Laboratory safety internet-of-things intelligent early warning system

Abstract

The invention discloses a laboratory safety IOT (Internet of Things) intelligent early warning system, which comprises a laboratory environment early warning system, a laboratory equipment early warning system, a laboratory personnel behavior early warning system, a self-adaptive learning system and an integrated control system, wherein the laboratory environment early warning system, the laboratory equipment early warning system, the laboratory personnel behavior early warning system and the adaptive learning system are connected with the integrated control system through a wireless communication network; the adaptive learning system is connected with the laboratory environment early warning system, the laboratory equipment early warning system and the laboratory personnel behavior early warning system through a wireless communication network; the self-adaptive learning system is used for optimizing early warning models of the subsystems, and the integrated control system is used for efficient integration of the subsystems to ensure data sharing and cooperative work among the subsystems. According to the invention, the safety and management efficiency of the laboratory are effectively improved, and the potential risk is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of early warning systems, and in particular to an intelligent early warning system for laboratory safety Internet of Things. Background Art

[0002] With the continuous development of laboratory scientific research, laboratory safety management has become an increasingly important issue. There are a series of potential risks in the chemicals and equipment operations involved in the laboratory, so timely monitoring and early warning of laboratory safety are particularly important. At present, there are some laboratory safety monitoring systems, however, these systems still have some shortcomings in terms of intelligence and Internet of Things integration.

[0003] Environmental monitoring system: Existing laboratory safety monitoring systems mainly focus on monitoring environmental parameters such as temperature, humidity, gas concentration, etc. However, these systems usually fail to cover comprehensive monitoring of experimental equipment and personnel behavior.

[0004] Equipment monitoring system: Some laboratories use equipment monitoring systems to detect the working status of equipment. However, most of these systems operate independently and lack effective integration with environmental monitoring systems, making it difficult to share information and form a comprehensive laboratory safety monitoring plan.

[0005] Personnel behavior monitoring: There are also some personnel behavior monitoring systems in the existing technology, but they often rely on equipment such as cameras and lack intelligent behavior analysis. These systems can usually only provide limited information on laboratory personnel behavior and have limited ability to identify and warn of abnormal operations.

[0006] In summary, existing technologies in laboratory safety monitoring have problems such as insufficient intelligence and poor system integration, and fail to meet the needs of comprehensive, efficient and intelligent safety management in complex laboratory environments. Summary of the invention

[0007] The purpose of the present invention is to provide a laboratory safety IoT intelligent early warning system that integrates various sensors, video monitoring, adaptive learning and integrated control to achieve intelligent data analysis and response, and improve the level and efficiency of laboratory safety management.

[0008] In order to solve the above technical problems, the present invention adopts the following technical solutions:

[0009] A laboratory safety IoT intelligent early warning method, comprising:

[0010] Laboratory environment early warning system, laboratory equipment early warning system, laboratory personnel behavior early warning system, adaptive learning system and integrated control system.

[0011] Among them, the laboratory environment warning system, laboratory equipment warning system, laboratory personnel behavior warning system and adaptive learning system are all connected to the integrated control system through the wireless communication network; the adaptive learning system is connected to the laboratory environment warning system, laboratory equipment warning system and laboratory personnel behavior warning system through the wireless communication network.

[0012] The laboratory environment early warning system is used to extract features from the real-time environmental parameters collected by sensors, obtain fusion features, analyze the fusion features using advanced real-time pattern recognition algorithms, obtain early warning signals, and adjust the environmental parameters using a control feedback mechanism; and transmit the real-time collected environmental parameters to the integrated control system.

[0013] The laboratory equipment early warning system is used to pre-process the laboratory equipment operation data and build a time domain multi-scale feature sharing fusion-domain adaptive learning model to extract features with different time scales in the data and integrate them. The Lasso regression model is used to predict the integrated features, identify potential faults, and perform corresponding processing.

[0014] The laboratory personnel behavior warning system is used to obtain personnel activity data and analyze the data using a convolutional neural network visual adaptive learning model. When abnormal behavior is identified, the warning mechanism is triggered and the laboratory administrator is notified; and the personnel activity data is transmitted to the integrated control system.

[0015] The adaptive learning system is used to optimize the laboratory environment warning system, laboratory equipment warning system and laboratory personnel behavior warning system using machine learning technology, and feed back the optimized warning system and its obtained characteristics to the integrated control system.

[0016] The integrated control system is used to integrate the data of the laboratory environment early warning system, laboratory equipment early warning system and laboratory personnel behavior early warning system using data analysis and processing technology, and feed back the processing results to the laboratory environment early warning system, laboratory equipment early warning system, laboratory personnel behavior early warning system and adaptive learning system to achieve real-time monitoring and early warning functions.

[0017] Furthermore, the laboratory environment early warning system includes advanced sensor technology components, environmental parameter transmission modules, real-time pattern recognition modules, safety threshold setting modules and real-time monitoring and feedback modules.

[0018] The advanced sensing technology component includes a sensor network that is used to collect laboratory environmental parameters in real time, including temperature, humidity, air quality, noise level, entry and exit channels, and the opening and closing status of ventilation windows.

[0019] The environmental parameter transmission module is used to receive the parameters collected by the advanced sensing technology components through the wireless communication network, pre-process the parameters and perform feature extraction to obtain spatial features and temporal features, fuse the features, and transmit the fused features to the real-time pattern recognition module through the wireless communication network.

[0020] The real-time pattern recognition module is used to analyze the fusion features using the real-time pattern recognition algorithm, obtain the changing trend of laboratory environmental parameters, generate early warning signals, and pass them to the safety threshold setting module and the real-time monitoring and feedback module for corresponding regulation and processing; the early warning signals can also be notified to the laboratory administrator through real-time alarm notifications, text messages or emails.

[0021] The safety threshold setting module is used to provide a user interface, allowing laboratory administrators to set safety thresholds according to experimental needs and special environmental requirements, and pass the set safety threshold information to the real-time pattern recognition module and the real-time monitoring and feedback module.

[0022] The real-time monitoring and feedback module is used to start the real-time control and feedback mechanism after receiving the early warning signal, and control the relevant equipment to adjust the laboratory environmental parameters.

[0023] Furthermore, the environment parameter transmission module is configured to perform the following actions:

[0024] Step 1: Preprocessing includes cleaning the parameters, removing duplicate data, filling missing values ​​and eliminating outliers; using signal processing technology to filter noise on the cleaned parameters; and standardizing the filtered parameters.

