Real-time monitoring method of drug production line based on cloud-edge computing

Abstract

The present invention discloses a real-time monitoring method for a drug production line based on cloud-edge computing, which relates to the field of drug monitoring technology. The method includes: dividing the production monitoring area according to the distribution of the drug production line and the cleanliness requirements; constructing a graph model according to the production monitoring area and its cleanliness requirement level; deploying a pressure difference sensor based on the production monitoring area to monitor the pressure difference of the area in real time; building a cloud platform to perform a global trend analysis based on the received real-time pressure difference of the area, including: performing air flow direction analysis on the pressure difference of each area based on the graph model and fitting the spatial flow direction; judging the trend matching relationship between the spatial flow direction and the cleanliness requirement level, and regulating the production pressure difference. The present invention solves the technical problem that the pressure difference control in the prior art is not accurate enough, which makes it difficult to ensure the stability and cleanliness of the production environment, and achieves the technical effect of improving the stability and cleanliness of the production environment through accurate pressure difference regulation.

Description

Technical Field

[0001] The present invention relates to the technical field of drug monitoring, and in particular to a real-time monitoring method for a drug production line based on cloud-edge computing. Background Art

[0002] In the field of drug production, it is crucial to maintain a high degree of stability and cleanliness in the production environment. However, traditional production environment control methods have many shortcomings. In the past, the pressure difference control of the drug production environment usually adopted a relatively extensive method, which lacked accuracy and real-time performance. This resulted in the inability of the pressure difference control to accurately adapt to the dynamic changes in the production process, making it difficult to maintain the stability of the production environment. At the same time, due to the limitations of technical means, the cleanliness monitoring of the production area is not comprehensive and detailed enough, and potential cleanliness problems cannot be discovered in time, making it difficult to meet the strict requirements of drug production for high cleanliness. In addition, the pressure difference coordination between the various production areas is not optimized enough, and an effective overall control strategy cannot be formed, which affects the stability and cleanliness of the entire production environment.

[0003] The existing technology has the technical problem that the pressure difference control is not precise enough, which makes it difficult to ensure the stability and cleanliness of the production environment. Summary of the invention

[0004] The present application provides a real-time monitoring method for a drug production line based on cloud-edge computing, which is used to solve the technical problem in the prior art that the pressure difference control is not accurate enough, resulting in difficulty in ensuring the stability and cleanliness of the production environment.

[0005] In view of the above problems, the present application provides a real-time monitoring method for a drug production line based on cloud-edge computing, the method comprising:

[0006] According to the distribution of drug production lines and clean requirements, the production monitoring areas are divided, and the production monitoring areas have clean requirement levels; according to the production monitoring areas and their clean requirement levels, a graph model is constructed; based on the production monitoring areas, pressure difference sensors are arranged, and edge operators are deployed to the corresponding areas of the graph model, and the data channels of the pressure difference sensors are connected to the edge operators for real-time pressure difference monitoring of the areas; a cloud platform is built to connect the edge operators of each area, and a global trend analysis is performed based on the received real-time pressure difference of the area, including: air flow direction analysis of the pressure difference of each area based on the graph model, and fitting of the spatial flow direction; judging the trend matching relationship between the spatial flow direction and the clean requirement level, and when there is no match, the production pressure difference is regulated based on the real-time pressure difference of the area and the clean requirement level, wherein the matching relationship is that the clean requirement level of the area of ​​the spatial flow path node is continuously decreasing.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0008] According to the distribution of drug production lines and cleanliness requirements, production monitoring areas are divided; according to the production monitoring areas and their cleanliness requirements, a graph model is constructed; based on the production monitoring areas, differential pressure sensors are deployed to monitor the pressure differences of the areas in real time; a cloud platform is built to conduct global trend analysis, including: air flow analysis of the pressure differences of each area based on the graph model, fitting the spatial flow direction; judging the trend matching relationship between the spatial flow direction and the cleanliness requirement level, and regulating the production pressure difference when there is a mismatch. The technical effect of improving the stability and cleanliness of the production environment through precise pressure difference regulation is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0010] Figure 1 A schematic diagram of a real-time monitoring method for a drug production line based on cloud-edge computing provided in an embodiment of the present application;

[0011] Figure 2 A schematic diagram of the process of constructing a graphical model for a real-time monitoring method of a drug production line based on cloud-edge computing provided in an embodiment of the present application. DETAILED DESCRIPTION

[0012] This application provides a real-time monitoring method for a drug production line based on cloud-edge computing, which is used to solve the technical problem in the prior art that the pressure difference control is not accurate enough, resulting in difficulty in ensuring the stability and cleanliness of the production environment.

[0013] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0014] Embodiment 1

[0015] like Figure 1 As shown, the present application provides a real-time monitoring method for a drug production line based on cloud-edge computing, the method comprising:

[0016] Step S100: dividing the production monitoring area according to the distribution of drug production lines and cleanliness requirements, wherein the production monitoring area has a cleanliness requirement level.

[0017] Specifically, in the process of drug production, it is necessary to divide the production monitoring area according to the specific distribution of the drug production line and the specific requirements for cleanliness. This process must strictly follow the requirements of GMP (Good Manufacturing Practice) and the characteristics of drug production. First, a cleanliness requirement level is assigned to each area based on its importance in the production process, the nature of the drugs involved, and the risk of possible contamination. Then, clustering and segmentation are performed according to the location and span relationship of the distribution. For example, adjacent areas that are closely related in the production process and have similar cleanliness requirements will be classified into one category. Areas that are far away, have large differences in production links, and have significantly different cleanliness requirements will be divided into different categories. Through this clustering and segmentation method, multiple areas with different cleanliness requirement levels can be obtained. The division of these areas will help to carry out more targeted monitoring and management, ensuring that the cleanliness of each area can meet the requirements of drug production, thereby ensuring the quality and safety of drugs.

[0018] Step S200: constructing a graphical model according to the production monitoring area and its cleanliness requirement level.

