A Method and System for Laboratory Safety Status Assessment Based on Multi-Source Data Fusion

Abstract

This application discloses a method and system for assessing laboratory safety status based on multi-source data fusion, relating to the field of laboratory safety management. The method includes: collecting multi-source sensor data from the laboratory and aligning the data to a unified time axis to construct a laboratory sensor data matrix; introducing grey relational analysis to calculate the risk correlation between each sensor's data and preset hazard modes, thereby identifying potential risk modes; dynamically determining reliability weights based on the sensors' historical stability, baseline deviation, and calibration cycle to suppress data distortion caused by aging and drift; and comprehensively calculating a weighted risk energy by combining the risk correlation, reliability weights, and risk severity coefficients, and generating a laboratory risk field energy through spatiotemporal coupling to calculate the corresponding laboratory risk status level. This achieves dynamic quantification and hierarchical judgment of the overall laboratory safety status, improving the ability to identify risks in complex laboratory environments.

Description

Technical Field

[0001] This application relates to the field of laboratory safety management technology, and in particular to a method and system for assessing laboratory safety status based on multi-source data fusion. Background Technology

[0002] Modern laboratories, especially those in chemistry, biology, and materials science, typically contain multiple risk sources, including toxic and hazardous chemicals, flammable and explosive substances, precision instruments, and personnel activities. In recent years, with the development of the Internet of Things (IoT) and Industrial Internet technologies, an increasing number of hardware and software sensors have been applied to monitor industrial processes and laboratory environments, enabling real-time tracking of key variables to improve product quality and ensure process safety.

[0003] In chemical production processes, multiple sensors simultaneously monitor variables such as temperature, pressure, and component concentration. However, individual sensors are susceptible to errors due to various factors. Similar problems exist in laboratories as well. For example, temperature sensors are easily affected by airflow, gas sensors experience decreased sensitivity due to aging, and personnel may not record operations in a timely manner.

[0004] Currently, IoT-based laboratory hazard monitoring systems use sensors for gases, smoke, temperature, and humidity to detect risks in real time and issue alarms when thresholds are exceeded. However, these systems still largely rely on simple threshold judgments, and cannot reliably determine the current safety status of the laboratory based on a single data source. They also lack in-depth analysis of the correlation and relative reliability between multi-source sensor data. Summary of the Invention

[0005] This application provides a laboratory safety status assessment method, system, storage medium, computer program product, and electronic device based on multi-source data fusion, which at least solves the problems in current related technologies where laboratory safety monitoring relies on single sensor data, has low accuracy in risk assessment, and cannot fully reflect the state of complex environments.

[0006] In a first aspect, embodiments of this application provide a laboratory safety status assessment method based on multi-source data fusion. The method includes: collecting multi-source sensor data corresponding to the laboratory environment based on sensor modules, and constructing a laboratory sensor data matrix under a unified time axis based on the multi-source sensor data; processing the laboratory sensor data matrix using grey relational analysis to calculate the risk correlation degree corresponding to each sensor; the risk correlation degree is used to indicate the degree of proximity between the current sensor sampling characteristics and preset hazard mode characteristics; obtaining baseline deviation information corresponding to each sensor, and determining the corresponding reliability weight based on the historical stability, baseline deviation information, and calibration cycle of each sensor; the historical stability is defined as the standard deviation of the data collected by the sensor under historical normal conditions, and the baseline deviation information is defined as the degree of deviation of the current measurement value of the sensor relative to the baseline mean of the time period; multiplying the risk correlation degree of each sensor by the reliability weight and the risk severity coefficient to obtain the weighted risk energy of each sensor; each sensor has a unique corresponding risk severity coefficient; spatiotemporally coupling and superimposing the weighted risk energies of all sensors to generate laboratory risk field energy, and calculating the corresponding laboratory risk status level based on the laboratory risk field energy.

[0007] Secondly, embodiments of this application provide a laboratory safety status assessment system based on multi-source data fusion. The system includes: a data acquisition unit, used to acquire multi-source sensor data corresponding to the laboratory environment based on sensor modules, and construct a laboratory sensor data matrix under a unified time axis based on the multi-source sensor data; a risk analysis unit, used to process the laboratory sensor data matrix using grey relational analysis to calculate the risk correlation degree corresponding to each sensor; the risk correlation degree is used to indicate the degree of proximity between the current sensor sampling characteristics and preset hazard mode characteristics; and a sensor reliability analysis unit, used to obtain baseline deviation information corresponding to each sensor and, based on the historical stability of each sensor... The system uses the following parameters to determine the reliability weights: reliability, baseline deviation information, and calibration cycle; historical stability is defined as the standard deviation of data collected by the sensor under historical normal conditions, and baseline deviation information is defined as the degree of deviation of the sensor's current measurement value from the baseline mean of its time period; a sensor risk energy analysis unit is used to multiply the risk correlation of each sensor by the reliability weight and risk severity coefficient to obtain the weighted risk energy of each sensor; each sensor has a unique corresponding risk severity coefficient; a laboratory risk field energy generation unit is used to spatiotemporally couple and superimpose the weighted risk energies of all sensors to generate laboratory risk field energy, and calculate the corresponding laboratory risk status level based on the laboratory risk field energy.

[0008] Thirdly, an electronic device is provided, comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the steps of the laboratory safety status assessment method based on multi-source data fusion according to any embodiment of this application.

[0009] Fourthly, embodiments of this application provide a storage medium storing a computer program thereon, characterized in that, when the program is executed by a processor, it implements the steps of the laboratory safety status assessment method based on multi-source data fusion according to any embodiment of this application.

[0010] Fifthly, embodiments of this application provide a computer program product, including a computer program / instructions, which, when executed by a processor, implement the steps of the laboratory safety status assessment method based on multi-source data fusion according to any embodiment of this application.

[0011] The laboratory safety status assessment method and system based on multi-source data fusion provided in this application can achieve at least the following technical effects:

[0012] (1) By constructing a multi-source sensor data matrix under a unified time axis, the collaborative expression of environmental, equipment, and personnel behavior data was realized, enabling the dynamic relationship between different physical quantities to be quantitatively presented within the same time-series framework. The grey relational analysis method was introduced to calculate the relative change trends among multi-dimensional sensor features, obtaining the correlation degree between each sensor and potential hazard patterns. This allows the system to move beyond relying on a single threshold trigger and instead make comprehensive judgments based on multi-dimensional trend coupling, thereby capturing potential abnormal situations before risks become apparent and achieving a shift from passive alarm to proactive prediction. This not only improves the sensitivity and foresight of risk identification but also significantly enhances the accuracy and continuity of the overall laboratory safety status assessment.

