solution, through three layers of monitoring equipment, monitors the temperature and humidity of the reagent storage facility from multiple aspects and dimensions, achieving high monitoring accuracy and enabling targeted control of reagents through the use of a deep learning anomaly prediction model.
[0019] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit the invention. Attached Figure Description
[0020] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.
[0021] Figure 1 This is a schematic diagram illustrating the steps of an intelligent temperature and humidity monitoring method for an IVD reagent library according to an exemplary embodiment; Figure 2 This is a schematic block diagram illustrating an intelligent temperature and humidity monitoring system for an IVD reagent library according to an exemplary embodiment. Detailed Implementation
[0022] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Rather, they are merely examples of apparatuses and methods consistent with some aspects of the invention as detailed in the appended claims.
[0023] In one embodiment, Figure 1 This is a schematic diagram illustrating the steps of an intelligent temperature and humidity monitoring method for an IVD reagent library according to an exemplary embodiment. See also... Figure 1 This invention provides an intelligent temperature and humidity monitoring method for IVD reagent libraries, applied to temperature and humidity monitoring equipment, including: Step S11: When a monitoring instruction is received, the initial data acquisition frequency is dynamically set based on the sensitivity coefficient and remaining validity period identified before the reagents are stored.
[0024] Upon receiving a monitoring command, the temperature and humidity monitoring equipment dynamically determines the initial data collection frequency based on the information identification and data storage completed before the reagents are stored. During the reagent storage phase, the system can identify the sensitivity coefficient and remaining expiration date of the reagents. When a monitoring command is received, the system calculates the sensitivity coefficient and remaining expiration date of the reagents according to preset risk assessment weights, classifies the reagent storage risk level, and then matches the collection frequency according to the risk level. For high-risk reagents with high sensitivity coefficients and nearing their expiration date, a high-frequency collection mode is activated, while a low-frequency collection mode is used for regular low-risk reagents. This ensures that the collection frequency is accurately matched with the characteristics of the reagents themselves and the storage risk, solving the problem that traditional fixed collection frequencies cannot meet the personalized monitoring needs of different reagents.
[0025] In a preferred embodiment, the initial data collection frequency is dynamically set based on the sensitivity coefficient and remaining shelf life identified before the reagents are stored, including: The reagent's unique identifier is identified by scanning a code; based on the unique identifier, the sensitivity coefficient and remaining shelf life of the reagent are automatically retrieved from the IVD reagent characteristic database; according to the preset risk assessment weight, the sensitivity coefficient and the remaining shelf life are weighted and calculated to obtain the risk level of the reagent within the current storage period, and the initial data collection frequency is dynamically matched according to the risk level.
[0026] In practical applications, the calculated initial acquisition frequency command is sent through the transmission layer to the underlying micro-embedded temperature and humidity sensor associated with the reagent storage location. After receiving the command, the sensor automatically adjusts its working mode and begins to acquire data at the frequency of the new configuration information. At the same time, the configuration information is synchronized in real time to the reagent association record on the cloud server to ensure that all subsequent monitoring data matches the real-time risk status of the reagent.
[0027] Step S12: Start the three-layer monitoring device deployed in the reagent library to collect data according to the initial data collection frequency, and send the data collected by the three-layer monitoring device to the processing layer edge computing node and cloud server of the temperature and humidity monitoring device.
[0028] In a preferred embodiment, a three-layer monitoring device deployed within a reagent storage facility includes: Miniature embedded temperature and humidity sensors, either embedded in reagent packaging boxes or in direct contact with reagent packaging, are used to collect temperature and humidity data of the reagent storage microenvironment. Shelf level monitoring nodes are deployed at different levels of each shelf to collect temperature and humidity data for local areas of the shelf. Environmental monitoring terminals deployed in different physical areas within the reagent storage facility are used to collect baseline data on temperature and humidity of the overall environment of the storage facility.
[0029] In practical applications, all monitoring devices operate in real time according to the data acquisition frequency set in step S11, and the collected temperature and humidity data are encrypted and sent to the processing layer edge computing node and cloud server of the temperature and humidity monitoring device through the redundant transmission network of the transmission layer using the AES encryption algorithm. At the same time, the device supports the function of resuming interrupted transmission to ensure the stability and integrity of data transmission.