[0025] Step 2: Process the parameters preprocessed in step 1, extract the relationship between the operating status of the equipment in the laboratory and the environmental parameters, and obtain the corresponding spatial characteristics, which include the location of the equipment in the laboratory and the distribution of sensors; use time series analysis technology to analyze the time series data of the parameters preprocessed in step 1 to obtain time characteristics, which include periodic changes, trend changes and mutation points of the data.

[0026] Step 3: Standardize the spatial features and temporal features, use feature fusion technology to assign different weights to the processed features, and fuse them to obtain fused features.

[0027] Furthermore, the laboratory environment early warning system also includes a historical data storage module for recording the time, location and specific parameter values ​​of abnormal environmental parameter events.

[0028] Furthermore, the laboratory equipment early warning system includes an equipment condition monitoring module, a data transmission module, an equipment monitoring adaptive module and an abnormal event reporting module.

[0029] The equipment condition monitoring module is used for the sensor network to monitor the operating data of all equipment in the laboratory in real time, including power consumption, temperature, vibration and current.

[0030] The data transmission module is used to transmit the operation data of the equipment to the integrated control system and the equipment monitoring adaptive module, and update the data in the integrated control system.

[0031] The equipment monitoring adaptive module is used to preprocess the equipment operation data, including denoising, standardization and grayscale processing; construct a time domain multi-scale feature sharing fusion-domain adaptive learning model, input the preprocessed equipment status data into the model, extract features with different time scales, and integrate them, and use the Lasso regression model to predict the integrated features to obtain equipment fault information.

[0032] The abnormal event reporting module is used to collect data related to abnormal events based on equipment failure information, including the time of abnormal occurrence, equipment location and specific abnormal parameter values, automatically generate detailed abnormal event reports, and send alerts to laboratory administrators; the generated abnormal event reports are archived in the system.

[0033] Furthermore, the device monitoring adaptation module is configured to perform the following actions:

[0034] The standardized expression is:

[0035]

[0036] Among them, X represents the original data, μ represents the mean, σ represents the standard deviation, and X std Represents the normalized data.

[0037] The time-domain multi-scale feature sharing fusion-domain adaptive learning model includes an input layer, a feature sharing layer and a fusion layer; the domain adaptive algorithm is used to train the time-domain multi-scale feature sharing fusion-domain adaptive learning model; the pre-processed equipment status data is input into the trained model, and the time-domain multi-scale feature extraction method is used to extract features with different time scales, which include vibration signals, temperature changes and pressure fluctuations of the equipment. The features are integrated using feature weighted average to obtain an integrated feature vector, and the specific expression is:

[0038]

[0039] Among them, F fusion represents the fused feature vector, w i represents the feature weight of the j-th time scale feature, F w represents the original features of the w-th device state data, and n represents the total number of time scale features.

[0040] The Lasso regression model is used to perform fault feature recognition and prediction analysis on the integrated feature vector to obtain the predicted equipment failure information, and intelligent diagnosis of equipment failures under changing working conditions is achieved through optimization.

[0041] The expression of the Lasso regression model is:

[0042] Y=X·β

[0043] Among them, Y represents the target variable and β represents the regression coefficient.

[0044] The expression of the loss function with L1 regularization term is:

[0045]

[0046] Among them, N represents the number of samples, λ represents the regularization parameter, and L(β) represents the loss function.

[0047] Minimize the loss function to get the optimal regression coefficient. The specific expression is:

[0048]

[0049] in, represents the estimated regression coefficient.

[0050] After Lasso regression training, the non-zero coefficients are the weights of the selected features.

[0051] The prediction results are verified and evaluated by calculating the mean square error between the fusion results and the actual observation values. The specific formula is as follows:

[0052]

[0053] Among them, MSE represents the mean square error, y i represents the actual observed value of the i-th sample, Represents the corresponding predicted value of the i-th sample.

[0054] Furthermore, the laboratory equipment early warning system also includes an equipment health history record module for recording equipment health history data, including operating time, maintenance history, and fault records.

[0055] Furthermore, the laboratory personnel behavior early warning system includes a personnel positioning and identification module, a behavior monitoring adaptive module and a real-time alarm notification module.

[0056] The personnel positioning and identification module is used to use UWB personnel positioning technology to achieve high-precision positioning of the position of personnel in the laboratory by sending and receiving short pulse signals, and use identity recognition technology to confirm the identity of personnel in the laboratory. When unauthorized personnel enter sensitive areas, security measures are automatically triggered; in addition, the information of personnel in the laboratory is recorded, including identity recognition data and entry and exit time, to obtain personnel activity data, which includes personnel location information, movement trajectory and behavior pattern.

[0057] The behavior monitoring adaptive module is used to process the convolutional neural network visual adaptive learning model based on the personnel behavior image to obtain the processed model, extract the features of the pre-processed personnel activity data to obtain spatial features and temporal features; compare the extracted features with normal behavior and abnormal behavior, identify any abnormal behavior occurring in the laboratory, and issue an early warning; the specific contents are:

[0058] The convolutional neural network visual adaptive learning model includes multiple convolutional layers, pooling layers, and fully connected layers;

[0059] Obtaining images of personnel behavior through identity recognition technology, historical records or warning records;

[0060] The personnel behavior images are divided into training set, validation set and test set. The training set is input into the convolutional neural network visual adaptive learning model. The convolution operation is performed through the convolution kernel in the convolution layer to extract the features in the training set. The specific expression is:

[0061]

[0062] Among them, I represents the training set matrix, K represents the convolution kernel matrix, r represents the row position of the output feature, c represents the column position of the output feature, m represents the row index of the convolution kernel, and b represents the column index of the convolution kernel;

[0063] The fully connected layer is used to classify and learn the extracted features, and then identify normal and abnormal behaviors. The specific expression is:

[0064] z=Wx+g

[0065] Among them, W represents the weight matrix, x represents the input feature, g represents the bias term, and z represents the output;

[0066] Using the gradient descent method and Adam optimization algorithm, the model parameters are continuously adjusted to minimize the loss function;

[0067] The calculation formula of the gradient descent method is as follows:

[0068]

[0069] Among them, P represents the updated model parameters, θ represents the current model parameters, and η represents the learning rate. represents the gradient of the loss function;

[0070] The calculation formula of Adam optimization algorithm is as follows:

[0071] First-order momentum estimate of the gradient:

