A COB light source detection system and method based on the Internet of Things

Abstract

The application discloses a COB light source detection system and method based on an Internet of Things, relates to the technical field of fault monitoring data analysis, and specifically comprises the following steps: acquiring sensor data and production equipment operation data; constructing an Internet of Things network inside a production line for data transmission; performing data preprocessing, extracting key characteristic variables by analyzing the preprocessed data, and performing correlation analysis; establishing a multivariate control model according to the correlation analysis result, and calculating a joint control threshold according to historical data; collecting real-time data, calculating the deviation degree of the real-time data and the historical data according to the multivariate control model, and performing abnormality judgment by comparing the joint control threshold; performing contribution degree analysis on the abnormal deviation degree according to the correlation analysis result, and performing positioning inspection according to the contribution degree size. The application establishes a detection model and a control standard by means of data driving, and improves the production management level.

Description

Technical Field

[0001] This invention relates to the field of fault monitoring data analysis technology, specifically to a COB light source detection system and method based on the Internet of Things. Background Technology

[0002] The COB (Chip-on-Board) spectral splitting and micro-illumination packaging fully automated production line is an integrated automated production line that combines automatic feeding, automatic spectral splitting, micro-illumination, and packaging (blister packing). It is equipped with a high-precision industrial camera that automatically identifies the position and angle of materials on the tray and transmits this information back to the control system. A robotic arm automatically picks up the COB light source and places it on the transfer track. The high-precision spectral splitting system detects the product data and transmits it back to the control system, then transfers the material to the next station for low-voltage illumination. Traditional COB light source testing methods often rely on the detection of a single indicator, such as luminous flux or current, which cannot comprehensively reflect the actual performance of the COB light source and potential problems in the production process. Furthermore, the relationship between equipment operating status and light source quality during production lacks systematic research, making it difficult to accurately identify and locate key factors affecting light source quality. Summary of the Invention

[0003] The purpose of this invention is to provide a COB light source detection system and method based on the Internet of Things to solve the problems raised in the prior art.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a COB light source detection method based on the Internet of Things, wherein the COB light source detection method specifically includes the following steps:

[0005] Step S100: Install light sensors, current sensors, and temperature sensors at the spectral separation station and micro-lighting station of the COB spectral separation and micro-lighting packaging fully automated production line; acquire sensor data.

[0006] Step S200: Obtain production equipment operation data through the control terminal of the COB spectral micro-illumination packaging fully automated production line;

[0007] Step S300: Construct an Internet of Things (IoT) network within the production line to transmit the sensor data and production equipment operation data to the local data center;

[0008] Step S400: Perform data preprocessing, extract key feature variables by analyzing the preprocessed data, and perform correlation analysis on the key feature variables;

[0009] Step S500: Based on the key feature variables, establish a multivariate control model according to the results of the correlation analysis, and calculate the joint control threshold based on historical data;

[0010] Step S600: Collect real-time data of key feature variables, calculate the deviation between the real-time data and historical data according to the multivariate control model, and compare with the joint control threshold to make anomaly judgment;

[0011] Step S700: Based on the results of the correlation analysis, perform a contribution analysis on the abnormal deviation and perform a location check based on the magnitude of the contribution.

[0012] In step S100, light sensors, current sensors, and temperature sensors are installed at the spectral splitting and micro-illumination stations of the COB spectral splitting and micro-illumination packaging fully automated production line; sensor data is acquired, specifically as follows:

[0013] The optical sensor collects luminous flux data of the light source;

[0014] The current sensor monitors the drive current;

[0015] The temperature sensor acquires the light source temperature and the ambient temperature;

[0016] The light source temperature includes the temperature of the beam-splitting track and the temperature of the micro-illumination track.

[0017] In step S200, the operating data of the production equipment is obtained through the control terminal of the COB spectral micro-illumination packaging fully automated production line, specifically as follows:

[0018] The production equipment operation data includes the operating time, number of actions, and motor speed of each workstation in the production line.

