IoT-based method for monitoring the condition and lifecycle management of cleanroom panels for aluminum plates

Abstract

This invention relates to the field of data processing technology, specifically to a method for monitoring the status and lifecycle management of cleanroom aluminum plates based on the Internet of Things (IoT). The method includes: acquiring time-series data of the plate surface temperature field and auxiliary monitoring indicators; acquiring infrared images of the cleanroom aluminum plate and filtering abnormal pixels; merging temperature anomaly regions and calculating the region anomaly deviation index; for any auxiliary monitoring indicator, calculating the indicator anomaly deviation index, determining the abnormal growth period and recovery period of the auxiliary indicator; determining the temperature anomaly growth period of the cleanroom aluminum plate and calculating the response index; determining the probability that the temperature anomaly region is dominated by a real performance defect; calculating the warning priority and outputting warning information for each temperature anomaly region. This invention can improve the accuracy, reliability, and lifecycle management efficiency of cleanroom aluminum plate status monitoring.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and more specifically to a method for monitoring the status and lifecycle management of cleanroom aluminum plates based on the Internet of Things. Background Technology

[0002] In modern cleanroom engineering, aluminum cleanroom panels serve as a core component of the enclosure structure. Their airtightness, thermal stability, and surface cleanliness directly impact the operational safety and product quality of the clean area. Condition monitoring data for aluminum cleanroom panels is collected throughout their entire lifecycle, from production and installation to daily operation, maintenance, and replacement. This data is crucial for determining panel performance degradation and developing maintenance strategies. Aluminum cleanroom panels are characterized by large-area installation, complex service environments, and a gradual degradation process. Early performance deterioration (such as internal delamination, air leakage, and dust accumulation) often manifests as weak temperature field anomalies, which are easily confused with normal environmental disturbances such as air conditioning start-up and shutdown, personnel movement, and equipment heat dissipation.

[0003] Currently, existing methods typically rely on single-point threshold alarms, such as using infrared thermometers to sample or a single humidity sensor to determine the condition of the board material.

[0004] However, in actual monitoring of cleanroom aluminum panels, environmental disturbances (such as the operation of equipment in the cleanroom and interference from personnel behavior) may cause temperature change zones on the surface of the cleanroom aluminum panels that are similar to actual performance abnormalities. Existing methods cannot distinguish between transient environmental disturbances and actual performance degradation of the panels, resulting in high false alarm rates, missed detection of real defects, and an inability to achieve a closed loop from condition monitoring to graded early warning and lifecycle management. Summary of the Invention

[0005] This invention provides an IoT-based method for monitoring the status and lifecycle management of cleanroom aluminum plates to solve existing problems.

[0006] The IoT-based method for monitoring the status and lifecycle management of cleanroom aluminum plates of the present invention adopts the following technical solution: One embodiment of the present invention provides a method for monitoring the status and lifecycle management of cleanroom panels for aluminum plates based on the Internet of Things, the method comprising the following steps: Acquire time-series data of surface temperature field and auxiliary monitoring indicators of aluminum cleanroom panels throughout their entire life cycle; Infrared images of clean aluminum plates are obtained based on time-series data of plate surface temperature field. Abnormal pixels are filtered out from all pixels based on the degree of temperature change abnormality of each pixel in the infrared image over time. Abnormal pixels that meet the spatial adjacency condition and temperature consistency condition are merged into a temperature abnormality region, and the regional abnormality deviation index of the temperature abnormality region is calculated. For any auxiliary monitoring indicator, calculate its abnormal deviation index, and determine the period of abnormal growth and recovery of each auxiliary monitoring indicator. Based on the regional abnormal deviation index, the period of abnormal temperature growth in the aluminum plate cleanroom is determined. When the period of abnormal growth of any auxiliary monitoring indicator overlaps with the period of abnormal temperature growth, the response index of the auxiliary monitoring indicator to the period of abnormal temperature growth is calculated based on the proportion of the overlap duration and the difference between the growth coefficient of the indicator abnormal deviation index of the auxiliary monitoring indicator during the period of abnormal temperature growth and the growth coefficient of the regional abnormal deviation index. Based on the mean and standard deviation of the response index of all auxiliary monitoring indicators and the mean duration of the recovery period of each auxiliary indicator, the possibility that the temperature anomaly area is dominated by a real performance defect is determined. The monitoring tags of the aluminum cleanroom plate at each stage of its life cycle are extracted, and the state vector is generated by word vector method. Combined with the probability dominated by real performance defects, the warning priority is calculated and the warning information for each temperature abnormality area is output.

[0007] Furthermore, infrared images of the clean aluminum plate are obtained based on the time-series data of the plate surface temperature field. Based on the degree of temperature change anomaly of each pixel in the infrared image over time, abnormal pixels are filtered from all pixels, specifically including: Infrared images of clean aluminum plates are generated based on the time-series data of the temperature field on the plate surface. The infrared images are composed of pixels, and each pixel corresponds to a spatial location and its temperature time-series value. For each pixel in the infrared image, calculate the absolute value of the temperature difference between adjacent time points within the current working cycle and take the average value to obtain the average temperature difference of that pixel. Calculate the absolute value of the temperature difference between adjacent time points of the pixel within the current working cycle and calculate the standard deviation to obtain the standard deviation of the temperature difference of the pixel. Divide the mean temperature difference of the pixel by the average of the mean temperature differences of other pixels, and then divide by the standard deviation of the temperature difference of the pixel to obtain the degree of abnormal temperature change of the pixel. The degree of temperature change abnormality of each pixel is normalized, and pixels whose normalization result is greater than the preset temperature abnormality threshold are identified as abnormal pixels.

[0008] Furthermore, anomalous pixels that satisfy both spatial adjacency and temperature consistency conditions are merged into a temperature anomaly region, and the regional anomaly deviation index of this temperature anomaly region is calculated, specifically including: For any two anomalous pixels, if their spatial positions in the infrared image are adjacent, it is determined that the two anomalous pixels satisfy the spatial adjacency condition. Obtain the degree of temperature change abnormality of the two abnormal pixels in the current working cycle and calculate the difference; obtain the grayscale mean of the two abnormal pixels in the current working cycle and calculate the difference. Multiply the absolute value of the difference in the degree of temperature change abnormality by the absolute value of the difference in the grayscale mean, and take the negative value of the exponential function to obtain the probability that the two abnormal pixels belong to the same temperature abnormality region. When the probability of these two abnormal pixels belonging to the same temperature anomaly region is greater than the preset merging threshold, it is determined that these two abnormal pixels meet the temperature consistency condition. Abnormal pixels that meet the spatial adjacency condition and the temperature consistency condition are merged into a temperature abnormality region. For each temperature anomaly region, the average edge gradient magnitude of the edge pixels in that temperature anomaly region at the current time and the average edge gradient magnitude at the previous time are obtained, and the ratio of the two is calculated to obtain the gradient magnitude ratio. Obtain the cosine similarity between the gradient direction of all pixels in the temperature anomaly region at the current time and the gradient direction at the previous time to obtain the direction consistency. The product of the gradient magnitude ratio and the direction consistency is determined as the regional anomaly deviation index of the temperature anomaly region.