[0072] Second-order momentum estimate of the gradient:

[0073] Bias correction for the first-order momentum estimate:

[0074] Bias correction for the second-order momentum estimate:

[0075] Parameter update formula:

[0076] Among them, m t represents the first-order momentum estimate of the gradient at the tth iteration, β1 represents the momentum decay coefficient, and m t-1 represents the first-order momentum estimate of the gradient at the t-1th iteration, v t represents the second-order momentum estimate of the gradient at the tth iteration, β2 represents the momentum decay coefficient, and v t-1 represents the second-order momentum estimate of the gradient at the t-1th iteration, represents the bias-corrected first-order momentum estimate, represents the bias-corrected second-order momentum estimate, represents the momentum decay coefficient of the first-order momentum estimate, represents the momentum decay coefficient of the second-order momentum estimation, P' represents the updated model parameters after one iteration, θ' represents the updated model parameters, and ∈ represents a constant;

[0077] After each training phase, the model is evaluated using the validation set, and the loss function and accuracy of the model on the validation set are calculated; if the change in the validation set is less than the preset threshold in several consecutive training cycles, the model is considered to have reached the optimal state and training is stopped. The specific formula is:

[0078] Loss val,t -Loss val,t-1 <λ

[0079] Among them, λ represents a positive number, Loss val,t Represents the validation set loss value obtained in the tth validation cycle, Loss val,t-1 Represents the validation set loss value obtained in the t-1th validation cycle;

[0080] After training is completed, the test set is input into the model, and the loss and accuracy of the model on the test set are calculated. The specific formula is:

[0081]

[0082] Among them, L represents the loss function, y e represents the true label of the g-th test set sample, represents the model prediction value of the g-th test set sample, and G represents the number of test set samples.

[0083] The real-time alarm notification module is used to send the received warnings to the laboratory administrator so that the administrator can take necessary measures to deal with abnormal situations in a timely manner.

[0084] Furthermore, the laboratory personnel behavior warning system also includes a personnel activity history recording module, which is used to record the activity history data of laboratory personnel, including entry and exit time, stay location and activity pattern.

[0085] Furthermore, the integrated control system includes a unified control platform module, a real-time data integration module, a system linkage control module, a remote monitoring and operation module, an intelligent decision support module and an expandable module.

[0086] The unified control platform module is used to provide an intuitive user interface to centrally monitor and control the laboratory environment early warning system, laboratory equipment early warning system and laboratory personnel behavior early warning system, making it convenient for laboratory administrators to operate and manage.

[0087] The real-time data integration module is used to integrate data from the laboratory environment early warning system, laboratory equipment early warning system and laboratory personnel behavior early warning system in real time. Through data integration, the laboratory's operating status can be fully understood.

[0088] The system linkage control module is used to automatically trigger associated control operations when an abnormal situation is detected.

[0089] Remote monitoring and operation module, used for remote monitoring and operation, allows laboratory administrators to remotely access the system via the Internet in real time.

[0090] Intelligent decision support module, used to provide decision suggestions through data analysis.

[0091] Expandable modules are used to integrate new subsystems or update existing systems to optimize the system.

[0092] Compared with the prior art, the present invention adopts the above technical solution and has the following technical effects:

[0093] The present invention introduces an integrated control system, including multiple modules such as laboratory environment warning, equipment warning, and personnel behavior warning, to comprehensively monitor various parameters inside and outside the laboratory. Through the intelligent decision support module, a large amount of data can be analyzed in real time to provide more comprehensive decision support for laboratory managers. This comprehensive monitoring and early warning mechanism enables the laboratory to detect and take corresponding measures earlier when abnormal situations occur, thereby effectively reducing potential risks.

[0094] In addition, the present invention also introduces an adaptive learning algorithm, which can autonomously learn and improve the accuracy of abnormal behavior according to the specific environment and historical data of the laboratory. The introduction of the intelligent decision support module further enhances the system's ability to respond to abnormal situations, making laboratory management more intelligent and efficient.

[0095] Therefore, compared with the prior art, the present invention has significant advantages in comprehensive monitoring and early warning, adaptive learning and intelligent control in the field of laboratory safety, which effectively improves the safety and management efficiency of the laboratory. BRIEF DESCRIPTION OF THE DRAWINGS

[0096] Figure 1 It is the overall structural diagram of the system of the present invention.

[0097] Figure 2 It is a structural diagram of the integrated control system of the present invention. DETAILED DESCRIPTION

[0098] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments.

[0099] It should be noted that, in the description of the present invention, unless otherwise specified, "plurality" means two or more than two; the terms "upper", "lower", "left", "right", "inner", "outer", "front end", "rear end", "head", "tail" and the like indicate the orientation or position relationship based on the orientation or position relationship shown in the drawings, which is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention.

[0100] Furthermore, the terms “first,” “second,” “third,” etc. are used for descriptive purposes only and are not to be construed as indicating or implying relative importance.

[0101] At the same time, in the description of the present invention, unless otherwise clearly specified and limited, the terms "connected" and "connection" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0102] To achieve the above purpose, the present invention proposes a laboratory safety IoT intelligent early warning method, which is used in a chemical laboratory. Figure 1 As shown, including:

[0103] Laboratory environment early warning system, laboratory equipment early warning system, laboratory personnel behavior early warning system, adaptive learning system and integrated control system.

[0104] Among them, the laboratory environment warning system, laboratory equipment warning system, laboratory personnel behavior warning system and adaptive learning system are all connected to the integrated control system through the wireless communication network; the adaptive learning system is connected to the laboratory environment warning system, laboratory equipment warning system and laboratory personnel behavior warning system through the wireless communication network to receive environmental parameters and equipment operation data in real time for comprehensive analysis and optimization.

[0105] The laboratory environment early warning system is used to extract the features of the real-time environmental parameters collected by sensors, obtain fusion features, analyze the fusion features using advanced real-time pattern recognition algorithms, obtain early warning signals, quickly and accurately identify potential risk factors, and adjust the environmental parameters according to the control feedback mechanism to keep the environment within a safe range; and transmit the real-time collected environmental parameters to the integrated control system to ensure the timely sharing of environmental data, so as to facilitate the system to conduct comprehensive analysis and decision-making; the specific contents are:

[0106] The system includes advanced sensing technology components, environmental parameter transmission modules, real-time pattern recognition modules, safety threshold setting modules, real-time monitoring and feedback modules, and historical data storage modules.