[0019] In step S300, an Internet of Things (IoT) network is constructed within the production line to transmit the sensor data and production equipment operation data to a local data center, specifically as follows:

[0020] The local data center integrates a database and has data storage and retrieval functions.

[0021] In step S400, data preprocessing is performed. By analyzing the preprocessed data, key feature variables are extracted, and correlation analysis is conducted on the key feature variables. Specifically, this includes the following steps:

[0022] Step S401: Obtain historical sensor data and production equipment operation data through data retrieval;

[0023] Step S402: Perform data preprocessing on the acquired historical sensor data and production equipment operation data. The data preprocessing includes data cleaning, outlier removal, and data standardization.

[0024] Preferably, the data standardization method chosen is the Z-score standardization method, specifically:

[0025] z = (x - μ) / σ; where z represents the standardized data; x represents the original data; μ represents the mean of the original data; and σ represents the standard deviation of the original data.

[0026] Step S403: Using the preprocessed historical sensor data and production equipment operation data as references, perform normal simulation. Based on the results of the normal simulation, determine the sensor data and production equipment operation data that conform to the normal distribution as key feature variables.

[0027] Preferably, the specific steps of the normal simulation are as follows:

[0028] Step 1: Using the preprocessed historical sensor data and production equipment operation data as references, calculate the Mahalanobis distance, specifically as follows: ;in, Z represents the Mahalanobis distance; Z represents the vector composed of standardized reference values. This represents a vector composed of the standardized mean values ​​of the reference quantities. represents the matrix transpose; S represents the covariance matrix, whose dimensions are the same as the dimensions of the vector composed of reference values; The inverse matrix of the covariance matrix;

[0029] Step 2, Sort the data in descending order and calculate the quantiles of the chi-square distribution, specifically:

[0030] Where n represents the number of reference quantities, and n is a positive integer; i represents the rank of the reference quantities after descending order, and i is a positive integer, i∈[1,n]; denoted by quantile of the chi-square distribution of reference quantity rank i; p represents the dimension of the vector composed of reference quantities. It is represented by a chi-square distribution;

[0031] Step 3, Drawing Ranked values ​​and quantiles of the chi-square distribution Scatter plot;

[0032] Step 4: When the points in the scatter plot are approximately distributed on a straight line, it is determined that the reference quantity conforms to a multivariate normal distribution, and the reference quantity that conforms to the multivariate normal distribution is determined as the key feature variable.

[0033] Step S404: Perform correlation analysis on the key feature variables to obtain the correlation magnitude between different key feature variables.

[0034] Preferably, the calculation formula for the correlation analysis is as follows:

[0035] ;

[0036] in, Key feature variables and The correlation magnitude is given by , where a represents the number of key feature variables; m represents the number of historical observations, where m is a positive integer; and j represents the number of historical observations, where j is a positive integer, ∈ [1, m]. Key feature variables The j-th historical observation; Key feature variables The j-th historical observation; Key feature variables Historical average observations; Key feature variables The historical average value.

[0037] In step S500, a multivariate control model is established based on the key feature variables and the results of the correlation analysis. A joint control threshold is calculated based on historical data. Specifically:

[0038] Step S501: Obtain historical key feature variable data, and use the data standardization method described in step S402 to standardize the key feature variable data;

[0039] Step S502: Based on the fact that the key feature variables conform to a normal distribution, establish a multivariate control model by combining the standardized key feature variable data with the correlation analysis results;

[0040] Preferably, the characterization formula of the multivariable control model is:

[0041] ;in, Indicates the deviation between real-time data and historical data; A vector representing the standardized key feature variables; The inverse matrix representing the magnitude of the correlation between key feature variables;

[0042] Step S503: Based on historical key feature variable data, set confidence intervals and calculate joint control thresholds.