[0013] (2) By comprehensively considering the historical stability of sensors, baseline deviation information, and calibration cycle, the reliability weight is dynamically calculated, enabling different sensors to have differentiated influence when contributing risk information. The introduction of a weighting mechanism allows the system to automatically reduce the interference of drifting, aging, or signal-abnormal sensors, maintaining a stable response of the assessment model to actual risk changes. In addition, the risk correlation, reliability weight, and risk severity coefficient are fused and calculated to form a weighted risk energy, which is then spatiotemporally superimposed in the spatial dimension to construct the risk energy distribution field of the laboratory. Through this energy field model, risk information is transformed into visualized and quantifiable continuous distribution characteristics, enabling intuitive identification of local high-risk areas and achieving spatial analysis and refined management of the laboratory's risk structure.

[0014] This technical solution organically combines multi-source heterogeneous sensor data fusion, trend correlation analysis, and reliability weighting mechanisms to establish an evolvable risk energy field expression mechanism. This overcomes the discrete and single-point limitations of traditional monitoring methods, enabling laboratory safety assessment to form a closed loop from real-time fusion of multi-dimensional data and trend perception to spatial energy mapping. Consequently, it achieves dynamic quantification and hierarchical judgment of the overall safety status of the laboratory, significantly improving the ability to identify risks in complex laboratory environments. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 A flowchart illustrating an example of a laboratory safety status assessment method based on multi-source data fusion according to an embodiment of this application is shown;

[0017] Figure 2 A flowchart illustrating an example of calculating risk correlation degree according to an embodiment of this application is shown;

[0018] Figure 3 A flowchart illustrating an example of calculating the reliability weights of a sensor according to an embodiment of this application is shown.

[0019] Figure 4 A flowchart illustrating an example of generating laboratory risk field energy according to an embodiment of this application is shown.

[0020] Figure 5 A flowchart illustrating an example of determining a laboratory risk status level based on laboratory risk field energy according to an embodiment of this application is shown.

[0021] Figure 6 The comparison results of risk score curves of different algorithms changing over time in a simulation environment are shown;

[0022] Figure 7 The performance comparison results of different algorithms under the main risk assessment indicators are shown;

[0023] Figure 8 A structural block diagram of an example of a laboratory safety status assessment system based on multi-source data fusion according to an embodiment of this application is shown. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0025] It should be noted that multi-sensor data fusion technology has been widely studied in fields such as remote sensing imagery, aerospace, and industrial process monitoring. Algorithms can be categorized into four levels based on their fusion hierarchy: signal-level, pixel-level, feature-level, and decision-level. Signal-level fusion directly combines the raw signals from different sensors to improve the signal-to-noise ratio; pixel-level fusion fuses multi-spectral information at the pixel level of the image; feature-level fusion extracts and fuses features from different source data; and decision-level fusion synthesizes the decisions made independently by each sensor. However, these classic methods are mostly designed for visual images and lack consideration for multimodal data (such as time series, event logs, and equipment status) in laboratory environments. Furthermore, standard image fusion algorithms require high registration accuracy of the input data and are prone to introducing spectral distortion.

[0026] In recent years, some scholars and researchers have proposed IoT-based laboratory safety systems, such as the paper "Laboratory Safety System using IoT," which uses flame, gas, and personnel detectors to build a smart laboratory that can send alarms to the person in charge when measured parameters exceed set values. However, this system mainly addresses situations where a single sensor exceeds its limit, without considering the superposition of multiple hazards or the correlation between sensor data, and it does not provide a quantitative assessment model for the degree of risk. Another study, "IoT-Based Real-Time Monitoring System for Laboratory Hazards," proposes using sensors such as gas, smoke, temperature, and humidity sensors combined with a cloud platform for real-time monitoring and notification. This paper acknowledges that traditional manual supervision methods are inefficient and error-prone, but its data processing still mainly relies on setting thresholds, lacking modeling of differences in sensor data quality, and failing to consider the impact of daily laboratory activity patterns on thresholds.

[0027] Furthermore, most theoretical research in multi-sensor fusion in academia is represented by the Dempster–Shafer (DS) evidence theory, which allows the fusion of multiple uncertain pieces of information and the resulting confidence distribution. However, research indicates that when conflicting evidence is caused by sensor malfunctions, traditional improvement methods struggle to remove erroneous information, leading to unreliable combined results in high-conflict scenarios. Laboratory environment sensors may suffer from aging, drift, and other malfunctions; without differentiating sensor reliability, relying solely on the DS algorithm may result in misjudgments or failure to identify anomalies.

[0028] More specifically, the system proposed in "IoT-Based Real-Time Monitoring System for Laboratory Hazards" uses gas, smoke, temperature, and humidity sensors to collect environmental data. This data is then transmitted in real-time to a cloud platform via ESP32 or NodeMCU, and alerts are sent to mobile applications or emails when thresholds are exceeded. The paper points out that traditional manual inspections are inefficient and prone to missing detections, and that continuous data collection and historical data logs can improve maintenance strategies. However, the system's fusion algorithm remains a simple threshold fusion. Although the paper mentions that "the mathematical model supports hazardous threshold setting, sensor data fusion, anomaly detection, probability estimation, etc.," it does not provide specific algorithms or methods for handling data quality. Experiments show that this fixed-threshold-based system has some responsiveness to sudden events, but its tolerance for complex multi-factor scenarios, sensor failures, and interference is weak.

[0029] For chemical process monitoring, the paper "Monitoring Chemical Processes Using Judicious Fusion of Multi-Rate Sensor Data" proposes a multi-rate sensor fusion scheme based on Bayesian inference to improve the accuracy and reliability of process monitoring. The paper points out that laboratory analyzers have low sampling frequencies but high accuracy, while online analyzers have high sampling frequencies but limited accuracy. This study combines measurements from laboratory analyzers, online analyzers, and soft sensors, using Kalman filtering to achieve state estimation. While this method effectively fuses multi-rate data, it assumes that the sampling rates of each sensor are fixed and reliable, whereas the sampling frequency and accuracy of actual laboratory sensors vary with environmental changes. Furthermore, the fusion algorithm primarily addresses variable estimation and does not consider risk assessment and alarm requirements. Additionally, when sensors malfunction or drift, the Bayesian method struggles to detect and remove abnormal data, potentially leading to erroneous estimates.

[0030] DS evidence theory has advantages in solving multi-sensor data fusion problems, handling uncertain and incomplete information. However, researchers have found that standard DS fusion yields unreasonable results in cases of strong conflicts caused by sensor failures. To overcome this problem, scholars have proposed improved combination rules or preprocessing methods, such as weighted averaging based on evidence distance or entropy weighting. However, these improvements mainly weight conflicts at the combination rule level, rarely considering the interaction structure between multi-source data and lacking business implications specific to the scenario. In laboratory security scenarios, sensor failures often manifest as continuous offsets or signal jamming, situations that simple DS fusion struggles to identify.

