A method and system for monitoring air particles in a clean room

Abstract

The embodiments of this specification provide a method and system for monitoring air particles in a clean room. The method for monitoring air particles in a clean room includes: obtaining activity data of at least one subspace in the clean room, the activity data being the activity trajectory of personnel and the location of equipment in at least one subspace; based on the activity data of at least one subspace, evaluating the future diffusion characteristics of at least one subspace; based on the future diffusion characteristics and the activity data, determining the cumulative distribution characteristics of multiple future time points; based on the cumulative distribution characteristics, determining the particle accumulation area and controlling the clean equipment to regulate the clean room.

Description

Technical Field

[0001] The present application relates to the field of clean room environment control, and in particular to a clean room air particle monitoring method and system. Background Art

[0002] A clean room, also known as a dust-free workshop, clean room or clean room, refers to a well-enclosed space where parameters such as air cleanliness, temperature, humidity, pressure, and noise are controlled as needed. Clean rooms are particularly widely used in high-end manufacturing fields such as the precision machinery industry and the semiconductor industry. By removing pollutants such as particulate matter, harmful air, and bacteria from the air, the risk of contamination in the clean room can be effectively reduced, and production personnel and products can be protected from particulate contamination. However, the current environmental control in the clean room often involves detecting the current number of particles in the clean room and cleaning it, and does not involve predicting the particle diffusion trend in the clean room in advance and cleaning it.

[0003] Therefore, it is necessary to provide a clean room air particle monitoring system and method, which can predict in advance the areas where particles may accumulate in the clean room, so as to regulate the clean room in advance and disperse the particles that are about to accumulate. Summary of the invention

[0004] One or more embodiments of the present specification provide a method for monitoring air particles in a clean room, the method comprising: acquiring activity data of at least one subspace in the clean room, the activity data being the activity trajectory of personnel and the location of equipment in the at least one subspace; based on the activity data of the at least one subspace, evaluating the future diffusion characteristics of the at least one subspace; based on the future diffusion characteristics and the activity data, determining the accumulated distribution characteristics of multiple future time points; based on the accumulated distribution characteristics, determining the particle accumulation area and controlling the clean equipment to regulate the clean room.

[0005] One or more embodiments of the present specification provide a clean room air particle monitoring system, the system comprising: a processor, clean equipment, the processor being communicatively connected with the clean equipment, the processor being configured to: obtain activity data of at least one subspace in the clean room, the activity data being the activity trajectory of personnel and the location of equipment in the at least one subspace; based on the activity data of the at least one subspace, evaluate future diffusion characteristics of the at least one subspace; based on the future diffusion characteristics and the activity data, determine cumulative distribution characteristics of multiple future time points; based on the cumulative distribution characteristics, determine a particle accumulation area and control the clean equipment to regulate the clean room.

[0006] One or more embodiments of this specification provide a clean room air particle monitoring device. The device includes at least one storage medium and at least one processor, wherein the at least one storage medium is used to store computer instructions; and the at least one processor is used to execute the computer instructions to implement the clean room air particle monitoring method described in the above embodiment.

[0007] One or more embodiments of the present specification provide a computer-readable storage medium, wherein the storage medium stores computer instructions, and when the computer instructions are executed by a computer, the method for monitoring clean room air particles described in the above embodiment is implemented. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] This specification will be further described in the form of exemplary embodiments, which will be described in detail by the accompanying drawings. These embodiments are not restrictive, and in these embodiments, the same number represents the same structure, wherein:

[0009] Figure 1 is an exemplary schematic diagram of a clean room air particle monitoring system according to some embodiments of the present specification;

[0010] Figure 2 is an exemplary flow chart of a clean room air particle monitoring method according to some embodiments of the present specification;

[0011] Figure 3 is an exemplary flow chart for evaluating future diffusion characteristics according to some embodiments of the present specification;

[0012] Figure 4 It is a schematic diagram of a feature determination model according to some embodiments of this specification. DETAILED DESCRIPTION

[0013] In order to more clearly illustrate the technical solutions of the embodiments of this specification, the following briefly introduces the drawings required for describing the embodiments. The drawings do not represent all implementation methods.

[0014] It should be understood that the "system", "device", "unit" and / or "module" used herein are a method for distinguishing different components, elements, parts, portions or assemblies at different levels. If other words can achieve the same purpose, the words can be replaced by other expressions.

[0015] Unless the context clearly indicates an exception, the words "a", "an", "an" and / or "the" do not refer to the singular and may also include the plural. Generally speaking, the terms "include" and "comprise" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0016] When the operations are performed according to the step description in the embodiments of this specification, unless otherwise specified, the order of the steps is interchangeable, the steps may be omitted, and other steps may be included in the operation process.

[0017] Figure 1 is an exemplary schematic diagram of a clean room air particle monitoring system according to some embodiments of the present specification.

[0018] In some embodiments, the clean room air particle monitoring system 100 may include a processor 110 and a clean equipment 120. The processor 110 is in communication connection with the clean equipment 120.