[0043] Preferably, a 95% confidence interval is set, and the joint control threshold is calculated as follows:

[0044] ;in, Indicates the joint control threshold; This indicates that the F-distribution with degrees of freedom p and mp is at a significance level. The critical value below; based on the set 95% confidence interval, .

[0045] In step S600, real-time data of key feature variables are collected, the deviation between the real-time data and historical data is calculated according to the multivariate control model, and anomaly judgment is made by comparing the data with the joint control threshold. Specifically:

[0046] Step S601: Collect real-time data of key feature variables;

[0047] Step S602: Standardize the real-time data of the key feature variables;

[0048] Step S603: Calculate the deviation between the real-time data and the historical data based on the multivariate control model;

[0049] Step S604: Compare the deviation between the real-time data and the historical data with the joint control threshold, and determine the deviation that is greater than the joint control threshold as abnormal.

[0050] In step S700, based on the correlation analysis results, a contribution analysis is performed on the abnormal deviation, and a location check is performed based on the magnitude of the contribution, specifically:

[0051] Based on the correlation between different key feature variables, the contribution of each key feature variable to the degree of abnormal deviation is calculated.

[0052] Preferably, the calculation of the contribution of different key feature variables to the degree of abnormal deviation is specifically as follows: ;in, This represents the contribution of the a-th key feature variable to the degree of abnormal deviation; This represents the real-time data of the a-th key feature variable after data standardization; Representing vectors The a-th element;

[0053] Sort the contributions in descending order;

[0054] The key feature variables are located and checked based on the descending order.

[0055] A COB light source detection system based on the Internet of Things (IoT) includes a data acquisition module, an IoT network transmission module, a local data center module, a data preprocessing module, a feature extraction and analysis module, a data modeling module, an anomaly detection module, and a contribution analysis and location module.

[0056] The data acquisition module is used to collect sensor data and production equipment operation data;

[0057] The IoT network transmission module is used to build an IoT network inside the production line and is responsible for transmitting the data collected by the data acquisition module to the local data center.

[0058] The local data center module integrates a database for data storage and retrieval;

[0059] The data preprocessing module is used to perform data preprocessing operations on the collected data, including data cleaning, outlier removal, and data standardization.

[0060] The feature extraction and analysis module is used to perform normal simulation with the preprocessed data as a reference, determine key feature variables based on the simulation results, and perform correlation analysis on the key feature variables to obtain the correlation between different key feature variables.

[0061] The data modeling module is used to establish a multivariate control model based on key feature variables and correlation analysis results, and to calculate the joint control threshold based on historical key feature variable data.

[0062] The anomaly detection module is used to collect real-time data of key feature variables, calculate the deviation between real-time data and historical data according to the multivariate control model, and compare it with the joint control threshold to determine whether an anomaly has occurred.

[0063] The contribution analysis and localization module is used to perform contribution analysis on abnormal deviation based on the correlation analysis results, calculate the contribution of different key feature variables to abnormal deviation, sort the contribution in descending order, and then perform localization checks on the key feature variables.

[0064] The output of the data acquisition module is connected to the input of the IoT network transmission module; the output of the IoT network transmission module is connected to the input of the local data center module; the output of the local data center module is connected to the input of the data preprocessing module; the output of the data preprocessing module is connected to the input of the anomaly detection module and the feature extraction and analysis module, respectively; the output of the feature extraction and analysis module is connected to the input of the data modeling module and the contribution analysis and location module, respectively; the output of the data modeling module is connected to the input of the anomaly detection module; and the output of the anomaly detection module is connected to the input of the contribution analysis and location module.

[0065] The data acquisition module includes a sensor data acquisition unit, a runtime data acquisition unit, and a time alignment unit;

[0066] The sensor data acquisition unit is used to acquire sensor data through the optical sensor, current sensor and temperature sensor installed at the spectral separation station and micro-lighting station of the COB spectral micro-lighting packaging fully automated production line.