[0009] Furthermore, for any auxiliary monitoring indicator, its abnormal deviation index is calculated, and the periods of abnormal growth and recovery for each auxiliary monitoring indicator are determined, specifically including: For each auxiliary monitoring indicator, obtain the indicator value at the current monitoring location at the current time, the mean and standard deviation of all indicator values ​​of the auxiliary monitoring indicator at the current monitoring location up to the current time within the current working cycle, and the sum of the differences between the indicator value at the current monitoring location and the indicator values ​​of the same auxiliary monitoring indicator at all other monitoring locations. Add one to the absolute value of the difference between the current indicator value and the mean, multiply by the sum of the differences, and then multiply by the standard deviation to obtain the abnormal deviation index of the auxiliary monitoring indicator at the current monitoring location at the current time. The abnormal deviation index of each auxiliary monitoring indicator at each monitoring location is arranged in time sequence, the arrangement result is smoothed by interpolation, and the valley points and peak points in the smoothed curve are marked. The time period between adjacent valley points and peak points is defined as an abnormal growth period of this auxiliary monitoring indicator; The time period between an adjacent peak and the next trough is defined as a recovery period for this auxiliary monitoring indicator.

[0010] Furthermore, based on the regional anomaly deviation index, the periods of abnormal temperature increase in the aluminum cleanroom plate were determined, specifically including: The regional anomaly deviation index of each temperature anomaly region is arranged in time sequence, the arrangement result is smoothed by interpolation, and the valley points and peak points in the smoothed curve are obtained. The time period between adjacent valley points and peak points is defined as a period of abnormal temperature growth corresponding to the temperature anomaly region.

[0011] Furthermore, based on the proportion of overlapping durations and the difference between the growth coefficient of the abnormal deviation index of this auxiliary monitoring indicator during periods of abnormal temperature growth and the growth coefficient of the regional abnormal deviation index, the response index of this auxiliary monitoring indicator to periods of abnormal temperature growth is calculated, specifically including: Obtain the abnormal deviation index of the auxiliary monitoring indicator at the beginning and end of the abnormal temperature growth period, and determine the growth coefficient of the auxiliary monitoring indicator by the ratio of the abnormal deviation index at the end to the abnormal deviation index at the beginning. Obtain the regional anomaly deviation index at the start time and the regional anomaly deviation index at the end time of the temperature anomaly growth period. The ratio of the regional anomaly deviation index at the end time to the regional anomaly deviation index at the start time is determined as the regional growth coefficient. The absolute value of the difference between the growth coefficient of the indicator and the growth coefficient of the region is used to obtain the growth coefficient deviation. Obtain the intersection duration between the abnormal growth period of the auxiliary monitoring indicator and the abnormal growth period of temperature, calculate the proportion of the intersection duration to the total duration of the abnormal growth period of temperature, and obtain the intersection duration percentage. Divide the percentage of the intersection duration by the sum of the growth coefficient deviation and the preset small constant, and add one to obtain the initial response index of this auxiliary monitoring indicator to the period of abnormal temperature growth. The initial response index is normalized to obtain the response index of this auxiliary monitoring indicator to periods of abnormal temperature growth.

[0012] Furthermore, based on the mean and standard deviation of the response indices of all auxiliary monitoring indicators, as well as the mean duration of the recovery period for each auxiliary indicator, the likelihood that the temperature anomaly area is dominated by a true performance defect is determined, specifically including: Obtain the response index of all auxiliary monitoring indicators corresponding to the temperature anomaly area to each period of temperature anomaly growth, and calculate the arithmetic mean and standard deviation of all response indices; Obtain the recovery period of each auxiliary monitoring indicator within each abnormal temperature growth period corresponding to the abnormal temperature area, calculate the duration of each recovery period, and calculate the arithmetic mean of the durations of all recovery periods; Divide the average response index by the standard deviation of the response index, and then multiply by the average recovery period to obtain the initial value of the probability that the temperature anomaly region is dominated by real performance defects. Normalize the initial probability value to obtain the probability that the temperature anomaly region is dominated by real performance defects.

[0013] Furthermore, monitoring tags for aluminum cleanroom panels at each stage of their lifecycle are extracted. State vectors are generated using word vectorization, and the probability of actual performance defects is considered to calculate warning priorities and output warning information for each temperature anomaly area. Specifically, this includes: The monitoring tags recorded during the manufacturing, transportation and installation stages of the aluminum cleanroom plate are extracted. Each monitoring tag is converted into a word vector using the word vector method and then concatenated in a fixed dimension order to generate the state vector of the aluminum cleanroom plate. Obtain the state vectors of the other aluminum clean panels in the cleanroom where the aluminum clean panel is located, and calculate the average cosine similarity between the state vector of the aluminum clean panel and the state vectors of the other aluminum clean panels, as the mean state similarity value. Obtain the deterioration index of the cleanroom plate of this aluminum plate; Multiply the average probability that all temperature anomaly areas of the aluminum cleanroom plate are dominated by actual performance defects by the deterioration index, and then divide by the sum of the average state similarity and the preset small constant to obtain the initial value of the warning priority of the aluminum cleanroom plate. The initial value of the warning priority is normalized to obtain the warning priority, and the corresponding warning information is output according to the preset level of the warning priority.

[0014] Furthermore, the deterioration index of the cleanroom plate of the aluminum plate is obtained, specifically including: Obtain the total area of ​​all temperature abnormal areas on the clean aluminum plate as the total area of ​​abnormal areas; For each temperature anomaly region, the maximum number of consecutive positive differences in the temporal sequence of the probability that the temperature anomaly region is dominated by a real performance defect is obtained as the deterioration trend value. Obtain the total duration of the temperature anomaly region as the region's duration. Divide the average duration of all temperature anomaly areas by the total current monitoring duration to obtain the time percentage coefficient. The deterioration index of the aluminum cleanroom plate is obtained by multiplying the total area of ​​the abnormal area, the deterioration trend value, and the time proportion coefficient.

[0015] This invention proposes an IoT-based aluminum plate cleanroom condition monitoring and lifecycle management system, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the computer program is executed by the processor, it implements the steps of the IoT-based aluminum plate cleanroom condition monitoring and lifecycle management method.

[0016] The beneficial effects of the technical solution of the present invention are: In this embodiment of the invention, by analyzing the temporal variation characteristics of the infrared temperature field on the surface of the aluminum cleanroom panel, and combining spatial coherence with the coordinated abnormal responses of multi-dimensional sensors such as humidity, pressure difference, and surface impedance, it is possible to effectively distinguish between instantaneous environmental disturbances such as air conditioning start-up and shutdown and personnel activities, and actual performance degradation of the panel such as internal delamination, air leakage, and dust accumulation, significantly reducing false alarm rate and missed detection rate. On this basis, the warning priority is calculated according to the deterioration trend and life cycle tag vector, and hierarchical management instructions are output. For high-level warnings, repair or replacement is recommended; for medium-level warnings, manual inspection is recommended; and for cases without warnings, the baseline is updated and data is recorded. After maintenance, tags are added to form closed-loop management, thereby improving the accuracy, reliability, and full life cycle management efficiency of aluminum cleanroom panel status monitoring. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A flowchart illustrating an IoT-based method for monitoring the condition and lifecycle management of cleanroom aluminum plates, provided in one embodiment of the present invention; Figure 2 This is a structural diagram of an IoT-based aluminum plate cleanroom status monitoring and lifecycle management system provided in one embodiment of the present invention. Detailed Implementation

[0019] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of the IoT-based aluminum cleanroom plate status monitoring and lifecycle management method proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0021] The following description, in conjunction with the accompanying drawings, details the specific solution of the IoT-based aluminum plate cleanroom status monitoring and lifecycle management method provided by this invention.