[0019] The processor 110 is used to process data from at least one component of the clean room air particle monitoring system 100 or an external data source (e.g., a cloud data center). For example, the processor 110 can obtain activity data of at least one subspace in the clean room, and evaluate the future diffusion characteristics of the at least one subspace based on the activity data. For example, the processor 110 can determine the cumulative distribution characteristics of multiple future time points based on the future diffusion characteristics and the activity data. For another example, the processor 110 can determine the particle accumulation area based on the cumulative distribution characteristics and control the clean equipment to regulate the clean room. A clean room refers to a closed space where air cleanliness, temperature, humidity, pressure, noise, etc. need to be controlled.

[0020] In some embodiments, the processor 110 may include a central processing unit (CPU), an application specific integrated circuit (ASIC), an application specific instruction set processor (ASIP), a graphics processing unit (GPU), a controller, a microcontroller unit, a microprocessor, etc. or any combination thereof. In some embodiments, the processor 110 may be a single server or a server group. In some embodiments, the processor 110 may be local or remote.

[0021] In some embodiments, processor 110 may be further configured to estimate future diffusion characteristics based on the activity data and the airflow distribution characteristics.

[0022] In some embodiments, the processor 110 can be further configured to obtain the number of air particles at the activation time of the device through a detection device based on the activation time of the device; determine the hotspot distribution of at least one subspace based on activity data; and evaluate future diffusion characteristics based on the number of air particles, hotspot distribution, and airflow distribution characteristics.

[0023] In some embodiments, the processor 110 may be further configured to determine the cumulative distribution characteristics based on the future diffusion characteristics, the activity data, and the spatial structure data.

[0024] In some embodiments, the processor 110 may be further configured to construct a spatial map based on future diffusion characteristics, activity data, and spatial structure data; and determine the cumulative distribution characteristics through a feature determination model based on the spatial map.

[0025] The clean equipment 120 is configured to control the clean room. The clean equipment can deliver airflow to the clean room to disperse the particles in the particle accumulation area. For example, the clean equipment can include a hair dryer, a hair dryer, etc., which can deliver air to the clean room to disperse the particles.

[0026] In some embodiments, at least one clean device may be deployed in each subspace.

[0027] In some embodiments, the clean room air particle monitoring system 100 may further include at least one of a camera device, a detection device, a storage device, a network, and the like.

[0028] The network may include any suitable network that can facilitate information and / or data exchange of the clean room air particle monitoring system 100. In some embodiments, the processor 110, the clean equipment 120, the camera device, the detection device, and the storage device may exchange information and / or data through the network. For example, the processor 110 may establish a connection with the clean equipment 120 through the network. In some embodiments, the network 120 may be any one or more of a wired network or a wireless network.

[0029] The storage device may be used to store data, instructions, and / or any other information. In some embodiments, the storage device may store data and / or information obtained from at least one component of the clean room air particle monitoring system 100 or an external data source. For example, the storage device may store historical accident data. In some embodiments, the storage device may store data and / or instructions that the processor 110 uses to execute or use to complete the exemplary methods described in this specification. In some embodiments, the storage device is integrated into the processor 110.

[0030] The camera device is configured to collect activity data of at least one subspace. In some embodiments, the camera device can capture images of personnel and / or equipment activities in the clean room and send them to the processor. In some embodiments, at least one camera device can be deployed in at least one subspace.

[0031] The detection device is configured to collect the number of particles at different positions in at least one subspace. In some embodiments, the detection device may include a particle counter, a particle counter, etc. In some embodiments, at least one detection device may be deployed in at least one subspace.

[0032] It should be understood that Figure 1The system and its modules shown can be implemented in various ways. It should be noted that the above description of the clean room air particle monitoring system and its equipment is only for the convenience of description and cannot limit the present specification to the scope of the embodiments given.

[0033] Figure 2 FIG. 200 is an exemplary flow chart of a method for monitoring air particles in a clean room according to some embodiments of the present specification. In some embodiments, process 200 is executed by processor 110. Figure 2 As shown, process 200 includes the following steps:

[0034] Step 210: Acquire activity data of at least one subspace in the clean room.

[0035] A subspace refers to a portion of space in a clean room. In some embodiments, the subspace can be pre-set based on the layout of the clean room. For example, there are multiple spaces for placing equipment and / or personnel in the clean room, and each independent space for placing equipment and / or personnel is a subspace. Clean rooms have different uses and different equipment is placed. For example, a clean room is used to produce integrated circuits, and the equipment may include etching equipment, photolithography equipment, etc.

[0036] Activity data refers to data used to reflect the activities of people and / or the location of equipment in a subspace. In some embodiments, the activity data may include the trajectory of people's activities and the location of equipment. The trajectory of people's activities refers to data related to people's activities. In some embodiments, the trajectory of people's activities may include the starting point, the end point, and the trajectory between the starting point and the end point of the people's activities. The location of the equipment may include the location of the equipment.

[0037] In some embodiments, the processor may acquire images of people and / or equipment in each subspace at a preset frequency through camera devices corresponding to different subspaces, and determine activity data through image recognition algorithms, etc. The preset frequency may be preset based on historical experience, such as acquiring images of people and / or equipment in each subspace every 60 seconds.