[0030] Step S13: The edge computing node of the processing layer runs the pre-trained deep learning anomaly prediction model. The data collected in real time by the three-layer monitoring equipment, the sensitivity coefficient, the historical temperature and humidity change curve, the reservoir environmental parameters and external meteorological data are used as model inputs to obtain the probability and prediction deviation of the temperature and humidity of each monitoring point exceeding the tolerance threshold of its associated reagent within the future preset time window, and classify the anomaly level.
[0031] The deep learning anomaly prediction model uses a long short-term memory neural network as its core architecture and is deployed locally after being trained on historical data.
[0032] In practical applications, the deep learning anomaly prediction model decrypts, verifies the integrity of, and constructs feature vectors for the input data. It then outputs the probability and degree of deviation of the temperature and humidity at each monitoring point exceeding the tolerance threshold of its associated reagent within a preset time window. Based on the prediction results, it automatically classifies the anomaly levels, including mild warning, moderate warning, and severe warning, each corresponding to different response actions. It automatically sends pre-adjustment instructions to the control layer of the temperature and humidity monitoring equipment and synchronizes the prediction results and response actions to the cloud server.
[0033] In a preferred embodiment, the edge computing nodes of the processing layer run a pre-trained deep learning anomaly prediction model, including: After the temperature and humidity data collected by the sensing layer is encrypted and transmitted through the redundant network of the transmission layer, the edge computing nodes of the processing layer decrypt and verify the integrity of the received real-time data, and then parse it to restore the real-time temperature and humidity values, timestamps and sensitivity coefficients of the corresponding reagents of each monitoring point of the three-layer monitoring equipment.
[0034] The edge computing nodes in the processing layer obtain historical temperature and humidity variation curves, reservoir environmental parameters, and external meteorological data from the cloud server.
[0035] The real-time data stream obtained from the analysis, historical temperature and humidity change curves, reservoir environmental parameters, and external meteorological data are fused together to construct a time series feature vector in the input format.
[0036] The edge computing node loads a pre-trained deep learning anomaly prediction model, and uses the deep learning anomaly prediction model to perform forward inference calculation on the time series feature vector, outputting the probability and prediction deviation of the temperature and humidity of each monitoring point exceeding the tolerance threshold of its associated reagent within a future preset time window. Based on the prediction results output by the deep learning anomaly prediction model, the anomaly levels of each reagent are classified as follows: The comprehensive anomaly index is defined according to the following formula:
[0037] Among them, A i,t For comprehensive anomaly indicators, α, β, and γ are preset weighting coefficients that satisfy α + β + γ = 1, p i,t δ represents the probability, as output by the model, that the temperature and humidity at monitoring point i exceed a threshold within a preset time window in the future. i,t To predict the degree of temperature and humidity deviation, i,t ΔT represents the degree of deviation between the currently measured temperature and humidity and the threshold. tol,iThis represents the allowable deviation range of the reagent associated with monitoring point i.
[0038] Then, based on the comprehensive abnormality indicators, the abnormality level L is determined. i,t :
[0039] Among them, the preset abnormality level thresholds θ1<θ2<θ3 are used to divide different warning levels.
[0040] The node determines the anomaly level based on the prediction results and real-time data, triggers the corresponding response action, and synchronizes the results to the cloud server.
[0041] In a preferred embodiment, after the edge computing nodes of the processing layer classify the reagents into anomaly levels, the method further includes: When a minor warning is detected, a reminder message is pushed to the administrator terminal; when a moderate warning is detected, a pre-adjustment command is automatically issued to the control equipment, and a warning message is pushed to the administrator terminal at the same time; when a severe warning is detected, an audible and visual alarm is immediately triggered, the backup control equipment is activated, and an emergency alarm message is pushed to the preset multi-level administrators and monitoring terminals to lock the abnormal area.
[0042] Step S14: Input the prediction results of the deep learning anomaly prediction model into the preset reagent-equipment-environment linkage model, and dynamically adjust the operating parameters of the corresponding area control equipment according to the output of the linkage model.