[0107] The advanced sensing technology component includes a high-precision, high-sensitivity, and multi-functional sensor network for real-time acquisition of laboratory environmental parameters, including temperature, humidity, air quality, noise level, entry and exit channels, and the opening and closing status of ventilation windows.

[0108] The environmental parameter transmission module is used to receive the parameters collected by the advanced sensor technology components through the wireless communication network, pre-process the parameters and extract features to obtain spatial features and temporal features, fuse the features, and transmit the fused features to the real-time pattern recognition module through the wireless communication network to ensure timely and accurate data transmission and better capture the dynamic changes of environmental parameters. The specific contents are:

[0109] Step 1: Preprocessing includes cleaning the parameters, removing duplicate data, filling missing values ​​and removing outliers to ensure the accuracy and completeness of the data set; using signal processing technology (such as low-pass filtering, high-pass filtering, etc.) to filter the noise of the cleaned parameters to eliminate the impact of environmental interference on data analysis, making subsequent feature extraction more reliable; standardizing the filtered parameters to convert data of different dimensions to the same standard, which helps to improve the training effect of the model and reduce the negative impact caused by differences in data range.

[0110] Step 2: Use regression analysis or other machine learning techniques to process the parameters preprocessed in step 1, extract the correlation between the operating status of the equipment in the laboratory and the environmental parameters, and obtain the corresponding spatial features. The features include the location of the equipment in the laboratory and the distribution of sensors, which help to understand the correlation between the equipment status and the environment; use time series analysis technology to analyze the time series data of the preprocessed parameters in step 1 to obtain time features. The features include periodic changes, trend changes and mutation points of the data, which reflect the operating status of the equipment in different time periods and help identify potential abnormal situations.

[0111] Step 3: Standardize the spatial and temporal features to ensure that different features are within the same range, which helps reduce the impact of different feature scales. Use feature fusion technology to assign different weights to the processed features to emphasize the features that are more important to the analysis results, and then fuse them to obtain fused features, which can more comprehensively capture the dynamic changes of environmental parameters and provide a more accurate analysis basis.

[0112] The fused feature vector is verified, and the error between the fusion result and the actual observation value is calculated to evaluate the fusion effect to ensure that it can effectively capture the dynamic changes of environmental parameters.

[0113] The real-time pattern recognition module is used to analyze the fusion features using real-time pattern recognition algorithms (including support vector machines, decision trees, random forests and neural networks) to obtain the changing trends of laboratory environmental parameters (changing trends include upward or downward trends, periodic changes, mutations, etc.), quickly identify potential risk factors (for example, when temperature parameters show an abnormal upward trend, the system can identify possible fire hazards), generate early warning signals, and pass them to the safety threshold setting module and the real-time monitoring and feedback module for corresponding regulation and processing; the early warning signals can also be notified to laboratory administrators through real-time alarm notifications, text messages or emails.

[0114] Among them, the upward or downward trend is to determine whether a parameter is increasing or decreasing and to evaluate its rate of change; the periodic change is to identify whether the parameter has periodic fluctuations, such as daytime temperature changes; the mutation is to identify whether the parameter has a sudden change, such as a sharp rise or fall in temperature, which may indicate potential danger.

[0115] The safety threshold setting module is used to provide a user interface, allowing laboratory administrators to set safety thresholds according to experimental needs and special environmental requirements, ensuring that environmental parameters are always within the preset safety range, and passing the set safety threshold information to the real-time pattern recognition module and the real-time monitoring and feedback module for comparison and adjustment.

[0116] The real-time monitoring and feedback module is used to activate the real-time control and feedback mechanism after receiving the early warning signal, and adjust the laboratory environmental parameters by controlling related equipment, including adjusting the ventilation system, adjusting the air-conditioning temperature, and opening or closing windows to restore the environment to a safe level.

[0117] The historical data storage module is used to record the time, location and specific parameter values ​​of abnormal environmental parameter events to support subsequent analysis and improve the system's learning ability.

[0118] The laboratory equipment early warning system is used to pre-process the laboratory equipment operation data and build a time domain multi-scale feature sharing fusion-domain adaptive learning model to extract features with different time scales in the data and integrate them. The Lasso regression model is used to predict the integrated features and identify potential faults in advance. Abnormal conditions will immediately trigger early warnings to ensure the safe operation of laboratory equipment. The specific contents are as follows:

[0119] The system includes an equipment condition monitoring module, a data transmission module, an equipment monitoring adaptive module, an abnormal event reporting module and an equipment health history recording module.

[0120] The equipment condition monitoring module is used to monitor the operating data of all equipment in the laboratory in real time, including power consumption, temperature, vibration and current, through a high-precision, high-sensitivity and multi-functional sensor network.

[0121] The data transmission module is used to transmit the operation data of the equipment to the integrated control system and the equipment monitoring adaptive module, and ensure timely and accurate update of the data in the integrated control system.

[0122] The equipment monitoring adaptive module is used to pre-process the equipment operation data, including denoising, standardization and grayscale processing, to ensure the consistency and quality of the data; build a time domain multi-scale feature sharing fusion-domain adaptive learning model, input the pre-processed equipment status data into the model, extract features with different time scales, and integrate them, and use the Lasso regression model to predict the integrated features to obtain equipment fault information. The specific contents are:

[0123] The standardized expression is:

[0124]

[0125] Among them, X represents the original data, μ represents the mean, σ represents the standard deviation, and X std Represents the standardized data;

[0126] The time-domain multi-scale feature sharing fusion-domain adaptive learning model includes an input layer, a feature sharing layer, and a fusion layer. The model is trained using a domain adaptive algorithm to reduce the distribution differences between domains and improve its accuracy and robustness in fault diagnosis tasks. The preprocessed equipment status data is input into the trained model, and the time-domain multi-scale feature extraction method is used to extract features with different time scales. The features include the vibration signal, temperature change, and pressure fluctuation of the equipment, representing the operating status of the equipment in different time periods. The features are integrated using feature weighted averaging to obtain an integrated feature vector, so as to comprehensively utilize information at various scales. The specific expression is:

[0127]

[0128] Among them, F fusion represents the fused feature vector, w i represents the feature weight of the j-th time scale feature, F w represents the original features of the w-th device status data, and n represents the total number of time scale features;

[0129] The Lasso regression model is used to identify and predict the fault characteristics of the integrated feature vector to obtain fault prediction results and health status assessment, which can accurately identify and predict equipment fault information and realize intelligent diagnosis of equipment faults under changing working conditions through optimization;

[0130] The expression of the Lasso regression model is:

[0131] Y=X·β

[0132] Among them, Y represents the target variable, β represents the regression coefficient;

[0133] The expression of the loss function with L1 regularization term is:

[0134]

[0135] Where N represents the number of samples, λ represents the regularization parameter, and L(β) represents the loss function;

[0136] Minimize the loss function to get the optimal regression coefficient. The specific expression is:

[0137]

[0138] in, represents the estimated regression coefficient;

[0139] After Lasso regression training, the non-zero coefficients are the weights of the selected features;

[0140] By calculating the mean square error between the fusion result and the actual observation value, the prediction result is verified and evaluated to ensure that the fusion result meets expectations and can be effectively applied to subsequent tasks. The formula is as follows:

[0141]

[0142] Among them, MSE represents the mean square error, y i represents the actual observed value of the i-th sample, Represents the corresponding predicted value of the i-th sample.