[0031] It should be understood that the above description of the relevant technologies is intended only to help the public better understand the inventive spirit and motivation of this application, and is not intended to limit this application. Furthermore, the technical solutions described in the above-mentioned relevant technologies are not prior art, and may also be undisclosed technical solutions, such as those under research or in the laboratory stage.

[0032] The technical solutions in this application, including the collection, storage, use, processing, transmission, provision, and disclosure of users' personal information, comply with relevant laws and regulations and do not violate public order and good morals.

[0033] Figure 1 A flowchart illustrating an example of a laboratory safety status assessment method based on multi-source data fusion according to an embodiment of this application is shown.

[0034] Regarding the execution subject of the method in the embodiments of this application, it can be any controller or processor with computing or processing capabilities, such as a laboratory risk early warning platform controller, which transforms laboratory safety assessment from a single threshold judgment to a multi-dimensional and interpretable comprehensive assessment system, realizing collaborative judgment of multi-source information and dynamic risk perception of complex laboratory environments.

[0035] In some examples, it may be integrated into an electronic device or terminal through software, hardware, or a combination of both, and the type of terminal or electronic device may be diverse.

[0036] like Figure 1 As shown, in step S110, multi-source sensor data corresponding to the laboratory environment is collected based on the sensor module, and a laboratory sensor data matrix under a unified time axis is constructed based on the multi-source sensor data.

[0037] In some implementations, multi-source data is collected synchronously by sensor modules deployed in the laboratory environment. These sensor modules include environmental sensors, equipment status sensors, and personnel behavior sensors. Specifically, environmental sensors (such as temperature, humidity, gas concentration, smoke, and illuminance sensors) are used to collect the macroscopic environmental conditions of the laboratory; equipment status sensors (such as current, voltage, vibration, and pressure sensors) are used to monitor the operating status and energy consumption characteristics of key instruments; and personnel behavior sensors (such as infrared sensors, ultrasonic ranging, access control recording, or video behavior analysis nodes) are used to perceive the spatial distribution and frequency of operational behaviors.

[0038] Furthermore, due to differences in sampling frequencies and clock skew among different sensors, the system uses timestamp synchronization and interpolation resampling algorithms to align various data streams onto a unified time reference, forming a multi-dimensional time series matrix. For example, the laboratory sensor data matrix maintains equal-interval sampling in the time dimension and includes multi-source sensor variables in the feature dimension, achieving a temporal representation of the entire laboratory state and avoiding misjudgments caused by sampling delays or asynchrony.

[0039] In some examples of embodiments of this application, data preprocessing operations can be performed before constructing the laboratory sensor data matrix. Specifically, time synchronization and interpolation are performed on sensor data with different sampling frequencies. It should be noted that the sampling period of environmental temperature and humidity sensors can reach several seconds, while electrical power monitoring or infrared motion detection modules may sample at millisecond-level frequencies. Inconsistent sampling frequencies will cause alignment deviations in the time dimension of multi-source data. In some implementations, the original sampling timestamp sequences of various sensors can be obtained, and a unified time axis can be constructed using integer multiples of the global minimum sampling period as the time reference for the laboratory sensor data matrix, ensuring consistency in time scale.

[0040] In addition, a sliding window filtering algorithm is used to smooth the high-frequency sensor data to remove transient noise interference. It should be noted that during the sampling process of high-frequency sensors (such as current, voltage, or noise intensity sensors), they are often affected by transient interference, electromagnetic noise, or human-induced disturbances, leading to spikes or abrupt changes in the measurement data. In some implementations, the sensor data stream can be scanned in a sliding window manner on the time axis, and mean filtering, median filtering, or weighted moving average algorithms can be applied to the data sequence within the window to smooth local fluctuations and effectively reduce high-frequency noise caused by short-term disturbances.

[0041] Furthermore, when missing sensor data is detected, statistical reconstruction is performed using observations from adjacent time periods and related sensors to fill in the missing samples, thereby ensuring the continuity and integrity of the laboratory sensor data matrix under a unified time axis. It should be noted that laboratory monitoring equipment may experience temporary communication interruptions, sensor aging, or network packet loss, leading to missing samples in the data stream. In some implementations, missing sample points are detected based on a unified time axis, and the missing locations are marked; then, statistical reconstruction is performed using observations from similar or spatially adjacent sensors. For example, when data from a certain temperature sensor is missing, compensation can be made using the average trend and spatial correlation of other temperature sensors in the same area, thereby ensuring the temporal continuity and spatial integrity of the laboratory sensor data matrix.

[0042] In step S120, the laboratory sensor data matrix is ​​processed by grey relational analysis to calculate the risk correlation degree corresponding to each sensor. The risk correlation degree is used to indicate the degree of proximity between the current sensor sampling characteristics and the preset hazard mode characteristics.

[0043] In some implementations, after obtaining a multi-source data matrix with a unified time axis, the system invokes Grey Relational Analysis (GRA) to perform correlation calculations on the laboratory's multidimensional time-series characteristic sequences. Specifically, a preset hazard mode feature is used as a reference sequence, representing the characteristic change trajectory under typical hazardous conditions (e.g., a linkage pattern of temperature increase-gas concentration increase-current fluctuation); the real-time sampling sequences of each sensor are used as comparison sequences, and the relative similarity between the two is calculated based on the GRA model, thereby capturing the trend similarity with the preset hazard mode feature. Grey Relational Analysis is used to analyze the laboratory sensor data matrix to quantitatively measure the similarity between the data change trajectory of each sensor and the known hazard mode feature, thereby obtaining the risk correlation degree of each sensor.

[0044] It should be noted that grey relational analysis is robust to small samples and uncertain data, and can adapt to situations in laboratory settings where data is incomplete, noise interference, and local anomalies exist.

[0045] In this embodiment of the application, grey relational analysis is used to identify the proximity between the current environmental state and known hazard patterns, instead of simply relying on numerical threshold triggering mechanisms. This enables the risk assessment to have the ability to recognize patterns and predict trends, and can still stably reflect the overall risk change trend even when some data is abnormal or missing.

[0046] In step S130, baseline deviation information corresponding to each sensor is obtained, and the corresponding reliability weight is determined based on the historical stability, baseline deviation information and calibration cycle of each sensor.

[0047] It should be noted that different sensors are affected by environmental drift, component aging, maintenance delays, and other factors during long-term operation, leading to inconsistencies in the reliability of measurement data. To prevent data from low-reliability sensors from excessively influencing the overall evaluation results, the reliability weight of each sensor is dynamically determined by comprehensively calculating its historical stability, current baseline deviation information, and calibration cycle.