[0043] In a preferred embodiment, the reagent-equipment-environment linkage model integrates the real-time temperature and humidity requirements of the reagents stored in the integrated area, the deviation between the current temperature and humidity and the target values, changes in the external environment, and the energy efficiency curve of the equipment itself. Based on the reagent-equipment-environment linkage model, the following dynamic control algorithm formula is constructed:
[0044] Among them, P For optimal equipment operating power; T set The target temperature is set based on the characteristics of the reagents stored in the area; H set The target humidity is set based on the characteristics of the reagents stored in the area; T act This is the current measured temperature; H act The current measured humidity; ΔT tol The allowable temperature deviation range for the reagent; ΔH tol λ1 represents the allowable humidity deviation range of the reagent; E(P) represents the energy consumption function of the equipment under power P; λ1, λ2, and λ3 are the weighting coefficients between control accuracy, humidity stability, and energy consumption, respectively.
[0045] Preferably, once reagents from a certain area have been shipped out, the temperature and humidity control thresholds for that area are automatically adjusted to energy-saving mode, and the monitoring frequency and equipment control intensity for that area are reduced.
[0046] Step S15: Use a cloud server to encrypt and store all received data using blockchain technology. Based on a preset time period, use the data stored on the cloud server to automatically generate an audit trail report for a specific reagent batch.
[0047] The cloud server encrypts and stores all received raw temperature and humidity data, equipment operation logs, abnormal alarm records, control instructions, and administrator handling records using blockchain technology, forming an immutable data chain. Simultaneously, the temperature and humidity monitoring equipment automatically captures all data related to a specified time period or specific reagent batch according to regulatory requirements. The extracted data includes raw monitoring values, equipment logs, alarm records, control instructions, and administrator handling records, generating an audit trail report compliant with GMP, ISO13485, and FDA21CFRPart11. This report fully includes temperature and humidity change curves, closed-loop processing records of abnormal events from occurrence to resolution, detailed information on reagent batches involved, and proof of all equipment calibration status related to the monitoring period, and supports one-click export.
[0048] It is understood that the technical solution presented in this invention addresses the shortcomings of being unable to adapt to the personalized needs of different reagents by constructing an IVD reagent characteristic database and automatically identifying its sensitivity coefficient and remaining shelf life when the reagents are put into storage, and dynamically setting the initial sampling frequency based on the reagent risk level; by deploying a three-layer monitoring architecture consisting of miniature embedded sensors, shelf-level nodes, and warehouse terminals, it eliminates all-dimensional monitoring blind spots from the inside of the reagent packaging to the entire warehouse; and by running a deep learning prediction model with a long short-term memory neural network on edge computing nodes, it achieves early prediction and response to future abnormal temperature and humidity trends, upgrading passive alarms to proactive early warnings.
[0049] In a preferred embodiment, before receiving the monitoring instruction, the method further includes: Temperature and humidity tolerance thresholds, storage periods, and sensitivity coefficients of different types of IVD reagents are pre-entered into temperature and humidity monitoring equipment to construct an IVD reagent characteristic database. A reagent information acquisition unit deployed at the reagent storage entrance automatically reads the identification mark on the reagent packaging when reagents are received, obtaining the type, expiration date, and temperature and humidity storage requirements of the current batch of reagents. The information collected by the reagent information acquisition unit is compared and verified with the parameters in the IVD reagent characteristic database. If no abnormalities are found, the reagent, its storage location, and the monitoring equipment deployed at that location are automatically logically associated.
[0050] In a preferred embodiment, step S15 utilizes data stored on a cloud server to automatically generate an audit trail report for a specific reagent batch, including: The system obtains the query conditions input by the user and automatically extracts all related data for a specified time period or a specific reagent batch from the blockchain storage unit based on the query conditions. The extracted data range includes the original temperature and humidity monitoring values, equipment operation logs, abnormal alarm records, control command issuance records, and on-site handling operation records of the administrator. The extracted data is processed and converted into formats to generate an audit trail report that meets the requirements of GMP, ISO13485 and FDA21CFRPart11 standards. The audit trail report includes temperature and humidity change curves, closed-loop processing records of abnormal events from occurrence to completion, detailed information on reagent batches, and verification status certificates of all equipment related to the monitoring period.
