Intelligent safety early warning system and method for chemical laboratory based on multi-modal perception

Abstract

The invention provides a chemical laboratory intelligent safety early warning system and method based on multi-mode perception, and relates to the technical field of safety monitoring. The early warning system comprises a multi-dimensional sensing module, an intelligent judgment and decision module, a graded early warning response module, a cloud server communication module, an image report module and a log backup and optimization module, and the monitoring accuracy is greatly improved by fusing multi-modal sensing data of four dimensions of environment, equipment, personnel and hazardous chemicals; through graded early warning response, the early warning precision is greatly improved; remote control and intelligent interconnection are supported, and unattended operation and automatic disposal can be achieved; meanwhile, by recording and analyzing historical data and response schemes, continuous updating and iteration can be performed on a security risk feature database and a judgment recognition threshold value, the response schemes are optimized, the accuracy of the system is continuously improved, the system can be continuously compatible with future application scenes, and long-term technical suitability is achieved.

Description

Technical Field

[0001] This application relates to the field of safety monitoring technology, and more specifically, to an intelligent safety early warning system and method for chemical laboratories based on multimodal perception. Background Technology

[0002] Chemistry laboratories, as important locations for scientific research and teaching, often store large quantities of toxic, hazardous, flammable, and explosive chemical reagents, and involve complex experimental procedures such as high-temperature, high-pressure, and precision instrument operation. Traditional laboratory safety monitoring has significant deficiencies:

[0003] Single detection dimension: Relying on manual inspections or a single sensor (such as a gas alarm) cannot simultaneously monitor multiple dimensions of risks, such as the environment, equipment, personnel behavior, and the status of hazardous chemicals. According to statistics, personnel behavior is the main cause of laboratory safety accidents. By adding personnel behavior as a primary detection dimension, this system can effectively improve the early warning effect of laboratory safety and truly reduce the occurrence of laboratory safety accidents.

[0004] High false alarm rate: Single sensors are easily affected by environmental interference (such as temperature and humidity fluctuations causing gas sensors to be falsely triggered), and frequent false alarms affect monitoring efficiency.

[0005] Delayed response: Manual monitoring has a delay in response, making it impossible to initiate emergency measures the moment a risk occurs, which can easily lead to the escalation of the accident.

[0006] Lack of intelligent decision-making: Traditional systems cannot comprehensively determine risk levels based on multi-source data, making it difficult to achieve graded early warning and precise response.

[0007] This invention aims to provide a chemical laboratory intelligent safety early warning system and method based on multimodal perception. By integrating data from four dimensions—environment, equipment, personnel, and hazardous chemicals—through multimodal perception, and combining intelligent judgment algorithms with a graded early warning mechanism, it achieves comprehensive monitoring, precise early warning, and intelligent control of laboratory safety, thus solving the core pain points of traditional technologies. Summary of the Invention

[0008] The purpose of this application is to provide a chemical laboratory intelligent safety early warning system and method based on multimodal perception, which can solve the technical problems of comprehensive monitoring, accurate early warning and intelligent control of laboratory safety.

[0009] This application provides an intelligent safety early warning system for chemical laboratories based on multimodal perception, including:

[0010] The multi-dimensional sensing module integrates laboratory environment sensing unit, equipment operation sensing unit, personnel behavior capture unit and hazardous chemical status sensing unit to collect multi-dimensional data of the laboratory in real time.

[0011] The intelligent judgment and decision-making module includes a laboratory environment risk identification and judgment unit, an equipment operation risk identification and judgment unit, a personnel behavior risk identification and judgment unit, a hazardous chemical risk identification and judgment unit, a laboratory risk level judgment unit, and a laboratory safety risk identification feature download unit. It calculates a comprehensive risk value through a weighted multi-source risk fusion algorithm and compares it with a preset threshold to dynamically determine the risk source and risk level.

[0012] The tiered early warning and response module is used to output corresponding measures based on the risk level.

[0013] The cloud server communication module includes a laboratory safety risk characteristic database, a cloud data interoperability unit, and a communication interconnection unit. It is used to record and store laboratory safety risk characteristics, data upload / reception between devices, and ensure that the system can be linked and controlled to realize 24-hour remote monitoring and real-time automatic handling of the laboratory.

[0014] The image reporting module is used to mark the types and locations of hazards in the laboratory's multi-dimensional data according to risk level, and to generate graphic and textual early warning response reports through the report export unit;

[0015] The log backup and optimization module is used to record historical multi-dimensional laboratory data and graphic warning response reports stored in the backup unit, and to analyze historical data through iterative optimization unit, and to regularly update the laboratory safety risk characteristic database and judgment thresholds.

[0016] Furthermore, the multi-dimensional sensing module is equipped with multiple sets of gas sensors, temperature sensors, humidity sensors, infrared thermal imaging cameras, and RFID tags to monitor and collect multi-dimensional data from the laboratory in real time.

[0017] Furthermore,

[0018] The integrated laboratory environment sensing unit is used to collect environmental data of the laboratory, including the concentration of toxic and harmful gases, temperature and humidity, open flame, smoke and air pressure data.