[0048] Here, historical stability is defined as the standard deviation of data collected by the sensor under historical normal conditions, and baseline deviation is defined as the degree of deviation of the sensor's current measurement value relative to the baseline mean of its time period. In some examples, the historical stability factor can be calculated by analyzing the standard deviation of the sensor's output under historical normal conditions, reflecting the degree of signal fluctuation; a smaller standard deviation indicates higher stability. The baseline deviation factor can be the degree of deviation of the current measurement value relative to the baseline mean of the same time period; a larger deviation indicates that the sensor's condition may be abnormal. The calibration cycle factor can adjust the reliability based on the time elapsed since the sensor's last calibration, while devices that have not been calibrated for a long time automatically have their corresponding reliability weight reduced. Furthermore, by integrating the above three factors, for example through nonlinear weighting functions or neural network processing, the reliability weight of each sensor can be dynamically determined, enabling it to adaptively reflect changes in the sensor's health status.

[0049] Therefore, by introducing a dynamic reliability mechanism, the impact of abnormal sensors can be reduced, preventing a single abnormal sensor from causing overall judgment distortion, maintaining the stability and accuracy of the risk analysis model during long-term operation, and possessing adaptive suppression capabilities against sensor drift, aging, and noise interference.

[0050] In step S140, the risk correlation degree of each sensor is multiplied by the reliability weight and the risk severity coefficient to obtain the weighted risk energy of each sensor.

[0051] Here, each sensor has a unique risk severity coefficient, which represents the potential hazard level that may be caused if the physical quantity monitored by the sensor is abnormal. For example, the severity coefficient of a combustible gas leak sensor is higher than that of a temperature and humidity sensor.

[0052] Specifically, by multiplying and fusing the risk correlation, reliability weight, and risk severity coefficient of each sensor, a corresponding weighted risk energy is obtained. This achieves weighted fusion of multi-source risk information under a unified measurement system, making the contributions of different types of risks comparable. Furthermore, this calculation process can employ time window aggregation, smoothing the weighted risk energy through a sliding time window to filter out transient interference.

[0053] By calculating weighted risk energy, the system has been upgraded from qualitative alarms to quantitative risks, enabling it to simultaneously measure the intensity and credibility of multiple risk sources. Furthermore, by utilizing a unified energy scale, risk analysis has shifted from decentralized monitoring of heterogeneous indicators to centralized assessment.

[0054] In step S150, the weighted risk energies of all sensors are spatiotemporally coupled and superimposed to generate laboratory risk field energy, and the corresponding laboratory risk status level is calculated based on the laboratory risk field energy.

[0055] In some implementations, based on laboratory spatial layout information (such as sensor coordinates and equipment area division) and sampling time series, interpolation mapping and convolution smoothing are performed on each weighted risk energy to form a two-dimensional or three-dimensional risk field distribution map. Subsequently, by calculating the total energy or energy gradient of this risk field within a given time window, the overall risk intensity index of the laboratory is extracted. Furthermore, according to preset grading standards (such as quantile thresholds or adaptive thresholds based on historical data), the risk field energy is converted into a risk status level (such as "safe," "warning," "high risk," etc.) and output to the monitoring interface or upper control system.

[0056] Here, the weighted risk energies of all sensors are spatiotemporally superimposed based on their spatial coordinate information to generate the risk field energy distribution of the laboratory. For example, using the laboratory's two-dimensional plane or three-dimensional structure as a reference space, a thermal diffusion model or Gaussian kernel function can be used to diffuse and superimpose the risk energy of each sensor within its neighborhood, forming a continuous energy distribution field. This risk field spatially reflects the risk distribution density and is dynamically updated over time through a sliding window. Furthermore, based on the total energy of the risk field and the changes in local energy gradients, the laboratory's risk status level can be classified to quantify the overall safety situation.

[0057] As a preferred embodiment of this application, a two-dimensional plane can be used as the laboratory reference space to reduce system resource consumption and improve the timeliness of status monitoring. Specifically, although the laboratory space is essentially three-dimensional, considering that the sensor distribution characteristics and risk diffusion characteristics of most laboratories are mainly concentrated on the same working plane (such as the ground floor or operating table plane), various environmental sensors (temperature and humidity, gas concentration), equipment status sensors, and personnel behavior sensors are mostly installed within a similar height range. The differences in their data changes in the vertical direction (Z-axis) are much smaller than the differences in their distribution in the plane (XY direction). Therefore, using a two-dimensional plane as the reference space for calculation is more engineering-reasonable and computationally stable when modeling the risk field. The heat diffusion model or Gaussian kernel function established based on the two-dimensional reference space can fully describe the diffusion and superposition process of risk energy in the laboratory plane, realize efficient spatial coupling calculation, and significantly reduce modeling complexity and real-time computing load.

[0058] Figure 2 A flowchart illustrating an example of calculating risk correlation degree according to an embodiment of this application is shown.

[0059] like Figure 2 As shown, in step S210, for each sensor, a hazard reference feature sequence corresponding to the hazard mode is defined.

[0060] In laboratory safety monitoring scenarios, to measure the proximity of the current sampling state of each sensor to a potential hazardous state, it is necessary to establish a reference characteristic sequence that can represent typical hazard modes in the laboratory. .

[0061] In some implementations, based on historical safety incidents or records of abnormal laboratory operations, time-series data of several key indicators (such as the rate of temperature increase, decrease in oxygen concentration, sudden increase in equipment power, increase in smoke concentration, etc.) are extracted from the database. Sensor data from these abnormal periods are then normalized and clustered to extract common temporal trends. Finally, the pattern sequence with the most significant trend is used as the reference feature sequence for hazard patterns. This sequence can reflect the multidimensional temporal characteristics of typical danger evolution processes.

[0062] Thus, a unified time-series benchmark for risk assessment was established, enabling data from different sensors to be compared within the same feature space. Furthermore, by introducing a hazard pattern reference sequence, grey relational analysis no longer relies on absolute value differences but instead judges risk similarity based on changing trends, thereby enhancing the sensitivity and universality of anomaly identification.

[0063] In step S220, the grey correlation coefficient between the current sensor sampling sequence indicated by the laboratory sensor data matrix and the hazard reference feature sequence is calculated.