[0051] This technical solution utilizes a reagent characteristic-related dynamic acquisition frequency mechanism and a three-layer monitoring architecture to improve the monitoring frequency of highly sensitive or near-expiry reagents to the minute level, achieving a microenvironment temperature and humidity capture accuracy of ±0.1℃. A deep learning prediction model combined with multi-source data fusion technology, along with tiered responses for mild, moderate, and severe warnings, reduces the anomaly response time from tens of minutes of traditional manual intervention to seconds of automatic adjustment, effectively preventing irreversible damage to reagents caused by response lag. The reagent-equipment-environment linkage control algorithm optimizes equipment operating power and duration in real time, based on actual... Deploying tests can reduce the energy consumption of refrigeration and dehumidification equipment by 20%-30%. At the same time, it automatically switches to energy-saving mode after reagents are released from the warehouse, realizing the organic unity of on-demand energy supply and precise control. The data encryption storage and automated audit report generation function based on blockchain not only ensures the immutability and traceability of the entire life cycle from data collection to destruction, but also shortens the data preparation time for audits from several days of manual work to minutes of one-click export. The audit report fully includes temperature and humidity curves, abnormal closed-loop records, reagent batch information and equipment calibration status, meeting the increasingly stringent regulatory requirements of the IVD industry.
[0052] In another embodiment, see Figure 2 A smart temperature and humidity monitoring system for IVD reagent storage is provided, comprising: The control module 101 is used to dynamically set the initial data acquisition frequency based on the sensitivity coefficient and remaining shelf life identified before the reagents are put into storage when a monitoring instruction is received. The sensing module 102 is used to start the three-layer monitoring device deployed in the reagent library to collect data according to the initial data collection frequency, and send the data collected by the three-layer monitoring device to the processing layer edge computing node and cloud server of the temperature and humidity monitoring device. The transmission module 103 is used to encrypt and send the collected data to the processing module and the cloud server through a redundant transmission network; Processing module 104 is used to run a pre-trained deep learning anomaly prediction model using the edge computing nodes of the processing layer. The model takes the real-time data collected by the three-layer monitoring equipment, the sensitivity coefficient, the historical temperature and humidity change curve, the reservoir environmental parameters and external meteorological data as input to obtain the probability and prediction deviation of the temperature and humidity of each monitoring point exceeding the tolerance threshold of its associated reagent within a future preset time window, and classifies the anomaly level. The control module 105 is used to input the prediction results of the deep learning anomaly prediction model into the preset reagent-equipment-environment linkage model, and dynamically adjust the operating parameters of the corresponding area control equipment according to the output of the linkage model. The storage and traceability module 106 is used to encrypt and store all received data using blockchain technology on a cloud server, and automatically generate audit trail reports for specific reagent batches based on the data stored on the cloud server according to a preset time period.
[0053] Preferably, the sensing module includes: Miniature embedded temperature and humidity sensor, embedded in reagent packaging box or in direct contact with reagent packaging, is used to collect temperature and humidity data of the microenvironment in which the reagent is stored; Shelf layer monitoring nodes are deployed at different levels of each shelf to collect temperature and humidity data for localized areas of the shelf. The environmental monitoring terminal is deployed in different physical areas within the reagent storage facility to collect baseline data on temperature and humidity of the overall environment.
[0054] This technical solution addresses the shortcomings of inability to adapt to the personalized needs of different reagents by constructing an IVD reagent characteristic database and automatically identifying the sensitivity coefficient and remaining shelf life of reagents upon receipt. It dynamically sets the initial sampling frequency based on the reagent risk level. A three-layer monitoring architecture consisting of miniature embedded sensors, shelf-level nodes, and warehouse terminals eliminates blind spots in comprehensive monitoring from inside the reagent packaging to the entire warehouse. A deep learning prediction model combined with multi-source data fusion technology, along with tiered responses for minor, moderate, and severe warnings, reduces the anomaly response time from tens of minutes of traditional manual intervention to seconds of automatic adjustment, effectively preventing irreversible damage to reagents caused by delayed response.
[0055] It is understood that the same or similar parts in the above embodiments can be referred to each other, and the contents not described in detail in some embodiments can be referred to the same or similar contents in other embodiments.
[0056] It should be noted that in the description of this invention, the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance. Furthermore, in the description of this invention, unless otherwise stated, "a plurality of" means at least two.
[0057] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing a particular logical function or process, and the scope of the preferred embodiments of the invention includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as will be understood by those skilled in the art to which embodiments of the invention pertain.
[0058] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.
[0059] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.