[0022] This invention provides a method for monitoring the status and lifecycle management of cleanroom panels on aluminum plates based on the Internet of Things. Please refer to [link / reference]. Figure 1 The diagram illustrates a flowchart of an IoT-based method for monitoring the condition and lifecycle management of cleanroom aluminum plates, according to an embodiment of the present invention. The method includes the following steps: S101. Obtain time-series data of surface temperature field and auxiliary monitoring indicators of aluminum cleanroom panels throughout their entire life cycle.

[0023] In this embodiment, firstly, an RFID tag or QR code is attached to the surface of each clean aluminum plate. A reader / writer is used to record the plate's unique identification code, as well as the inspection tags generated at each stage of its lifecycle (such as factory qualification tags, transportation process tags, installation acceptance tags, etc.). This information is stored in the data processing unit for subsequent construction of the lifecycle state vector.

[0024] Secondly, an infrared thermal imager is installed at a fixed position inside the cleanroom, perpendicular to the plane of the aluminum cleanroom panel. The optical axis of the thermal imager is perpendicular to the center of the panel, and the acquisition frequency is set to once every 10 minutes to continuously collect time-series data of the panel's temperature field. Each frame of the infrared image contains several pixels, and each pixel carries a temperature value. Using existing calibration algorithms, each pixel is mapped to the actual spatial physical coordinates of the aluminum cleanroom panel to obtain the correspondence between the temperature distribution and the spatial position of the panel, forming a temperature distribution matrix.

[0025] Furthermore, multiple types of sensors for auxiliary monitoring indicators are evenly distributed on the surface of each aluminum clean panel, including: a humidity sensor, used to collect the air humidity around the panel and assess the risk of condensation; a micro-differential pressure sensor, whose probes are placed between the clean area and the non-clean area to monitor the pressure difference on both sides of the panel and reflect changes in airtightness; and a surface capacitance sensor, whose electrodes are in close contact with the panel surface to measure the capacitance changes caused by surface dust accumulation and reflect the degree of dust accumulation.

[0026] All sensor data acquisition times are synchronized with infrared thermal image acquisition times, and timestamps are uniformly aligned via Network Time Protocol (NTP) to ensure the accuracy of subsequent time series analysis.

[0027] Finally, the infrared images, time-series data from each sensor, and identity information and tag data read by RFID are transmitted to the data processing unit (e.g., an industrial computer or cloud server) via wired or wireless network to complete data cleaning (removing obvious bad values ​​and filling in missing points), timestamp alignment, and spatial registration, thereby obtaining the time-series data of the plate temperature field and auxiliary monitoring indicators.

[0028] S102. Obtain infrared images of clean aluminum plates based on the time-series data of plate surface temperature field. Based on the degree of abnormal temperature change of each pixel in the infrared image over time, filter out abnormal pixels from all pixels.

[0029] In this embodiment, an infrared image of the clean aluminum plate is obtained based on the time-series data of the plate surface temperature field. Based on the degree of temperature change anomaly of each pixel in the infrared image over time, abnormal pixels are filtered from all pixels. Specifically, this includes: Infrared images of clean aluminum plates are generated based on the time-series data of the temperature field on the plate surface. The infrared images are composed of pixels, and each pixel corresponds to a spatial location and its temperature time-series value. For each pixel in the infrared image, calculate the absolute value of the temperature difference between adjacent time points within the current working cycle and take the average value to obtain the average temperature difference of that pixel. Calculate the absolute value of the temperature difference between adjacent time points of the pixel within the current working cycle and calculate the standard deviation to obtain the standard deviation of the temperature difference of the pixel. Divide the mean temperature difference of the pixel by the average of the mean temperature differences of other pixels, and then divide by the standard deviation of the temperature difference of the pixel to obtain the degree of abnormal temperature change of the pixel. The degree of temperature change abnormality of each pixel is normalized, and pixels whose normalization result is greater than the preset temperature abnormality threshold are identified as abnormal pixels.

[0030] For example, in actual cleanroom working scenarios, the temperature field of aluminum cleanroom panels is not static. Normal environmental disturbances such as air conditioning start-up and shutdown, personnel movement, and equipment operation cause the panel surface temperature to fluctuate regularly over time. When the aluminum cleanroom panel suffers from performance degradation such as internal delamination or air leakage, its thermal inertia changes, resulting in a significant difference in the rate of temperature change in that area compared to the normal area. This embodiment is based on this physical principle and quantifies the degree of anomaly by analyzing the temporal difference characteristics of the temperature of individual pixels.

[0031] First, for each clean aluminum plate, a complete work cycle is selected (e.g., one workday, 8 hours in total, with data collected every 10 minutes, yielding 48 temperature field data points). For any pixel in the infrared image... Extract the temperature time series of this point throughout the entire working cycle. .

[0032] Then, the absolute difference between the temperature values ​​of the pixel at adjacent times within the current working cycle is calculated, and their average is obtained to get the average temperature difference of the pixel, denoted as . It reflects the pixel. The average rate of temperature change within the period. Simultaneously, the standard deviation of the temperature difference at this pixel is calculated. The smaller the standard deviation, the more stable the rate of temperature change at that point, and the more likely the abnormal deviation is caused by a persistent defect rather than random noise.

[0033] In addition, the average value of the temperature difference of all other pixels in the current infrared image is obtained and denoted as . That is, removing pixels Then, the arithmetic mean of the temperature differences of the other pixels. This value represents the average rate of temperature change of the normal board surface.

[0034] Then, the pixel points are calculated. abnormal temperature change The calculation formula can be: ; The physical meaning of this formula is as follows: the larger the mean temperature difference in the numerator, the more drastic the temperature change of that pixel; the larger the average value of the mean temperature differences of the remaining pixels in the denominator, the greater the overall background fluctuation of the board surface, and the higher the anomaly threshold can be accordingly; the smaller the standard deviation of the temperature difference of that pixel in the denominator, the more stable the abnormal deviation of that pixel is, and the less likely it is to be an instantaneous disturbance. Therefore, the calculated degree of temperature change anomaly can comprehensively measure the degree of temperature change anomaly of that pixel relative to the normal board surface.

[0035] After obtaining the degree of temperature change anomaly for all pixels, the minimum and maximum value normalization method is used to map it to... The range is used to obtain the normalized anomaly value. A preset temperature anomaly threshold is set to 0.5. Pixels with a normalization result greater than 0.5 are marked as abnormal pixels, and the remaining pixels are considered normal or areas of environmental disturbance.

[0036] S103. Merge the abnormal pixels that meet the spatial adjacency condition and the temperature consistency condition into a temperature abnormal region, and calculate the regional abnormality deviation index of the temperature abnormal region.

[0037] In this embodiment, abnormal pixels that meet the spatial adjacency condition and the temperature consistency condition are merged into a temperature abnormality region, and the regional abnormality deviation index of the temperature abnormality region is calculated, specifically including: For any two anomalous pixels, if their spatial positions in the infrared image are adjacent, it is determined that the two anomalous pixels satisfy the spatial adjacency condition. Obtain the degree of temperature change abnormality of the two abnormal pixels in the current working cycle and calculate the difference; obtain the grayscale mean of the two abnormal pixels in the current working cycle and calculate the difference. Multiply the absolute value of the difference in the degree of temperature change abnormality by the absolute value of the difference in the grayscale mean, and take the negative value of the exponential function to obtain the probability that the two abnormal pixels belong to the same temperature abnormality region. When the probability of these two abnormal pixels belonging to the same temperature anomaly region is greater than the preset merging threshold, it is determined that these two abnormal pixels meet the temperature consistency condition. Abnormal pixels that meet the spatial adjacency condition and the temperature consistency condition are merged into a temperature abnormality region. For each temperature anomaly region, the average edge gradient magnitude of the edge pixels in that temperature anomaly region at the current time and the average edge gradient magnitude at the previous time are obtained, and the ratio of the two is calculated to obtain the gradient magnitude ratio. Obtain the cosine similarity between the gradient direction of all pixels in the temperature anomaly region at the current time and the gradient direction at the previous time to obtain the direction consistency. The product of the gradient magnitude ratio and the direction consistency is determined as the regional anomaly deviation index of the temperature anomaly region.