[0064] Here, to characterize the similarity between each sensor's sampling sequence and the hazardous reference sequence, the grey relational coefficient model from grey system theory is used for calculation. For each sensor in the laboratory sensor data matrix... Calculate the temporal difference between the sensor sampling sequence and the hazard mode sequence within a sliding time window. Therefore, the grey relational coefficient formula is adopted:

[0065] Equation (1)

[0066] The numerator of the right-hand side of equation (1) is a weighted combination of the global minimum and maximum differences, and the denominator is the actual difference between the current sensor and the reference mode. The smaller the value, the closer the sensor sampling trend is to the dangerous reference sequence, the corresponding... The closer the value is to 1, the better; conversely, if the difference is large, The value is close to 0. Therefore, it can effectively reflect the dynamic similarity of different sensors over time series.

[0067] Unlike traditional correlation analysis based on mean squared error, the grey relational analysis method does not require assumptions about data distribution characteristics and can handle common nonlinear, short-sequence, and small-sample data in laboratories. This allows the system to identify potential risk trends in the early stages, rather than only issuing an alarm after the anomaly has fully manifested, thus significantly improving the foresight of early warning.

[0068] In step S230, the mean value of the grey relational coefficient is calculated within the time window to obtain the risk relational degree.

[0069] Because the laboratory environment is highly volatile, transient anomalies may originate from temporary disturbances (such as opening doors for ventilation, starting up equipment, etc.). Therefore, by smoothing the risk characteristics within a time window, a stable risk measurement can be obtained.

[0070] Specifically, in the sliding time window (length is Within this range, the mean value of the grey relational coefficient for each sensor is calculated:

[0071] Equation (2)

[0072] In the formula, and For sensor indexing, , This represents the total number of sensors in the sensor module. and For discrete-time indexes within a time window, where and , The window length is the time window length. and These are the reference sequences for the dangerous mode at time [time]. and time The value, For the first Each sensor at time The sampled values, For the resolution coefficient, satisfying ; For the first Each sensor at time The sampled values; For the first Each sensor at time The grey relational coefficient; For the first Risk correlation of individual sensors.

[0073] In Equation (2), through the mean calculation process, short-term fluctuation signals (noise) are smoothed, while persistent deviation trends are amplified, effectively filtering out the interference of instantaneous disturbances on risk assessment, and retaining risk information that reflects the systematic change trend.

[0074] In some implementations, the window length can be dynamically adjusted based on the sensor sampling frequency and laboratory response characteristics: for sensors with fast response speeds (such as electrical equipment current and voltage), a shorter window is used to quickly capture risks; for environmental variables (such as temperature, humidity, and gas concentration), a longer window is used to stabilize the calculation results.

[0075] In some examples of embodiments of this application, the hazard reference feature sequence can also be used to describe the semantics of hazardous experimental situations, enabling the system to possess situational awareness capabilities. Specifically, the hazard reference feature sequence not only serves as a benchmark for numerical feature patterns but is also extended into a hazard situation vector model with semantic hierarchical descriptive capabilities. During the training phase, historical laboratory accident reports, sensor anomaly data, and operation logs are jointly encoded to abstract hazardous events into semantic labels (such as "solvent leak," "high temperature overload," "combustible gas diffusion," etc.) and mapped to the feature space in a multimodal manner. Thus, the hazard reference feature sequence can accommodate both numerical trends and semantic contextual information, expanding the risk assessment process from simple feature similarity to situational understanding and reasoning.

[0076] As a further preferred implementation, when a high similarity is detected between the current sensor state and a certain semantic hazard pattern, the system not only outputs a risk correlation score but also triggers a scenario inference operation. Specifically, by using predefined causal relationship chains in the knowledge graph, such as "solvent leakage → vapor accumulation → electric spark ignition → deflagration risk," the system automatically infers potential chain reaction processes. During this process, by dynamically querying the status information of other sensor nodes (such as temperature, humidity, current, smoke concentration, etc.), the system verifies whether there are any signs consistent with the inferred path, thereby improving the accuracy of predicting complex risks.

[0077] For example, when a hazardous mode of flammable solvent leakage is detected, the system will automatically deduce potential secondary risk sources and check for high-temperature equipment or electrical devices in the vicinity. If such potential triggering sources are detected, the system can further estimate the vapor cloud diffusion range based on airflow models and Gaussian diffusion functions, and mark the affected area in real time on the laboratory floor plan. Administrators can then visually observe the risk propagation trend on the interface and take preventative measures such as isolation, cooling, or power outages.

[0078] By introducing the aforementioned semantic inference and multi-layer perception mechanisms, the function of the hazard reference feature sequence is transformed from passive pattern matching to proactive risk evolution prediction. This not only enhances the system's contextual awareness and adaptive decision-making capabilities but also achieves a leap from "event detection" to "risk reasoning." Ultimately, the laboratory safety monitoring system can identify risk sources, predict risk chains, and locate key triggering nodes at an early stage, thereby achieving more intelligent laboratory safety control.

[0079] It should be noted that the risk correlation values ​​of each sensor... This reflects the degree of proximity between the monitoring dimensions represented by the sensor (temperature, humidity, gas, equipment power, behavioral activity, etc.) and potential hazard patterns. Through risk correlation, multi-source heterogeneous sensor data can be fused within the same risk measurement framework, enabling laboratory safety assessments to possess quantitative, dynamic, and intelligent characteristics. In some preferred embodiments, for high-risk sensor areas (i.e.... If a value remains high for an extended period, the system can automatically mark it as a key monitoring point, triggering enhanced sampling and local anomaly tracking mechanisms.

[0080] Figure 3 A flowchart illustrating an example of calculating the reliability weights of a sensor according to an embodiment of this application is shown.

[0081] like Figure 3 As shown, in step S310, the standardized deviation of each sensor is calculated.

[0082] During laboratory operations, the physical quantities collected by different sensors exhibit differences in dimensions and distribution. To achieve a unified risk assessment standard across sensors, it is necessary to standardize the real-time sampling values ​​of each sensor.

[0083] Specifically, for each sensor Based on historical normal operation data, a corresponding baseline period statistical model is established, including the baseline mean. and baseline standard deviation Collect the sensor's measurement value at the current moment. Calculate its standardized deviation relative to the baseline:

[0084] Equation (3)

[0085] In the formula, For the first The current measurement value of each sensor, For the first The baseline mean of each sensor during its respective time period. For the first The baseline standard deviation of each sensor over its respective time period; For the first The standardized deviation of each sensor represents baseline deviation information.

[0086] Specifically, when A large deviation indicates a significant deviance between the current measurement and the normal baseline, which is considered a potential sign of sensor anomaly. Introducing standardized deviation allows for comparison of the degree of anomaly in different physical quantities (such as temperature, humidity, current, vibration, etc.) on a uniform scale. Simultaneously, it introduces... This reflects the historical stability of the sensor, allowing noisier sensors to automatically reduce their sensitivity and improve the robustness of risk analysis.