[0019] Specifically, a graph model is constructed based on the previously divided production monitoring areas and the cleanliness requirement levels determined for each area. This graph model is a network structure composed of nodes and lines. The nodes represent the various production monitoring areas, and the lines represent the relationship between the areas. Each node will be labeled with relevant information of the corresponding production monitoring area, the key of which is the cleanliness requirement level of the area. Through this labeling, the difference in cleanliness requirements of different areas can be intuitively seen in the graph model, and the lines can represent the physical connection between areas, the direction of air flow, or the information transmission relationship. Building such a graph model helps to more clearly present the distribution and relationship between the various areas in the entire production environment, as well as their characteristics in terms of cleanliness requirements, which provides an intuitive and convenient infrastructure for subsequent monitoring, analysis, and regulation.

[0020] Step S300: Based on the production monitoring area, a pressure difference sensor is arranged, and an edge operator is deployed to the area corresponding to the graph model. The data channel of the pressure difference sensor is connected to the edge operator for real-time pressure difference monitoring of the area.

[0021] Specifically, based on the production monitoring area that has been determined, the layout of the differential pressure sensor begins. The differential pressure sensor is usually installed in pairs, one of which is installed inside the clean room to measure the pressure inside the room; the other is installed outside the clean room or in an adjacent lower cleanliness area to measure the pressure outside or in an adjacent area. At the same time, the edge operator with specific functions is deployed to the position in the graph model that precisely corresponds to these production monitoring areas. The deployment position of the edge operator needs to be completely matched with the actual production monitoring area in space and logic to ensure the accuracy and relevance of the data. Then, the differential pressure sensor is closely connected to the edge operator through the set data channel. The establishment of the data channel must ensure stable, fast and accurate data transmission to avoid data loss or delay. In this way, the edge operator can obtain the accurate data measured by the differential pressure sensor in real time and uninterruptedly. Through this real-time data acquisition, the pressure difference changes in each area can be continuously and accurately monitored. Whether it is a small pressure fluctuation or a significant pressure difference change, it can be detected and recorded in time. Real-time and accurate monitoring of the pressure difference between the two areas is achieved to ensure that the pressure stability and cleanliness of the environment meet the requirements.

[0022] Step S400: Build a cloud platform, connect the edge operators in each area, and perform global trend analysis based on the received real-time pressure difference in the area.

[0023] Specifically, we will start to build a powerful cloud platform, which will serve as the core hub for data processing and analysis. Through network communication technology, edge operators distributed in different regions will be connected to the cloud platform. This connection ensures that data can be smoothly transmitted from the edge operators to the cloud platform. When the edge operators collect the pressure difference data of each region in real time, these data will be quickly transmitted to the cloud platform. After the cloud platform receives the real-time pressure difference data from different regions, it will conduct a global trend analysis on these data. Global trend analysis not only focuses on the pressure difference changes in a single region, but also takes into account the pressure difference of all regions from an overall perspective. It will analyze the change trend of pressure difference data over time, such as whether there is a trend of gradual increase or decrease, and whether there is a periodic fluctuation pattern. It will also compare the relationship and difference between the pressure differences between different regions to determine whether the pressure difference changes in some regions have an impact on other regions. Through this global trend analysis, potential systemic problems or trend changes can be discovered, not just local anomalies.

[0024] Step S500: performing air flow analysis on the pressure difference in each area based on the graphical model, and fitting the spatial flow direction.

[0025] Specifically, we must first rely on the carefully constructed graphical model, which clearly shows the layout of each production monitoring area and the relationship between them. Obtain the pressure difference data corresponding to each area in the graphical model. These pressure difference data are the key basis for air flow analysis. For the analysis of air flow, we start from the basic principle that pressure difference causes air flow, combined with specific environmental conditions and boundary constraints. For example, consider the influence of obstacles between areas, the location and characteristics of vents, the shape and size of the room and other factors on air flow. When fitting the spatial flow direction, the following specific methods and techniques will be used: Based on the calculation method of fluid dynamics, the relevant theoretical formulas and numerical simulation techniques are used to calculate the flow speed and direction of air between different areas according to the pressure difference data. Data interpolation and fitting technology, for discrete pressure difference measurement points, the pressure difference of unmeasured points is estimated through interpolation methods (such as linear interpolation, spline interpolation, etc.), so as to have a more comprehensive understanding of the pressure difference distribution of the entire area, and then infer the air flow direction. Probabilistic statistical methods, analyze the distribution characteristics and statistical laws of pressure difference data to assist in judging the possibility and trend of air flow direction. By comprehensively applying the above methods and technologies, the spatial paths and directions of air flow between various areas are gradually constructed. This fitted spatial flow direction is not a simple straight line or single-direction flow, but a complex flow pattern that fully considers various practical factors. For example, in a pharmaceutical production area with multiple rooms and passages, through analysis and fitting, it is found that air flows from a high-cleanliness room through specific vents and passages in a circuitous way to a low-cleanliness room, rather than directly in a straight line. The final fitted spatial flow results will be presented in the form of intuitive diagrams or data, providing important basis and guidance for subsequent cleanliness evaluation and regulation.

[0026] Step S600: Determine the trend matching relationship between the spatial flow direction and the cleanliness requirement level. When there is no match, control the production pressure difference based on the regional real-time pressure difference and the cleanliness requirement level, wherein the matching relationship is that the regional cleanliness requirement level of the spatial flow path node is continuously decreasing.

[0027] Specifically, the previously fitted spatial flow direction is compared and judged with the trend of the pre-set cleanliness requirement level to determine whether there is a matching relationship between them, and the areas on the path nodes passed by the spatial flow direction are checked one by one. According to the definition of the matching relationship, if the spatial flow direction is from the area with a high cleanliness requirement level to the area with a gradually decreasing cleanliness requirement level, then it is considered to be matched. For example, if the air flows from the area with a cleanliness requirement of Class A to the Class B area, and then to the Class C area, this is the normal flow direction that meets the matching relationship. However, if it is found that the spatial flow direction does not conform to the law of the continuous decrease of the cleanliness requirement level, such as the situation of flowing from the area with a low cleanliness requirement level to the area with a high cleanliness requirement level, it is determined to be mismatched. When it is determined to be mismatched, it is necessary to regulate the production pressure difference based on the real-time pressure difference of the area and the established cleanliness requirement level. The regulation process may include the following operations: adjust the operating parameters of the ventilation equipment, such as the speed and air volume of the fan, to change the pressure difference between the areas. Check and repair existing sealing loopholes or ventilation duct blockages to ensure that the air flows in the desired direction. Re-optimize the layout and isolation measures between areas to improve the pressure difference distribution and air flow. Through such regulation, the spatial flow can meet the trend requirements of the cleanliness demand level, thereby ensuring that the cleanliness of the pharmaceutical production environment meets relevant standards and regulations.