[0067] The operation data acquisition unit is used to collect operation data of the production equipment; the operation data of the production equipment includes the operation time, number of actions and motor speed of each station in the production line;

[0068] The time alignment unit is used to align the sensor data with the acquisition time of the production equipment operation data;

[0069] The output of the sensor data acquisition unit is connected to the input of the time alignment unit; the output of the running data acquisition unit is connected to the input of the time alignment unit.

[0070] Compared with existing technologies, the beneficial effects of this invention are as follows: By installing multiple sensors at key workstations on the production line and acquiring production equipment operation data, this invention achieves multi-dimensional real-time monitoring, breaking through the limitations of traditional single-indicator detection and more comprehensively and accurately reflecting the performance and production status of COB light sources; with the help of data preprocessing, feature extraction, multivariate control models, and joint control thresholds, anomalies can be detected in a timely and accurate manner, and key variables can be quickly located through contribution analysis, significantly improving the accuracy of anomaly detection and the efficiency of fault diagnosis, reducing quality risks and downtime; by utilizing the Internet of Things to build a network to transmit data to a local data center for in-depth analysis, data-driven detection models and control standards are established, providing a scientific basis for light source quality control and improving production management. Attached Figure Description

[0071] Figure 1 This is a schematic diagram illustrating the steps of a COB light source detection method based on the Internet of Things according to the present invention.

[0072] Figure 2 This is a schematic diagram of the structure of a COB light source detection system based on the Internet of Things according to the present invention;

[0073] Figure 3 This is a schematic diagram of a normalized simulated scatter plot in an embodiment of the IoT-based COB light source detection method of the present invention. Detailed Implementation

[0074] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0075] Example: Figures 1-3 As shown, this invention provides a technical solution, a COB light source detection method based on the Internet of Things, such as... Figure 1 As shown, the specific steps include the following:

[0076] Step S100: Install light sensors, current sensors, and temperature sensors at the spectral separation station and micro-lighting station of the COB spectral separation and micro-lighting packaging fully automated production line; acquire sensor data.

[0077] Step S200: Obtain production equipment operation data through the control terminal of the COB spectral micro-illumination packaging fully automated production line;

[0078] Step S300: Build an Internet of Things (IoT) network within the production line to transmit sensor data and production equipment operation data to the local data center;

[0079] Step S400: Perform data preprocessing, extract key feature variables by analyzing the preprocessed data, and perform correlation analysis on the key feature variables;

[0080] Step S500: Based on the key feature variables, establish a multivariate control model according to the results of correlation analysis, and calculate the joint control threshold based on historical data;

[0081] Step S600: Collect real-time data of key feature variables, calculate the deviation between real-time data and historical data based on the multivariate control model, and compare with the joint control threshold to make anomaly judgment;

[0082] Step S700: Based on the results of the correlation analysis, perform a contribution analysis on the abnormal deviation and conduct a location check based on the magnitude of the contribution.

[0083] In step S100, light sensors, current sensors, and temperature sensors are installed at the spectral splitting and micro-illumination stations of the COB spectral splitting and micro-illumination packaging fully automated production line; sensor data is acquired, specifically as follows:

[0084] Sensor Deployment: Install the following sensors at the beam splitting station and the micro-lighting station:

[0085] Optical sensor: Collects luminous flux (unit: lm, range 0-3000lm, accuracy ±2%).

[0086] Current sensor: monitors drive current (unit: A, range 0-1A, accuracy ±0.5%).

[0087] Temperature sensor: measures the temperature of the light source (temperature of the beam splitter track and the micro-illumination track, unit: C, accuracy ±0.5°C);

[0088] In step S200, the operating data of the production equipment is obtained through the control terminal of the COB spectral micro-illumination packaging fully automated production line, specifically as follows:

[0089] The PLC collects the running time, number of actions, and motor speed (unit: rpm) of each workstation.