[0038] For example, after obtaining anomalous pixels, it is necessary to merge spatially adjacent anomalous pixels with similar temperature change characteristics into connected temperature anomalous regions. This embodiment uses a region growing method to achieve this.

[0039] Specifically, for any two abnormal pixels, the spatial adjacency condition is first determined: if the two pixels are adjacent in the infrared image (e.g., within the eight-neighborhood), then the spatial adjacency condition is satisfied.

[0040] For two anomalous pixels that meet the spatial adjacency condition, the degree of temperature change anomaly between them within the current working cycle is obtained, and their difference is calculated; simultaneously, the average grayscale value of each pixel within the current working cycle is obtained, and their difference is calculated. Then, the probability that these two anomalous pixels belong to the same temperature anomaly region is calculated. The calculation formula is: ; in, This represents the difference in the degree of temperature aberration between two abnormal pixels within the current period. and These are the average gray values ​​of the two pixels within that period.

[0041] The physical meaning of this formula is that the closer the abnormal temperature changes of two pixels are (the smaller the difference), and the closer their grayscale mean values ​​are (the smaller the difference), the closer the exponent term is to 0 and the closer the exponent function value is to 1, indicating that the two are more likely to belong to the same region; conversely, if either difference is large, the probability decreases rapidly.

[0042] After calculating the probabilities, the minimum-maximum normalization method is used to map them to... Range. Set a preset merging threshold (0.5 in this embodiment), and merge neighboring points with a normalization result greater than 0.5 into the same region with the current point. Then repeat the above operation for newly added pixels, traversing all abnormal pixels in their neighborhood until no new pixels can be added, thus obtaining a complete temperature anomaly region. By traversing all abnormal pixels that have not yet been merged in this way, all temperature anomaly regions within the current working cycle can be determined.

[0043] It should be noted that the preset merging threshold can be adjusted appropriately according to the cleanliness level of the cleanroom, the accuracy of the sensor, and the actual monitoring requirements. For example, it can be reduced to 0.4 in scenarios with higher accuracy requirements and increased to 0.6 in scenarios with less stringent requirements.

[0044] When the performance of an aluminum cleanroom plate deteriorates, the temperature change rate and magnitude in the abnormal area accelerate, increasing the temperature gradient between this area and the surrounding normal plate surface. This results in a sharper gradient feature at the edge of the area in the infrared image. Simultaneously, the deterioration trend is relatively stable, meaning the gradient direction changes less compared to the previous moment. Based on this physical principle, this embodiment calculates the regional anomaly deviation index for each temperature anomaly area.

[0045] For a certain temperature anomaly region The average gradient magnitude of the edge pixels at the current time is obtained and denoted as . Simultaneously, the average gradient magnitude of the edge pixels at the previous time step is obtained, denoted as... Calculate the ratio of the two to obtain the gradient magnitude ratio. Furthermore, obtain the cosine similarity between the gradient direction vector of all pixels in the region at the current time step and the gradient direction vector at the previous time step, denoted as . This value reflects the stability of the gradient direction in the region; a larger value indicates that the direction changes are smaller and more concentrated.

[0046] Calculate the regional anomaly deviation index of this region at the current time. The calculation formula used can be: ; The physical meaning of this formula is as follows: the gradient magnitude ratio reflects the changing trend of gradient intensity at the edge of the region; a ratio greater than 1 indicates that the edge is becoming increasingly sharp, and the anomaly is intensifying. The cosine similarity reflects the temporal stability of the gradient direction; the closer the value is to 1, the more persistent the deterioration trend and the less likely it is to be a random perturbation. The product of these two factors comprehensively characterizes the severity of the anomaly in the temperature anomaly region at the current moment. The larger the value, the more significant the temperature anomaly deviation in the region, and the more likely it is to be related to a real performance defect.

[0047] Using the method described above, the regional anomaly deviation index is calculated for each temperature anomaly region at each data acquisition time.

[0048] S104. For any auxiliary monitoring indicator, calculate its abnormal deviation index and determine the period of abnormal growth and recovery of each auxiliary monitoring indicator.

[0049] In this embodiment, for any auxiliary monitoring indicator, its abnormal deviation index is calculated, and the abnormal growth period and recovery period of each auxiliary monitoring indicator are determined, specifically including: For each auxiliary monitoring indicator, obtain the indicator value at the current monitoring location at the current time, the mean and standard deviation of all indicator values ​​of the auxiliary monitoring indicator at the current monitoring location up to the current time within the current working cycle, and the sum of the differences between the indicator value at the current monitoring location and the indicator values ​​of the same auxiliary monitoring indicator at all other monitoring locations. Add one to the absolute value of the difference between the current indicator value and the mean, multiply by the sum of the differences, and then multiply by the standard deviation to obtain the abnormal deviation index of the auxiliary monitoring indicator at the current monitoring location at the current time. The abnormal deviation index of each auxiliary monitoring indicator at each monitoring location is arranged in time sequence, the arrangement result is smoothed by interpolation, and the valley points and peak points in the smoothed curve are marked. The time period between adjacent valley points and peak points is defined as an abnormal growth period of this auxiliary monitoring indicator; The time period between an adjacent peak and the next trough is defined as a recovery period for this auxiliary monitoring indicator.

[0050] For example, in actual monitoring, when the performance of a real aluminum cleanroom panel becomes abnormal, its multidimensional auxiliary monitoring indicators will typically show varying degrees of response; while environmental disturbances (such as air conditioning start-up and shutdown, personnel movement) often do not cause abnormal linkage of auxiliary monitoring indicators. Based on this principle, this embodiment calculates the index of abnormal deviation of each auxiliary monitoring indicator at each monitoring location.

[0051] Using a certain auxiliary monitoring indicator At a certain monitoring location For example, within the current work cycle, obtain the indicator value at the current moment. At the same time, the following statistics are obtained: : The arithmetic mean of all values ​​of this indicator at this monitoring location at the cutoff time within the current work cycle; The standard deviation of all the above indicator values ​​reflects the historical fluctuation range of the indicator. The sum of the differences between the current monitoring location's index value and the index values ​​of the same auxiliary monitoring index at all other monitoring locations, i.e., the degree of deviation of the monitoring point from the average level of the index for the entire cleanroom.

[0052] Calculate the abnormal deviation index of this auxiliary monitoring indicator at the current monitoring location at the current time. The calculation formula can be: ; The physical meaning of this formula is: It reflects the degree to which the current value deviates from its own historical average; the greater the deviation, the larger the value. This reflects the consistency between the current monitoring point and other monitoring points. The greater the difference between the current point and other points, the more significant the local anomaly at that point. This represents the historical standard deviation of the indicator. A larger standard deviation indicates greater volatility in the indicator itself, and its weight in anomaly detection increases accordingly. The product of these three factors comprehensively measures the degree of abnormal deviation of this auxiliary monitoring indicator at the current moment.

[0053] For each auxiliary monitoring indicator and each monitoring location, the abnormal deviation index of the indicator is calculated time-by-time using the method described above to obtain the time series.