[0038] Step 220 , based on the activity data of the at least one subspace, estimating the future diffusion characteristics of the at least one subspace.

[0039] The future diffusion feature is used to characterize the diffusion trend of particles at multiple points in the subspace in a preset future time period. In some embodiments, the diffusion trend may include the diffusion direction and diffusion speed of the particles. The preset future time period refers to a period of time in the future after the activity data is acquired. In some embodiments, the preset future time period can be preset based on historical experience, such as the next 5 seconds, the next 10 seconds, etc.

[0040] The multiple points are used to represent multiple positions of the subspace. In some embodiments, the processor may grid the subspace according to the unit space size, and use each grid as a point. The unit space size may be preset based on historical experience.

[0041] The diffusion speed of the particles in the preset future period refers to the average diffusion speed of the particles in the preset future period. The diffusion direction of the particles in the preset future period is assumed to be constant.

[0042] In some embodiments, the duration of the preset future period may be related to the personnel activity time, device usage time, etc. of the preset past period in the subspace. For example, the longer the personnel activity time and / or device usage time, the shorter the duration of the preset future period. The preset past period refers to a period of time in the past when the activity data is obtained. In some embodiments, the preset past period can be preset based on historical experience, such as the past 1 day, etc.

[0043] Personnel activity time refers to the length of time that a person is active in a subspace. Equipment usage time refers to the length of time that an equipment is used in a subspace. In some embodiments, the processor may obtain images of persons and / or equipment in each subspace in a preset past period from a storage device, and determine the personnel activity time and equipment usage time in each subspace in a preset past period by means of an image recognition algorithm or the like.

[0044] In some embodiments, the processor may determine the duration of the preset future period by a preset formula based on the personnel activity time and device usage time of the preset past period. For example, the preset formula may be as shown in formula (1):

[0045] T=t1 / (w1*s1 / t2+w2*s2 / t2) (1)

[0046] Wherein, T represents the duration of the preset future period, t1 represents the length of the reference period, s1 represents the equipment usage time, s2 represents the personnel activity time, t2 represents the preset past period, w1 and w2 represent the coefficient of the equipment usage time and the coefficient of the personnel activity time respectively. The coefficient can be preset based on historical experience.

[0047] The reference period length refers to the length of time used to determine the length of a preset future period. In some embodiments, the processor may count the preset historical period with the most occurrences and the lowest accident frequency in the historical data, and use the length of the preset historical period as the reference period length. The accident frequency refers to the frequency of accidents in the clean room. Accidents may include fires in the clean room or product scrapping.

[0048] In some embodiments of the present specification, personnel activity time and equipment usage time are taken into consideration when determining the preset future time period, so that clean rooms that are used frequently can have shorter preset future time periods, which is conducive to more frequent determination of future diffusion characteristics and improved accuracy of the determined future diffusion characteristics.

[0049] In some embodiments, the processor can evaluate the future diffusion characteristics of each subspace in at least one subspace in a variety of ways based on the activity data of each subspace in at least one subspace. For example, the processor can determine the tangent direction of the activity trajectory and determine the direction as the diffusion direction of the particles whose points are on the activity trajectory of the personnel, and determine the average speed of the activity trajectory as the diffusion speed of the particles at the point; the direction perpendicular to the device surface can be determined as the diffusion direction of the particles whose points are around the device, and the default diffusion speed can be determined as the diffusion speed of the particles at the point; the default direction can be determined as the diffusion direction of the particles whose points are at other positions in the subspace, and the diffusion speed of the particles at the point is determined to be 0. Among them, the default diffusion speed and the default direction can be preset based on historical experience. The average speed of the activity trajectory refers to the average speed of the movement of the personnel on the activity trajectory. In some embodiments, the processor can obtain one or more images of the activities of the personnel in the subspace through a camera device, and determine the average speed of the activity trajectory through an image recognition algorithm and other methods.

[0050] It can be understood that by determining the diffusion direction of particles at other positions in the subspace as the default direction and the diffusion speed as 0, the influence of particles at other positions on the diffusion trend of particles on the activity track of personnel and around the equipment can be eliminated.

[0051] In some embodiments, the future diffusion characteristic may be related to the airflow distribution characteristic of the at least one subspace. The processor may estimate the future diffusion characteristic based on the activity data and the airflow distribution characteristic.

[0052] The airflow distribution feature is used to characterize the airflow in the subspace. In some embodiments, the airflow distribution feature may include the airflow direction and airflow velocity at each point in the subspace.

[0053] In some embodiments, the airflow distribution characteristics of at least one subspace can be determined by manual monitoring. For example, a technician can monitor the airflow direction and airflow velocity at each point when there are no people in the clean room to obtain the airflow distribution characteristics of at least one subspace. Manual monitoring can be performed by any feasible method such as using a gas monitor.