[0028] In a possible implementation, step S600 further includes:

[0029] Step S610: configuring the pressure difference gradient of each area according to the real-time pressure difference of the area and the cleanliness requirement level of each area.

[0030] Step S620: Obtain the forced pressure difference of each area.

[0031] Step S630: taking the forced pressure difference as a constraint, performing path search based on the pressure difference gradients of the various regions, and determining a pressure difference control strategy.

[0032] Specifically, first obtain the real-time pressure difference data of each area, which reflects the current pressure difference between the areas. At the same time, consider the cleanliness requirement level set for each area. The higher the cleanliness requirement level, the more stringent the control requirements for the pressure difference in the area are, so as to better maintain its high cleanliness environment. Then, the pressure difference gradient of each area is configured based on the real-time pressure difference and the cleanliness requirement level. For example, if the cleanliness requirement level of an area is the highest level, and the real-time pressure difference between it and the adjacent area is small, in order to achieve a better cleaning effect, a larger pressure difference gradient will be configured so that the area always maintains a higher positive pressure difference relative to the adjacent area, thereby ensuring that air flows from the low cleanliness area to the high cleanliness requirement area, reducing the risk of pollutants entering.

[0033] Obtain the mandatory differential pressure value that must be achieved in each area from relevant standards and specifications, regulatory requirements, specific production process requirements, or previously set system parameters. The determination of mandatory differential pressure is usually based on the following considerations: Industry standards and specifications, such as the pharmaceutical industry has clear minimum requirements for the differential pressure between areas of different cleanliness levels to ensure the maintenance of cleanliness and prevent cross contamination. Product characteristics, some special drug production processes may have specific strict requirements for differential pressure to protect product quality. Risk assessment, based on the assessment of contamination risk, product failure risk, etc., determine the mandatory differential pressure required for a specific area. For example, in a sterile drug production workshop, according to relevant regulations, the mandatory differential pressure between the sterile filling area and the adjacent buffer area is stipulated to be no less than 15 Pascals. Or, for areas producing highly active drugs, in order to avoid the spread of drug dust, the mandatory differential pressure between it and the outside world is set to a higher value. By accurately obtaining the mandatory differential pressure of each area, clear goals and constraints are provided for the subsequent formulation of differential pressure control strategies.

[0034] The previously obtained forced pressure difference of each area is set as a constraint that must be followed, and the path search is carried out based on the previously configured pressure difference gradient of each area. Path search can be imagined as a process of finding the optimal solution in a complex possibility space. Various adjustment methods should be considered, such as changing the operating parameters of the ventilation equipment (such as fan speed, air volume), adjusting the valve opening of the ventilation duct, changing the working mode of the air filtration system, etc. For each adjustment method, its impact on the pressure difference of each area will be calculated to see whether a reasonable pressure difference gradient distribution can be achieved under the premise of meeting the forced pressure difference constraint. Through the evaluation and comparison of many possibilities, the pressure difference control strategy is finally determined. This strategy will clarify specific operating instructions, such as increasing the speed of a fan to a specific value, or adjusting the opening of a valve to a certain extent. For example, suppose there are three areas A, B, and C, the forced pressure difference between area A and area B is 10Pa, and the forced pressure difference between area B and area C is 8Pa. The configured pressure difference gradient indicates that the pressure difference between area A and area B should gradually increase, and the pressure difference between area B and area C should be relatively stable. When searching for a path, different combinations of fan speeds and valve opening settings will be tried. For example, the air inlet fan speed in area A is first increased by 10%, and the pressure difference at this time is calculated to see whether it meets the requirements. If not, the exhaust valve opening in area B is reduced by 5% at the same time. This process is repeated and evaluated until the best combination that meets both the mandatory pressure difference constraint and the pressure difference gradient requirement is found, thereby determining the final pressure difference control strategy.

[0035] In one possible implementation, Figure 2 As shown, step S200 also includes:

[0036] Step S210: setting the distribution position relationship of the nodes according to the distribution position of the production monitoring area.

[0037] Step S220: Obtain the interval distance of the production monitoring area and set the connection edges between nodes.

[0038] Step S230: Calculate the regional level difference according to the cleanliness requirement level of the production monitoring area and configure the weight of the connecting edge.

[0039] Step S240: Establishing a graph relationship among the nodes, connecting edges and their weights, and constructing the graph model.

[0040] Specifically, we must first have a clear understanding of the specific distribution locations of production monitoring areas in the actual physical space, which are determined based on factors such as the process layout of the production line, the functional division of different areas, and the geometric structure of the space. Then, these actual distribution location information is mapped to the constructed graph model. When setting the distribution location relationship of the nodes, certain principles and logic will be followed. For example, if two production monitoring areas are adjacent and closely related in the actual space, the corresponding nodes in the graph model will also be set to relatively close positions to intuitively reflect the direct connection or frequent interaction that may exist between them. Conversely, if the two areas are far apart in the physical space and the interaction is relatively weak, the corresponding node positions in the graph model will also be relatively far apart. In this way, the node distribution location relationship in the graph model can accurately reflect the layout of the production monitoring area in the actual space, thereby providing a more realistic basis for subsequent analysis and decision-making based on the graph model.

[0041] Accurately obtain the actual distance between each production monitoring area through measuring tools, architectural drawings or related space planning data. After obtaining the distance information, set the connection edges between the nodes in the graph model according to these distances. If the distance between two production monitoring areas is close, a connection edge will be directly established between their corresponding nodes in the graph model, indicating that there is a direct connection or mutual influence between the two areas. On the contrary, if the distance is far, it is necessary to determine whether to establish a connection edge based on the specific situation. For example, if there is still an important connection in the production process or function despite the long distance, a connection edge will be established; but if the connection is weak, no connection edge will be established to avoid the graph model being too complicated and affecting subsequent analysis and processing.