[0090] Some examples are as follows: Among the spectroscopic orbital temperature and the micro-illumination orbital temperature, those with the largest differences from the average temperature are recorded;

[0091]

[0092] In step S300, an Internet of Things (IoT) network is constructed within the production line to transmit sensor data and production equipment operation data to the local data center. Specifically:

[0093] The local data center integrates a database, providing data storage and retrieval capabilities.

[0094] In step S400, data preprocessing is performed. By analyzing the preprocessed data, key feature variables are extracted, and correlation analysis is conducted on the key feature variables. Specifically, this includes the following steps:

[0095] Step S401: Obtain historical sensor data and production equipment operation data through data retrieval;

[0096] Step S402: Perform data preprocessing on the acquired historical sensor data and production equipment operation data, including data cleaning, outlier removal, and data standardization;

[0097] Step S403: Using the preprocessed historical sensor data and production equipment operation data as references, perform normal simulation. Based on the results of the normal simulation, determine the sensor data and production equipment operation data that conform to the normal distribution as key feature variables.

[0098] Step 1: Using the preprocessed historical sensor data and production equipment operation data as references, calculate the Mahalanobis distance, specifically as follows: ;in, Z represents the Mahalanobis distance; Z represents the vector composed of standardized reference values. This represents a vector composed of the standardized mean values ​​of the reference quantities. represents the matrix transpose; S represents the covariance matrix, whose dimensions are the same as the dimensions of the vector composed of reference values; The inverse matrix of the covariance matrix;

[0099] Step 2, Sort the data in descending order and calculate the quantiles of the chi-square distribution, specifically:

[0100] Where n represents the number of reference quantities, and n is a positive integer; i represents the rank of the reference quantities after descending order, and i is a positive integer, i∈[1,n]; denoted by quantile of the chi-square distribution of reference quantity rank i; p represents the dimension of the vector composed of reference quantities. It is represented by a chi-square distribution;

[0101] Step 3, Drawing Ranked values ​​and quantiles of the chi-square distribution Scatter plot;

[0102] Step 4: When the points in the scatter plot are approximately distributed on a straight line, it is determined that the reference quantity conforms to a multivariate normal distribution, and the reference quantity that conforms to the multivariate normal distribution is determined as the key feature variable.

[0103] Three reference quantities (luminous flux, driving current, and light source temperature) of the simulated COB light source are collected under stable operating conditions. The average value vector is obtained as follows: μ=[1500,0.6,35].

[0104] Luminous flux variance: 50 2 ;

[0105] Drive current variance: 0.05 2 ;

[0106] Temperature variance: 2 2 ;

[0107] Correlation coefficient between luminous flux and current: 0.8;

[0108] Correlation coefficient between luminous flux and temperature: −0.3;

[0109] Current-temperature correlation coefficient: 0.2;

[0110] The covariance matrix S is obtained as follows:

[0111] ;

[0112] Where the number of reference values ​​n=1000, the dimension of the vector composed of the reference values ​​p=3, and the plotting... Ranked values ​​and quantiles of the chi-square distribution Scatter plot, such as Figure 3 As shown;

[0113] Therefore, it can be concluded that luminous flux, driving current, and light source temperature approximately conform to a multivariate normal distribution, and luminous flux, driving current, and light source temperature are identified as key characteristic variables.

[0114] Step S404: Perform correlation analysis on the key feature variables to obtain the correlation magnitude between different key feature variables.

[0115] The formula for calculating correlation analysis is:

[0116] ;

[0117] in, Key feature variables and The correlation magnitude is given by , where a represents the number of key feature variables; m represents the number of historical observations, where m is a positive integer; and j represents the number of historical observations, where j is a positive integer, ∈ [1, m]. Key feature variables The j-th historical observation; Key feature variables The j-th historical observation; Key feature variables Historical average observations; Key feature variables Historical average observations;

[0118] Calculations show that:

[0119] The correlation between luminous flux and current is 0.82;

[0120] The correlation between luminous flux and temperature is -0.28;

[0121] The correlation between current and temperature is 0.19;

