Full-link temperature and humidity intelligent monitoring method and system for pharmaceutical production

Abstract

The invention relates to the field of environment monitoring, and discloses a full-link temperature and humidity intelligent monitoring method and system for pharmaceutical production, and the system comprises a fixed monitoring node which is used for collecting environment temperature and humidity data at a first frequency in a monitoring area; the mobile monitoring node moves in the monitoring area to collect environment temperature and humidity data at a second frequency; the central processing module comprises an identification unit for identifying a failed fixed monitoring node based on the data flow of the fixed monitoring node; the analysis unit is used for judging whether the invalid fixed monitoring node is temporarily covered or directly replaced by the mobile monitoring node based on the identification result of the identification unit; the scheduling unit is used for sending an instruction to the mobile monitoring node and executing the part or replacement operation determined by the analysis unit; the digital twin module constructs a three-dimensional model consistent with the physical monitoring area; according to the method, temporary position covering or replacement can be carried out according to actual analysis conditions when the existing environment or the monitoring point fails.

Description

Technical Field

[0001] This invention relates to the field of environmental monitoring, specifically to a method and system for intelligent monitoring of temperature and humidity across the entire pharmaceutical production chain. Background Technology

[0002] In the pharmaceutical industry, environmental temperature and humidity during the production, storage, and distribution of drugs are key parameters for ensuring drug safety, efficacy, and quality stability. Furthermore, monitoring temperature and humidity throughout the entire drug lifecycle has clear and stringent requirements. Therefore, establishing a reliable, compliant, and end-to-end intelligent temperature and humidity monitoring system is a core infrastructure and a rigid requirement for pharmaceutical companies.

[0003] Existing temperature and humidity monitoring systems typically rely on a network of fixed monitoring points deployed in critical locations. This technical solution usually includes fixed sensors with wired or wireless connections, data acquisition units, a central monitoring platform, and alarm modules.

[0004] Although this type of fixed network can achieve basic continuous monitoring, it still has some shortcomings in practical applications. For example, once the existing system is initially deployed, the location of the monitoring points remains fixed. However, the pharmaceutical production environment is dynamic, with factors such as adjustments to the layout of goods in the warehouse, the addition of new production equipment or HVAC systems, and changes in airflow patterns due to seasonal climate changes. This may cause the previously representative monitoring points to no longer accurately reflect the true environmental conditions of their respective areas, creating blind spots or distorted monitoring areas that could pose a risk to drug quality. While increasing the density of fixed points can alleviate this problem, it will bring new challenges such as wiring difficulties, a sharp increase in costs, and increased system complexity.

[0005] In summary, there is an urgent need in this field for an intelligent temperature and humidity monitoring solution that can overcome the aforementioned shortcomings. Summary of the Invention

[0006] The purpose of this invention is to provide a method and system for intelligent monitoring of temperature and humidity throughout the entire pharmaceutical production process, thereby solving the above-mentioned technical problems.

[0007] The objective of this invention can be achieved through the following technical solutions: A smart temperature and humidity monitoring system for the entire pharmaceutical manufacturing process includes: Multiple fixed monitoring nodes are set up in a distributed manner within the monitoring area and collect ambient temperature and humidity data at the highest frequency. A mobile monitoring node is used to move within a monitored area and collect ambient temperature and humidity data at a second frequency, wherein the second frequency is higher than or equal to the first frequency. A central processing module, used for communication connection with the fixed monitoring node and the mobile monitoring node, includes: The identification unit identifies failed fixed monitoring nodes based on the data stream from the fixed monitoring nodes. The analysis unit, based on the identification result of the identification unit, determines whether to temporarily replace the failed fixed monitoring node with a mobile monitoring node or to directly replace it. The scheduling unit sends instructions to the mobile monitoring node to execute the replacement or substitution operation determined by the analysis unit; The digital twin module is used to construct a three-dimensional model consistent with the physical monitoring area, to visualize the real-time location, status and data of all fixed and mobile monitoring nodes, and to support the path planning of the scheduling unit.