[0143] The abnormal event reporting module is used to collect data related to abnormal events based on equipment failure information, including the time when the abnormality occurred, the location of the equipment and the specific abnormal parameter values, and automatically generate a detailed abnormal event report. The report includes a description of the abnormal event, an impact assessment, an analysis of possible causes, and recommended solutions. It also sends an alert to the laboratory administrator, including a report summary and urgency level, to ensure that the experimenter can take necessary measures in a timely manner. The generated abnormal event report is archived in the system for subsequent analysis and improvement of the system's learning capabilities. The report archive includes a timestamp and detailed information on report generation for traceability and reference.

[0144] The equipment failure information includes:

[0145] Abnormal power consumption: The power consumption of the device increases or decreases suddenly.

[0146] Abnormal temperature: The device is running too high or too low.

[0147] Abnormal vibration: The vibration amplitude of the equipment increases abnormally.

[0148] Abnormal current: The current value of the device changes abnormally.

[0149] Abnormal running time: The equipment running time is too long or too short.

[0150] Abnormal maintenance history: The equipment is not maintained on time or the maintenance record is abnormal.

[0151] Abnormal fault records: The frequency of equipment failures increases and the failures are not recorded in a timely manner.

[0152] Automatic adjustment of related equipment includes the following:

[0153] ① Adjust equipment operating parameters: Modify the equipment's operating parameters, such as lowering the current, voltage or temperature settings.

[0154] ② Start backup equipment: When the main equipment fails, start the backup equipment to ensure the normal operation of the laboratory.

[0155] ③ Reduce the load: When the equipment is overloaded, reduce the pressure on the equipment by reducing the load to prevent further damage.

[0156] The equipment health history record module is used to record the equipment health history data, including operating time, maintenance history, and fault records, to support in-depth analysis of equipment conditions and provide more accurate predictions.

[0157] The laboratory personnel behavior early warning system is used to obtain personnel activity data and analyze the data using a convolutional neural network visual adaptive learning model. When abnormal behavior is identified, the early warning mechanism is triggered and the laboratory administrator is notified to ensure laboratory safety. The personnel activity data is transmitted to the integrated control system to provide the system with real-time personnel behavior data support so that potential safety risks can be identified and responded to in a timely manner. The specific contents are as follows:

[0158] The system includes a personnel positioning and identification module, a behavior monitoring and adaptive module, a real-time alarm notification module and a personnel activity history recording module.

[0159] The personnel positioning and identification module is used to use advanced UWB personnel positioning technology to achieve high-precision positioning of personnel in the laboratory by sending and receiving short pulse signals, avoiding signal interference and position errors, and using identity recognition technology (such as cameras) to confirm the identity of personnel in the laboratory. When unauthorized personnel enter sensitive areas, the system can automatically trigger security measures such as alarm systems and security door locks to ensure the safety of the laboratory and prevent unauthorized personnel from entering sensitive areas; at the same time, the information of personnel in the laboratory, including identity recognition data and entry and exit time, is recorded for subsequent activity history recording and analysis, thereby obtaining personnel activity data, which includes personnel location information, movement trajectory and behavior patterns.

[0160] The behavior monitoring adaptive module is used to process the convolutional neural network visual adaptive learning model based on the personnel behavior image to obtain the processed model, extract the features of the pre-processed personnel activity data, and obtain the spatial features and temporal features to better capture the dynamics and changes of personnel behavior; compare the extracted features with normal behavior and abnormal behavior to identify any abnormal behavior occurring in the laboratory; once an abnormality is detected, the system will issue a timely warning; the specific contents are:

[0161] The convolutional neural network visual adaptive learning model includes multiple convolutional layers, pooling layers, and fully connected layers;

[0162] The sources of human behavior images include the following:

[0163] (1) Real-time monitoring camera: The monitoring camera installed in the laboratory captures the behavior images of personnel in real time. These image data will be automatically stored in the system database for subsequent processing and analysis.

[0164] (2) Historical records: These images may be obtained from routine monitoring of the laboratory or from specific experiments or tests.

[0165] (3) Warning records: When the system detects abnormal behavior, it will record relevant images as warning records, which can be used to train and optimize the model.

[0166] The personnel behavior images are divided into training set, validation set and test set. The training set is input into the convolutional neural network visual adaptive learning model. The convolution operation is performed through the convolution kernel (filter) in the convolution layer to extract the low-level features (edges, corners) and high-level features (shape, texture) in the training set. The specific expression is:

[0167]

[0168] Among them, I represents the training set matrix, K represents the convolution kernel matrix, r represents the row position of the output feature, c represents the column position of the output feature, m represents the row index of the convolution kernel, and b represents the column index of the convolution kernel;

[0169] The fully connected layer is used to classify and learn the extracted features, and then identify normal and abnormal behaviors. The specific expression is:

[0170] z=Wx+g

[0171] Among them, W represents the weight matrix, x represents the input feature, g represents the bias term, and z represents the output;

[0172] Using the gradient descent method and Adam optimization algorithm, the model parameters are continuously adjusted to minimize the loss function and improve the classification accuracy;

[0173] The calculation formula of the gradient descent method is as follows:

[0174]

[0175] Among them, P represents the updated model parameters, θ represents the current model parameters, and η represents the learning rate. represents the gradient of the loss function;

[0176] The calculation formula of Adam optimization algorithm is as follows:

[0177] First-order momentum estimate of the gradient:

[0178] Second-order momentum estimate of the gradient:

[0179] Bias correction for the first-order momentum estimate:

[0180] Bias correction for the second-order momentum estimate:

[0181] Parameter update formula:

[0182] Among them, m t represents the first-order momentum estimate of the gradient at the tth iteration; β1 represents the momentum decay coefficient, which is usually set to 0.9; m t-1 represents the first-order momentum estimate of the gradient at the t-1th iteration; v t represents the second-order momentum estimate of the gradient at the tth iteration; β2 represents the momentum decay coefficient, which is usually set to 0.999; v t-1 represents the second-order momentum estimate of the gradient at the t-1th iteration; represents the bias-corrected first-order momentum estimate; represents the bias-corrected second-order momentum estimate; represents the momentum decay coefficient of the first-order momentum estimate, usually set to 0.9; represents the momentum decay coefficient of the second-order momentum estimation, which is usually set to 0.999; P' represents the updated model parameters after one iteration, θ' represents the updated model parameters, and ∈ represents a constant.