[0122] In step S500, a multivariate control model is established based on the key feature variables and the results of correlation analysis. The joint control threshold is calculated based on historical data, specifically as follows:

[0123] Step S501: Obtain historical key feature variable data, and use the data standardization method in step S402 to standardize the key feature variable data;

[0124] Step S502: Based on the fact that the key feature variables conform to a normal distribution, establish a multivariate control model by combining the standardized key feature variable data with the results of correlation analysis;

[0125] The representation formula for a multivariable control model is:

[0126] ;in, Indicates the deviation between real-time data and historical data; A vector representing the standardized key feature variables; The inverse matrix representing the magnitude of the correlation between key feature variables;

[0127] Standardized data matrix: 1000×3 dimensions (light: flux, current, temperature):

[0128] Perform batch calculations using Python's computation libraries:

[0129] ;

[0130] Step S503: Based on historical key feature variable data, set confidence intervals and calculate joint control thresholds:

[0131] Joint control threshold: With a 95% confidence level, degrees of freedom p=3, m=1000, the calculation yields:

[0132] t lim 2 =(1000−1)×3 / (1000−3)×F 0.05 (3,997)≈8.03;

[0133] In step S600, real-time data of key feature variables are collected, the deviation between real-time data and historical data is calculated based on the multivariate control model, and anomaly judgment is made by comparing the data with the joint control threshold. Specifically:

[0134] Step S601: Collect real-time data of key feature variables;

[0135] Step S602: Standardize the real-time data of key feature variables;

[0136] Step S603: Calculate the deviation between real-time data and historical data based on the multivariate control model;

[0137] Step S604: Compare the deviation between real-time data and historical data with the joint control threshold, and determine the deviation that is greater than the joint control threshold as abnormal.

[0138] In step S700, based on the results of the correlation analysis, a contribution analysis is performed on the abnormal deviation, and a location check is conducted based on the magnitude of the contribution. Specifically:

[0139] Based on the correlation between different key feature variables, the contribution of each key feature variable to the degree of abnormal deviation is calculated.

[0140] Get real-time data:

[0141] Luminous flux (lm): 1420.5;

[0142] Drive current (A): 0.55;

[0143] Temperature (°C): 37.8;

[0144] Perform data standardization to eliminate dimensions; the Z-score standardization method should be selected.

[0145] Real-time data after standardization: W=[−1.59,−1.0,1.4];

[0146] Deviation calculation:

[0147] ;

[0148] have The result is considered normal.

[0149] like Figure 2 As shown, an IoT-based COB light source detection system includes a data acquisition module, an IoT network transmission module, a local data center module, a data preprocessing module, a feature extraction and analysis module, a data modeling module, an anomaly detection module, and a contribution analysis and location module.

[0150] The data acquisition module is used to collect sensor data and production equipment operation data;

[0151] The IoT network transmission module is used to build an IoT network within the production line and is responsible for transmitting the data collected by the data acquisition module to the local data center.

[0152] The local data center module integrates a database for data storage and retrieval;

[0153] The data preprocessing module is used to perform data preprocessing operations on the collected data. Data preprocessing includes data cleaning, outlier removal, and data standardization.

[0154] The feature extraction and analysis module is used to perform normal simulation with the preprocessed data as a reference, determine key feature variables based on the simulation results, and perform correlation analysis on the key feature variables to obtain the correlation between different key feature variables.

[0155] The data modeling module is used to build a multivariate control model based on the results of key feature variables and correlation analysis, and to calculate the joint control threshold based on historical key feature variable data;

[0156] The anomaly detection module is used to collect real-time data of key feature variables, calculate the deviation between real-time data and historical data based on the multivariate control model, and compare it with the joint control threshold to determine whether an anomaly has occurred.

[0157] The contribution analysis and localization module is used to perform contribution analysis on abnormal deviation based on the results of correlation analysis, calculate the contribution of different key feature variables to abnormal deviation, sort the contribution in descending order, and then perform localization checks on the key feature variables.