[0054] In some embodiments, the processor can determine the superposition direction of the tangent direction of the activity trajectory and the airflow direction of the intersection point as the diffusion direction of the particles at the point on the activity trajectory of the personnel, and determine the superposition speed of the average speed of the activity trajectory and the airflow speed of the intersection point as the diffusion speed of the particles at the point; the superposition direction of the direction perpendicular to the device surface and the airflow direction of the device surface can be determined as the diffusion direction of the particles at the point around the device, and the superposition speed of the default diffusion speed and the airflow speed on the device surface can be determined as the diffusion speed of the particles at the point; the airflow direction of the point where the particle is located can be determined as the diffusion direction of the particles at other positions in the subspace, and the airflow speed of the point where the particle is located can be determined as the diffusion speed of the particles at the point. Among them, the superposition direction refers to the synthesis of two directions into one direction by methods such as vector addition. The superposition speed refers to the synthesis of two speeds into one speed by methods such as vector addition.

[0055] In some embodiments of the present specification, by considering the airflow conditions in at least one subspace within the clean room, future diffusion characteristics that better reflect the actual diffusion trend can be obtained, which is conducive to determining more accurate cumulative distribution characteristics.

[0056] In some embodiments, the processor can estimate future diffusion characteristics based on the number of air particles, hot spot distribution, and airflow distribution characteristics. Figure 3 and its related description.

[0057] Step 230, based on the future diffusion characteristics and the activity data, determine the cumulative distribution characteristics of multiple future time points.

[0058] Multiple future time points refer to some time points after the activity data is obtained. In some embodiments, multiple future time points can be preset based on historical experience. For example, the 10th second, 13th second, 16th second, etc. after the activity data is obtained.

[0059] In some embodiments, the number of multiple future time points may be related to historical accident data. Historical accident data refers to data related to accidents that occurred at historical times. In some embodiments, the historical accident data may include historical accident times. For relevant descriptions of accidents, please refer to the relevant description above.

[0060] The historical accident time refers to the time when the accident occurred in the historical time. In some embodiments, the processor can remove the date information in the historical accident time, retain the time information in the historical accident time, select the time information that is later than the preset time interval of the current time information, construct a clustering vector based on the time information for clustering, and use the time points corresponding to the multiple cluster centers formed by clustering as future time points. Among them, the preset time interval can be preset, such as 1 hour. The date information includes the year, month and day in the historical accident time. The time information includes the hours and minutes in the historical accident time. The clustering vector constructed based on the time information can be a clustering vector constructed based on hours and minutes. In some embodiments, the processor can obtain the pre-stored historical accident time through a storage device.

[0061] Exemplarily, the current time information is 12:00, and the time information later than the preset time interval of the current time information includes 12:03, 12:10, 12:10, 12:08, 12:08 and 12:03. The processor constructs a clustering vector based on the time information for clustering, and uses the time points 12:03, 12:08 and 12:10 corresponding to the multiple cluster centers formed by clustering as future time points. It should be noted that this example is only used to illustrate the process of determining multiple future time points, and is not an actual clustering process.

[0062] In some embodiments of the present specification, determining multiple future time points through the time when accidents occur most frequently is helpful in determining more important future time points, and then obtaining the cumulative distribution characteristics of the time when accidents occur most frequently, so as to more accurately judge whether to control the clean equipment to provide regulation and control for the clean room.

[0063] The cumulative distribution feature is used to characterize the distribution of the particle cumulative value in each subspace. In some embodiments, the cumulative distribution feature can be represented by a sequence or the like. For example, the cumulative distribution feature can be [(preset position 1, particle cumulative value), (preset position 2, particle cumulative value) ...]. The preset position can be pre-set based on historical experience. The preset position can cover multiple points in the subspace.

[0064] The particle accumulation value refers to the total number of particles at multiple points within a preset position at multiple future time points.

[0065] In some embodiments, the processor can determine the cumulative distribution characteristics of multiple future time points in a variety of ways based on future diffusion characteristics and activity data. For example, the processor can construct Newton's motion equation based on the diffusion direction and diffusion speed of the particles in the future diffusion characteristics, bring a future time point into Newton's motion equation, obtain the point position of each particle at the future time point, and count the point position of each particle at the future time point to obtain the distribution position of the particles in the subspace; cluster the distribution positions of the particles to form multiple cluster centers, and determine the preset position corresponding to the cluster center where the number of particles is greater than the number threshold and the number of particles in the preset position as the cumulative distribution characteristics of the future time point, and so on to obtain the cumulative distribution characteristics of multiple future time points. Among them, the preset position can correspond to one or more cluster centers. The number threshold can be pre-set based on historical experience. Newton's motion equation refers to the equation that describes the movement of particles.

[0066] In some embodiments, the processor may obtain, through the detection device, the number of particles at multiple points within a preset position at the time point of acquiring the activity data.

[0067] In some embodiments, the processor may cluster the distribution positions of the particles in any feasible manner. For example, the processor may construct a spatial coordinate system and determine the coordinates of the points where the particles are located, and cluster the coordinates of the points where the particles are located as clustering vectors to form a preset number of cluster centers. The preset number may be pre-set based on historical experience.

[0068] In some embodiments, the cumulative distribution characteristics may be related to the spatial structure data, and the processor may determine the cumulative distribution characteristics based on the future diffusion characteristics, the activity data, and the spatial structure data.