[0008] As a further technical solution, the identification unit identifies the failed fixed monitoring node in at least one of the following ways: The first identification method is to identify data patterns that are rigid, drift beyond the range, or conflict with physical logic through data anomaly pattern analysis. The second identification method is to calculate the correlation coefficient between the current fixed monitoring node and the preset adjacent node group through correlation decay analysis, and trigger an early warning when the correlation coefficient is continuously lower than the first threshold. The third identification method is to conduct mobile verification divergence analysis. When the data collected by the mobile monitoring node within a preset range of any fixed monitoring node differs from the data of that fixed monitoring node by more than the allowable error, an early warning is triggered for that fixed monitoring node.

[0009] As a further technical solution, the data rigidity in the first identification method is that within a preset time window, the standard deviation of the data of the fixed monitoring node is lower than the second threshold of the standard deviation benchmark value of the fixed monitoring node.

[0010] As a further technical solution, the judgment method of the analysis unit is as follows: If the cause of failure is determined to be temporary environmental disturbance or short-term recoverable calibration of fixed monitoring nodes, it is determined to be a temporary replacement; If the cause of failure is determined to be damage to the fixed monitoring node or a permanent change in the environment, it is determined to be a direct replacement.

[0011] As a further technical solution, the scheduling unit also includes: When it is determined to be a temporary replacement, the mobile monitoring point is dispatched to conduct periodic patrol monitoring of the area where the failed fixed monitoring point is located; When it is determined to be a direct substitution: Dispatch the mobile monitoring node to the location of the fixed monitoring node for long-term monitoring; A fixed monitoring node migration task is generated, the migration task including a new installation location determined based on data collected by the mobile monitoring node.

[0012] As a further technical solution, the system is communicatively connected to the warehouse management system and / or facility management system; The identification unit further includes: when receiving cargo layout change information from the warehouse management system or equipment modification information from the facility management system, marking the status of the fixed monitoring node related to the changed area as pending evaluation; The analysis unit further includes: in response to a fixed monitoring node being marked as to be evaluated, dispatching at least one mobile monitoring node to the changed area to collect confirmatory data; The scheduling unit uses the verification data collection results to perform a third identification method to determine whether the fixed monitoring node has failed.

[0013] As a further technical solution, the process of determining the new installation location is as follows: S1. Based on the location of the failed fixed monitoring node, the scheduling unit delineates a sub-area to be evaluated in the digital twin module and plans a data acquisition path covering the sub-area to be evaluated for the mobile monitoring node. The mobile monitoring node moves along the data acquisition path and collects temperature and humidity data at a third frequency higher than the second frequency; then the data with precise location information is transmitted to the central processing module in real time; the digital twin module generates a dynamic temperature and humidity field distribution map in the three-dimensional model of the sub-region to be evaluated based on the data with precise location information. S2. Within the sub-region to be evaluated defined by the digital twin module, a two-dimensional coordinate grid is established, and the data points collected by the mobile monitoring node are mapped to the nearest grid vertex to form a temperature and humidity data matrix covering the sub-region to be evaluated. S3. For each candidate location point in the grid, obtain representative indicators, stability indicators, and risk proximity indicators based on temperature and humidity data. Representative indicators and stability indicators are first weighted and summed, and then risk proximity indicators are used as index terms. After comprehensive calculation, the overall installation suitability score is obtained. S4. Sort the comprehensive installation suitability scores of all candidate locations in descending order to generate a candidate location list; S5. From the candidate location list, check each candidate point in turn from top to bottom to see if it meets the preset engineering installation conditions in the digital twin module; determine the first candidate location point that meets all engineering installation conditions as the final new installation location.

[0014] As a further technical solution, the representative indicators are obtained as follows: Calculate the average of all temperature and humidity readings for each candidate location point during the scanning period, and compare it with the overall average of all data points in the sub-region to be evaluated. Take the reciprocal of the sum of the absolute values ​​of the temperature and humidity differences between the two, so that the smaller the difference, the greater the representative index. The stability index is obtained by calculating the standard deviation of all temperature and humidity readings during the scanning period for each candidate location point. The stability index is the reciprocal of the standard deviation, i.e., stability index = 1 / (1 + standard deviation). The smaller the fluctuation, the larger the stability index. The risk proximity index is obtained as follows: based on the dynamic temperature and humidity field distribution map, all high temperature, low temperature or high fluctuation risk points exceeding the threshold are identified; the reciprocal of the sum of the Euclidean distances from the candidate location point to all identified risk points is calculated, so that the closer to the risk point, the larger the risk proximity index.