[0042] The cleanliness requirement level of each production monitoring area is clarified. The cleanliness requirement level is usually determined according to the requirements of the production process, the characteristics of the product and the relevant industry standards. Then, the cleanliness requirement level difference between adjacent or related production monitoring areas is calculated. The calculation of the level difference is a simple numerical subtraction. The weight is configured for the connecting edge according to the calculated regional level difference. The size of the weight reflects the importance and influence of the difference in the cleanliness requirement level of adjacent areas. For example, if the cleanliness requirement level difference between two areas is large, it means that the cleanliness transition between them requires stricter control and management, then the weight of the edge connecting the two areas will be set higher, which means that in the subsequent analysis and decision-making, the relationship between the areas represented by this edge will be given greater attention. On the contrary, if the cleanliness requirement levels of the two areas are close and the level difference is small, the weight of the connecting edge will be relatively low, indicating that the cleanliness transition between them is relatively easy or less critical. Assume that a food processing workshop is divided into a raw material processing area (high cleanliness requirement), a processing area (medium cleanliness requirement) and a packaging area (low cleanliness requirement). The cleanliness requirement level difference between the raw material processing area and the processing area is 2, and the level difference between the processing area and the packaging area is 1. Then the weight of the edge connecting the raw material handling area and the processing area can be set to 8, while the weight of the edge connecting the processing area and the packaging area can be 4.

[0043] The nodes, connecting edges and weights configured for the connecting edges determined in the previous steps are integrated to establish a complete graph relationship, clarify the attributes and characteristics of each node, including the relevant information of the production monitoring area it represents, such as location, name, cleanliness requirement level, etc. Then, the connecting edges are accurately connected to the corresponding nodes to form a clear network structure. At the same time, the weights previously calculated and configured for the connecting edges are assigned to the corresponding connecting edges. In this way, a graph model that can comprehensively and accurately reflect the positional relationship, connection relationship and importance of the cleanliness requirement level differences between production monitoring areas is constructed. This graph model can be presented in an intuitive graphical form or stored and processed in the form of a data structure. It is possible to clearly understand the differences in cleanliness requirement levels in different production monitoring areas and the relationship between them, thereby achieving more accurate monitoring and management of cleanliness and ensuring the production quality of drugs or products.

[0044] In a possible implementation, step S200 further includes:

[0045] Step S250: Based on the personnel flow position in the production monitoring area, deploy monitoring sensor equipment, obtain regional personnel flow information, set flow path nodes, and fit them into associated nodes.

[0046] Step S260: Generate flow connection edges according to the setting positions of the flow path nodes, and the flow connection edges represent the paths of personnel flow.

[0047] Step S270: Configure flow weights according to the frequency of personnel flow and the impact on regional purity.

[0048] Step S280: Add the flow path nodes, flow connection edges and flow weights thereof to the graph model, and dynamically change the graph model according to the regional passenger flow information obtained in real time.

[0049] Specifically, the usual flow positions of personnel in the production monitoring area are carefully analyzed, including the entrances and exits of various work areas, passages between different functional areas, and the connection between the public rest area and the work area. Based on these determined personnel flow positions, monitoring sensor devices are deployed in a targeted manner. These devices can be sensors that can detect the passage of personnel, or cameras that can capture images of personnel movement. Through these monitoring sensor devices, the flow information of the area is obtained in real time, including the entry time, departure time, stay time and flow direction of personnel. According to the obtained flow information, flow path nodes are set. These nodes represent the key positions or turning points in the flow process of personnel. Then, these flow path nodes are fitted to other nodes associated with them. The associated nodes are the corresponding production area nodes or other nodes closely related to the flow of personnel. Through this fitting, a close connection between the flow of personnel and the production monitoring area is established, so that the impact of the flow of personnel on the production environment and cleanliness can be analyzed more comprehensively in the future.

[0050] Flow connection edges are generated based on the specific locations of the set flow path nodes. For example, if two flow path nodes, "entrance node" and "processing area node", are set, a flow connection edge will be generated between them. This connection edge clearly represents the possible path of personnel flow from the entrance to the processing area. Flow connection edges are not just simple line connections, they also contain information about the direction of personnel flow. For example, the connection edge from the "rest area node" to the "production area node" clearly indicates that personnel are flowing from the rest area to the production area. By generating these flow connection edges, a complete personnel flow path network can be constructed in the graph model, intuitively showing the flow routes of personnel between different areas.

[0051] The weights of flow connections are configured by considering two key factors. The first is the frequency of personnel flow. If the personnel flow on a certain flow connection is very frequent, it means that there are many people moving around on this path. For example, in a workshop, the connection from the locker room to the main production area will be assigned a higher weight because employees pass by many times a day and the personnel flow is frequent. By monitoring the personnel flow data in real time, the pressure difference control strategy can be dynamically adjusted to ensure that the cleanliness requirements of each area are met. For example, the pressure difference can be increased in areas with frequent personnel flow to reduce the spread of pollution. The second is the impact of personnel flow on the purity of the area. Some personnel flow paths are more likely to bring external pollutants into a specific area, thereby having a greater impact on the purity of the area. For example, the flow connection from an area with a high risk of contamination (such as a raw material handling area) to an area with high purity requirements (such as a sterile operation area) poses a greater threat to purity. Real-time monitoring of personnel flow can detect abnormal situations in a timely manner, such as personnel entering sensitive areas or not flowing along the prescribed path, triggering an early warning mechanism, and responding quickly, which will also be assigned a higher weight. The specific value of the weight can be determined according to a pre-set evaluation standard. The weight reflects the importance of the flow connection edge in the entire personnel flow system and the degree of attention it requires.

[0052] The previously determined flow path nodes, flow connection edges, and flow weights configured for them are added to the constructed graph model. In this way, the graph model not only contains relevant information about the production area, but also incorporates detailed data on personnel flow, making it more complete and comprehensive. Then, by obtaining real-time personnel flow information in each area, such as the real-time location, flow direction, and number of personnel, the graph model is dynamically adjusted and changed. If at a certain moment, the flow of personnel in a certain area suddenly increases or decreases, or a new flow path for personnel appears, the graph model will update the properties of the flow path nodes and flow connection edges accordingly, such as adjusting the size of the weights, or adding new nodes and connection edges. This dynamic change can ensure that the graph model always accurately reflects the current personnel flow status and provides the latest and most accurate information support for drug production monitoring and management.