[0087] In step S320, the historical stability of each sensor is calculated.

[0088] Specifically, the standard deviation of each sensor during the stable operation phase is calculated in advance. and with a certain reference standard deviation As a benchmark, the basic reliability is calculated using an exponential decay method:

[0089] Equation (4)

[0090] In the formula, For the first The basic reliability of a sensor represents its historical stability. For the first The stability and sensitivity coefficient of each sensor This is a reference stability constant.

[0091] Specifically, the smaller the historical fluctuations (i.e.) The lower the value), the corresponding The closer the value is to 1, the more stable and reliable the sensor is in the long term; for sensors with high noise or long-term instability, its... This significantly reduces the risk and enables adaptive hierarchical adjustment of reliability, effectively suppressing the misleading influence of abnormal sensors on the overall risk assessment.

[0092] In step S330, the calibration cycle attenuation factor of each sensor is calculated.

[0093] After long-term operation, the measurement accuracy of the sensor will gradually deviate from the factory standard. Therefore, the system should consider the impact of the sensor's last calibration time and calibration cycle on reliability.

[0094] Specifically, record the moment when each sensor last completed calibration. and its nominal calibration cycle Then, the calibration overdue ratio at the current moment is calculated:

[0095] Equation (5)

[0096] Equation (6)

[0097] In the formula, Indicates the current assessment time. For the first The calibration cycle of each sensor, For the first Calibration overdue ratio of individual sensors For the first The most recent calibration completion time for each sensor To calibrate the overdue sensitivity coefficient, Indicates the first Calibration cycle attenuation factor for each sensor.

[0098] In equation (5), when When this occurs, it indicates that the sensor is within its effective calibration period and the calibration attenuation factor is equal to 1; when When the sensor has exceeded its nominal calibration period, its reliability declines exponentially. Therefore, ensuring the system automatically identifies expired sensors and reduces their weight improves the overall system's long-term reliability and security.

[0099] In step S340, the sensor reliability weight of the corresponding sensor is determined by combining the deviation threshold and the modulation coefficient.

[0100] In some implementations, to comprehensively consider the current deviation, historical stability, and the impact of the calibration cycle, these three factors are coupled through a nonlinear function to obtain the final reliability weight. Specifically, a deviation threshold is introduced. This is used to distinguish between slight fluctuations and significant anomalies, while the standardized deviation... At that time, reliability will rapidly decline:

[0101] Equation (7)

[0102] In the formula, For deviation from the threshold, This is the attenuation modulation coefficient, used to control the sensitivity of weight descent; For the first Reliability weights for each sensor.

[0103] when When the sensor is considered to be operating within the normal range, the reliability weight remains at a high level. This weighting mechanism enables dynamic control based on a three-dimensional perspective of sensor physical deviation, historical fluctuations, and time aging. It can automatically suppress the negative impact of abnormal, drifting, and expired sensors on the system risk analysis results.

[0104] Figure 4 A flowchart illustrating an example of generating laboratory risk field energy according to an embodiment of this application is shown.

[0105] like Figure 4 As shown, in step S410, for each sensor within the time window The weighted risk energy within is integrated to obtain the time aggregated energy of a single sensor.

[0106] In a laboratory environment, the instantaneous weighted risk energy collected by various sensors (such as temperature and humidity sensors, gas concentration sensors, power sensors, smoke sensors, etc.) changes dynamically over time. To reflect the accumulation and sustained effect of risk energy over time, it is necessary to perform integral calculations on it within a certain time window to obtain the time-aggregated energy of a single sensor.

[0107] In some implementations, the system at each time step Push forward a length of time window This ensures that energy calculations include both historical risk evolution information and retain the response characteristics at the current moment. Specifically, for the first... Instantaneous weighted risk energy of individual sensors Integrate within the time window:

[0108] Equation (8)

[0109] Equation (9)

[0110] In the formula, Represents the time integral variable; Indicates the first Each sensor at time Instantaneous weighted risk energy, Indicates the first Risk severity coefficient of each sensor Indicates the first Each sensor at time Time aggregates energy.

[0111] Here, the time integration process can be implemented through sliding window accumulation to improve the real-time responsiveness of the system. The time-aggregated energy comprehensively considers risk trends, signal credibility, and risk severity, realizing dynamic energy accumulation in the time-series dimension; through integration calculation, it can smooth out false anomalies caused by instantaneous fluctuations and improve the stability of risk assessment.

[0112] In step S420, based on the spatial positional relationship of the sensors, nuclear diffusion is carried out with each sensor position as the center to construct a spatial energy field with the sensors as field points.

[0113] It should be noted that sensors in the laboratory are usually distributed in multiple areas (such as workbenches, vents, around equipment, near medicine cabinets, etc.), and risk signals from different areas need to be fused in a spatial dimension.

[0114] In some implementations, a spatial energy field is established using a nuclear diffusion method based on the spatial location relationships of the sensors. Specifically, a neighborhood set is constructed centered on the sensor's location coordinates in the laboratory space. The spatial diffusion effect of risk energy is modeled using a spatial Gaussian kernel function:

[0115] Equation (10)

[0116] Spatial kernel function Choose the Gaussian kernel function:

[0117] Equation (11)

[0118] In the formula, Sensor indexes for spatial field sampling points. ; For sensors The spatial neighborhood set centered on, This represents the sensor index within the spatial neighborhood set. Indicates the first Each sensor at time Time-based energy aggregation and They represent the first The spatial coordinates of the first sensor and the first The spatial coordinates of each sensor; It is a Gaussian kernel bandwidth (or spatial diffusion bandwidth), which can be adaptively adjusted according to the laboratory area and sensor spacing to support transferable modeling of spaces of different sizes; For distance Let Gaussian kernel function be the independent variable. For in the sensor The spatial energy field value at the field point.

[0119] By using nuclear diffusion, a smooth energy distribution surface can be constructed in a continuous space, allowing risks to naturally decay in the spatial dimension, rather than simply dividing the space into regions. This enables continuous spatial modeling of laboratory risks, allowing the risk energy from different sensors to form a global distribution pattern. In addition, the smoothing properties of Gaussian kernels effectively avoid abrupt changes in local risk values ​​caused by differences in sensor density.

[0120] In step S430, the spatial energy of all field points is summed to obtain the laboratory risk field energy.

[0121] To comprehensively reflect the overall risk level of the laboratory, the energy field values ​​of all spatial sampling points are summed to obtain the real-time total energy of the laboratory risk field.

[0122] Specifically, the spatial energy fields at all points are summed:

[0123] Equation (12)

[0124] In the formula, Indicates time Laboratory risk field energy.