[0015] A method for intelligent monitoring of temperature and humidity across the entire pharmaceutical manufacturing process is disclosed, which is implemented based on the aforementioned intelligent monitoring system for temperature and humidity across the entire pharmaceutical manufacturing process.

[0016] The beneficial effects of this invention are: (1) This invention introduces mobile monitoring nodes and fixed nodes to form a dynamic and static monitoring system, and adopts temporary replacement or direct replacement strategies based on failure analysis results. It can deal with temporary environmental disturbances at the lowest cost, and completely solve the problem of permanent failure through long-term monitoring of mobile nodes and relocation process. While ensuring the continuity of monitoring across the entire link, it avoids the high cost caused by frequent modification of fixed networks, and achieves the best balance between operation and maintenance costs and monitoring effect. (2) This invention uses a triple mechanism of data rigidity analysis, correlation decay analysis and mobile verification deviation analysis to cross-verify the status of fixed monitoring nodes, which significantly improves the accuracy of failure identification and early warning capability. The analysis unit makes refined classification decisions based on the root cause of failure, while the scheduling unit realizes accurate scheduling and path planning of mobile nodes through a digital twin model. Attached Figure Description

[0017] The invention will now be further described with reference to the accompanying drawings.

[0018] Figure 1 This is a logical schematic diagram of the present invention. Detailed Implementation

[0019] 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.

[0020] Please see Figure 1 As shown, this invention is an intelligent temperature and humidity monitoring system for the entire pharmaceutical production chain, comprising: Multiple fixed monitoring nodes are set up in a distributed manner within the monitoring area and collect ambient temperature and humidity data at the highest frequency. A mobile monitoring node is used to move within a monitored area and collect ambient temperature and humidity data at a second frequency, wherein the second frequency is higher than or equal to the first frequency. A central processing module, used for communication connection with the fixed monitoring node and the mobile monitoring node, includes: The identification unit identifies failed fixed monitoring nodes based on the data stream from the fixed monitoring nodes. The analysis unit, based on the identification result of the identification unit, determines whether to temporarily replace the failed fixed monitoring node with a mobile monitoring node or to directly replace it. The scheduling unit sends instructions to the mobile monitoring node to execute the replacement or substitution operation determined by the analysis unit; The digital twin module is used to construct a three-dimensional model consistent with the physical monitoring area, to visualize the real-time location, status and data of all fixed and mobile monitoring nodes, and to support the path planning of the scheduling unit.

[0021] In this embodiment, to address the issue of fixed monitoring points losing accurate ambient temperature and humidity data due to environmental factors or other influences, and the significant costs associated with adding or modifying fixed monitoring points, a central processing module, fixed monitoring nodes, and mobile monitoring nodes are established. Fixed monitoring nodes are pre-positioned based on the status of the pharmaceutical factory or warehouse, while mobile monitoring nodes are used for daily inspections. Data from all monitoring nodes is transmitted to the central processing module. First, the identification unit analyzes and identifies the fixed monitoring nodes. If a potentially faulty fixed monitoring node is detected, it is promptly marked. All marked fixed monitoring nodes undergo further evaluation. This evaluation determines whether the faulty node is due to equipment damage or a permanent, significant change in the surrounding environment leading to inaccurate data collection, or whether it's a temporary change in the environment or normal equipment malfunction requiring simple maintenance. After accurately classifying these two scenarios, the analysis unit determines whether temporary replacement or direct substitution is necessary based on the specific circumstances. Finally, the scheduling unit schedules the mobile monitoring nodes for execution. The reason for setting up the above two states is that simple fixed monitoring nodes experiencing errors due to long-term operation or temporary changes in the surrounding environment are short-term problems that can be temporarily overcome. This can be achieved by isolating the data from the fixed monitoring node and replacing it with data from a new mobile monitoring node. Once restored, the fixed monitoring node can function normally. However, if the equipment is permanently damaged or the surrounding environment has permanently changed, the fixed monitoring node becomes completely ineffective in that location. Since modifying the fixed monitoring network is too costly, a mobile monitoring node is dispatched to the area to find a new location with monitoring significance to replace it. The data from the fixed monitoring node is also isolated in this case. The difference between the two is the duration of the data isolation for the fixed monitoring node. The above technical solutions, including temporary replacement and direct replacement, can address two types of monitoring failures. By supplementing with mobile monitoring nodes, the system can minimize changes to the temperature and humidity data monitoring network while improving the ability to quickly repair local failures.