[0053] In a possible implementation, step S500 further includes:

[0054] Step S510: determining the flow direction starting position according to the graphical model, the regional position relationship and the pressure difference distribution;

[0055] Step S520: performing neighborhood pressure difference flow direction analysis based on the flow direction starting position to determine a first pressure difference flow direction, where the first pressure difference flow direction includes one or more flow directions;

[0056] Step S530: iteratively analyzing the neighborhood pressure difference flow direction according to the first pressure difference flow direction to obtain a second pressure difference flow direction, and continuously obtaining an Nth pressure difference flow direction, where N is a positive integer not less than 2;

[0057] Step S540: Obtain the spatial flow direction according to the first pressure difference flow direction, the second pressure difference flow direction, ... until the Nth pressure difference flow direction.

[0058] Specifically, the constructed graph model is carefully analyzed. The graph model clearly shows the relative positions of each area and the connections between them, which is the regional position relationship. At the same time, the pressure difference distribution data between each area should be analyzed. The pressure difference distribution reflects the specific value and change trend of the pressure difference between different areas. The two aspects are considered comprehensively to determine the starting position of the flow direction. For example, if some areas are in a relatively central position in the graph model and are directly connected to multiple areas, and the pressure difference in this area is larger than that in the surrounding areas, then this area is determined as the starting position of the flow direction, because the larger pressure difference will cause the air to flow from this area to the adjacent areas with lower pressure, and its central and multi-connected location characteristics make it more critical to the entire flow direction. By comprehensively and meticulously analyzing the regional position relationship and pressure difference distribution, it is possible to accurately find the starting position that is most likely to guide the air flow direction, laying the foundation for the subsequent accurate analysis of the flow direction of the entire space.

[0059] After the starting position of the flow direction is determined, the shortest path search algorithm is used to analyze the pressure difference flow direction of its neighborhood to determine the first pressure difference flow direction. The shortest path search algorithm calculates the "cost" from the starting position to each of its neighborhoods. The "cost" here is defined based on the pressure difference, the distance between regions, or other relevant factors. The algorithm will find the path with the lowest cost, that is, the direction in which the pressure difference is most conducive to air flow. Assume that the starting position is node A, and its neighborhood includes nodes B, C, and D. The algorithm will calculate the cost from A to B, A to C, and A to D respectively. If the pressure difference from A to B is the largest and the distance between the regions is relatively close, then the path cost from A to B is the lowest, and the first pressure difference flow direction is from A to B. In this way, the shortest path search algorithm can quickly and accurately determine the first pressure difference flow direction containing one or more flow directions, providing a basis for subsequent more in-depth flow analysis, and also ensuring that the determined flow direction is the most likely and effective under the comprehensive consideration of factors such as pressure difference and distance.

[0060] After determining the first pressure difference flow direction, this step performs iterative analysis on this basis. Assuming that the first pressure difference flow direction is from node A to node B, next, with node B as the new starting point, the shortest path search algorithm is used again to analyze the pressure difference flow direction of node B's neighborhood. The algorithm recalculates the cost from node B to its neighboring nodes (such as C, D, and E). If the cost from node B to node C is the lowest, then the flow direction from node B to node C is determined, which is the second pressure difference flow direction. Then, continue this process with node C as the new starting point and iterate continuously. Repeat this operation for each iteration until the Nth pressure difference flow direction is obtained. Here, N is a positive integer not less than 2, indicating the number of continuously deepening and expanding analysis. Through this continuous iteration and the use of the shortest path search algorithm, the complex pressure difference flow relationship in the entire space can be gradually, comprehensively and accurately revealed, so as to more accurately grasp the path and trend of air flow.

[0061] Comprehensively consider the first pressure difference flow direction, the second pressure difference flow direction, and the Nth pressure difference flow direction determined previously. Connect these pressure difference flow directions determined in sequence to form a continuous flow path. By integrating and combing this series of pressure difference flow directions, the flow trajectory of air in the entire space can be clearly depicted. For example, the first pressure difference flow direction is from area A to area B, the second pressure difference flow direction is from area B to area C, the third pressure difference flow direction is from area C to area D, and so on. By connecting these flow directions in series, we can get the complete spatial flow direction from area A through area B, area C to area D, etc. In this way, we can fully and accurately grasp the overall trend and specific path of air flow in the entire space, providing an important basis for subsequent analysis and decision-making, such as optimizing the ventilation system and controlling cleanliness.

[0062] In a possible implementation, step S600 further includes:

[0063] Step S610: According to the graph model, the weight of each area and path is configured. The weight is determined according to the pressure difference gradient and the cleanliness requirement level, and the expression is: ,in, represents the weight from the i-th region to the j-th region, is the pressure gradient from the ith region to the jth region, is the weight factor of the cleanliness requirement level difference, is the cleanliness requirement level of the ith area, is the cleanliness requirement level of the jth area.

[0064] Step S620: identifying a plurality of pressure difference paths from a high cleanliness requirement level to a low cleanliness requirement level, wherein the plurality of pressure difference paths are straight line paths from the high cleanliness requirement level to the low cleanliness requirement level.

[0065] Step S630: Based on the multiple pressure difference paths, set a starting node, select a node with the smallest current distance, and update the distances of adjacent nodes.

[0066] Step S640: Mark the processed node with the smallest distance, and select a new unprocessed node with the smallest current distance, until all nodes are completed and the ideal path is determined.

[0067] Step S650: Based on the ideal path and the forced pressure difference, ensure that the actual path is consistent with the ideal path, and obtain the pressure difference control strategy, wherein the pressure difference adjustment formula is: ,in, is the pressure difference adjustment value of the i-th area, is the adjustment factor, is the forced pressure difference in the ith region, is the real-time pressure difference of the ith area.