[0125] In equation (12), the summation process can be viewed as performing a global energy integration over the laboratory space to obtain the total risk energy at the current moment. Through the analysis of... Time series analysis can further monitor the trend of risk energy changes, thereby identifying stages of risk increase, stabilization, or decline. This yields a quantitative index of the overall risk energy of the laboratory, which can be directly used for risk level classification and status determination. For example, when... When the safety threshold is exceeded, the system can automatically issue an alarm and lock the high-energy accumulation area, enabling rapid source tracing and emergency response.

[0126] In this embodiment, a complete laboratory risk field energy modeling mechanism is constructed through time integration, spatial diffusion, and global aggregation. This preserves the risk evolution trend in the time domain, achieves smooth energy diffusion in the spatial domain, and realizes spatiotemporal fusion risk field modeling based on multi-dimensional monitoring signals, enhancing the system's comprehensive perception capability. Furthermore, by using sliding time windows and continuous field modeling, the system can reflect the laboratory's operating status in real time while providing physically meaningful risk intensity indicators.

[0127] Figure 5 A flowchart illustrating an example of determining the laboratory risk status level based on the laboratory risk field energy according to an embodiment of this application is shown.

[0128] like Figure 5As shown, in step S510, a baseline model of laboratory activities is constructed based on historical data of laboratory risk field energy within a preset historical time period.

[0129] In laboratory safety assessment systems, historical risk field energy data can reflect typical energy distribution patterns at different experimental stages. For example, a historical sample set of laboratory risk field energy data collected within a preset historical time period (e.g., the past 30 days or several recent experimental cycles) is compiled. The samples are derived from the fusion results of multiple sensor sources, covering multi-dimensional information such as ambient temperature and humidity, gas concentration, equipment current, voltage, and personnel behavior. Through statistical analysis and normalization, a time-ordered risk energy feature set is formed. This feature set is then input into the laboratory activity baseline model to characterize the typical energy state distribution during experimental operation. This achieves multi-dimensional feature summarization of the laboratory's historical operating status, avoiding the subjective biases arising from traditional threshold settings that rely solely on experience.

[0130] In step S520, the laboratory activity baseline model is divided into several activity clusters using an unsupervised clustering method, with each activity cluster corresponding to a typical experimental runtime segment.

[0131] In some implementations, based on the established baseline model, unsupervised clustering algorithms (such as K-means or Gaussian mixture models) are used to perform cluster analysis on the historical risk field energy. Furthermore, each cluster center represents a typical experimental operation mode, such as: equipment startup phase, stable operation phase, peak reaction phase, or cooling-off phase. The clustering process automatically divides periods with similar risk energy into the same activity cluster based on the risk energy distribution density and fluctuation characteristics within each time window.

[0132] Therefore, it is possible to achieve adaptive classification of experimental states, automatically identify key stages in the experimental process without human intervention, and improve the model's universality and transferability to complex experimental scenarios.

[0133] In step S530, the target activity cluster to which the current activity cluster belongs is determined based on the similarity between the current multi-source sensor state vector and the center vector of each activity cluster.

[0134] In some implementations, during the real-time operation phase, the similarity between the current multi-source sensor state vector (e.g., various environmental parameters, electrical parameters, and behavioral feature vectors) and the center vector of each activity cluster is calculated using the Euclidean distance or cosine similarity. Once the cluster with the highest similarity is determined, the current time is assigned to the corresponding target activity cluster.

[0135] Through this matching operation, the system can quickly identify the current experimental operating mode (e.g., high-energy reaction or routine monitoring phase), achieving automatic recognition of the contextual semantics of the current environment. Compared to static models, the embodiments of this application can flexibly switch judgment criteria according to the risk characteristics of different experimental stages, significantly improving the sensitivity and accuracy of risk identification.

[0136] In step S540, the risk field energy distribution parameters of the target active cluster are calculated. To dynamically set risk level thresholds.

[0137] In some implementations, for a given target activity cluster, the system statistically analyzes the risk field energy distribution parameters in its historical samples, including the mean and standard deviation. Based on these two parameters, the system dynamically generates multi-level risk thresholds.

[0138] Equation (13)

[0139] Equation (14)

[0140] Equation (15)

[0141] In the formula, , and These represent the safety threshold, warning threshold, and danger threshold, respectively. These are preset coefficients; and represents the mean risk energy and standard deviation of the current activity cluster, respectively.

[0142] By introducing statistical distribution parameters to calculate dynamic thresholds, the risk assessment thresholds are adapted to the energy fluctuation levels of the current experimental activities. This effectively solves the problem of high misjudgment rates in traditional fixed threshold models under different experimental types or equipment conditions, and achieves adaptive adjustment of risk level assessment.

[0143] In step S550, the current risk field energy is compared with the dynamically set risk level threshold, and the corresponding laboratory risk status level is output based on the comparison result.

[0144] Here, the laboratory risk status level includes any one of the following: safe status, low-risk status, alert status, and dangerous status. More specifically, the risk field energy value calculated at the current moment is compared with the above dynamic threshold; when the energy value is lower than... When the energy value is between [a certain value], it is considered a safe state; when the energy value is between [a certain value], it is considered a safe state. and When the energy value is between [a certain value], it is considered a low-risk state; when the energy value is between [a certain value], it is considered a low-risk state. and When the energy level is between [a certain value], it is considered to be in a state of alert; when the energy value exceeds [a certain value], it is considered to be in a state of alert. When this occurs, the situation is determined to be dangerous. Preferably, the final determination result is output to the laboratory risk management interface or alarm module.

[0145] Through the embodiments of this application, an active baseline model and clustering mechanism are introduced to achieve adaptive identification of the experimental operation stage; then, using the risk energy distribution parameter, active clusters are obtained by unsupervised clustering and dynamic thresholds are set according to the statistics within the clusters, so that the judgment is adaptively matched with the current operation stage, thereby enhancing the robustness and adaptability of risk level judgment.

[0146] Regarding the details of using laboratory risk field energy for laboratory risk location, in some examples of embodiments of this application, the contribution of each sensor to the total risk field energy is calculated.

[0147] Because of differences in sensor type and placement, their contributions to the overall risk energy are not consistent. Therefore, the time-aggregated energy of each sensor is calculated at the current moment and normalized to the sum of the time-aggregated energy of all sensors to obtain the risk energy contribution.

[0148] Equation (16)

[0149] in, and For sensor indexing, For the first Risk energy contribution of each sensor and They represent the first Each sensor at time Time-based energy aggregation and the first Each sensor at time Time aggregates energy.

[0150] In equation (16), the relative importance of each sensor in the global risk energy is quantified by normalization. For example, if a certain environmental sensor is continuously in a high-risk correlation state, then its The value will be significantly higher than that of other sensors, thus having a dominant influence on the risk in the region in terms of spatial distribution.