[0022] Example: A large temperature-controlled warehouse of a pharmaceutical company has 50 fixed temperature and humidity sensors, i.e., fixed monitoring nodes, deployed in accordance with GMP requirements. At the same time, it is equipped with two AGVs equipped with temperature and humidity sensors, i.e., mobile monitoring nodes, as a mobile monitoring force. All nodes are connected to the central processing module through a LoRaWAN network, and a digital twin model exactly the same as the physical warehouse is built based on a 3D modeling engine. The identification unit identifies failed fixed monitoring nodes using at least one of the following methods: The first identification method is to identify data patterns that are rigid, drift beyond the range, or conflict with physical logic through data anomaly pattern analysis. The second identification method is to calculate the correlation coefficient between the current fixed monitoring node and the preset adjacent node group through correlation decay analysis, and trigger an early warning when the correlation coefficient is continuously lower than the first threshold. The third identification method is to conduct mobile verification divergence analysis. When the data collected by the mobile monitoring node within a preset range of any fixed monitoring node differs from the data of that fixed monitoring node by more than the allowable error, an early warning is triggered for that fixed monitoring node.

[0023] In this embodiment, a three-tiered diagnostic logic is used to identify fixed monitoring nodes through three identification methods to determine whether a failure has occurred. Specifically, data anomaly pattern analysis is a preliminary screening used to discover obvious equipment failures; correlation decay analysis is a deep insight used to discover hidden failures where the equipment itself is normal but fails due to environmental changes; and mobile verification deviation analysis is the final decision, using the mobility of mobile nodes to verify the results of the first two analyses on-site.

[0024] Data stagnation: This means that the sensor readings do not fluctuate at all or fluctuate far below the historical normal level within a set time, such as 1 hour. This may mean that the sensor is damaged or blocked.

[0025] Out-of-range drift: This means that the reading continuously and stably deviates from other measuring points in the same area, but does not reach the alarm threshold, which may be due to sensor drift.

[0026] Logical conflict: This refers to a contradiction in physical logic between a reading at a certain point and data from adjacent points, air conditioning supply / return air temperature, external ambient temperature, etc. For example, the temperature at a point farther away from the heat source may be higher. Through cross-validation using multiple methods, the misjudgment of a single method is avoided to the greatest extent, ensuring that subsequent scheduling operations are only triggered when there is a true failure.

[0027] Example: Data anomaly mode: The system found that the temperature reading of the Numb15 sensor located in the east corner of the warehouse remained at 5.0℃ for 2 hours without any fluctuation, which is called data stagnation. The identification module judged that it was suspected of failure. Correlation decay: System calculations revealed that the temperature correlation coefficient between the sensor Numb08 located in the main channel and its five surrounding nodes dropped from over 0.95 in history to 0.3, triggering an early warning. Mobile verification discrepancy: When the AGV patrolled near Number08, the instantaneous temperature collected was 3.5℃, while Number08 itself displayed 5.1℃, a difference far exceeding the allowable error of ±0.5℃, triggering a check. By cross-checking through multiple analysis methods, the accuracy and reliability of failure identification were greatly improved, avoiding false alarms or missed alarms due to a single judgment method.

[0028] The data rigidity in the first identification method means that within a preset time window, the standard deviation of the data of the fixed monitoring node is lower than the second threshold of the standard deviation benchmark value of the fixed monitoring node.

[0029] In this embodiment, rigidity is defined by comparing the difference between short-term data volatility within a preset time window and the long-term historical volatility of the standard deviation benchmark value. This transforms the commonly used absolute threshold into a relative value judgment. For example, sensors near the doorway naturally have larger fluctuations, allowing this judgment to adapt to sensors in different locations. This improves the sensitivity and specificity of sensor fault identification because it can effectively distinguish between environmental stability and sensor malfunction, reducing false alarms. At the same time, it enables the system to adapt to the characteristics of different microenvironments within the monitored area.