[0054] The abnormal deviation indices of each auxiliary monitoring indicator at each monitoring location are arranged chronologically. Interpolation methods (e.g., cubic spline interpolation) are used to smooth the curves to eliminate random noise. Then, the valleys and peaks in the smoothed curves are marked.

[0055] The period between adjacent troughs and peaks is defined as an abnormal growth period for this auxiliary monitoring indicator. The period between adjacent peaks and the next trough is defined as a recovery period for this auxiliary monitoring indicator.

[0056] It should be noted that the interpolation smoothing method and the valley and peak identification algorithm can be adjusted according to the actual data characteristics, and this embodiment does not limit them.

[0057] S105. Based on the regional abnormal deviation index, determine the period of abnormal temperature growth in the aluminum plate cleanroom. When there is an intersection between the period of abnormal temperature growth of any auxiliary monitoring indicator and the period of abnormal temperature growth, calculate the response index of the auxiliary monitoring indicator to the period of abnormal temperature growth based on the proportion of the intersection duration and the difference between the growth coefficient of the indicator abnormal deviation index of the auxiliary monitoring indicator and the growth coefficient of the regional abnormal deviation index during the period of abnormal temperature growth.

[0058] In this embodiment, the period of abnormal temperature increase in the clean aluminum plate is determined based on the regional abnormal deviation index, specifically including: The regional anomaly deviation index of each temperature anomaly region is arranged in time sequence, the arrangement result is smoothed by interpolation, and the valley points and peak points in the smoothed curve are obtained. The time period between adjacent valley points and peak points is defined as a period of abnormal temperature growth corresponding to the temperature anomaly region.

[0059] Based on the proportion of overlapping durations and the difference between the growth coefficient of the abnormal deviation index of this auxiliary monitoring indicator during periods of abnormal temperature growth and the growth coefficient of the regional abnormal deviation index, the response index of this auxiliary monitoring indicator to periods of abnormal temperature growth is calculated, specifically including: Obtain the abnormal deviation index of the auxiliary monitoring indicator at the beginning and end of the abnormal temperature growth period, and determine the growth coefficient of the auxiliary monitoring indicator by the ratio of the abnormal deviation index at the end to the abnormal deviation index at the beginning. Obtain the regional anomaly deviation index at the start time and the regional anomaly deviation index at the end time of the temperature anomaly growth period. The ratio of the regional anomaly deviation index at the end time to the regional anomaly deviation index at the start time is determined as the regional growth coefficient. The absolute value of the difference between the growth coefficient of the indicator and the growth coefficient of the region is used to obtain the growth coefficient deviation. Obtain the intersection duration between the abnormal growth period of the auxiliary monitoring indicator and the abnormal growth period of temperature, calculate the proportion of the intersection duration to the total duration of the abnormal growth period of temperature, and obtain the intersection duration percentage. Divide the percentage of the intersection duration by the sum of the growth coefficient deviation and the preset small constant, and add one to obtain the initial response index of this auxiliary monitoring indicator to the period of abnormal temperature growth. The initial response index is normalized to obtain the response index of this auxiliary monitoring indicator to periods of abnormal temperature growth.

[0060] For example, the regional anomaly deviation index of each temperature anomaly region is arranged chronologically, and interpolation smoothing is performed on it. The valleys and peaks in the smoothed curve are then marked. The time period between adjacent valleys and peaks is defined as a temperature anomaly growth period corresponding to that temperature anomaly region.

[0061] The specific implementation method of this step is the same as the method for determining the abnormal growth period of auxiliary indicators in S104, and will not be repeated here.

[0062] For a certain temperature anomaly region A certain period of abnormal temperature increase is defined as the start time of that period. The end time is For any auxiliary monitoring indicator If at least one auxiliary indicator shows an abnormal growth period that intersects with the abnormal temperature growth period on the time axis, then the proportion of the time within the intersection to the total number of time periods of the abnormal temperature growth period is recorded as follows: (If there is no intersection, then) Simultaneously, the growth coefficient of this auxiliary monitoring indicator during the period of abnormal temperature increase was calculated. This is the ratio of the abnormal deviation index of the indicator at the end of the period to the abnormal deviation index of the indicator at the beginning of the period.

[0063] Similarly, obtain the regional growth coefficient of this temperature anomaly region during the same time period. It is the ratio of the regional anomaly deviation index at the end time to the regional anomaly deviation index at the beginning time.

[0064] Then, the initial response index of this auxiliary monitoring indicator to the period of abnormal temperature increase can be calculated according to the following formula. : ; Here, 0.01 is a preset small constant used to prevent the denominator from being zero. This embodiment uses 0.01, but this constant can be adjusted according to the actual data volume. The physical meaning of this formula is: the larger the overlap ratio, the more the abnormal growth period of the auxiliary monitoring indicator overlaps with the abnormal temperature growth period; the smaller the difference in growth coefficients, the more synchronous the deterioration trends of the two; therefore, the larger the response index. Finally, the minimum-maximum value normalization method is used to map the initial response index to... The interval is used to obtain the response index of this auxiliary monitoring indicator to the period of abnormal temperature increase. .

[0065] S106. Based on the mean and standard deviation of the response index of all auxiliary monitoring indicators and the mean duration of the recovery period of each auxiliary indicator, determine the likelihood that the temperature anomaly area is dominated by a real performance defect.

[0066] In this embodiment, based on the mean and standard deviation of the response index of all auxiliary monitoring indicators and the mean duration of the recovery period of each auxiliary indicator, the likelihood that the temperature anomaly area is dominated by a real performance defect is determined, specifically including: Obtain the response index of all auxiliary monitoring indicators corresponding to the temperature anomaly area to each period of temperature anomaly growth, and calculate the arithmetic mean and standard deviation of all response indices; Obtain the recovery period of each auxiliary monitoring indicator within each abnormal temperature growth period corresponding to the abnormal temperature area, calculate the duration of each recovery period, and calculate the arithmetic mean of the durations of all recovery periods; Divide the average response index by the standard deviation of the response index, and then multiply by the average recovery period to obtain the initial value of the probability that the temperature anomaly region is dominated by real performance defects. Normalize the initial probability value to obtain the probability that the temperature anomaly region is dominated by real performance defects.

[0067] For example, in order to distinguish between environmental disturbances and actual plate performance defects, this step uses the response consistency of multiple auxiliary monitoring indicators and the abnormal recovery time to calculate the probability that the temperature anomaly area is dominated by actual performance defects.

[0068] For areas with abnormal temperatures Considering all corresponding periods of abnormal temperature increase, the response indices of various auxiliary monitoring indicators have been obtained for each such period. The arithmetic mean of all response indices is now calculated. and standard deviation The larger the average value of the response index, the stronger the overall response; the smaller the standard deviation, the more consistent the responses of the various auxiliary monitoring indicators. This is often related to systematic anomalies caused by real performance defects, while environmental disturbances usually only cause short-term fluctuations in sporadic indicators.

[0069] In addition, for each auxiliary monitoring indicator, its recovery period (i.e., the period from the peak to the next trough) within each of the aforementioned periods of abnormal temperature growth is obtained. The duration of each recovery period is calculated, and the arithmetic mean of the durations of all recovery periods is obtained. The longer the recovery period, the less likely the anomaly is to disappear naturally after it occurs. This is usually related to the continuous degradation caused by real performance defects. Anomalies caused by environmental disturbances tend to recover more quickly after the disturbances disappear.

[0070] The initial probability that this temperature anomaly region is dominated by a true performance defect can be calculated using the following formula. : ; The physical meaning of this formula is: It reflects the concentration intensity of the response index; the larger the ratio, the stronger and more consistent the responses of each index. This reflects the persistence of the anomaly. The larger the product of the two, the higher the likelihood that the temperature anomaly region is dominated by performance defects in the actual aluminum cleanroom panel.