[0019] The equipment operation sensing unit is used to collect data from laboratory equipment and report operational data such as laboratory equipment temperature / pressure, fume hood fan speed, power supply current, and reagent cabinet temperature.

[0020] The personnel behavior capture unit is used to capture behavioral data of experimental personnel, including data on failure to wear protective equipment as required, violation of regulations, and trespassing into restricted areas.

[0021] The hazardous chemical status sensing unit is used to collect hazardous chemical status data, including hazardous chemical inventory, hazardous chemical storage location, hazardous chemical temperature, and other hazardous chemical status data.

[0022] Furthermore,

[0023] The laboratory environment risk identification and judgment unit is used to identify and judge the laboratory environment data output by the multi-dimensional sensing module based on the safety risk characteristics caused by the laboratory environment, and to determine the types of toxic and harmful gases, temperature and humidity, open flame, smoke and air pressure risks in the laboratory environment and the corresponding risk values.

[0024] The equipment operation risk identification and judgment unit is used to identify and judge the equipment operation data output by the multi-dimensional perception module based on the safety risk characteristics caused by equipment operation, and to determine whether there are risks related to equipment temperature / pressure, fume hood wind speed, power current, reagent cabinet temperature and corresponding risk values ​​during equipment operation.

[0025] The personnel behavior risk identification and judgment unit is used to identify and judge the personnel behavior data output by the multi-dimensional perception module based on the safety risk characteristics caused by personnel behavior, and to determine whether the personnel behavior has the risk of not wearing protective equipment as required, violating regulations, or entering restricted areas, as well as the corresponding risk value.

[0026] The hazardous chemical risk identification and judgment unit is used to identify and judge the hazardous chemical status data output by the multi-dimensional perception module based on the safety risk characteristics caused by the status of hazardous chemicals, and to determine the inventory quantity, storage location, temperature status risk and corresponding risk value of hazardous chemicals.

[0027] The laboratory risk level determination unit is used to statistically determine the cumulative risk values ​​of the laboratory environmental risk identification and determination unit, equipment operation risk identification and determination unit, personnel behavior risk identification and determination unit, and hazardous chemical risk identification and determination unit, to determine the corresponding risk level, and issue a warning signal of the corresponding level.

[0028] The laboratory safety risk identification feature download unit is used to download laboratory risk identification feature data for comparison and verification of laboratory safety risk categories and risk values.

[0029] Furthermore, the risk levels determined by the intelligent judgment and decision-making module include Level III, Level II, and Level I; the graded early warning response module includes Level III, Level II, and Level I early warning response units corresponding to the risk levels.

[0030] Among them, the Level III early warning response unit is used to output a Level III early warning response signal, issue a green light warning, output a linkage signal, and send a text warning to the laboratory intelligent management terminal, prompting "minor abnormality, inspection recommended";

[0031] Among them, the Level II early warning response unit is used to output a Level II early warning response signal, issue a yellow light and sound horn warning, output a linkage signal, and send a graphic warning to the laboratory intelligent management terminal and the experimenter's mobile APP, prompting "moderate risk, handle in time";

[0032] The Level I early warning response unit is used to output a Level I early warning response signal, which emits a red light and a high-decibel siren to warn of the danger. It also triggers a mobile APP alarm and a fire linkage module (activating emergency call, starting the sprinkler system, and cutting off the power), prompts "Serious risk, take immediate action", and pushes alarm information to the school's safety management department or fire department.

[0033] Furthermore,

[0034] The cloud-based laboratory safety risk characteristic database records and stores laboratory safety risk characteristic data in the form of numbers, and records and stores the risk level of the corresponding laboratory safety risk characteristics in the form of risk values;

[0035] The cloud data communication unit is used to receive and transmit data and commands between devices, enabling the uploading and receiving of data and commands;

[0036] The communication interconnection unit enables interconnection between the intelligent safety early warning system and the laboratory intelligent management terminal, mobile APP, intelligent devices and cloud server through Bluetooth, Wi-Fi, GSM and ZigBee, achieving 24-hour remote monitoring and automatic handling.

[0037] Furthermore, the image reporting module includes:

[0038] The image labeling unit is used to identify and mark the types and locations of hazards in the laboratory multidimensional data output by the multidimensional perception module. Hazards are identified according to their risk levels and marked with different colors to improve their recognizability.

[0039] The report export unit is used to export hazard sources, risk levels, and early warning data, and generate graphic and textual early warning response reports.

[0040] Furthermore, the logging and optimization module includes:

[0041] The log recording and backup unit is used to record and back up historical monitoring data and early warning response reports in order to summarize early warning experience;

[0042] The iterative optimization unit is used to analyze historical monitoring data and historical early warning response plans, and regularly update and iterate the laboratory safety risk characteristic database to optimize the identification threshold of the intelligent judgment and decision-making module.