[0068] Specifically, according to the graph model, weights are set for each region and the paths connecting these regions. The weight calculation method is determined by an expression, namely , in this expression Represents the weight from the i-th region to the j-th region. is the pressure difference gradient from the i-th area to the j-th area, reflecting the change amplitude or steepness of the pressure difference from the i-th area to the j-th area. The larger the pressure difference gradient, the more significant the pressure difference change between the two areas, which is of great significance in air flow and pressure difference control. The weight factor of the cleanliness requirement level difference is used to adjust the importance of the cleanliness requirement level difference in the weight calculation. and Respectively represent the cleanliness requirement levels of the i-th area and the j-th area, by calculating the absolute value of their difference and multiplying it by the weight factor , plus the pressure gradient , and get the final weight ,Such weight configuration comprehensively considers the differences in the pressure difference gradient and the ,cleanliness requirement level, so that the weight can accurately ,reflect the importance of each area and path in the pressure difference ,control strategy.

[0069] Identify multiple pressure difference paths from high cleanliness requirement level to low cleanliness requirement level. The characteristics of these paths are that they are straight paths, which means that air will flow from areas with high cleanliness requirement level to areas with low cleanliness requirement level along these direct paths. This identification process is based on the previously constructed graph model and the cleanliness requirement level set for each area. The connection relationship and pressure difference distribution of each area in the graph model are analyzed to find those straight paths that meet the flow direction from high cleanliness requirement level to low cleanliness requirement level. For example, in the graph model of a pharmaceutical production workshop, straight paths from the sterile operation room with extremely high cleanliness requirements to the packaging area with relatively low cleanliness requirements will be identified. These paths are pressure difference paths. By identifying these pressure difference paths, the flow direction of air in the production environment can be better understood and controlled to ensure that the air quality in areas with higher cleanliness is not polluted. At the same time, it also helps to optimize the pressure difference control strategy and improve the stability and reliability of the production environment.

[0070] Based on multiple pressure difference paths, the shortest path search algorithm is used to set the starting node, select the node with the shortest current distance, and update the distance of the neighboring nodes. First, add the starting node to the set of nodes to be processed, then select the node with the shortest distance from the set of nodes to be processed and mark it as a processed node. Next, for the neighboring nodes of this node, calculate their distance to the starting node according to the shortest path search algorithm. If the distance to the neighboring node through the current node is smaller than the originally known distance, update the distance of the neighboring node. By repeatedly selecting the node with the shortest distance and updating the distance of the neighboring nodes, the shortest path search algorithm can gradually find the shortest path from the starting node to other nodes, thereby providing a basis for determining the ideal path.

[0071] Mark the nodes with the smallest distance that have been processed, which means that these nodes have been considered and the shortest distance between them and the starting node has been determined. Then, select the new node with the smallest current distance from the nodes that have not been processed. This distance is determined based on the previous calculation and update, and represents the shortest path length from the starting node to the node. Then, repeat this process and continuously select new unprocessed nodes with the smallest current distance until all nodes have been processed. In this process, the distance information between nodes is continuously updated to ensure that the path found is the shortest. Finally, by completing the processing of all nodes, the ideal path can be determined. This ideal path is a collection of the shortest paths from the starting node to all other nodes, which satisfies the optimal solution under the given pressure difference path and node distance conditions. Through such steps, it can be ensured that the ideal path found is the best under the current conditions, thus providing an important basis for the formulation of the pressure difference control strategy.

[0072] A pressure differential control strategy is developed based on the previously determined ideal path and forced pressure differential to ensure that the actual path is consistent with the ideal path. First, the ideal path is the optimal path calculated through the previous steps. It takes into account factors such as the pressure differential path and node distance, and can meet the requirements for air flow and cleanliness in the production environment. Forced pressure differential is the pressure differential target that each area should achieve based on production requirements and cleanliness standards. In order to make the actual path consistent with the ideal path, it needs to be achieved through pressure differential adjustment. The pressure differential adjustment formula is: ,in is the pressure difference adjustment value of the ith area, which determines how the pressure difference of the area needs to be adjusted. It is an adjustment factor used to control the amplitude and speed of adjustment. is the mandatory pressure difference of the ith area, that is, the target pressure difference that the area should reach. is the real-time pressure difference of the ith area, that is, the actual pressure difference of the area at present. Through this formula, the pressure difference adjustment value required for each area can be calculated. If the real-time pressure difference is less than the forced pressure difference, then the adjustment value is a positive number, and the pressure difference needs to be increased; if the real-time pressure difference is greater than the forced pressure difference, then the adjustment value is a negative number, and the pressure difference needs to be reduced. Through such adjustments, the actual pressure difference can be gradually approached to the forced pressure difference, so that the actual path gradually becomes consistent with the ideal path. Through such a pressure difference control strategy, the pressure difference in the production environment can be effectively controlled to ensure that the air flows along the ideal path, thereby improving the cleanliness and stability of the production environment.

[0073] In a possible implementation, step S600 further includes:

[0074] Step S660: monitor the pressure difference and airflow of each path in real time, and monitor the disturbance information between the paths.

[0075] Step S670: Based on the disturbance information, configure the collaborative path, and determine the path priority according to the path weight and the disturbance information.

[0076] Step S680: adjusting the pressure difference according to the path priority, and determining a coordinated adjustment strategy as the pressure difference control strategy.

[0077] Specifically, through high-precision sensors and monitoring equipment, the pressure difference and airflow of each path are monitored in real time and uninterruptedly. These sensors will accurately measure the pressure difference between each area, as well as the flow speed and direction of the air in the path. At the same time, the disturbance information between the paths will be closely monitored, including but not limited to sudden fluctuations in pressure, unstable changes in airflow, and possible interference factors from the outside world. For example, when a certain device starts or stops, it will cause changes in the pressure and airflow of nearby paths. The monitoring system will capture these changes in time. By monitoring this disturbance information, potential problems or abnormal conditions can be discovered in time, so that appropriate measures can be taken quickly to adjust and control, ensuring the stability and cleanliness of the production environment.