[0030] Example: The system's preset time window is 30 minutes, and the second threshold is 20%. The standard deviation of sensor Numb5's data within 30 minutes is 0.05℃, while its historical standard deviation baseline is 0.3℃. Calculations show that 0.05 / 0.3 ≈ 16.7% < 20%, therefore the system determines that data stagnation has occurred.

[0031] The judgment method of the analysis unit is as follows: If the cause of failure is determined to be temporary environmental disturbance or short-term recoverable calibration of fixed monitoring nodes, it is determined to be a temporary replacement; If the cause of failure is determined to be damage to the fixed monitoring node or a permanent change in the environment, it is determined to be a direct replacement.

[0032] This embodiment provides a specific logic for determining the cause of failure. After identifying a failure of a fixed monitoring node, instead of blindly taking uniform action, it delves into diagnosing the root cause of the failure. First, it determines whether the failure is reversible or temporary. If it's a temporary obstruction or short-term calibration, it selects a temporary replacement solution with minimal disruption to the current business process and the lowest cost. If the failure is determined to be irreversible or permanent, such as hardware damage or a permanent change in the environment, it decisively initiates a permanent replacement process to fundamentally solve the problem. Compared to indiscriminately replacing all failures, which incurs huge and unnecessary hardware and labor costs, and conversely, if only replacements are performed, the system will accumulate a large number of permanent fault points, leading to a continuous deterioration of the monitoring network quality. This invention's configuration enables precise allocation of operational resources.

[0033] Example: Temporary replacement: The diagnosis revealed that Numb08 malfunctioned because newly arrived goods temporarily blocked the air vent; the analysis unit determined this to be a temporary environmental disturbance and decided to temporarily replace it.

[0034] Permanent Replacement: Diagnosis revealed that the Numb15 failure was due to damage to the sensor's own circuitry. The analysis unit determined this to be permanent hardware damage, decided on permanent replacement, and initiated the migration process. By differentiating handling methods through this process, the waste of resources caused by a one-size-fits-all approach is avoided. Minor issues are addressed temporarily, while major issues are completely eradicated, achieving optimal operational costs. Furthermore, it ensures that monitoring coverage will not experience prolonged gaps under any circumstances, guaranteeing the continuity of monitoring data and meeting GMP compliance requirements.

[0035] The scheduling unit also includes: When it is determined to be a temporary replacement, the mobile monitoring point is dispatched to conduct periodic patrol monitoring of the area where the failed fixed monitoring point is located; When it is determined to be a direct substitution: Dispatch the mobile monitoring node to the location of the fixed monitoring node for long-term monitoring; A fixed monitoring node migration task is generated, the migration task including a new installation location determined based on data collected by the mobile monitoring node.

[0036] In this embodiment, under the traditional model, there is a maintenance window between the removal of a failed fixed point and the installation of a new one, during which monitoring data is missing, posing a significant risk to GMP compliance and product quality traceability. Therefore, by scheduling mobile monitoring nodes to this location for long-term monitoring, a dynamic backup mechanism is introduced. The mobile nodes are no longer temporary inspection points, but temporary fixed points, immediately taking over the monitoring responsibilities of the area. This means that the system's monitoring coverage capability achieves perpetual continuity in both space and time, and its data link is never interrupted regardless of how the fixed point network is adjusted or maintained.

[0037] Example: Temporary replacement: The scheduling module commands an AGV to patrol the aisle of the shelf where Numb08 is located every 2 hours until the goods are moved.

[0038] Permanent Replacement: The scheduling module immediately commands another AGV to move to position Number15 for long-term monitoring. Simultaneously, the system automatically generates a migration work order (Work Order No. Move001) and initiates the new location search process. Through the long-term monitoring of the mobile node, seamless monitoring transition from the failure of the fixed point to the activation of the new point is achieved, completely eliminating monitoring blind spots.

[0039] The system is communicatively connected to the warehouse management system and / or facility management system; The identification unit further includes: when receiving cargo layout change information from the warehouse management system or equipment modification information from the facility management system, marking the status of the fixed monitoring node related to the changed area as pending evaluation; The analysis unit further includes: in response to a fixed monitoring node being marked as to be evaluated, dispatching at least one mobile monitoring node to the changed area to collect confirmatory data; The scheduling unit uses the verification data collection results to perform a third identification method to determine whether the fixed monitoring node has failed.