[0043] Furthermore, the intelligent judgment and decision-making module integrates multi-dimensional laboratory data, sets weighting coefficients based on the risk values ​​and occurrence frequencies of different hazard sources, performs weighted fusion of multiple variables, and then dynamically judges the risk level based on the accumulated risk value. Specifically:

[0044] The weighting coefficients are dynamically adjusted based on historical accident data statistics from the laboratory, experimental time, and experimental type, and satisfy the following conditions: , The weights are set;

[0045] In the process of weighted fusion of multiple variables and dynamic judgment of risk level based on cumulative risk value, the risk value of each hazard source is first calculated, and the detection values ​​of different hazard sources are then combined. With safety threshold Compare and combine with early warning thresholds The normalized risk contribution value is calculated using the following formula:

[0046]

[0047] Next, calculate the total risk value of the hazard source. The formula is Then calculate the weighted sum of various risk values, Rtotal, using the following formula: Finally, risk assessment is conducted based on the established risk thresholds.

[0048] Based on the same inventive concept, this application also provides a method for intelligent safety early warning in chemical laboratories based on multimodal perception, comprising the following steps:

[0049] The multi-dimensional sensing module collects real-time data on the laboratory environment, equipment, personnel behavior, and the status of hazardous chemicals.

[0050] The intelligent judgment and decision-making module identifies and judges risks based on multi-dimensional data, and outputs a comprehensive risk value and risk level.

[0051] The tiered early warning response module activates corresponding early warning measures based on the risk level.

[0052] The image reporting module generates and exports hazard identification and early warning response reports;

[0053] The log backup and optimization module records historical data and regularly optimizes the risk characteristic database and judgment thresholds.

[0054] The beneficial effects of this invention are:

[0055] This invention provides a multimodal perception-based intelligent safety early warning system for chemical laboratories. By integrating multimodal perception data from four dimensions—environment, equipment, personnel, and hazardous chemicals—and combining it with intelligent database term comparison technology, it achieves rapid identification of risk sources and accurate determination of risk levels, significantly improving monitoring accuracy. This solves the technical problems of high missed detection rates and frequent false alarms in traditional single-dimensional monitoring. The invention avoids management chaos caused by "one-size-fits-all" alarms through tiered early warning responses, greatly improving early warning accuracy and enabling early risk prediction and intervention to effectively prevent the risk from escalating. Furthermore, this invention supports remote control and intelligent interconnection. Users can view the laboratory safety status anytime, anywhere through intelligent management terminals and mobile apps, achieving 24-hour remote monitoring. The system can be linked to other equipment, automatically triggering the fire protection system in case of severe risks, with a fast response speed, achieving "unmanned operation + automatic handling." Simultaneously, this invention can continuously update and iterate the safety risk characteristic database and judgment and identification thresholds by recording and analyzing historical data and response plans, optimizing response plans, continuously improving system accuracy, and ensuring long-term technical adaptability to future application scenarios. Attached Figure Description

[0056] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0057] Figure 1 This is a flowchart illustrating the intelligent security early warning process in an embodiment of the present invention;

[0058] Figure 2 This is a flowchart illustrating the calculation process of the comprehensive risk value in an embodiment of the present invention. Detailed Implementation

[0059] To facilitate understanding of the present invention, it will be described more fully below through embodiments, and preferred embodiments are given below. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. Any other implementation schemes obtained by modifying or equivalently substituting the technical solutions of the present invention without inventive step are all within the protection scope of the present invention.

[0060] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art. The terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention.

[0061] The numerical values ​​disclosed in the embodiments of this invention are approximate values, not definitive values. Where error or experimental conditions permit, all values ​​within the error range may be included, and the specific numerical values ​​disclosed in the embodiments of this invention are not limited to those specified.

[0062] See below Figure 1 As shown, this embodiment provides a multimodal sensing-based intelligent safety early warning system for chemical laboratories, including:

[0063] The multi-dimensional sensing module integrates laboratory environment sensing unit, equipment operation sensing unit, personnel behavior capture unit and hazardous chemical status sensing unit to collect multi-dimensional data of the laboratory in real time.

[0064] The intelligent judgment and decision-making module includes a laboratory environment risk identification and judgment unit, an equipment operation risk identification and judgment unit, a personnel behavior risk identification and judgment unit, a hazardous chemical risk identification and judgment unit, a laboratory risk level judgment unit, and a laboratory safety risk identification feature download unit. It calculates a comprehensive risk value through a weighted multi-source risk fusion algorithm and compares it with a preset threshold to dynamically determine the risk source and risk level.

[0065] The graded early warning response module includes a local audible and visual early warning device, a laboratory management terminal, a mobile APP, and a fire linkage module. It is used to output corresponding measures according to the risk level. For example, it outputs the corresponding early warning mode according to the judgment result (Level III, Level II, Level I) of the intelligent judgment and decision module. When the risk is Level I, the fire linkage module is automatically triggered (such as calling for rescue, starting the sprinkler system, cutting off the power, etc.).

[0066] The cloud server communication module includes a laboratory safety risk characteristic database, a cloud data interoperability unit, and a communication interconnection unit. It is used to record and store laboratory safety risk characteristics, data upload / reception between devices, and ensure that the system can be linked and controlled to realize 24-hour remote monitoring and real-time automatic handling of the laboratory.

[0067] The image reporting module is used to mark the types and locations of hazards in the laboratory's multi-dimensional data according to risk level, and to generate graphic and textual early warning response reports through the report export unit;

[0068] The log backup and optimization module is used to record historical multi-dimensional laboratory data and graphic warning response reports stored in the backup unit, and to analyze historical data through iterative optimization unit, and to regularly update the laboratory safety risk characteristic database and judgment thresholds.