[0069] The spatial structure data refers to data related to the spatial structure of the clean room. In some embodiments, the spatial structure data may include enclosure information within at least one subspace.

[0070] The enclosure information refers to information related to the enclosure wall and / or partition board in the subspace. In some embodiments, the enclosure information may include the location of the enclosure wall and / or partition board in the subspace and the size of the area surrounded by the enclosure wall and / or partition board. In some embodiments, the enclosure information may be determined by user input or the like.

[0071] In some embodiments, the enclosure information may also include a variety of information related to the partitioned space, such as information related to the anti-static door curtain in the subspace.

[0072] In some embodiments, the processor may construct a first target vector based on future diffusion characteristics, activity data, and spatial structure data, match the first target vector with the first standard vector, and select the first standard vector that satisfies a preset matching condition with the first target vector as the first standard vector that successfully matches, and determine the cumulative distribution feature corresponding to the first standard vector that successfully matches as the cumulative distribution feature of the first target vector. The preset matching condition includes the highest similarity. The similarity of vectors is negatively correlated to the distance between vectors, and the distance between vectors may include a cosine distance, etc.

[0073] In some embodiments, the processor may construct a clustering vector based on the historical diffusion characteristics, historical activity data, historical spatial structure data, and actual accumulated distribution characteristics in the historical data for clustering, and construct the historical diffusion characteristics, historical activity data, and historical spatial structure data corresponding to the cluster center formed by clustering into a first standard vector, and use the actual accumulated distribution characteristics corresponding to the cluster center formed by clustering as the accumulated distribution characteristics corresponding to the first standard vector. Among them, the types of clustering algorithms may include multiple types, for example, the clustering algorithm may include K-Means clustering, density-based clustering method (DBSCAN), etc. The method of determining the cluster center may be carried out in any manner known to those skilled in the art.

[0074] The actual cumulative distribution feature is used to characterize the actual distribution of particles at multiple second historical time points relative to the first historical time point in the subspace. In some embodiments, the processor can obtain the actual positions of the particles at multiple second historical time points in the subspace through a detection device, and determine the actual cumulative distribution feature based on the actual positions of the particles at multiple second historical time points in the subspace. The first historical time point refers to the time point at which historical activity data is obtained. The second historical time point is later than the first historical time point and is a future time point of the first historical time point.

[0075] In some embodiments of the present specification, when determining the accumulation distribution characteristics, the influence of the walls and / or isolation panels in the subspace on the diffusion of particles and the accumulation of particles are considered, so that more accurate accumulation distribution characteristics can be determined, which is conducive to precise control of the clean room.

[0076] In some embodiments, the processor may construct a spatial map based on future diffusion characteristics, activity data, and spatial structure data; based on the spatial map, the accumulated distribution characteristics may be determined by a feature determination model. For an explanation of this part, please refer to Figure 4 and its related description.

[0077] Step 240, based on the accumulation distribution characteristics, determine the particle accumulation area and control the clean equipment to regulate the clean room.

[0078] The particle accumulation area refers to an area in the clean room where particles may accumulate in the future. In some embodiments, the processor can count the distribution positions of particles in the clean room at the future time based on the accumulation distribution characteristics of multiple future time points in at least one subspace, and determine the distribution positions of particles in the clean room at the future time as the particle accumulation area.

[0079] In some embodiments, in response to determining the particle accumulation area, the processor can control the clean equipment to regulate the clean room. For example, in response to determining the particle accumulation area, the processor can control the clean equipment to deliver airflow to the particle accumulation area to blow away the particles in the particle accumulation area.

[0080] In some embodiments of the present specification, by predicting the distribution of particles in advance, it is possible to know in advance the areas where particles may accumulate, so as to control the clean room in advance to avoid risks such as fire or product scrapping caused by particle accumulation.

[0081] It should be noted that the above description of process 200 is only for illustration and description, and does not limit the scope of application of this specification. For those skilled in the art, various modifications and changes can be made to the process hand-eye calibration under the guidance of this specification. However, these modifications and changes are still within the scope of this specification.

[0082] Figure 3 FIG. 300 is an exemplary flow chart of evaluating future diffusion characteristics according to some embodiments of the present specification. In some embodiments, process 300 is executed by processor 110. Figure 3 As shown, process 300 includes the following steps:

[0083] Step 310: Based on the activation time point of the device, the number of air particles at the activation time point of the device is obtained through a detection device. For a description of the detection device, see Figure 1 and its related description.

[0084] In some embodiments, the activity data may include an activation time point of a device in at least one subspace. The activation time point of a device refers to a time point when the device starts to work. In some embodiments, the processor may communicate with the device to obtain the activation time point of the device.

[0085] The number of air particles refers to the total number of particles in the subspace at the time point when the device is activated. In some embodiments, the processor can obtain the number of air particles at the time point when the device is activated through a detection device.

[0086] Step 320: Determine the hotspot distribution of at least one subspace based on the activity data.

[0087] For a description of activity data, see Figure 2and its related description.