[0071] Finally, the minimum-maximum normalization method is used to map the initial probability values ​​to... The range represents the probability that a true performance defect dominates the temperature anomaly area. The closer the value is to 1, the more likely the area is caused by actual degradation of the panel (such as internal delamination, air leakage, dust accumulation, etc.) rather than temporary environmental disturbances. It should be noted that the maximum and minimum values ​​during normalization can be dynamically determined based on the initial probability values ​​of all temperature anomaly areas in the current clean room, or they can be preset based on historical data. This embodiment does not impose any limitations on this.

[0072] S107. Extract monitoring tags for aluminum cleanroom plates at each stage of their life cycle, generate state vectors using word vector methods, combine the probability of actual performance defects, calculate warning priorities, and output warning information for each temperature anomaly area.

[0073] In this embodiment, monitoring tags for the cleanroom aluminum plate at each stage of its life cycle are extracted. A state vector is generated using word vectorization, and the probability of actual performance defects is considered to calculate the warning priority. Warning information for each temperature anomaly area is then output, specifically including: The monitoring tags recorded during the manufacturing, transportation and installation stages of the aluminum cleanroom plate are extracted. Each monitoring tag is converted into a word vector using the word vector method and then concatenated in a fixed dimension order to generate the state vector of the aluminum cleanroom plate. Obtain the state vectors of the other aluminum clean panels in the cleanroom where the aluminum clean panel is located, and calculate the average cosine similarity between the state vector of the aluminum clean panel and the state vectors of the other aluminum clean panels, as the mean state similarity value. Obtain the deterioration index of the cleanroom plate of this aluminum plate; Multiply the average probability that all temperature anomaly areas of the aluminum cleanroom plate are dominated by actual performance defects by the deterioration index, and then divide by the sum of the average state similarity and the preset small constant to obtain the initial value of the warning priority of the aluminum cleanroom plate. The initial value of the warning priority is normalized to obtain the warning priority, and the corresponding warning information is output according to the preset level of the warning priority.

[0074] The deterioration index of the cleanroom plate was obtained, specifically including: Obtain the total area of ​​all temperature abnormal areas on the clean aluminum plate as the total area of ​​abnormal areas; For each temperature anomaly region, the maximum number of consecutive positive differences in the temporal sequence of the probability that the temperature anomaly region is dominated by a real performance defect is obtained as the deterioration trend value. Obtain the total duration of the temperature anomaly region as the region's duration. Divide the average duration of all temperature anomaly areas by the total current monitoring duration to obtain the time percentage coefficient. The deterioration index of the aluminum cleanroom plate is obtained by multiplying the total area of ​​the abnormal area, the deterioration trend value, and the time proportion coefficient.

[0075] For example, before the aluminum cleanroom panel is put into use, the inspection tags generated at each stage of its life cycle (such as factory qualification tags, transportation process tags, installation acceptance tags, etc.) are recorded. The text of these tags is extracted and input into a Word2Vec model to obtain the word vector corresponding to each tag. Then, all word vectors are concatenated in a fixed-dimensional order to form the state vector of the aluminum cleanroom panel, and this state vector is aligned with the RFID identification code of the aluminum cleanroom panel and stored.

[0076] It should be noted that the vector dimension of the Word2Vec model can be adjusted according to the number of labels and data features. This embodiment uses the commonly used 100-dimensional word vectors, but it is not limited to this.

[0077] To quantify the performance degradation of the aluminum cleanroom panel, this step calculates a degradation index. First, each temperature anomaly region on the aluminum cleanroom panel is matched with temperature anomaly regions at adjacent times (e.g., the previous or next time point). The matching principle is to select the region with the largest overlap area as the temporal continuation of the same region. The duration of continuous existence of each temperature anomaly region can be obtained from the matching results.

[0078] For each temperature anomaly region, obtain its temporal probability sequence dominated by true performance defects, and calculate the first-order difference value of this sequence. Mark consecutive positive difference values ​​as a deterioration subsequence, and count the maximum number of consecutive positive difference values ​​in each temperature anomaly region, denoted as . This represents the longest number of steps in which the temperature in that region continues to deteriorate. Then, take all temperature anomaly regions... The maximum value is denoted as .

[0079] Simultaneously, obtain the total area of ​​all temperature abnormal areas on the cleanroom plate of the aluminum plate, and denot it as... (Unit: pixel area or actual physical area). Calculate the arithmetic mean of the duration of occurrence of each temperature anomaly region, denoted as... Let the total monitoring time up to the current monitoring moment be . .

[0080] The deterioration index of this aluminum cleanroom panel can be calculated using the following formula. : ; The physical meaning of this formula is: The larger the value, the wider the anomaly coverage. The larger the value, the more likely it is to indicate an area of ​​continued deterioration. This reflects the proportion of time the abnormal area spends during the entire monitoring period. The larger the product of these three factors, the more severe the performance degradation of the aluminum cleanroom plate.

[0081] Assuming that, up to the current monitoring time, the arithmetic mean of the true probability of performance defects dominating in all temperature anomaly areas of the aluminum cleanroom plate is: Simultaneously, the arithmetic mean of the cosine similarity between the state vector of this aluminum cleanroom plate and the state vectors of other aluminum cleanroom plates in the same cleanroom is obtained, denoted as . This value reflects the consistency of the life stage of this plate with other normal plates.

[0082] Calculate the initial value of the early warning priority using the following formula: ; Here, 0.01 is a preset small constant used to prevent the denominator from being zero. In this embodiment, 0.01 is used, and this constant can be adjusted according to the magnitude of the similarity values.

[0083] The physical meaning of this formula is: The larger the value, the higher the probability that the plate defect is dominant; The larger the value, the more severe the deterioration. The smaller the value, the greater the difference in condition between this board and other normal boards, and the more attention it warrants. Therefore, the warning priority... This comprehensively reflects the degree of abnormality and urgency of the cleanroom plate.

[0084] Finally, the minimum-maximum normalization method is used to... Mapped to The interval is used to obtain the final warning priority.

[0085] The infrared thermal images, multi-dimensional monitoring data, and RFID identification information captured by the data acquisition unit are transmitted to the data processing unit (including a processor) via a bus. The data processing unit combines the prior tag vectors (i.e., state vectors) of the aluminum cleanroom panels from the factory to the installation stage with the actual performance anomaly trends of each area obtained during monitoring to calculate the warning priority of each aluminum cleanroom panel and output the results according to the following rules: When the warning priority When necessary, advanced early warning information is output, including: location markings of abnormal temperature areas in the aluminum cleanroom plate, a mask image of the abnormal area, and relevant management suggestions, such as "included in the next planned shutdown window for repair or replacement." In this embodiment, the threshold of 0.7 can be adjusted according to actual safety requirements; for example, it can be reduced to 0.6 for high-level cleanrooms.

[0086] When the warning priority When the system is active, it outputs a medium-level warning message, including the location label of the corresponding clean aluminum plate, the warning priority value, and management suggestions, such as "manual inspection is recommended." This threshold is also adjustable; 0.4 can be optimized based on historical data.

[0087] When the warning priority At that time, no warning is issued; only relevant data is recorded and incorporated into the baseline update of the next monitoring cycle to dynamically adjust the normal fluctuation range.