[0069] Specifically, the multi-dimensional sensing module is equipped with multiple gas sensors, temperature sensors, humidity sensors, infrared thermal imaging cameras, and RFID tags to conduct real-time comprehensive monitoring of the laboratory's "environment-equipment-personnel-hazardous chemicals" and collect multi-dimensional data from the laboratory.

[0070] Specifically,

[0071] The integrated laboratory environmental sensing unit is used to collect environmental data in the laboratory, including the concentration of toxic and harmful gases, temperature and humidity, open flame, smoke, and air pressure. Specifically, a gas sensor array is installed 0.5m directly above the fume hood, containing three independent sensor groups. Each group integrates a CO concentration sensor (range 0-500ppm), a VOCs sensor (range 0-100ppm), and a toluene sensor (range 0-100ppm) to ensure that a level-three warning is triggered within 30 seconds in the event of a toxic gas leak. Four temperature and humidity sensors, open flame sensors, and smoke sensors are evenly distributed above the laboratory ceiling to monitor the laboratory environmental data in real time.

[0072] The equipment operation sensing unit is used to collect data from laboratory equipment and report data such as temperature / pressure, fume hood airflow, power supply current, and reagent cabinet temperature. Specifically, temperature and pressure sensors are installed on the laboratory equipment to monitor the status of the experimental equipment in real time.

[0073] The personnel behavior capture unit is used to capture behavioral data of experimental personnel, including data on failure to wear protective equipment as required, violation of regulations, and trespassing into restricted areas.

[0074] For example:

[0075] First, two sets of infrared monitoring cameras are distributed about 2 meters above the experimental workbench to monitor the behavior data of the experimental personnel, such as whether they are wearing protective equipment correctly or violating regulations.

[0076] Secondly, because researchers operate the equipment directly for extended periods, their sense of smell can become dulled. This often leads to situations where toxic gas concentrations are already high, but researchers fail to detect them in advance, causing harm or even loss of consciousness. Therefore, at least three independent gas sensors should be installed above the experimental workbench. Different gas sensors can be selected depending on the type of experiment, such as toluene sensors, ether sensors, and formaldehyde sensors. Since researchers operate the equipment directly in front of the equipment, the safety detection threshold should be set lower than the generally stored safety threshold. For example, the safety threshold for an ether sensor could be set to 1 g / m³. 3 about;

[0077] Third, temperature and humidity sensors, smoke sensors, and open flame sensors are installed about 2 meters above the experimental operating table to ensure that an alarm is triggered within 30 seconds if abnormal temperature and humidity occur during the experimental operation, or if smoke or open flame occurs.

[0078] Fourth, restricted access areas in the laboratory, such as high-temperature, high-pressure, or radiation areas, should be equipped with infrared monitoring cameras 2 meters above them to prevent unauthorized personnel from causing direct harm.

[0079] Fifth, place one set of infrared monitoring cameras about 2 meters above the laboratory entrance and exit to monitor whether laboratory personnel bring prohibited items (such as food, beverages, fireworks, etc.) into the laboratory or take medicines and reagents out of the laboratory.

[0080] The hazardous chemical status sensing unit is used to collect hazardous chemical status data, including hazardous chemical inventory, storage location, and temperature. Specifically, an RFID reader can identify RFID tags on hazardous chemical reagents to collect inventory and storage location data. In addition, temperature and humidity sensors and toxic gas sensors are installed inside the hazardous chemical storage cabinet to identify the temperature and humidity status of the hazardous chemicals and the concentration of volatile gases emitted by the hazardous chemicals.

[0081] Specifically,

[0082] The laboratory environment risk identification and judgment unit is used to identify and judge the laboratory environment data output by the multi-dimensional sensing module based on the safety risk characteristics caused by the laboratory environment, and to determine the types of toxic and harmful gases, temperature and humidity, open flame, smoke and air pressure risks in the laboratory environment and the corresponding risk values.

[0083] The equipment operation risk identification and judgment unit is used to identify and judge the equipment operation data output by the multi-dimensional perception module based on the safety risk characteristics caused by equipment operation, and to determine whether there are risks related to equipment temperature / pressure, fume hood wind speed, power current, reagent cabinet temperature and corresponding risk values ​​during equipment operation.

[0084] The personnel behavior risk identification and judgment unit is used to identify and judge the personnel behavior data output by the multi-dimensional perception module based on the safety risk characteristics caused by personnel behavior, and to determine whether the personnel behavior has the risk of not wearing protective equipment as required, violating regulations, or entering restricted areas, as well as the corresponding risk value.

[0085] The hazardous chemical risk identification and judgment unit is used to identify and judge the hazardous chemical status data output by the multi-dimensional perception module based on the safety risk characteristics caused by the status of hazardous chemicals, and to determine the inventory quantity, storage location, temperature status risk and corresponding risk value of hazardous chemicals.