[0060] Furthermore, the functional units in the various embodiments of the present invention can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.
[0061] The storage media mentioned above can be read-only memory, disk, or optical disk, etc.
[0062] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.
[0063] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.
Claims
1. A method for intelligent temperature and humidity monitoring in an IVD reagent library, characterized in that, Applications include temperature and humidity monitoring equipment, including: When a monitoring instruction is received, the initial data collection frequency is dynamically set based on the sensitivity coefficient and remaining shelf life identified before the reagents are stored. The three-layer monitoring equipment deployed in the reagent library is activated according to the initial data acquisition frequency to collect data, and the data collected by the three-layer monitoring equipment is sent to the processing layer edge computing node and cloud server of the temperature and humidity monitoring equipment. The edge computing nodes of the processing layer run a pre-trained deep learning anomaly prediction model. The model takes the real-time data collected by the three-layer monitoring equipment, the sensitivity coefficient, the historical temperature and humidity change curve, the reservoir environmental parameters, and the external meteorological data as input to obtain the probability and degree of deviation of the temperature and humidity of each monitoring point exceeding the tolerance threshold of its associated reagent within a future preset time window, and classifies the anomaly level. The prediction results of the deep learning anomaly prediction model are input into the preset reagent-equipment-environment linkage model, and the operating parameters of the corresponding regional control equipment are dynamically adjusted according to the output of the linkage model. All received data is encrypted and stored using blockchain technology on a cloud server. Based on a preset time period, an audit trail report for a specific reagent batch is automatically generated using the data stored on the cloud server.
2. The method according to claim 1, characterized in that, Before receiving the monitoring instruction, it also includes: The temperature and humidity tolerance thresholds, storage periods, and sensitivity coefficients of different types of IVD reagents were obtained in advance, and an IVD reagent characteristic database was constructed. By deploying a reagent information collection unit at the entrance of the reagent warehouse, the identification mark on the reagent packaging is automatically read when the reagent is put into the warehouse to obtain the type, expiration date and temperature and humidity storage requirements of the current batch of reagents; The information collected by the reagent information acquisition unit is compared and verified with the parameters in the IVD reagent characteristic database. If there is no abnormality, the reagent, the reagent storage location, and the monitoring equipment deployed at that location are automatically logically associated.
3. The method according to claim 1, characterized in that, Based on the sensitivity coefficient and remaining shelf life identified before reagents are stored, the initial data collection frequency is dynamically set, including: The reagent's unique identifier is determined by scanning a QR code. Based on the unique identifier, the sensitivity coefficient and remaining shelf life of the reagent are automatically retrieved from the IVD reagent characteristic database; Based on the preset risk assessment weights, the sensitivity coefficient and the remaining shelf life are weighted and calculated to obtain the risk level of the reagent within the current storage period, and the initial data collection frequency is dynamically matched according to the risk level.
4. The method according to claim 1, characterized in that, The three-layer monitoring system deployed within the reagent storage facility includes: Miniature embedded temperature and humidity sensors, either embedded in reagent packaging boxes or in direct contact with reagent packaging, are used to collect temperature and humidity data of the reagent storage microenvironment. Shelf level monitoring nodes are deployed at different levels of each shelf to collect temperature and humidity data for local areas of the shelf. Environmental monitoring terminals deployed in different physical areas within the reagent storage facility are used to collect baseline data on temperature and humidity of the overall environment of the storage facility.
5. The method according to claim 1, characterized in that, The edge computing nodes in the processing layer run a pre-trained deep learning anomaly prediction model, including: The edge computing nodes of the processing layer parse the received real-time data and restore the real-time temperature and humidity values, timestamps, and sensitivity coefficients of the corresponding reagents at each monitoring point of the three-layer monitoring equipment. The edge computing nodes in the processing layer obtain historical temperature and humidity variation curves, reservoir environmental parameters, and external meteorological data from the cloud server. The real-time data stream obtained from the analysis, historical temperature and humidity change curves, reservoir environmental parameters, and external meteorological data are fused to construct a time series feature vector in the input format; The edge computing node loads a pre-trained deep learning anomaly prediction model, uses the deep learning anomaly prediction model to perform forward inference calculation on the time series feature vector, and outputs the probability and prediction deviation of the temperature and humidity of each monitoring point exceeding the tolerance threshold of its associated reagent within a future preset time window. Based on the prediction results output by the deep learning anomaly prediction model, the anomaly level of each reagent is classified. The prediction results and anomaly levels are synchronized to the cloud server.