[0040] In this embodiment, the aforementioned solution can quickly identify and repair existing node failures, but it is essentially still a reactive logic. Therefore, by integrating with enterprise-level warehouse management systems (WMS) and facility management systems (FMCS), it can directly acquire planned events that indicate impending environmental changes, such as shelf adjustments and equipment modifications. This allows the system to proactively mark relevant area nodes as pending evaluation before actual physical environmental changes occur and immediately initiate mobile node verification. The system no longer passively waits for data anomalies to appear, but actively verifies and confirms the impact of environmental changes on the monitoring network, thus potentially completing diagnosis and response plan formulation before users even notice the problem.

[0041] Example: The Warehouse Management System (WMS) issued a notification announcing a planned large-scale shelving reorganization in Area A. This event was captured by the identification module, and the system automatically marked 10 fixed nodes related to Area A as pending evaluation. The scheduling unit then dispatched AGVs to Area A for a full-coverage scan. The scan revealed that the relevance of three nodes had significantly decreased; the system preemptively determined their failure and initiated the migration process. This approach shifts the focus of monitoring from post-event remediation to real-time awareness and even pre-event prediction, enabling predictive maintenance of the monitoring network.

[0042] The process of determining the new installation location is as follows: S1. Based on the location of the failed fixed monitoring node, the scheduling unit delineates a sub-area to be evaluated in the digital twin module and plans a data acquisition path covering the sub-area to be evaluated for the mobile monitoring node. The mobile monitoring node moves along the data acquisition path and collects temperature and humidity data at a third frequency higher than the second frequency; then the data with precise location information is transmitted to the central processing module in real time; the digital twin module generates a dynamic temperature and humidity field distribution map in the three-dimensional model of the sub-region to be evaluated based on the data with precise location information. S2. Within the sub-region to be evaluated defined by the digital twin module, a two-dimensional coordinate grid is established, and the data points collected by the mobile monitoring node are mapped to the nearest grid vertex to form a temperature and humidity data matrix covering the sub-region to be evaluated. S3. For each candidate location point in the grid, obtain representative indicators, stability indicators, and risk proximity indicators based on temperature and humidity data. Representative indicators and stability indicators are first weighted and summed, and then risk proximity indicators are used as index terms. After comprehensive calculation, the overall installation suitability score is obtained. S4. Sort the comprehensive installation suitability scores of all candidate locations in descending order to generate a candidate location list; S5. From the candidate location list, check each candidate point in turn from top to bottom to see if it meets the preset engineering installation conditions in the digital twin module; determine the first candidate location point that meets all engineering installation conditions as the final new installation location.

[0043] Example: When searching for a new location for the failed Numb15, the system delineates a 10m*10m area around it in the digital twin model; the AGV patrols this area for 5 minutes at a frequency of 1 time / 10 seconds; after generating a heat map, the system calculates the comprehensive installation suitability score for each candidate location point in the grid; finally, a candidate location point located in the middle of the shelf, which is far from the door (stable temperature), can cover the new airflow dead zone (risk proximity), and is easy to install (engineering feasibility) is selected with the highest score and automatically filled into the work order.

[0044] The representative indicators are obtained as follows: Calculate the average of all temperature and humidity readings for each candidate location point during the scanning period, and compare it with the overall average of all data points in the sub-region to be evaluated. Take the reciprocal of the sum of the absolute values ​​of the temperature and humidity differences between the two, so that the smaller the difference, the greater the representative index. The stability index is obtained by calculating the standard deviation of all temperature and humidity readings during the scanning period for each candidate location point. The stability index is the reciprocal of the standard deviation, i.e., stability index = 1 / (1 + standard deviation). The smaller the fluctuation, the larger the stability index. The risk proximity index is obtained as follows: based on the dynamic temperature and humidity field distribution map, all high temperature, low temperature or high fluctuation risk points exceeding the threshold are identified; the reciprocal of the sum of the Euclidean distances from the candidate location point to all identified risk points is calculated, so that the closer to the risk point, the larger the risk proximity index.