[0086] The laboratory risk level determination unit is used to statistically determine the cumulative risk values ​​of the laboratory environmental risk identification and determination unit, equipment operation risk identification and determination unit, personnel behavior risk identification and determination unit, and hazardous chemical risk identification and determination unit, to determine the corresponding risk level, and issue a warning signal of the corresponding level.

[0087] The laboratory safety risk identification feature download unit is used to download laboratory risk identification feature data for comparison and verification of laboratory safety risk categories and risk values.

[0088] Specifically, in this embodiment, the risk levels determined by the intelligent judgment and decision-making module are divided into Level III, Level II, and Level I; the graded early warning response module includes Level III early warning response units, Level II early warning response units, and Level I early warning response units corresponding to the risk levels.

[0089] Among them, the Level III early warning response unit is used to output a Level III early warning response signal, issue a green light warning, output a linkage signal, and send a text warning to the laboratory intelligent management terminal, prompting "minor abnormality, inspection recommended";

[0090] Among them, the Level II early warning response unit is used to output a Level II early warning response signal, issue a yellow light and sound horn warning, output a linkage signal, and send a graphic warning to the laboratory intelligent management terminal and the experimenter's mobile APP, prompting "moderate risk, handle in time";

[0091] The Level I early warning response unit is used to output a Level I early warning response signal, which emits a red light and a high-decibel siren to warn of the danger. It also triggers a mobile APP alarm and a fire linkage module (activating emergency call, starting the sprinkler system, and cutting off the power), prompts "Serious risk, take immediate action", and pushes alarm information to the school's safety management department or fire department.

[0092] Specifically,

[0093] The cloud-based laboratory safety risk characteristic database records and stores laboratory safety risk characteristic data in the form of numbers, and records and stores the risk level of the corresponding laboratory safety risk characteristics in the form of risk values;

[0094] The cloud data communication unit is used to receive and transmit data and commands between devices, enabling the uploading and receiving of data and commands;

[0095] The communication interconnection unit enables interconnection between the intelligent safety early warning system and the laboratory intelligent management terminal, mobile APP, intelligent devices and cloud server through Bluetooth, Wi-Fi, GSM and ZigBee, achieving 24-hour remote monitoring and automatic handling.

[0096] Specifically, the image reporting module includes:

[0097] The image labeling unit is used to identify and mark the types and locations of hazards in the laboratory multidimensional data output by the multidimensional perception module. Hazards are identified according to their risk levels and marked with different colors to improve their recognizability.

[0098] The report export unit is used to export hazard sources, risk levels, and early warning data, and generate graphic and textual early warning response reports.

[0099] Specifically, the logging and optimization module includes:

[0100] The log recording and backup unit is used to record and back up historical monitoring data and early warning response reports in order to summarize early warning experience;

[0101] The iterative optimization unit is used to analyze historical monitoring data and historical early warning response plans, and regularly update and iterate the laboratory safety risk characteristic database to optimize the identification threshold of the intelligent judgment and decision-making module.

[0102] Specifically, the intelligent judgment and decision-making module integrates multi-dimensional laboratory data (such as gas concentration, temperature, humidity, and personnel behavior), sets weighting coefficients based on the risk values ​​and frequency of occurrence of different hazards, performs weighted fusion of multiple variables, and then dynamically determines the risk level based on the accumulated risk value. See below. Figure 2 As shown, specifically:

[0103] The weighting coefficients are dynamically adjusted based on historical accident data statistics from the laboratory, experimental time (nighttime experiments), and experimental type (high-temperature experiments vs. routine experiments), and satisfy the following conditions: , The weights are set;

[0104] Example of weight setting:

[0105] Assume the laboratory has n types of risk sources, each of which contains several sub-items as shown in the table below:

[0106] Risk Category Sub-item Weight (Custom settings) illustrate Environmental risks Formaldehyde concentration, temperature and humidity, smoke 0.20 Environmental risks can easily lead to accidents. Equipment risk Fume hood airflow velocity, power supply current 0.20 Equipment failure can lead to a chain reaction. Personnel risks Unprotected persons entering restricted areas 0.40 Human factors are the main cause of accidents, and key areas of detection are being addressed. Hazardous chemical risks Inventory discrepancies, storage temperatures exceeding limits 0.20 Poor management of hazardous chemicals poses high risks

[0107] The risk sub-items in the table above are for illustrative purposes only and may be added to or modified.

[0108] In the process of weighted fusion of multiple variables and dynamic judgment of risk level based on cumulative risk value, the risk value of each hazard source is first calculated, and the detection values ​​of different hazard sources are then combined. (e.g., formaldehyde concentration) and safety threshold Compare and combine with early warning thresholds (e.g., Level I risk threshold), calculate the normalized risk contribution value using the following formula:

[0109]

[0110] Next, calculate the total risk value of the hazard source. The formula is Then calculate the weighted sum of various risk values, Rtotal, using the following formula: Finally, risk assessment is conducted based on the established risk thresholds.

[0111] Where Ri is the original risk value of the i-th type of risk source (such as hazardous chemical leakage, equipment overheating), Wi is the dynamic weighting coefficient, and Rtotal is the comprehensive risk value. A level one warning is triggered when Rtotal ≥ 0.8, and a level two warning is triggered when 0.5 ≤ Rtotal < 0.8.