[0078] According to the disturbance information monitored previously, the corresponding processing is carried out. The disturbance information reflects the unstable factors or interference in the path. Based on these disturbance information, the collaborative path is configured. The role of the collaborative path is to provide additional path selection when the original path is disturbed to ensure that the air can continue to flow stably. The priority of the path is determined according to the path weight and disturbance information. The path weight is determined in the previous step according to factors such as pressure gradient and clean demand level. It represents the importance of the path in the entire system. The disturbance information reflects the degree of interference currently suffered by the path. Taking these two factors into consideration, each path is evaluated. If a path has a higher weight and the disturbance information shows that it is less disturbed, then its priority will be higher. On the contrary, if a path has a lower weight or the disturbance information shows that it is more disturbed, then its priority will be lower. In this way, the optimal path priority is determined so that when performing pressure difference adjustment and air flow control, those important and relatively stable paths can be given priority, thereby improving the efficiency and stability of the entire system.

[0079] Analyze the order of path priority. For paths with higher priority, the pressure difference will be ensured to be stable and meet the requirements. If the pressure difference of some paths is found to be not as expected, measures are taken to increase or decrease the pressure difference on specific paths. For example, for paths that need to increase the pressure difference, the pressure difference of the path can be increased by adjusting related equipment, such as increasing the power of the fan, to promote the air to flow in the desired direction. At the same time, the airflow can be directly adjusted through valves or fans. By controlling the opening of the valve or the speed of the fan, the speed and flow of the airflow can be changed, thereby adjusting the pressure difference distribution on the path. In addition, in order to better adapt to changes in actual conditions, the path weights will be dynamically adjusted and the optimal path will be recalculated. If the disturbance of some paths changes or the production demand changes, the path weights will be adjusted accordingly to ensure that the pressure difference control strategy can always remain optimal. Finally, through the implementation of these collaborative adjustment strategies, the most appropriate pressure difference control strategy is determined to achieve effective pressure difference control for real-time monitoring of the drug production line, ensure the stability and cleanliness of the production environment, and meet production requirements.

[0080] In a possible implementation, step S670 further includes:

[0081] Step S671: Obtain the drug cleanliness constraint value of the area involved in the collaborative path.

[0082] Step S672: Obtain the disturbance direction of the disturbance information, where the disturbance direction includes incremental disturbance and decremental disturbance.

[0083] Step S673: Determine the weight correction direction according to the incremental disturbance and the decremental disturbance.

[0084] Step S674: Correct and adjust the path weight according to the weight correction direction.

[0085] Step S675: The drug cleanliness constraint value and the corrected path weight are fused, and the path priority is set according to the fused quantity value.

[0086] Specifically, the purpose of obtaining the drug cleanliness constraint value is to determine the correction direction of the path weight according to the disturbance direction obtained in the subsequent steps. If the disturbance direction is incremental disturbance, it means that it will have an adverse effect on the cleanliness of the area. In order to deal with this situation, the path weight needs to be increased to strengthen the control and adjustment of the path to ensure that the cleanliness can meet the requirements. On the contrary, if the disturbance direction is decremental disturbance, it means that the current situation has little impact on the cleanliness, and the path weight needs to be reduced to avoid excessive adjustment.

[0087] The disturbance direction of the disturbance information is obtained. The disturbance direction mainly includes incremental disturbance and decremental disturbance. Incremental disturbance refers to a change in a parameter or state caused by a certain factor. For example, due to equipment operation, personnel activities, etc., the pressure in a certain area increases, or the airflow speed increases. This is an incremental disturbance. Decremental disturbance is the opposite, that is, a certain factor causes a decrease in the parameter or state. For example, the pressure in a certain area decreases, or the airflow speed slows down. This is a decremental disturbance. By obtaining the disturbance direction of the disturbance information, the impact of these disturbances on cleanliness is further analyzed, so that corresponding measures can be taken to adjust the path weights and priorities to ensure the normal operation of the drug production line and product quality.

[0088] The direction of weight correction is determined based on the detected incremental disturbances and decremental disturbances. If an incremental disturbance is detected, it means that some factors have caused an increase in the relevant parameters, which will have a negative impact on stability and cleanliness. In order to deal with this situation, the direction of weight correction should be to increase, that is, to increase the weight of the relevant path. The purpose of this is to increase the importance of this path so that more attention can be paid to this path in subsequent adjustments, and appropriate measures can be taken to alleviate the adverse effects of incremental disturbances and meet the cleanliness requirements of drug production. On the contrary, if a decremental disturbance is detected, it means that some factors have caused a decrease in the relevant parameters. In this case, the direction of weight correction should be to decrease, that is, to reduce the weight of the relevant path.

[0089] According to the determined weight correction direction, the path weight is corrected and adjusted accordingly. If the weight correction direction is to increase, the weight of the path will be increased. For example, the original weight of the path is 5, and according to the correction direction, it will be increased to 7 or other appropriate values. Such an adjustment makes the path more important in subsequent calculations and decisions, and more attention will be paid to the situation of the path, and corresponding measures will be taken to ensure its stability and cleanliness. If the weight correction direction is to decrease, the weight of the path will be reduced. For example, the original weight is 8, which will be adjusted to 6 or lower. By reducing the weight, the relative importance of the path will be reduced, and the allocation of resources will be adjusted accordingly. Through such corrections and adjustments, the system can dynamically optimize the path weight according to the disturbance situation to better adapt to changes in the production environment and achieve more effective pressure difference control and cleanliness management.

[0090] Combine the obtained drug clean constraint value with the corrected path weight, and use the weighted average method to combine the drug clean constraint value and the corrected path weight in a certain proportion. The priority of the path is set according to the quantitative value obtained after fusion, and some standards will be formulated to judge the priority. For example, if the fused quantitative value is greater than a certain value T1, then the priority of the path is set to high; if the fused quantitative value is between T1 and another value T2, then its priority is set to medium; if the fused quantitative value is less than T2, then its priority is set to low. The purpose of this is to comprehensively consider the requirements of drug production for cleanliness and the changes in path weights, so as to more accurately judge the importance and priority of the path. Paths with high priority will receive more attention and priority when adjusting pressure differences and allocating resources to ensure that the cleanliness on these paths can meet the standards of drug production.

[0091] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0092] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

[0093] This specification and the drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.