[0088] The tiered early warning information and maintenance recommendations are transmitted to the operation and maintenance management terminal (such as a monitoring screen, mobile terminal, or MES system). After maintenance personnel complete the maintenance, they add the maintenance time, maintenance content, and information on replaced parts as new tags to the prior tag vector of the aluminum cleanroom plate via RFID or database. In this way, subsequent lifecycle management can use the updated state vector for more accurate assessment, thereby achieving closed-loop management of the entire lifecycle from state monitoring and tiered early warning to maintenance feedback.

[0089] In summary, this invention, through the construction of a complete technical chain from infrared temperature field anomaly pixel screening, spatial region merging, multi-dimensional index collaborative response analysis to lifecycle state vector comparison, achieves for the first time a quantitative distinction between the actual performance degradation of aluminum cleanroom panels and instantaneous environmental disturbances. Based on this, the temporal evolution characteristics of the regional anomaly deviation index and the length of the multi-sensor recovery period are used to quantify the probability of defect dominance. Combined with prior tag vectors from the factory to the installation stage, the warning priority is dynamically calculated, and three levels of management instructions (high-level, medium-level, and no warning) are output. After maintenance, maintenance information is appended to the state vector, forming continuous updates and closed-loop management of the entire lifecycle data. This method significantly reduces the false alarm rate and defect miss rate, improves the accuracy and traceability of aluminum cleanroom panel status assessment, and provides reliable technical support for predictive maintenance of cleanroom enclosure structures.

[0090] This invention also proposes an IoT-based aluminum plate cleanroom status monitoring and lifecycle management system. Please refer to [link / reference]. Figure 2 The diagram shows a structural diagram of an IoT-based aluminum plate cleanroom status monitoring and lifecycle management system provided in an embodiment of the present invention. The system includes: a data acquisition module 101, a data processing module 102, and an early warning and classification module 103.

[0091] Data acquisition module 101 is used to acquire time-series data of surface temperature field of aluminum cleanroom plate throughout its entire life cycle and auxiliary monitoring indicators; Data processing module 102 is used to obtain infrared images of clean aluminum plates based on time-series data of plate surface temperature field, and to filter out abnormal pixels from all pixels based on the degree of abnormal temperature change of each pixel in the infrared image over time. Abnormal pixels that meet the spatial adjacency condition and temperature consistency condition are merged into a temperature abnormality region, and the regional abnormality deviation index of the temperature abnormality region is calculated. For any auxiliary monitoring indicator, calculate its abnormal deviation index, and determine the period of abnormal growth and recovery of each auxiliary monitoring indicator. Based on the regional abnormal deviation index, the period of abnormal temperature growth in the aluminum plate cleanroom is determined. When the period of abnormal growth of any auxiliary monitoring indicator overlaps with the period of abnormal temperature growth, the response index of the auxiliary monitoring indicator to the period of abnormal temperature growth is calculated based on the proportion of the overlap duration and the difference between the growth coefficient of the indicator abnormal deviation index of the auxiliary monitoring indicator during the period of abnormal temperature growth and the growth coefficient of the regional abnormal deviation index. Based on the mean and standard deviation of the response index of all auxiliary monitoring indicators and the mean duration of the recovery period of each auxiliary indicator, the possibility that the temperature anomaly area is dominated by a real performance defect is determined. The early warning classification module 103 is used to extract the monitoring tags of the aluminum plate clean plate at each stage of its life cycle, generate state vectors through word vector method, combine the probability dominated by real performance defects, calculate the early warning priority, and output the early warning information for each temperature abnormality area.

[0092] It should be noted that the system provided in the above embodiments is only an example of the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the computer device can be divided into different functional modules to complete all or part of the functions described above. In addition, the IoT-based aluminum plate cleanroom plate status monitoring and life cycle management system and the IoT-based aluminum plate cleanroom plate status monitoring and life cycle management method embodiment provided in the above embodiments belong to the same concept. The specific implementation process is detailed in the method embodiment and will not be repeated here.

[0093] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0094] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0095] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for monitoring the status and managing the lifecycle of cleanroom aluminum plates based on the Internet of Things, characterized in that, include: Acquire time-series data of surface temperature field and auxiliary monitoring indicators of aluminum cleanroom panels throughout their entire life cycle; Infrared images of clean aluminum plates are obtained based on time-series data of plate surface temperature field. Abnormal pixels are filtered out from all pixels based on the degree of temperature change abnormality of each pixel in the infrared image over time. Abnormal pixels that meet the spatial adjacency condition and temperature consistency condition are merged into a temperature abnormality region, and the regional abnormality deviation index of the temperature abnormality region is calculated. For any auxiliary monitoring indicator, calculate its abnormal deviation index, and determine the period of abnormal growth and recovery of each auxiliary monitoring indicator. Based on the regional abnormal deviation index, the period of abnormal temperature growth in the aluminum plate cleanroom is determined. When the period of abnormal growth of any auxiliary monitoring indicator overlaps with the period of abnormal temperature growth, the response index of the auxiliary monitoring indicator to the period of abnormal temperature growth is calculated based on the proportion of the overlap duration and the difference between the growth coefficient of the indicator abnormal deviation index of the auxiliary monitoring indicator during the period of abnormal temperature growth and the growth coefficient of the regional abnormal deviation index. Based on the mean and standard deviation of the response index of all auxiliary monitoring indicators and the mean duration of the recovery period of each auxiliary indicator, the possibility that the temperature anomaly area is dominated by a real performance defect is determined. The monitoring tags of the aluminum cleanroom plate at each stage of its life cycle are extracted, and the state vector is generated by word vector method. Combined with the probability dominated by real performance defects, the warning priority is calculated and the warning information for each temperature abnormality area is output.

2. The method for monitoring the status and lifecycle management of cleanroom aluminum plates based on the Internet of Things according to claim 1, characterized in that, The process of obtaining infrared images of the clean aluminum plate based on the time-series data of the plate surface temperature field, and filtering out abnormal pixels from all pixels based on the degree of temperature change anomaly of each pixel in the infrared image over time, specifically includes: Infrared images of clean aluminum plates are generated based on the time-series data of the temperature field on the plate surface. The infrared images are composed of pixels, and each pixel corresponds to a spatial location and its temperature time-series value. For each pixel in the infrared image, calculate the absolute value of the temperature difference between adjacent time points within the current working cycle and take the average value to obtain the average temperature difference of that pixel. Calculate the absolute value of the temperature difference between adjacent time points of the pixel within the current working cycle and calculate the standard deviation to obtain the standard deviation of the temperature difference of the pixel. Divide the mean temperature difference of the pixel by the average of the mean temperature differences of other pixels, and then divide by the standard deviation of the temperature difference of the pixel to obtain the degree of abnormal temperature change of the pixel. The degree of temperature change abnormality of each pixel is normalized, and pixels whose normalization result is greater than the preset temperature abnormality threshold are identified as abnormal pixels.

3. The method for monitoring the status and lifecycle management of cleanroom aluminum plates based on the Internet of Things according to claim 1, characterized in that, The process of merging anomalous pixels that meet both spatial adjacency and temperature consistency conditions into a temperature anomalous region, and calculating the regional anomalous deviation index of this temperature anomalous region, specifically includes: For any two anomalous pixels, if their spatial positions in the infrared image are adjacent, it is determined that the two anomalous pixels satisfy the spatial adjacency condition. Obtain the degree of temperature change abnormality of the two abnormal pixels in the current working cycle and calculate the difference; obtain the grayscale mean of the two abnormal pixels in the current working cycle and calculate the difference. Multiply the absolute value of the difference in the degree of temperature change abnormality by the absolute value of the difference in the grayscale mean, and take the negative value of the exponential function to obtain the probability that the two abnormal pixels belong to the same temperature abnormality region. When the probability of these two abnormal pixels belonging to the same temperature anomaly region is greater than the preset merging threshold, it is determined that these two abnormal pixels meet the temperature consistency condition. Abnormal pixels that meet the spatial adjacency condition and the temperature consistency condition are merged into a temperature abnormality region. For each temperature anomaly region, the average edge gradient magnitude of the edge pixels in that temperature anomaly region at the current time and the average edge gradient magnitude at the previous time are obtained, and the ratio of the two is calculated to obtain the gradient magnitude ratio. Obtain the cosine similarity between the gradient direction of all pixels in the temperature anomaly region at the current time and the gradient direction at the previous time to obtain the direction consistency. The product of the gradient magnitude ratio and the direction consistency is determined as the regional anomaly deviation index of the temperature anomaly region.