[0088] The hotspot distribution is used to characterize the frequency of personnel activities at each preset position in the subspace. In some embodiments, the hotspot distribution can be represented by a sequence or the like. For example, the hotspot distribution can be [(preset position 1, activity frequency 1), (preset position 2, activity frequency 2) ...]. For a description of the preset positions, see Figure 2 and its related description.

[0089] Activity frequency can represent the frequency of personnel activities. The higher the activity frequency, the more frequent the personnel activities. For example, the more times a person moves back and forth, the higher the activity frequency.

[0090] In some embodiments, the processor may divide the activity trajectory of at least one person into multiple sub-trajectories according to a default unit length, and determine the point where the midpoint of each sub-trajectory is located as an activity point. An activity point refers to a point that characterizes the position of a sub-trajectory. The default unit length may be preset based on historical experience, such as 1 meter.

[0091] In some embodiments, the processor may cluster multiple activity points based on the location of the enabled device to form multiple cluster centers, determine the cluster centers formed by clustering as hot spots, and determine the hot spot distribution based on the cluster centers and the preset locations where the cluster centers are located by the correspondence between the number of hot spots and the activity frequency. The preset location may correspond to one or more cluster centers. The correspondence may be pre-set based on historical experience, including that the more hot spots there are in the preset location, the greater the activity frequency of the preset location.

[0092] In some embodiments, the processor may cluster the multiple activity points in any feasible manner. For example, the processor may construct a spatial coordinate system and determine the coordinates of the activity points, and cluster the coordinates of the activity points as clustering vectors to form a preset number of cluster centers. The preset number may be pre-set based on historical experience.

[0093] Step 330 , based on the number of airborne particles, hot spot distribution, and airflow distribution characteristics, future diffusion characteristics are evaluated.

[0094] For a description of the airflow distribution characteristics and future diffusion characteristics, see Figure 2 and its related description.

[0095] In some embodiments, the processor can evaluate the future diffusion characteristics in a variety of ways based on the number of air particles, hot spot distribution, and airflow distribution characteristics. For example, the processor can construct a second target vector based on the number of air particles, hot spot distribution, and airflow distribution characteristics, match the second target vector with the second standard vector, and select the second standard vector that meets the preset matching conditions with the second target vector as the second standard vector that successfully matches, and determine the actual future diffusion characteristics corresponding to the successfully matched second standard vector as the future diffusion characteristics of the second target vector. Among them, the preset matching conditions include the highest similarity. The similarity of vectors is negatively correlated to the distance between vectors, and the distance between vectors can include cosine distance, etc.

[0096] In some embodiments, the processor can construct a clustering vector for clustering based on the historical hotspot distribution, historical airflow distribution characteristics, historical air particle counts, and actual future diffusion characteristics in the historical data, and construct the historical hotspot distribution, historical airflow distribution characteristics, and historical air particle counts corresponding to the cluster center formed by clustering into a second standard vector, and use the actual future diffusion characteristics corresponding to the cluster center formed by clustering as the future diffusion characteristics corresponding to the second standard vector.

[0097] The actual future diffusion feature is used to characterize the actual diffusion of particles in the subspace at the third historical time point relative to the first historical time point. In some embodiments, the processor can obtain the actual diffusion direction and diffusion speed of the particles in the subspace at the third historical time point through the detection device, and determine the actual diffusion direction and diffusion speed of the particles as the actual future diffusion feature. The third historical time point is later than the first historical time point and earlier than the second historical time point. For the description of the first historical time point and the second historical time point, please refer to Figure 2 and its related description.

[0098] The diffusion assessment model refers to a model used to determine future diffusion characteristics. In some embodiments, the diffusion assessment model may be a machine learning model. For example, the diffusion assessment model may include any one or combination of a convolutional neural network (CNN) model, a neural network (NN) model, or other custom model structures.

[0099] In some embodiments, the processor may train the diffusion assessment model based on a large number of first training samples with first labels by gradient descent method or the like. The first training sample may include the number of sample air particles, sample hot spot distribution, and sample airflow distribution characteristics, and the first label of the first training sample may be the actual future diffusion characteristics corresponding to the sample hot spot distribution. In some embodiments, the first training sample may be obtained based on historical data. For an explanation of the actual future diffusion characteristics, please refer to the relevant description above.

[0100] In some embodiments, the diffusion assessment model can be trained in the following manner: multiple first training samples with first labels are input into the initial diffusion assessment model, a loss function is constructed by the first label and the prediction result of the initial diffusion assessment model, the initial diffusion assessment model is updated based on the iteration of the loss function, and the diffusion assessment model training is completed when the loss function of the initial diffusion assessment model meets the preset condition. The preset condition may be that the loss function converges, the number of iterations reaches a set value, etc.

[0101] In some embodiments of the present specification, the self-learning ability of the machine learning model is used to find patterns from a large amount of data, obtain the correlation between the number of air particles, hot spot distribution, and airflow distribution characteristics, and improve the accuracy and efficiency of determining future diffusion characteristics.