[0112] Example of specific risk level setting:

[0113] Assuming a four-level risk threshold (set according to actual circumstances), risk assessment is performed.

[0114] Level IV risk

[0115] The Level III risk threshold is A.

[0116] Level II risk threshold is B

[0117] The Level I risk threshold is C.

[0118]

[0119] That is, A ≤ R_total < B, Level III warning.

[0120] B ≤ R_total < C Level II Warning

[0121] R_total ≥ C Level I warning;

[0122] Specifically, the dynamic weighting coefficients can be determined through training with historical data, with an initial value of Wi=0.3, updated quarterly, and a first-level warning can be triggered when Rtotal≥0.8, and a second-level warning can be triggered when 0.5≤Rtotal<0.8.

[0123] The safety warning system for chemical laboratories based on multimodal perception provided in this embodiment includes the following steps:

[0124] The multi-dimensional perception module acquires multi-dimensional data from the laboratory in real time through multiple connected gas sensors, temperature and humidity sensors, monitoring cameras, equipment operation sensors, and RFID readers. This includes environmental data such as the concentration of toxic and harmful gases, temperature and humidity, open flames, smoke and air pressure, equipment operation data, personnel behavior data, and hazardous chemical status data. The acquired multi-dimensional data is then transmitted to the intelligent judgment and decision-making module.

[0125] The intelligent judgment and decision-making module identifies and judges risks based on multi-dimensional data, and outputs a comprehensive risk value and risk level.

[0126] The tiered early warning response module activates corresponding early warning measures based on the risk level, as follows:

[0127] Level III Warning: Issues a green light warning and sends a text warning to the laboratory's intelligent management terminal, indicating "Minor abnormality, inspection recommended";

[0128] Level II Warning: Issues a yellow light and sirens warning, sends a graphic warning (including a picture of the location of the hazard source) to the laboratory's intelligent management terminal and the laboratory personnel's mobile APP, and prompts "Moderate risk, handle it in time";

[0129] Level I Warning: Issues a red light and a high-decibel siren warning, simultaneously triggering a mobile APP alarm and fire linkage module (activating emergency call, sprinkler system activation, and power cut-off), displaying the message "Serious risk, take immediate action," and pushing alarm information to the school's safety management department or fire department;

[0130] The image reporting module receives monitoring and image data transmitted by the multi-dimensional perception module, as well as data from the intelligent judgment and decision-making module and the graded early warning response module. It identifies and marks the types and locations of hazards in the monitoring images, identifies the hazards according to the risk level (Level III, Level II, Level I) and marks them as green, yellow and red respectively to improve the recognition, and generates graphic early warning response reports.

[0131] The log backup and optimization module records and backs up historical monitoring data and early warning response reports, regularly updates and iterates the laboratory safety risk characteristic database, and optimizes the identification thresholds of the intelligent judgment and decision-making module.

[0132] The specific steps for determining risk sources and risk levels are as follows:

[0133] S1. Download the safety risk feature numbers and corresponding risk values ​​of laboratory environment, equipment operation, personnel behavior, and hazardous chemical status from the cloud-based laboratory safety risk feature database. Then, interpret and compare the safety risk feature numbers and corresponding risk values ​​according to the built-in numbering and risk value entries of the laboratory environment risk identification and judgment unit, equipment operation risk identification and judgment unit, personnel behavior risk identification and judgment unit, and hazardous chemical risk identification and judgment unit.

[0134] S2. Based on the laboratory environment risk identification and judgment unit, equipment operation risk identification and judgment unit, personnel behavior risk identification and judgment unit and hazardous chemical risk identification and judgment unit, the data output by the multi-dimensional perception module is identified and judged to determine the type of hazard source and output the corresponding risk value.

[0135] S3. If the risk source identification and judgment is successful once, output the risk source number and increase the corresponding risk value;

[0136] S4. Set Level III risk threshold A, Level II risk threshold B, and Level I risk threshold C.

[0137] S5. When A ≤ cumulative risk value < B, output a Level III risk warning signal and risk source type; when B ≤ cumulative risk value < C, output a Level II risk warning signal and risk source type; when the cumulative risk value ≥ C, output a Level I risk warning signal and risk source type.

[0138] Among them, S1 includes laboratory environmental safety risk characteristic numbers and corresponding risk values, including characteristic numbers and corresponding risk values ​​for toxic and harmful gases such as formaldehyde, toluene, chlorine, and VOCs, temperature and humidity, open flame, smoke and air pressure, etc.

[0139] Among them, S1 contains the equipment operation safety risk feature number and the corresponding risk value, including the feature number and corresponding risk value of equipment temperature / pressure, fume hood wind speed, power supply current, reagent cabinet temperature, etc.

[0140] Among them, S1 includes personnel behavior safety risk feature numbers and corresponding risk values, including feature numbers such as not wearing protective equipment as required, illegal operation, and entering restricted areas, and corresponding risk values ​​for different thresholds.

[0141] The hazardous chemical status safety risk feature number and corresponding risk value in S1 include the hazardous chemical inventory, the deviation of the hazardous chemical storage location, the hazardous chemical temperature, and other feature numbers and corresponding risk values ​​for different thresholds.