[0183] The validation set is used for parameter adjustment and early stopping to ensure the generalization ability of the model during training. The process is as follows:

[0184] ① Parameter adjustment: After each training phase, the model is evaluated using the validation set, and the loss function and accuracy of the model on the validation set are calculated to monitor the performance of the model;

[0185] ② Early stopping: If the change in the validation set is less than the preset threshold in several consecutive training cycles, the model is considered to have reached the optimal state and training is stopped. The specific formula is:

[0186] Loss val,t -Loss val,t-1 <λ

[0187] Among them, λ represents a small positive number (such as 0.0001), Loss val,t Represents the validation set loss value obtained in the tth validation cycle, Loss val,t-1 Represents the validation set loss value obtained in the t-1th validation cycle;

[0188] After training is completed, the test set is input into the model, and the loss and accuracy of the model on the test set are calculated. The specific formula is:

[0189]

[0190] Among them, L represents the loss function, y e represents the true label of the g-th test set sample, represents the model prediction value of the g-th test set sample, and G represents the number of test set samples.

[0191] Thus, the processed convolutional neural network visual adaptive learning model is obtained.

[0192] The real-time alarm notification module is used to send the received warnings to the laboratory administrator so that the administrator can take necessary measures to deal with abnormal situations in time to ensure laboratory safety.

[0193] Abnormal behavior includes:

[0194] Abnormal movement trajectory: When personnel frequently stay in sensitive areas in the laboratory and violate prescribed routes of action.

[0195] Abnormal stay: Staying in an area or near equipment for a long time, beyond the normal operating or working range.

[0196] Unauthorized entry into sensitive areas: Personnel enter sensitive areas without authorization or identity verification.

[0197] Abnormal operation of equipment: Unauthorized operation, modification or maintenance of equipment may cause equipment failure or safety hazards and will be considered dangerous behavior.

[0198] Violation of safety regulations: Failure to work in accordance with laboratory safety regulations and operating procedures may lead to laboratory safety accidents.

[0199] Carrying dangerous goods: Carrying dangerous goods such as flammable, explosive, corrosive substances into the laboratory may cause safety risks.

[0200] The personnel activity history recording module is used to record the activity history data of laboratory personnel, including entry and exit time, stay location, activity pattern, etc., to support subsequent analysis and improve the learning ability of the system.

[0201] The adaptive learning system is used to optimize the laboratory environment warning system, laboratory equipment warning system and laboratory personnel behavior warning system using machine learning technology, improve the intelligence level and response capability of the overall warning system to maintain the efficiency and adaptability of the system; and feed back the optimized warning system and its obtained characteristics to the integrated control system to improve the intelligence level and warning accuracy of the entire system.

[0202] The integrated control system is used to integrate the data of the laboratory environment early warning system, laboratory equipment early warning system and laboratory personnel behavior early warning system by using data analysis and processing technology, and feed back the processing results to the laboratory environment early warning system, laboratory equipment early warning system, laboratory personnel behavior early warning system and adaptive learning system to realize real-time monitoring and early warning functions; the specific contents are:

[0203] The system includes a unified control platform module, a real-time data integration module, a system linkage control module, a remote monitoring and operation module, an intelligent decision support module and an expandable module.

[0204] The unified control platform module is used to provide an intuitive user interface to centrally monitor and control the laboratory environment early warning system, laboratory equipment early warning system and laboratory personnel behavior early warning system, making it convenient for laboratory administrators to operate and manage.

[0205] The real-time data integration module is used to integrate data from the laboratory environment early warning system, laboratory equipment early warning system and laboratory personnel behavior early warning system in real time. Through data integration, the system can fully understand the operation status of the laboratory.

[0206] The system linkage control module is used to automatically trigger the associated control operations when an abnormal situation is detected; for example, the ventilation system can be adjusted when an environmental warning is issued, or the security door can be locked when abnormal behavior of personnel is detected.

[0207] The remote monitoring and operation module is used for remote monitoring and operation, allowing laboratory administrators to remotely access the system via the Internet. Administrators can monitor the laboratory status in real time no matter where they are, and perform remote operations when necessary.

[0208] The intelligent decision support module is used to provide decision recommendations by analyzing large amounts of data. The system can assist administrators in making more informed decisions based on previous historical data and learning models, such as what safety measures to take under certain environmental conditions.

[0209] Expandable modules are used to integrate new subsystems or update existing systems, allowing the system to be continuously upgraded and optimized as the laboratory grows and changes.

[0210] In this embodiment, at a certain moment, due to improper experimental operation, some chemical substances in the laboratory leaked, which quickly caused the air quality to deteriorate. The sensor network immediately captured this change and transmitted the data to the real-time pattern recognition module through the environmental parameter transmission module. Real-time pattern recognition technology analyzes the changing trend of these parameters, quickly identifies potential risk factors, and generates early warning signals. When the preset safety threshold in the safety threshold setting module is exceeded, the system immediately starts the real-time control feedback mechanism. By controlling the ventilation system in the laboratory, air circulation is accelerated, and at the same time, the laboratory administrator and relevant personnel are notified to take emergency measures.

[0211] During this process, the adaptive learning system optimized the early warning model based on this incident, improving the accuracy of early warnings for similar incidents. The integrated control system ensures data sharing and collaborative work between various subsystems, making the entire processing process efficient and smooth.

[0212] The application of the laboratory safety IoT intelligent early warning system proposed in the present invention in actual situations is feasible and can effectively improve the safety and management efficiency of the laboratory.