[0102] In some embodiments of the present specification, considering the concentrated areas of personnel activities in the clean room and the impact on particles in the air before and after the equipment is activated can improve the accuracy of the determined future diffusion characteristics, which is conducive to the subsequent determination of more accurate accumulation distribution characteristics.

[0103] It should be noted that the above description of process 300 is only for illustration and description, and does not limit the scope of application of this specification. For those skilled in the art, various modifications and changes can be made to the process hand-eye calibration under the guidance of this specification. However, these modifications and changes are still within the scope of this specification.

[0104] Figure 4 It is an exemplary schematic diagram of a feature determination model according to some embodiments of the present specification.

[0105] In some embodiments, the accumulation profile characteristics may include particle arrival time and diffusion duration.

[0106] The particle arrival time refers to the time when the particle arrives at the preset position. For the description of the preset position, please refer to Figure 2 and its related description.

[0107] Diffusion duration refers to the duration of diffusion of particles at a preset location.

[0108] In some embodiments, the processor may construct a spatial map 440 based on the future diffusion feature 410, the activity data 420, and the spatial structure data 430; based on the spatial map 440, the cumulative distribution feature 460 may be determined by the feature determination model 450. For more information on future diffusion features, activity data, spatial structure data, and cumulative distribution features, see Figure 2 and its related description.

[0109] The spatial graph 440 refers to a graph structure including connection relationships between subspaces. The graph structure is a data structure composed of nodes and edges, where the edges connect the nodes, and the nodes and edges may have features.

[0110] In some embodiments, the processor can construct a spatial map 440 based on the subspace area, future diffusion characteristics 410, activity data 420, spatial structure data 430, and channel area. The nodes of the spatial map 440 include subspace nodes (e.g., subspace nodes 411-1...). The edges of the spatial map (e.g., edges 412-1...) indicate that two subspaces are adjacent, and there is no edge between non-adjacent subspaces. Among them, the subspace node refers to the node representing the subspace.

[0111] In some embodiments, the node features of the subspace nodes in the spatial map are the subspace area, future diffusion features, and activity data corresponding to the subspace. The edge features are spatial structure data, channel area, etc. Among them, if there is a channel between the subspaces, the edge features include the channel area, which refers to the area of ​​the longitudinal section of the channel between the subspaces. The subspace area and channel area can be obtained by user input or other methods.

[0112] In some embodiments, the node features of the spatial map may also include the number of air particles and the location of the device. For an explanation of the number of air particles and the location of the device, see Figure 2 , Figure 3 and its related description.

[0113] In some embodiments of the present specification, since particles are generated when the equipment is turned on, considering the impact of the equipment turning on the number of particles in the subspace, the number of air particles and the location of the equipment are added to the node features of the spatial map. This can make the information contained in the spatial map more accurate and diverse, thereby improving the accuracy of the output of the feature determination model, which is conducive to further improving the accuracy of the determined cumulative distribution features, and thus facilitating the subsequent precise control of the clean room.

[0114] The feature determination model 450 refers to a model for determining cumulative distribution features. In some embodiments, the feature determination model 450 may be a machine learning model. For example, the feature determination model may include any one or combination of a graph neural network (GNN) model, a neural network (NN) model, or other custom model structures.

[0115] In some embodiments, the processor may determine the model 450 based on a large number of second training samples with second labels by training features such as gradient descent method. The second training sample may include a sample space map, and the second label of the second training sample may be the actual accumulated distribution feature corresponding to the sample space map. In some embodiments, the second training sample may be acquired based on historical data. The sample space map may include a historical space map determined based on historical data, and the nodes and attributes, edges and attributes of the historical space map are similar to the above description. For an explanation of the actual accumulated distribution feature, please refer to the relevant description above.

[0116] The training process of the feature determination model is similar to that of the diffusion assessment model. For details, see the implementation method of the training process of the diffusion assessment model.

[0117] In some embodiments of the present specification, by constructing a spatial map, the subspaces within the clean room and the data related to the subspaces can be effectively organized, the spatial map can be input into the feature determination model, and the self-learning ability of the machine learning model can be used to further improve the accuracy of the determined cumulative distribution characteristics, which is facilitating the subsequent precise control of the clean room.

[0118] Some embodiments of the present specification also provide a clean room air particle monitoring device, including at least one processor and at least one memory; the at least one memory is used to store computer instructions; the at least one processor is used to execute at least part of the computer instructions to implement the clean room air particle monitoring method described in any of the above embodiments.

[0119] Some embodiments of the present specification also provide a computer-readable storage medium, which stores computer instructions. When a computer reads the computer instructions in the storage medium, the computer executes the clean room air particle monitoring method described in any one of the above embodiments.

[0120] Furthermore, certain features, structures or characteristics in one or more embodiments of this specification may be appropriately combined.

[0121] In some embodiments, numbers describing the number of components and attributes are used. It should be understood that such numbers used in the description of the embodiments are modified by the modifiers "about", "approximately" or "substantially" in some examples. Unless otherwise specified, "about", "approximately" or "substantially" indicate that the numbers are allowed to vary by ±20%. Accordingly, in some embodiments, the numerical parameters used in the specification and claims are approximate values, which may change according to the required features of individual embodiments. In some embodiments, the numerical parameters should take into account the specified significant digits and adopt the general method of retaining digits. Although the numerical domains and parameters used to confirm the breadth of their range in some embodiments of this specification are approximate values, in specific embodiments, the setting of such numerical values ​​is as accurate as possible within the feasible range.