[0158] The output of the data acquisition module is connected to the input of the IoT network transmission module; the output of the IoT network transmission module is connected to the input of the local data center module; the output of the local data center module is connected to the input of the data preprocessing module; the output of the data preprocessing module is connected to the inputs of the anomaly detection module and the feature extraction and analysis module; the output of the feature extraction and analysis module is connected to the inputs of the data modeling module and the contribution analysis and location module; the output of the data modeling module is connected to the input of the anomaly detection module; and the output of the anomaly detection module is connected to the input of the contribution analysis and location module.

[0159] The data acquisition module includes a sensor data acquisition unit, a runtime data acquisition unit, and a time alignment unit;

[0160] The sensor data acquisition unit is used to acquire sensor data through optical sensors, current sensors, and temperature sensors installed at the spectral separation station and micro-lighting station of the COB spectral micro-lighting packaging fully automated production line.

[0161] The operation data acquisition unit is used to collect operation data of production equipment; the operation data of production equipment includes the running time, number of actions and motor speed of each station in the production line;

[0162] The time alignment unit is used to align the sensor data with the acquisition time of the production equipment operation data;

[0163] The output of the sensor data acquisition unit is connected to the input of the time alignment unit; the output of the running data acquisition unit is connected to the input of the time alignment unit.

[0164] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A COB light source detection method based on the Internet of Things, characterized in that: The COB light source detection method specifically includes the following steps: Step S100: Install light sensors, current sensors, and temperature sensors at the spectral separation station and micro-lighting station of the COB spectral separation and micro-lighting packaging fully automated production line; acquire sensor data. Step S200: Obtain production equipment operation data through the control terminal of the COB spectral micro-illumination packaging fully automated production line; Step S300: Construct an Internet of Things (IoT) network within the production line to transmit the sensor data and production equipment operation data to the local data center; Step S400: Perform data preprocessing, analyze the preprocessed data, extract key feature variables, and perform correlation analysis on the key feature variables; including: Using preprocessed historical sensor data and production equipment operation data as references, a normal simulation is performed to plot the Mahalanobis distance. The descending sorted values ​​and the quantiles of the chi-square distribution The scatter plot; when the points in the scatter plot are distributed on a straight line, it is determined that the reference quantity conforms to a multivariate normal distribution, and the reference quantity that conforms to the multivariate normal distribution is determined as the key feature variable; Correlation analysis was performed on the key feature variables to obtain the magnitude of the correlation between different key feature variables; Step S500: Using the key feature variables as a basis, establish a multivariate control model based on the results of the correlation analysis, and calculate the joint control threshold based on historical data; specifically: Step S501: Obtain historical key feature variable data and standardize the key feature variable data; Step S502: Based on the fact that the key feature variables conform to a normal distribution, establish a multivariate control model by combining the standardized key feature variable data with the results of the correlation analysis. The characterization formula for the multivariable control model is as follows: ;in, Indicates the deviation between real-time data and historical data; A vector representing the standardized key feature variables; The inverse matrix representing the magnitude of the correlation between key feature variables; Step S503: Based on historical key feature variable data, set confidence intervals and calculate joint control thresholds; set 95% confidence intervals and calculate joint control thresholds. Step S600: Collect real-time data of key feature variables, calculate the deviation between the real-time data and historical data according to the multivariate control model, and compare with the joint control threshold to make anomaly judgment; Step S700: Based on the results of the correlation analysis, perform a contribution analysis on the abnormal deviation, and conduct a location check based on the magnitude of the contribution; specifically: The contribution of different key feature variables to the degree of outlier deviation is calculated as follows: ;in, This represents the contribution of the a-th key feature variable to the degree of abnormal deviation; This represents the real-time data of the a-th key feature variable after data standardization; Representing vectors The a-th element; Sort the contributions in descending order; The key feature variables are located and checked based on the descending order.