[0213] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A laboratory safety IoT intelligent early warning system, characterized in that: include: Laboratory environment early warning system, laboratory equipment early warning system, laboratory personnel behavior early warning system, adaptive learning system and integrated control system; Among them, the laboratory environment warning system, laboratory equipment warning system, laboratory personnel behavior warning system and adaptive learning system are all connected to the integrated control system through the wireless communication network; the adaptive learning system is connected to the laboratory environment warning system, laboratory equipment warning system and laboratory personnel behavior warning system through the wireless communication network; The laboratory environment early warning system is used to extract the features of the real-time environmental parameters collected by sensors, obtain fusion features, analyze the fusion features using advanced real-time pattern recognition algorithms, obtain early warning signals, and adjust the environmental parameters using the control feedback mechanism; and transmit the real-time collected environmental parameters to the integrated control system; The laboratory equipment early warning system is used to pre-process the laboratory equipment operation data and build a time-domain multi-scale feature sharing fusion-domain adaptive learning model to extract features with different time scales in the data and integrate them. The Lasso regression model is used to predict the integrated features, identify potential faults, and perform corresponding processing; The laboratory personnel behavior early warning system is used to obtain personnel activity data and analyze the data using a convolutional neural network visual adaptive learning model. When abnormal behavior is identified, the early warning mechanism is triggered and the laboratory administrator is notified; and the personnel activity data is transmitted to the integrated control system; An adaptive learning system is used to optimize the laboratory environment warning system, laboratory equipment warning system, and laboratory personnel behavior warning system using machine learning technology, and feed back the optimized warning system and its obtained characteristics to the integrated control system; The integrated control system is used to integrate the data of the laboratory environment early warning system, laboratory equipment early warning system and laboratory personnel behavior early warning system using data analysis and processing technology, and feed back the processing results to the laboratory environment early warning system, laboratory equipment early warning system, laboratory personnel behavior early warning system and adaptive learning system to achieve real-time monitoring and early warning functions.

2. The laboratory safety IoT intelligent early warning system according to claim 1 is characterized in that: The laboratory environment early warning system includes advanced sensor technology components, environmental parameter transmission module, real-time pattern recognition module, safety threshold setting module and real-time monitoring and feedback module; The advanced sensing technology component includes a sensor network for real-time acquisition of laboratory environmental parameters, including temperature, humidity, air quality, noise level, entry and exit channels, and the opening and closing status of ventilation windows; An environmental parameter transmission module is used to receive parameters collected by the advanced sensor technology component through a wireless communication network, pre-process the parameters and perform feature extraction to obtain spatial features and temporal features, fuse the features, and transmit the fused features to the real-time pattern recognition module through the wireless communication network; The real-time pattern recognition module is used to analyze the fusion features using the real-time pattern recognition algorithm, obtain the change trend of the laboratory environmental parameters, generate early warning signals, and transmit them to the safety threshold setting module and the real-time monitoring and feedback module for corresponding regulation and processing; the early warning signals can also be notified to the laboratory administrator through real-time alarm notifications, SMS or emails; The safety threshold setting module is used to provide a user interface, allowing the laboratory administrator to set the safety threshold according to the experimental needs and special environment requirements, and transmit the set safety threshold information to the real-time pattern recognition module and the real-time monitoring and feedback module; The real-time monitoring and feedback module is used to start the real-time control and feedback mechanism after receiving the early warning signal, and control the relevant equipment to adjust the laboratory environmental parameters.

3. The laboratory safety IoT intelligent early warning system according to claim 2 is characterized in that: The environment parameter transmission module is configured to perform the following actions: Step 1: Preprocessing includes cleaning the parameters, removing duplicate data, filling missing values ​​and removing outliers; using signal processing technology to filter noise on the cleaned parameters; and standardizing the filtered parameters; Step 2: Process the parameters preprocessed in step 1, extract the correlation between the operating status of the equipment in the laboratory and the environmental parameters, and obtain the corresponding spatial characteristics, which include the location of the equipment in the laboratory and the distribution of sensors; use time series analysis technology to analyze the time series data of the parameters preprocessed in step 1 to obtain time characteristics, which include periodic changes, trend changes and mutation points of the data; Step 3: Standardize the spatial features and temporal features, use feature fusion technology to assign different weights to the processed features, and fuse them to obtain fused features.

4. The laboratory safety IoT intelligent early warning system according to claim 1 is characterized in that: The laboratory environment early warning system also includes a historical data storage module for recording the time, location and specific parameter values ​​of abnormal environmental parameter events.

5. The laboratory safety IoT intelligent early warning system according to claim 1 is characterized in that: The laboratory equipment early warning system includes equipment condition monitoring module, data transmission module, equipment monitoring adaptive module and abnormal event reporting module; Equipment condition monitoring module, which uses a sensor network to monitor the operating data of all equipment in the laboratory in real time, including power consumption, temperature, vibration and current; A data transmission module, used to transmit the operation data of the equipment to the integrated control system and the equipment monitoring adaptive module, and update the data in the integrated control system; The equipment monitoring adaptive module is used to pre-process the equipment operation data, including denoising, standardization and grayscale processing; construct a time domain multi-scale feature sharing fusion-domain adaptive learning model, input the pre-processed equipment status data into the model, extract and integrate features with different time scales, and use the Lasso regression model to predict the integrated features to obtain equipment fault information; The abnormal event reporting module is used to collect data related to abnormal events based on equipment failure information, including the time of abnormal occurrence, equipment location and specific abnormal parameter values, automatically generate detailed abnormal event reports, and send alerts to laboratory administrators; the generated abnormal event reports are archived in the system.