4. The method for monitoring the status and lifecycle management of cleanroom aluminum plates based on the Internet of Things according to claim 1, characterized in that, For any auxiliary monitoring indicator, the abnormal deviation index is calculated, and the abnormal growth period and recovery period of each auxiliary monitoring indicator are determined, specifically including: For each auxiliary monitoring indicator, obtain the indicator value at the current monitoring location at the current time, the mean and standard deviation of all indicator values ​​of the auxiliary monitoring indicator at the current monitoring location up to the current time within the current working cycle, and the sum of the differences between the indicator value at the current monitoring location and the indicator values ​​of the same auxiliary monitoring indicator at all other monitoring locations. Add one to the absolute value of the difference between the current indicator value and the mean, multiply by the sum of the differences, and then multiply by the standard deviation to obtain the abnormal deviation index of the auxiliary monitoring indicator at the current monitoring location at the current time. The abnormal deviation index of each auxiliary monitoring indicator at each monitoring location is arranged in time sequence, the arrangement result is smoothed by interpolation, and the valley points and peak points in the smoothed curve are marked. The time period between adjacent valley points and peak points is defined as an abnormal growth period of this auxiliary monitoring indicator; The time period between an adjacent peak and the next trough is defined as a recovery period for this auxiliary monitoring indicator.

5. The method for monitoring the status and lifecycle management of cleanroom aluminum plates based on the Internet of Things according to claim 1, characterized in that, The determination of the abnormal temperature increase period of the clean aluminum plate based on the regional abnormal deviation index specifically includes: The regional anomaly deviation index of each temperature anomaly region is arranged in time sequence, the arrangement result is smoothed by interpolation, and the valley points and peak points in the smoothed curve are obtained. The time period between adjacent valley points and peak points is defined as a period of abnormal temperature growth corresponding to the temperature anomaly region.

6. The method for monitoring the status and lifecycle management of cleanroom aluminum plates based on the Internet of Things according to claim 1, characterized in that, The method of calculating the response index of the auxiliary monitoring indicator to periods of abnormal temperature growth, based on the proportion of intersection duration and the difference between the growth coefficient of the indicator's abnormal deviation index and the growth coefficient of the regional abnormal deviation index during periods of abnormal temperature growth, specifically includes: Obtain the abnormal deviation index of the auxiliary monitoring indicator at the beginning and end of the abnormal temperature growth period, and determine the growth coefficient of the auxiliary monitoring indicator by the ratio of the abnormal deviation index at the end to the abnormal deviation index at the beginning. Obtain the regional anomaly deviation index at the start time and the regional anomaly deviation index at the end time of the temperature anomaly growth period. The ratio of the regional anomaly deviation index at the end time to the regional anomaly deviation index at the start time is determined as the regional growth coefficient. The absolute value of the difference between the growth coefficient of the indicator and the growth coefficient of the region is used to obtain the growth coefficient deviation. Obtain the intersection duration between the abnormal growth period of the auxiliary monitoring indicator and the abnormal growth period of temperature, calculate the proportion of the intersection duration to the total duration of the abnormal growth period of temperature, and obtain the intersection duration percentage. Divide the percentage of the intersection duration by the sum of the growth coefficient deviation and the preset small constant, and add one to obtain the initial response index of this auxiliary monitoring indicator to the period of abnormal temperature growth. The initial response index is normalized to obtain the response index of this auxiliary monitoring indicator to periods of abnormal temperature growth.

7. The method for monitoring the status and lifecycle management of cleanroom aluminum plates based on the Internet of Things according to claim 1, characterized in that, The method of determining the likelihood that the temperature anomaly region is dominated by a true performance defect based on the mean and standard deviation of the response index of all auxiliary monitoring indicators and the mean duration of the recovery period of each auxiliary indicator includes: Obtain the response index of all auxiliary monitoring indicators corresponding to the temperature anomaly area to each period of temperature anomaly growth, and calculate the arithmetic mean and standard deviation of all response indices; Obtain the recovery period of each auxiliary monitoring indicator within each abnormal temperature growth period corresponding to the abnormal temperature area, calculate the duration of each recovery period, and calculate the arithmetic mean of the durations of all recovery periods; Divide the average response index by the standard deviation of the response index, and then multiply by the average recovery period to obtain the initial value of the probability that the temperature anomaly region is dominated by real performance defects. Normalize the initial probability value to obtain the probability that the temperature anomaly region is dominated by real performance defects.

8. The method for monitoring the status and lifecycle management of cleanroom aluminum plates based on the Internet of Things according to claim 1, characterized in that, The monitoring tags of the extracted aluminum plate cleanroom at each stage of its life cycle are used to generate state vectors using word vector methods. Combined with the probability of actual performance defects, early warning priorities are calculated, and early warning information for each temperature anomaly area is output. Specifically, this includes: The monitoring tags recorded during the manufacturing, transportation and installation stages of the aluminum cleanroom plate are extracted. Each monitoring tag is converted into a word vector using the word vector method and then concatenated in a fixed dimension order to generate the state vector of the aluminum cleanroom plate. Obtain the state vectors of the other aluminum clean panels in the cleanroom where the aluminum clean panel is located, and calculate the average cosine similarity between the state vector of the aluminum clean panel and the state vectors of the other aluminum clean panels, as the mean state similarity value. Obtain the deterioration index of the cleanroom plate of this aluminum plate; Multiply the average probability that all temperature anomaly areas of the aluminum cleanroom plate are dominated by actual performance defects by the deterioration index, and then divide by the sum of the average state similarity and the preset small constant to obtain the initial value of the warning priority of the aluminum cleanroom plate. The initial value of the warning priority is normalized to obtain the warning priority, and the corresponding warning information is output according to the preset level of the warning priority.

9. The method for monitoring the status and lifecycle management of cleanroom aluminum plates based on the Internet of Things according to claim 8, characterized in that, The process of obtaining the deterioration index of the cleanroom aluminum plate specifically includes: Obtain the total area of ​​all temperature abnormal areas on the clean aluminum plate as the total area of ​​abnormal areas; For each temperature anomaly region, the maximum number of consecutive positive differences in the temporal sequence of the probability that the temperature anomaly region is dominated by a real performance defect is obtained as the deterioration trend value. Obtain the total duration of the temperature anomaly region as the region's duration. Divide the average duration of all temperature anomaly areas by the total current monitoring duration to obtain the time percentage coefficient. The deterioration index of the aluminum cleanroom plate is obtained by multiplying the total area of ​​the abnormal area, the deterioration trend value, and the time proportion coefficient.

10. An IoT-based cleanroom aluminum plate condition monitoring and lifecycle management system, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the computer program is executed by the processor, it implements the steps of the IoT-based aluminum plate cleanroom status monitoring and lifecycle management method as described in any one of claims 1-9.

Images