[0142] The early warning system in this embodiment is used for laboratory safety early warning. It has been verified that the false alarm rate is ≤3%, which is a significant reduction. When a risk occurs in the laboratory, the response speed is <3 seconds, and it can achieve "unattended operation + automatic handling", which can better meet the needs of laboratory safety early warning.

[0143] In summary, the intelligent safety early warning system for chemical laboratories based on multimodal perception provided by this invention integrates multimodal perception data from four dimensions: environment, equipment, personnel, and hazardous chemicals. Combined with intelligent database term comparison technology, it achieves rapid identification of risk sources and accurate determination of risk levels, significantly improving monitoring accuracy and solving the technical problems of high missed detection rates and frequent false alarms in traditional single-dimensional monitoring. This invention avoids management chaos caused by "one-size-fits-all" alarms through tiered early warning responses, greatly improving early warning accuracy and enabling early risk prediction and intervention, effectively preventing the further escalation of risks. Furthermore, this invention supports remote control and intelligent interconnection. Users can view the laboratory safety status anytime, anywhere through intelligent management terminals and mobile apps, achieving 24-hour remote monitoring. The system can be linked to other equipment, automatically triggering the fire protection system in case of severe risks, with a fast response speed, achieving "unmanned operation + automatic handling." Simultaneously, this invention can continuously update and iterate the safety risk characteristic database and judgment and identification thresholds by recording and analyzing historical data and response plans, optimizing response plans, continuously improving system accuracy, and ensuring long-term technical adaptability to future application scenarios.

[0144] The above are merely preferred embodiments of this application and are not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A chemical laboratory intelligent safety early warning system based on multimodal sensing, characterized in that, include: The multi-dimensional sensing module integrates laboratory environment sensing unit, equipment operation sensing unit, personnel behavior capture unit and hazardous chemical status sensing unit to collect multi-dimensional data of the laboratory in real time. The intelligent judgment and decision-making module includes a laboratory environment risk identification and judgment unit, an equipment operation risk identification and judgment unit, a personnel behavior risk identification and judgment unit, a hazardous chemical risk identification and judgment unit, a laboratory risk level judgment unit, and a laboratory safety risk identification feature download unit. It calculates a comprehensive risk value through a weighted multi-source risk fusion algorithm and compares it with a preset threshold to dynamically determine the risk source and risk level. The tiered early warning and response module is used to output corresponding measures based on the risk level. The cloud server communication module includes a laboratory safety risk characteristic database, a cloud data interoperability unit, and a communication interconnection unit. It is used to record and store laboratory safety risk characteristics, data upload / reception between devices, and ensure that the system can be linked and controlled to realize 24-hour remote monitoring and real-time automatic handling of the laboratory. The image reporting module is used to mark the types and locations of hazards in the laboratory's multi-dimensional data according to risk level, and to generate graphic and textual early warning response reports through the report export unit; The log backup and optimization module is used to record historical multi-dimensional laboratory data and graphic warning response reports stored in the backup unit, and to analyze historical data through iterative optimization units, and to regularly update the laboratory safety risk characteristic database and judgment thresholds.

2. The intelligent safety early warning system for chemical laboratories based on multimodal perception according to claim 1, characterized in that, The multi-dimensional sensing module is equipped with multiple gas sensors, temperature sensors, humidity sensors, infrared thermal imaging cameras, and RFID tags to monitor and collect multi-dimensional data from the laboratory in real time.

3. The intelligent safety early warning system for chemical laboratories based on multimodal perception according to claim 1, characterized in that, The integrated laboratory environment sensing unit is used to collect environmental data of the laboratory, including the concentration of toxic and harmful gases, temperature and humidity, open flame, smoke and air pressure data. The equipment operation sensing unit is used to collect data from laboratory equipment and report operational data such as laboratory equipment temperature / pressure, fume hood fan speed, power supply current, and reagent cabinet temperature. The personnel behavior capture unit is used to capture behavioral data of experimental personnel, including data on failure to wear protective equipment as required, violation of regulations, and trespassing into restricted areas. The hazardous chemical status sensing unit is used to collect hazardous chemical status data, including hazardous chemical inventory, hazardous chemical storage location, hazardous chemical temperature, and other hazardous chemical status data.

4. The intelligent safety early warning system for chemical laboratories based on multimodal perception according to claim 1, characterized in that, The laboratory environment risk identification and judgment unit is used to identify and judge the laboratory environment data output by the multi-dimensional sensing module based on the safety risk characteristics caused by the laboratory environment, and to determine the types of toxic and harmful gases, temperature and humidity, open flame, smoke and air pressure risks in the laboratory environment and the corresponding risk values. The equipment operation risk identification and judgment unit is used to identify and judge the equipment operation data output by the multi-dimensional perception module based on the safety risk characteristics caused by equipment operation, and to determine whether there are risks related to equipment temperature / pressure, fume hood wind speed, power current, reagent cabinet temperature and corresponding risk values ​​during equipment operation. The personnel behavior risk identification and judgment unit is used to identify and judge the personnel behavior data output by the multi-dimensional perception module based on the safety risk characteristics caused by personnel behavior, and to determine whether the personnel behavior has the risk of not wearing protective equipment as required, violating regulations, or entering restricted areas, as well as the corresponding risk value. The hazardous chemical risk identification and judgment unit is used to identify and judge the hazardous chemical status data output by the multi-dimensional perception module based on the safety risk characteristics caused by the status of hazardous chemicals, and to determine the inventory quantity, storage location, temperature status risk and corresponding risk value of hazardous chemicals. The laboratory risk level determination unit is used to statistically determine the cumulative risk values ​​of the laboratory environmental risk identification and determination unit, equipment operation risk identification and determination unit, personnel behavior risk identification and determination unit, and hazardous chemical risk identification and determination unit, to determine the corresponding risk level, and issue a warning signal of the corresponding level. The laboratory safety risk identification feature download unit is used to download laboratory risk identification feature data for comparison and verification of laboratory safety risk categories and risk values.