2. The method for detecting COB light sources based on the Internet of Things according to claim 1, characterized in that: In step S100, light sensors, current sensors, and temperature sensors are installed at the spectral splitting and micro-illumination stations of the COB spectral splitting and micro-illumination packaging fully automated production line; sensor data is acquired, specifically as follows: The optical sensor collects luminous flux data of the light source; The current sensor monitors the drive current; The temperature sensor acquires the light source temperature and the ambient temperature; The light source temperature includes the temperature of the beam-splitting track and the temperature of the micro-illumination track.

3. The method for detecting COB light sources based on the Internet of Things according to claim 2, characterized in that: In step S200, the operating data of the production equipment is obtained through the control terminal of the COB spectral micro-illumination packaging fully automated production line, specifically as follows: The production equipment operation data includes the operating time, number of actions, and motor speed of each workstation in the production line.

4. The method for detecting COB light sources based on the Internet of Things according to claim 1, characterized in that: In step S300, an Internet of Things (IoT) network is constructed within the production line to transmit the sensor data and production equipment operation data to a local data center, specifically as follows: The local data center integrates a database and has data storage and retrieval functions.

5. The method for detecting COB light sources based on the Internet of Things according to claim 4, characterized in that: In step S600, real-time data of key feature variables are collected, the deviation between the real-time data and historical data is calculated according to the multivariate control model, and anomaly judgment is made by comparing the deviation with the joint control threshold. Specifically: Step S601: Collect real-time data of key feature variables; Step S602: Standardize the real-time data of the key feature variables; Step S603: Calculate the deviation between the real-time data and the historical data based on the multivariate control model; Step S604: Compare the deviation between the real-time data and the historical data with the joint control threshold, and determine the deviation that is greater than the joint control threshold as abnormal.

6. An IoT-based COB light source detection system, employing the IoT-based COB light source detection method as described in any one of claims 1-5, characterized in that: The COB light source detection system includes a data acquisition module, an Internet of Things network transmission module, a local data center module, a data preprocessing module, a feature extraction and analysis module, a data modeling module, an anomaly detection module, and a contribution analysis and location module. The data acquisition module is used to collect sensor data and production equipment operation data; The IoT network transmission module is used to build an IoT network inside the production line and is responsible for transmitting the data collected by the data acquisition module to the local data center. The local data center module integrates a database for data storage and retrieval; The data preprocessing module is used to perform data preprocessing operations on the collected data, including data cleaning, outlier removal, and data standardization. The feature extraction and analysis module is used to perform normal simulation with the preprocessed data as a reference, determine key feature variables based on the simulation results, and perform correlation analysis on the key feature variables to obtain the correlation between different key feature variables. The data modeling module is used to establish a multivariate control model based on the results of key feature variables and correlation analysis, and to calculate the joint control threshold based on historical key feature variable data. The anomaly detection module is used to collect real-time data of key feature variables, calculate the deviation between real-time data and historical data according to the multivariate control model, and compare it with the joint control threshold to determine whether an anomaly has occurred. The contribution analysis and localization module is used to perform contribution analysis on abnormal deviation based on the results of correlation analysis, calculate the contribution of different key feature variables to abnormal deviation, sort the contribution in descending order, and then perform localization checks on the key feature variables.

7. The IoT-based COB light source detection system according to claim 6, characterized in that: The data acquisition module includes a sensor data acquisition unit, a runtime data acquisition unit, and a time alignment unit; The sensor data acquisition unit is used to acquire sensor data through the optical sensor, current sensor and temperature sensor installed at the spectral separation station and micro-lighting station of the COB spectral micro-lighting packaging fully automated production line. The operation data acquisition unit is used to collect operation data of the production equipment; the operation data of the production equipment includes the operation time, number of actions and motor speed of each station in the production line; The time alignment unit is used to align the sensor data with the acquisition time of the production equipment operation data; The output of the sensor data acquisition unit is connected to the input of the time alignment unit; the output of the running data acquisition unit is connected to the input of the time alignment unit.

Images