6. The method according to claim 1, characterized in that, After the edge computing nodes of the processing layer classify reagents into anomaly levels, the following are also included: When a minor warning is detected, a reminder message is pushed to the administrator's terminal; When a moderate warning is detected, a pre-adjustment command is automatically sent to the control equipment, and a warning message is simultaneously pushed to the administrator terminal. When a severe warning is detected, an audible and visual alarm is immediately triggered, backup control equipment is activated, and emergency alarm information is pushed to preset multi-level administrators and monitoring terminals to lock down the abnormal area.
7. The method according to claim 1, characterized in that, Also includes: The reagent-equipment-environment linkage model integrates the real-time temperature and humidity requirements of reagents stored in the comprehensive area, the deviation between the current temperature and humidity and the target values, changes in the external environment, and the energy efficiency curve of the equipment itself. Based on the reagent-equipment-environment linkage model, the following dynamic control algorithm formula is constructed: Among them, P For optimal equipment operating power; T set The target temperature is set based on the characteristics of the reagents stored in the area; H set The target humidity is set based on the characteristics of the reagents stored in the area; T act This is the current measured temperature; H act The current measured humidity; ΔT tol The allowable temperature deviation range for the reagent; ΔH tol λ1 represents the allowable humidity deviation range of the reagent; E(P) represents the energy consumption function of the equipment under power P; λ1, λ2, and λ3 are the weighting coefficients between control accuracy, humidity stability, and energy consumption, respectively.
8. The method according to claim 1, characterized in that, Using data stored on cloud servers, audit trail reports for specific reagent batches are automatically generated, including: The system obtains the query conditions input by the user and extracts all related data for a specified time period or a specific reagent batch based on the query conditions. The extracted data range includes the original temperature and humidity monitoring values, equipment operation logs, abnormal alarm records, control command issuance records, and on-site handling operation records of the administrator. The extracted data is processed and converted into formats automatically to generate an audit trail report that meets the preset standards. The audit trail report includes temperature and humidity change curves, closed-loop processing records of abnormal events from occurrence to completion, detailed information on reagent batches, and verification status certificates of all equipment related to the monitoring period.
9. An intelligent temperature and humidity monitoring system for an IVD reagent library, characterized in that, include: The control module is used to dynamically set the initial data acquisition frequency based on the sensitivity coefficient and remaining shelf life identified before the reagents are put into storage when a monitoring command is received. The sensing module is used to start the three-layer monitoring device deployed in the reagent library to collect data according to the initial data collection frequency, and send the data collected by the three-layer monitoring device to the processing layer edge computing node and cloud server of the temperature and humidity monitoring device. The transmission module is used to encrypt and send the collected data to the processing module and the cloud server through a redundant transmission network; The processing module is used to run a pre-trained deep learning anomaly prediction model using the edge computing nodes of the processing layer. It takes the real-time data collected by the three-layer monitoring equipment, the sensitivity coefficient, the historical temperature and humidity change curve, the reservoir environmental parameters and external meteorological data as the model input to obtain the probability and prediction deviation of the temperature and humidity of each monitoring point exceeding the tolerance threshold of its associated reagent within a future preset time window, and classifies the anomaly level. The control module is used to input the prediction results of the deep learning anomaly prediction model into the preset reagent-equipment-environment linkage model, and dynamically adjust the operating parameters of the corresponding area control equipment according to the output of the linkage model. The storage and traceability module is used to encrypt and store all received data using blockchain technology on a cloud server. Based on a preset time period, it automatically generates an audit trail report for a specific reagent batch using the data stored on the cloud server.
10. The system according to claim 9, characterized in that, The sensing module includes: Miniature embedded temperature and humidity sensor, embedded in reagent packaging box or in direct contact with reagent packaging, is used to collect temperature and humidity data of the microenvironment in which the reagent is stored; Shelf layer monitoring nodes are deployed at different levels of each shelf to collect temperature and humidity data for localized areas of the shelf. The environmental monitoring terminal is deployed in different physical areas within the reagent storage facility to collect baseline data on temperature and humidity of the overall environment.