5. The intelligent safety early warning system for chemical laboratories based on multimodal perception according to claim 1, characterized in that, The intelligent judgment and decision-making module determines the risk levels as Level III, Level II, and Level I; the graded early warning response module includes Level III, Level II, and Level I early warning response units corresponding to the risk levels. Among them, the Level III early warning response unit is used to output a Level III early warning response signal, issue a green light warning, output a linkage signal, and send a text warning to the laboratory intelligent management terminal, prompting "minor abnormality, inspection recommended"; Among them, the Level II early warning response unit is used to output a Level II early warning response signal, issue a yellow light and sound siren warning, output a linkage signal, and send graphic and text warnings to the laboratory intelligent management terminal and the experimenter's mobile APP, prompting "moderate risk, handle in time"; The Level I early warning response unit is used to output a Level I early warning response signal, which emits a red light and a high-decibel siren to warn of the danger. It also triggers a mobile APP alarm and a fire linkage module (activating emergency call, starting the sprinkler system, and cutting off the power), prompts "Serious risk, take immediate action", and pushes alarm information to the school's safety management department or fire department.

6. The intelligent safety early warning system for chemical laboratories based on multimodal perception according to claim 1, characterized in that, The cloud-based laboratory safety risk characteristic database records and stores laboratory safety risk characteristic data in the form of numbers, and records and stores the risk level of the corresponding laboratory safety risk characteristics in the form of risk values; The cloud data communication unit is used to receive and transmit data and commands between devices, enabling the uploading and receiving of data and commands; The communication interconnection unit enables interconnection between the intelligent safety early warning system and the laboratory intelligent management terminal, mobile APP, intelligent devices and cloud server through Bluetooth, Wi-Fi, GSM and ZigBee, achieving 24-hour remote monitoring and automatic handling.

7. The intelligent safety early warning system for chemical laboratories based on multimodal perception according to claim 1, characterized in that, The image reporting module includes: The image labeling unit is used to identify and mark the types and locations of hazards in the laboratory multidimensional data output by the multidimensional perception module. Hazards are identified according to their risk levels and marked with different colors to improve their recognizability. The report export unit is used to export hazard sources, risk levels, and early warning data, and generate graphic and textual early warning response reports.

8. The intelligent safety early warning system for chemical laboratories based on multimodal perception according to claim 1, characterized in that, The logging and optimization module includes: The log recording and backup unit is used to record and back up historical monitoring data and early warning response reports in order to summarize early warning experience; The iterative optimization unit is used to analyze historical monitoring data and historical early warning response plans, and regularly update and iterate the laboratory safety risk characteristic database to optimize the identification threshold of the intelligent judgment and decision-making module.

9. The intelligent safety early warning system for chemical laboratories based on multimodal perception according to claim 1, characterized in that, The intelligent judgment and decision-making module integrates multi-dimensional laboratory data, sets weighting coefficients based on the risk values ​​and occurrence frequencies of different hazards, performs weighted fusion of multiple variables, and then dynamically determines the risk level based on the accumulated risk value. Specifically: The weighting coefficients are dynamically adjusted based on historical accident data statistics from the laboratory, experimental time, and experimental type, and satisfy the following conditions: , The weights are set; In the process of weighted fusion of multiple variables and dynamic judgment of risk level based on cumulative risk value, the risk value of each hazard source is first calculated, and the detection values ​​of different hazard sources are then combined. With safety threshold Compare and combine with early warning thresholds The normalized risk contribution value is calculated using the following formula: ； Next, calculate the total risk value of the hazard source. The formula is Then calculate the weighted sum of various risk values, Rtotal, using the following formula: Finally, risk assessment is conducted based on the established risk thresholds.

10. A method for intelligent safety early warning in chemical laboratories based on multimodal perception, characterized in that, Includes the following steps: The multi-dimensional sensing module collects real-time data on the laboratory environment, equipment, personnel behavior, and the status of hazardous chemicals. The intelligent judgment and decision-making module identifies and judges risks based on multi-dimensional data, and outputs a comprehensive risk value and risk level. The tiered early warning response module activates corresponding early warning measures based on the risk level. The image reporting module generates and exports hazard identification and early warning response reports; The log backup and optimization module records historical data and regularly optimizes the risk characteristic database and judgment thresholds.

Images