Claims

1. Real-time monitoring method of drug production line based on cloud-edge computing, It is characterized in that Real-time monitoring methods for drug production lines based on cloud-edge computing include: According to the distribution of drug production lines and cleanliness requirements, the production monitoring areas are divided, and the production monitoring areas have cleanliness requirement levels; Constructing a graphical model according to the production monitoring area and its cleanliness requirement level; Based on the production monitoring area, a pressure difference sensor is arranged, and an edge operator is deployed to the area corresponding to the graph model, and the data channel of the pressure difference sensor is connected to the edge operator for real-time pressure difference monitoring of the area; Build a cloud platform to connect the edge operators in each region and perform global trend analysis based on the received regional real-time pressure difference, including: Based on the graphical model, air flow direction analysis is performed on the pressure difference in each area to fit the spatial flow direction; The trend matching relationship between the spatial flow direction and the cleanliness requirement level is determined. When there is no match, the production pressure difference is regulated based on the regional real-time pressure difference and the cleanliness requirement level, wherein the matching relationship is that the regional cleanliness requirement level of the spatial flow path node is continuously decreasing.

2. The real-time monitoring method for a drug production line based on cloud-edge computing as claimed in claim 1, It is characterized in that The production pressure difference control based on the regional real-time pressure difference and the cleanliness requirement level includes: According to the real-time pressure difference of the area and the cleanliness requirement level of each area, the pressure difference gradient of each area is configured; Get the forced pressure difference of each area; The forced pressure difference is used as a constraint, a path search is performed based on the pressure difference gradients of the various regions, and a pressure difference control strategy is determined.

3. The real-time monitoring method for a drug production line based on cloud-edge computing as claimed in claim 1, It is characterized in that According to the production monitoring area and its cleanliness requirement level, a graph model is constructed, including: According to the distribution position of the production monitoring area, setting the distribution position relationship of the nodes; Obtaining the interval distance of the production monitoring area and setting the connection edges between nodes; According to the cleanliness requirement level of the production monitoring area, the regional level difference is calculated, and the weight of the connecting edge is configured; A graph relationship among the nodes, connecting edges and their weights is established to construct the graph model.

4. The real-time monitoring method for a drug production line based on cloud-edge computing as claimed in claim 3, It is characterized in that Also includes: Based on the personnel flow position in the production monitoring area, monitoring sensor equipment is deployed to obtain regional personnel flow information, set flow path nodes, and fit them into the associated nodes; Generating flow connection edges according to the setting positions of the flow path nodes, wherein the flow connection edges represent the paths of personnel flow; Configure flow weights based on the frequency of personnel flow and the impact on regional purity; The flow path nodes, flow connection edges and flow weights thereof are added to the graph model, and the graph model is dynamically modified according to the regional passenger flow information obtained in real time.

5. The real-time monitoring method for a drug production line based on cloud-edge computing as claimed in claim 2, It is characterized in that Based on the graphical model, the air flow direction of each area pressure difference is analyzed and the spatial flow direction is fitted, including: According to the graphical model, the flow direction starting position is determined according to the regional position relationship and the pressure difference distribution; Performing neighborhood pressure difference flow direction analysis based on the flow direction starting position to determine a first pressure difference flow direction, wherein the first pressure difference flow direction includes one or more flow directions; Iteratively perform neighborhood pressure difference flow direction analysis according to the first pressure difference flow direction to obtain a second pressure difference flow direction, and continuously obtain an Nth pressure difference flow direction, where N is a positive integer not less than 2; The spatial flow direction is obtained according to the first pressure difference flow direction, the second pressure difference flow direction, ... until the Nth pressure difference flow direction.

6. The real-time monitoring method for a drug production line based on cloud-edge computing as claimed in claim 5, It is characterized in that Taking the forced pressure difference as a constraint, performing path search based on the pressure difference gradients of the various regions, and determining a pressure difference control strategy include: According to the graph model, the weight of each area and path is configured. The weight is determined according to the pressure difference gradient and the cleanliness requirement level. The expression is: ,in, represents the weight from the i-th region to the j-th region, is the pressure gradient from the ith region to the jth region, is the weight factor of the cleanliness requirement level difference, is the cleanliness requirement level of the ith area, is the cleanliness requirement level of the jth area; Identify a plurality of pressure difference paths from a high cleanliness requirement level to a low cleanliness requirement level, wherein the plurality of pressure difference paths are straight line paths from the high cleanliness requirement level to the low cleanliness requirement level; Based on the multiple pressure difference paths, a starting node is set, a node with the smallest current distance is selected, and distances of adjacent nodes are updated; Mark the processed node with the smallest distance, select the new unprocessed node with the smallest current distance, until all nodes are completed and the ideal path is determined; Based on the ideal path and the forced pressure difference, the actual path is ensured to be consistent with the ideal path, and the pressure difference control strategy is obtained, wherein the pressure difference adjustment formula is: ,in, is the pressure difference adjustment value of the i-th area, is the adjustment factor, is the forced pressure difference in the ith region, is the real-time pressure difference of the ith area.

7. The real-time monitoring method for a drug production line based on cloud-edge computing as claimed in claim 6, It is characterized in that Taking the forced pressure difference as a constraint, performing path search based on the pressure difference gradients of the various regions, and determining a pressure difference control strategy, further comprising: Monitor the pressure difference and airflow of each path in real time, and monitor the disturbance information between paths; Based on the disturbance information, configure a collaborative path, and determine a path priority according to the path weight and the disturbance information; The pressure difference is adjusted according to the path priority, and a coordinated adjustment strategy is determined as the pressure difference control strategy.

8. The real-time monitoring method for a drug production line based on cloud-edge computing as claimed in claim 7, It is characterized in that Determining the path priority according to the path weight and the disturbance information includes: Obtain drug cleanliness constraint values ​​for areas involved in collaborative paths; Acquiring a disturbance direction of disturbance information, wherein the disturbance direction includes an incremental disturbance and a decremental disturbance; Determining a weight correction direction according to the incremental disturbance and the decremental disturbance; Correcting and adjusting the path weight according to the weight correction direction; The drug cleanliness constraint value and the modified path weight are fused, and the path priority is set according to the fused quantity value.

Images