[0045] This embodiment provides a method for determining the specific details of new installation points, thereby achieving precise monitoring point layout and solving the monitoring blind spots or redundant coverage problems caused by the reliance on experience in traditional fixed-point installation. Specifically, by using a digital twin module to delineate the sub-area to be evaluated and planning the data acquisition path, combined with high-frequency data acquisition from mobile monitoring nodes, a high-resolution dynamic temperature and humidity field distribution map can be generated, completely restoring the spatial distribution characteristics of temperature and humidity within the area. Layout based on measured data can overcome the limitations of static planning, ensuring that the new installation location is highly matched with the actual environmental characteristics, especially suitable for temperature and humidity sensitive scenarios such as clean areas and storage areas in pharmaceutical production. Simultaneously, a multi-dimensional indicator system maximizes the effectiveness of monitoring points, where the representative indicator is the reciprocal of the sum of the absolute values ​​of temperature and humidity differences, ensuring that the point reflects the overall area. The system considers the overall physical condition of the sensors to avoid monitoring distortion caused by local anomalies. The stability index, the reciprocal of the standard deviation, prioritizes locations with minimal temperature and humidity fluctuations, reducing false alarms or sensor damage caused by drastic environmental changes and extending equipment lifespan. The risk proximity index, the reciprocal of the sum of distances to risk points, strengthens monitoring coverage of high-risk areas, aligning with the risk-first quality management principle in pharmaceutical production. The integration of these three indicators and the design of the index term ensure comprehensive monitoring while leveraging the exponential amplification effect of the risk proximity index to ensure focused monitoring of key risk areas, forming an optimized layout logic of global coverage + risk focus. Among the candidate locations with the highest comprehensive score, priority is given to those meeting pre-set engineering installation conditions, such as load-bearing capacity, power supply, and spatial interference, avoiding the infeasibility of purely data-driven solutions in actual construction.

[0046] A method for intelligent monitoring of temperature and humidity across the entire pharmaceutical manufacturing process is disclosed, which is implemented based on the aforementioned intelligent monitoring system for temperature and humidity across the entire pharmaceutical manufacturing process.

[0047] 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. A smart temperature and humidity monitoring system for the entire pharmaceutical manufacturing process, characterized in that, include: Multiple fixed monitoring nodes are set up in a distributed manner within the monitoring area and collect ambient temperature and humidity data at the highest frequency. A mobile monitoring node is used to move within a monitored area and collect ambient temperature and humidity data at a second frequency, wherein the second frequency is higher than or equal to the first frequency. A central processing module, used for communication connection with the fixed monitoring node and the mobile monitoring node, includes: The identification unit identifies failed fixed monitoring nodes based on the data stream from the fixed monitoring nodes. The analysis unit, based on the identification result of the identification unit, determines whether to temporarily replace the failed fixed monitoring node with a mobile monitoring node or to directly replace it. The scheduling unit sends instructions to the mobile monitoring node to execute the replacement or substitution operation determined by the analysis unit; The digital twin module is used to construct a three-dimensional model consistent with the physical monitoring area, to visualize the real-time location, status and data of all fixed and mobile monitoring nodes, and to provide support for path planning for the scheduling unit.

2. The intelligent temperature and humidity monitoring method for the entire pharmaceutical production chain according to claim 1, characterized in that, The identification unit identifies failed fixed monitoring nodes using at least one of the following methods: The first identification method is to identify data patterns that are rigid, drift beyond the range, or conflict with physical logic through data anomaly pattern analysis. The second identification method is to calculate the correlation coefficient between the current fixed monitoring node and the preset adjacent node group through correlation decay analysis, and trigger an early warning when the correlation coefficient is continuously lower than the first threshold. The third identification method is to conduct mobile verification divergence analysis. When the data collected by the mobile monitoring node within a preset range of any fixed monitoring node differs from the data of the fixed monitoring node by more than the allowable error, an early warning is triggered for the fixed monitoring node.

3. The intelligent temperature and humidity monitoring method for the entire pharmaceutical production chain according to claim 2, characterized in that, The data rigidity in the first identification method means that within a preset time window, the standard deviation of the data of the fixed monitoring node is lower than the second threshold of the standard deviation benchmark value of the fixed monitoring node.