6. The laboratory safety IoT intelligent early warning system according to claim 5 is characterized in that: The device monitoring adaptation module is configured to perform the following actions: The standardized expression is: Among them, X represents the original data, μ represents the mean, σ represents the standard deviation, and X std Represents the standardized data; The time-domain multi-scale feature sharing fusion-domain adaptive learning model includes an input layer, a feature sharing layer and a fusion layer; the domain adaptive algorithm is used to train the time-domain multi-scale feature sharing fusion-domain adaptive learning model; the pre-processed equipment status data is input into the trained model, and the time-domain multi-scale feature extraction method is used to extract features with different time scales, which include vibration signals, temperature changes and pressure fluctuations of the equipment. The features are integrated using feature weighted average to obtain an integrated feature vector, and the specific expression is: Among them, F fusion represents the fused feature vector, w i represents the feature weight of the th time scale feature, F w represents the original features of the w-th device status data, and n represents the total number of time scale features; Lasso regression model is used to identify and predict the fault characteristics of the integrated feature vector to obtain the predicted equipment fault information, and intelligent diagnosis of equipment faults under changing working conditions is achieved through optimization; The expression of the Lasso regression model is: Y=X·β Among them, Y represents the target variable, β represents the regression coefficient; The expression of the loss function with L1 regularization term is: Where N represents the number of samples, λ represents the regularization parameter, and L(β) represents the loss function; Minimize the loss function to get the optimal regression coefficient. The specific expression is: in, represents the estimated regression coefficient; After Lasso regression training, the non-zero coefficients are the weights of the selected features; The prediction results are verified and evaluated by calculating the mean square error between the fusion results and the actual observation values. The specific formula is as follows: Among them, MSE represents the mean square error, y i represents the actual observed value of the i-th sample, Represents the corresponding predicted value of the i-th sample.

7. The laboratory safety IoT intelligent early warning system according to claim 1 is characterized in that: The laboratory equipment early warning system also includes an equipment health history record module, which is used to record the health history data of the equipment, including operating time, maintenance history, and fault records.

8. The laboratory safety IoT intelligent early warning system according to claim 1 is characterized in that: The laboratory personnel behavior early warning system includes a personnel positioning and identification module, a behavior monitoring adaptive module, and a real-time alarm notification module; The personnel positioning and identification module is used to use UWB personnel positioning technology to achieve high-precision positioning of personnel in the laboratory by sending and receiving short pulse signals, and use identity recognition technology to confirm the identity of personnel in the laboratory. When unauthorized personnel enter sensitive areas, security measures are automatically triggered; in addition, the information of personnel in the laboratory, including identity recognition data and entry and exit time, is recorded to obtain personnel activity data, which includes personnel location information, movement trajectory and behavior pattern; The behavior monitoring adaptive module is used to process the convolutional neural network visual adaptive learning model based on the personnel behavior image to obtain the processed model, extract the features of the pre-processed personnel activity data to obtain spatial features and temporal features; compare the extracted features with normal behavior and abnormal behavior, identify any abnormal behavior occurring in the laboratory, and issue an early warning; the specific contents are: The convolutional neural network visual adaptive learning model includes multiple convolutional layers, pooling layers, and fully connected layers; Obtaining images of personnel behavior through identity recognition technology, historical records or warning records; The personnel behavior images are divided into training set, validation set and test set. The training set is input into the convolutional neural network visual adaptive learning model. The convolution operation is performed through the convolution kernel in the convolution layer to extract the features in the training set. The specific expression is: Among them, I represents the training set matrix, K represents the convolution kernel matrix, r represents the row position of the output feature, c represents the column position of the output feature, m represents the row index of the convolution kernel, and b represents the column index of the convolution kernel; The fully connected layer is used to classify and learn the extracted features, and then identify normal and abnormal behaviors. The specific expression is: z=Wx+g Among them, W represents the weight matrix, x represents the input feature, g represents the bias term, and z represents the output; Using the gradient descent method and Adam optimization algorithm, the model parameters are continuously adjusted to minimize the loss function; The calculation formula of the gradient descent method is as follows: Among them, P represents the updated model parameters, θ represents the current model parameters, and η represents the learning rate. represents the gradient of the loss function; The calculation formula of Adam optimization algorithm is as follows: First-order momentum estimate of the gradient: Second-order momentum estimate of the gradient: Bias correction for the first-order momentum estimate: Bias correction for the second-order momentum estimate: Parameter update formula: Among them, m t represents the first-order momentum estimate of the gradient at the tth iteration, β1 represents the momentum decay coefficient, and m t-1 represents the first-order momentum estimate of the gradient at the t-1th iteration, v t represents the second-order momentum estimate of the gradient at the tth iteration, β2 represents the momentum decay coefficient, and v t-1 represents the second-order momentum estimate of the gradient at the t-1th iteration, represents the bias-corrected first-order momentum estimate, represents the bias-corrected second-order momentum estimate, represents the momentum decay coefficient of the first-order momentum estimate, represents the momentum decay coefficient of the second-order momentum estimation, P′ represents the updated model parameters after one iteration, θ′ represents the updated model parameters, and ∈ represents a constant; After each training phase, the model is evaluated using the validation set, and the loss function and accuracy of the model on the validation set are calculated; if the change in the validation set is less than the preset threshold in several consecutive training cycles, the model is considered to have reached the optimal state and training is stopped. The specific formula is: Loss val，t -Loss val，t-1 <λ Among them, λ represents a positive number, Loss val，t Represents the validation set loss value obtained in the tth validation cycle, Loss val，t-1 Represents the validation set loss value obtained in the t-1th validation cycle; After training is completed, the test set is input into the model, and the loss and accuracy of the model on the test set are calculated. The specific formula is: Among them, L represents the loss function, y e represents the true label of the g-th test set sample, represents the model prediction value of the g-th test set sample, and G represents the number of test set samples; The real-time alarm notification module is used to send the received warnings to the laboratory administrator so that the administrator can take necessary measures to deal with abnormal situations in a timely manner.

9. The laboratory safety IoT intelligent early warning system according to claim 1 is characterized in that: The laboratory personnel behavior early warning system also includes a personnel activity history recording module, which is used to record the activity history data of laboratory personnel, including entry and exit time, stay location and activity pattern.

10. The laboratory safety IoT intelligent early warning system according to claim 1 is characterized in that: The integrated control system includes a unified control platform module, a real-time data integration module, a system linkage control module, a remote monitoring and operation module, an intelligent decision support module and an expandable module; The unified control platform module is used to provide an intuitive user interface to centrally monitor and control the laboratory environment early warning system, laboratory equipment early warning system, and laboratory personnel behavior early warning system, making it easier for laboratory administrators to operate and manage; Real-time data integration module, which is used to integrate data from the laboratory environment early warning system, laboratory equipment early warning system and laboratory personnel behavior early warning system in real time. Through data integration, the operation status of the laboratory can be fully understood; System linkage control module, used to automatically trigger related control operations when an abnormal situation is detected; Remote monitoring and operation module, used for remote monitoring and operation, allowing laboratory administrators to remotely access the system via the Internet in real time; Intelligent decision support module, used to provide decision suggestions through data analysis; Expandable modules are used to integrate new subsystems or update existing systems to optimize the system.

Images