[0122] If there is any inconsistency or conflict between the descriptions, definitions, and / or usage of terms in the materials cited in this specification and the contents described in this specification, the descriptions, definitions, and / or usage of terms in this specification shall prevail.

Claims

1. A method for monitoring air particles in a clean room, characterized in that: The method is executed by a processor, and the method includes: Acquire activity data of at least one subspace in the clean room, wherein the activity data is a trajectory of personnel activities and a location of equipment in the at least one subspace; estimating future diffusion characteristics of the at least one subspace based on the activity data of the at least one subspace; Determining cumulative distribution characteristics of a plurality of future time points based on the future diffusion characteristics and the activity data; the plurality of future time points are related to historical accident data; The cumulative distribution characteristics include particle arrival time and diffusion duration, and the determining of the cumulative distribution characteristics at multiple future time points based on the future diffusion characteristics and the activity data includes: The cumulative distribution feature is related to the spatial structure data, and the spatial structure data includes the enclosure information in the at least one subspace; and determining the cumulative distribution feature includes: Based on the future diffusion characteristics, the activity data and the spatial structure data, a spatial map is constructed; wherein the node characteristics of the subspace node in the spatial map are the subspace area corresponding to the subspace, the future diffusion characteristics, the activity data, the number of air particles in the subspace when the device is activated, and the position of the device in the subspace, and the edge characteristics are the spatial structure data and the channel area, and the channel area refers to the area of ​​the longitudinal section of the channel between the subspaces; Based on the spatial map, the accumulation distribution feature is determined by a feature determination model, The feature determination model is a graph neural network model; Based on the accumulation distribution characteristics, the particle accumulation area is determined and the clean equipment is controlled to regulate the clean room.

2. The method according to claim 1, characterized in that The future diffusion characteristic is related to the airflow distribution characteristic of the at least one subspace, and the evaluating the future diffusion characteristic of the at least one subspace based on the activity data of the at least one subspace includes: Based on the activity data and the airflow distribution characteristics, the future diffusion characteristics are estimated.

3. The method according to claim 2, characterized in that The activity data includes an activation time point of the device in the at least one subspace, and the evaluating the future diffusion characteristics based on the activity data and the airflow distribution characteristics includes: Based on the activation time point of the device, obtaining the number of air particles at the activation time point of the device through a detection device; Based on the activity data, determining a hotspot distribution of the at least one subspace; The future diffusion characteristics are estimated based on the air particle quantity, the hot spot distribution, and the airflow distribution characteristics.

4. A clean room air particle monitoring system, characterized in that: The system includes a processor and a clean device, wherein the processor is in communication connection with the clean device, and the processor is configured as follows: Acquire activity data of at least one subspace in the clean room, wherein the activity data is a trajectory of personnel activities and a location of equipment in the at least one subspace; estimating future diffusion characteristics of the at least one subspace based on the activity data of the at least one subspace; Determining cumulative distribution characteristics of a plurality of future time points based on the future diffusion characteristics and the activity data; the plurality of future time points are related to historical accident data; The cumulative distribution characteristics include particle arrival time and diffusion duration, and the determining of the cumulative distribution characteristics at multiple future time points based on the future diffusion characteristics and the activity data includes: The cumulative distribution feature is related to the spatial structure data, and the spatial structure data includes the enclosure information in the at least one subspace; and determining the cumulative distribution feature includes: Based on the future diffusion characteristics, the activity data and the spatial structure data, a spatial map is constructed; wherein the node characteristics of the subspace node in the spatial map are the subspace area corresponding to the subspace, the future diffusion characteristics, the activity data, the number of air particles in the subspace when the device is activated, and the position of the device in the subspace, and the edge characteristics are the spatial structure data and the channel area, and the channel area refers to the area of ​​the longitudinal section of the channel between the subspaces; Based on the spatial map, the accumulation distribution feature is determined by a feature determination model, The feature determination model is a graph neural network model; Based on the accumulation distribution characteristics, the particle accumulation area is determined and the clean equipment is controlled to regulate the clean room.

5. The system according to claim 4, characterized in that The future diffusion characteristic is related to the airflow distribution characteristic of the at least one subspace, and the processor is further configured to: Based on the activity data and the airflow distribution characteristics, the future diffusion characteristics are estimated.

6. A clean room air particle monitoring device, characterized in that: The apparatus comprises at least one processor and at least one memory; The at least one memory is used to store computer instructions; The at least one processor is configured to execute at least part of the computer instructions to implement the clean room air particle monitoring method according to any one of claims 1 to 3.

7. A computer-readable storage medium, characterized in that: The storage medium stores computer instructions. When the computer reads the computer instructions in the storage medium, the computer executes the clean room air particle monitoring method according to any one of claims 1 to 3.

Images