4. The intelligent temperature and humidity monitoring method for the entire pharmaceutical production chain according to claim 3, characterized in that, The judgment method of the analysis unit is as follows: If the cause of failure is determined to be temporary environmental disturbance or short-term recoverable calibration of fixed monitoring nodes, it is determined to be a temporary replacement; If the cause of failure is determined to be damage to the fixed monitoring node or a permanent change in the environment, it is determined to be a direct replacement.

5. The intelligent temperature and humidity monitoring method for the entire pharmaceutical production chain according to claim 4, characterized in that, The scheduling unit also includes: When it is determined to be a temporary replacement, the mobile monitoring point is dispatched to conduct periodic patrol monitoring of the area where the failed fixed monitoring point is located; When it is determined to be a direct substitution: Dispatch the mobile monitoring node to the location of the fixed monitoring node for long-term monitoring; A fixed monitoring node migration task is generated, the migration task including a new installation location determined based on data collected by the mobile monitoring node.

6. The intelligent temperature and humidity monitoring method for the entire pharmaceutical production chain according to claim 1, characterized in that, The system is communicatively connected to the warehouse management system and / or facility management system; The identification unit further includes: when receiving cargo layout change information from the warehouse management system or equipment modification information from the facility management system, marking the status of the fixed monitoring node related to the changed area as pending evaluation; The analysis unit further includes: in response to a fixed monitoring node being marked as to be evaluated, dispatching at least one mobile monitoring node to the changed area to collect confirmatory data; The scheduling unit uses the verification data collection results to perform a third identification method to determine whether the fixed monitoring node has failed.

7. The intelligent temperature and humidity monitoring method for the entire pharmaceutical production chain according to claim 5, characterized in that, The process of determining the new installation location is as follows: S1. Based on the location of the failed fixed monitoring node, the scheduling unit delineates a sub-area to be evaluated in the digital twin module and plans a data acquisition path covering the sub-area to be evaluated for the mobile monitoring node. The mobile monitoring node moves along the data acquisition path and collects temperature and humidity data at a third frequency higher than the second frequency; then the data with precise location information is transmitted to the central processing module in real time. The digital twin module generates a dynamic temperature and humidity field distribution map within the three-dimensional model of the sub-region to be evaluated, based on data with precise location information. S2. Within the sub-region to be evaluated defined by the digital twin module, a two-dimensional coordinate grid is established, and the data points collected by the mobile monitoring node are mapped to the nearest grid vertex to form a temperature and humidity data matrix covering the sub-region to be evaluated. S3. For each candidate location point in the grid, obtain representative indicators, stability indicators, and risk proximity indicators based on temperature and humidity data. Representative indicators and stability indicators are first weighted and summed, and then risk proximity indicators are used as index terms. After comprehensive calculation, the overall installation suitability score is obtained. S4. Sort the comprehensive installation suitability scores of all candidate locations in descending order to generate a candidate location list; S5. From the candidate location list, check each candidate point in turn from top to bottom to see if it meets the preset engineering installation conditions in the digital twin module; determine the first candidate location point that meets all engineering installation conditions as the final new installation location.

8. The intelligent temperature and humidity monitoring method for the entire pharmaceutical production chain according to claim 7, characterized in that, The representative indicators are obtained as follows: Calculate the average of all temperature and humidity readings for each candidate location point during the scanning period, and compare it with the overall average of all data points in the sub-region to be evaluated. Take the reciprocal of the sum of the absolute values ​​of the temperature and humidity differences between the two, so that the smaller the difference, the greater the representative index. The stability index is obtained by calculating the standard deviation of all temperature and humidity readings during the scanning period for each candidate location point. The stability index is the reciprocal of the standard deviation, i.e., stability index = 1 / (1 + standard deviation). The smaller the fluctuation, the larger the stability index. The risk proximity index is obtained by identifying all high temperature, low temperature or high fluctuation risk points that exceed the threshold based on the dynamic temperature and humidity field distribution map. Calculate the reciprocal of the sum of the Euclidean distances from the candidate location point to all identified risk points, such that the closer the location is to the risk point, the larger the risk proximity index.

9. A method for intelligent monitoring of temperature and humidity across the entire pharmaceutical manufacturing process, characterized in that, This method is implemented based on the end-to-end intelligent temperature and humidity monitoring system for pharmaceutical production as described in any one of claims 1-8.

Images