[0151] This avoids energy imbalances caused by differences in dimensions or sampling between different sensors. The normalized risk energy contribution makes laboratory risk assessment results more comparable and consistent, ensuring the model's fair response to multi-source heterogeneous data and enabling the influence of each sensor to be reasonably and dynamically quantified.

[0152] A laboratory risk thermal field is constructed based on the spatial coordinates of each sensor.

[0153] Specifically, after obtaining the risk energy contribution of each sensor, a risk thermal field is constructed based on the spatial distribution coordinates of the sensors in the laboratory plane. This process diffuses the risk energy of each sensor to its surrounding area through a two-dimensional Gaussian diffusion function.

[0154] Equation (17)

[0155] in, Represents the laboratory plane coordinates The risk intensity value at the location, Indicates the first The spatial coordinates of each sensor within the laboratory plane; This is the thermal diffusion radius coefficient, used to control the diffusion range of the two-dimensional Gaussian kernel function in space.

[0156] It should be noted that risks in laboratory environments typically exhibit spatial continuity. For example, the risks of chemical gas leaks, high-temperature equipment malfunctions, or personnel gatherings are not confined to a single point but rather exhibit spatial diffusion characteristics. Gaussian kernel function modeling naturally reflects the spatial attenuation effect of risks: areas closer to the sensor are more significantly affected by its energy, while the influence gradually weakens towards farther away areas. Furthermore, the use of a two-dimensional planar model rather than a three-dimensional spatial model is based on the actual situation where most risk monitoring points in the laboratory (environmental, equipment, and personnel behavior sensors) are arranged at similar heights, with minimal vertical differences. Planar modeling can accurately reflect the main spatial distribution characteristics. Thus, the goal of transforming discrete sensor signals into a continuous spatial risk distribution map is achieved, enabling the system to identify risk hotspots globally.

[0157] Display the laboratory's risk thermal field to indicate the main risk areas within the laboratory.

[0158] In some implementations, after constructing the laboratory risk thermal field, the thermal distribution is presented in the form of color gradients or contour maps, allowing administrators to intuitively understand the laboratory risk distribution through a graphical interface. This transforms multi-source sensor data from numerical signals into spatial risk energy distribution, enabling a leap from point-based monitoring of laboratory risks to field-wide understanding. Furthermore, the visualized results support rapid source tracing and regional control, significantly improving the operability and response efficiency of the laboratory safety monitoring system.

[0159] To verify the effectiveness of the method proposed in the embodiments of this application, a simulation-based experiment was designed. Since long-term data recorded in actual laboratories is difficult to obtain, comparative analysis was performed using sensor data from a simulated laboratory environment.

[0160] In terms of simulation data generation, the simulated environment includes four types of sensors: temperature, gas concentration, personnel flow, and humidity. Each type of sensor follows a normal distribution under normal conditions, with its mean and standard deviation being [missing values]. , , (Relative concentration unit) , (person / unit time) , , The sampling time was 1000 steps, simulating one day of operation. Three types of anomalies were introduced into the simulation:

[0161] Gas leak event: A gas concentration of 0.8 is superimposed on the gas sensor in 200-250 steps to simulate a leak accident;

[0162] High temperature and dryness: Increase the temperature to approximately [value missing] in 600-650 steps. Humidity decreased by 20%, and the simulation equipment overheated;

[0163] Crowding: Increase the number of people by 10 at 800-830 steps to simulate overcrowding in the laboratory.

[0164] In addition, a gas sensor malfunction was simulated in steps 500-550, causing its reading to be 0.7 higher than normal, but there was no real danger at this point.

[0165] Regarding the algorithm settings for the comparative experiment, the baseline algorithm uses a linear weighted sum of data from each sensor as the overall risk score:

[0166] Equation (18)

[0167] Equation (19)

[0168] In the formula, Indicates at time , Indicates the first Each sensor at time Standardized observations are used to eliminate differences in units and amplitudes between data from different sensors; Indicates the first Each sensor at time The original measurement value, Indicates the first The mean of each sensor during the baseline phase (e.g., taking the first 100 samples); Indicates the first The standard deviation of measurements taken by each sensor during the baseline phase; , , and These represent the standardized risk characteristics of temperature sensors, gas concentration sensors, electrical load sensors, and personnel behavior sensors, with weights selected empirically. The mean and standard deviation are calculated based on the first 100 samples from the baseline, and warning and danger thresholds are set. An alarm is triggered when the score exceeds the warning threshold.

[0169] Regarding the details of the Spatio-Temporal Energy Fusion Risk Mapping (STEFRM) algorithm proposed in the embodiments of this application, following the aforementioned method, the grey relational degree and risk energy of each sensor are first calculated, then the weighted risk energy is calculated according to the reliability weight function, and the risk field energy is obtained by summing. Thresholds were set using the mean and standard deviation of the baseline segment. The resolution coefficient was used in the experiment. The risk severity parameters for various sensors are as follows: , , , .

[0170] Figure 6 The comparison results of risk scores changing over time under different algorithms in a simulated environment are shown. The horizontal axis represents the time step, and the vertical axis represents the risk score. Among them, the blue curve corresponds to the risk score calculated by the STEFRM algorithm proposed in this paper, and the orange curve corresponds to the risk score of the linearly weighted baseline algorithm. The red shaded area represents the period of actual hazardous events in the laboratory (including gas leaks, high temperature and dryness, and personnel gathering), and the blue shaded area represents the period when the gas sensor experiences signal drift or intermittent failure but there is no actual environmental risk.

[0171] like Figure 6 As shown, the baseline algorithm's risk score increases significantly during the red period when hazardous events occur, reflecting the abnormal trend to some extent. However, during the sensor failure period indicated by the blue shade, the baseline algorithm's score also fluctuates significantly, leading to false alarms. This is because the baseline algorithm uses a fixed-weight linear weighting method (with gas having the highest weight), lacking reliability constraints when a single sensor output is abnormal, thus amplifying the impact of sensor drift on the overall risk.

[0172] In contrast, the STEFRM algorithm proposed in this application exhibits a stable and targeted risk increase trend during periods of hazardous events, accurately corresponding to real risk events; while during periods of sensor failure, the risk score remains at a low level without significant fluctuations. Specifically, STEFRM introduces sensor reliability weights and a spatiotemporal energy aggregation mechanism in risk modeling: on the one hand, it suppresses the interference of anomalous signals by evaluating the historical stability and baseline deviation of the sensor; on the other hand, it enhances multi-source consistency judgment through time integration and spatial diffusion, enabling the algorithm to distinguish between real hazardous events and pseudo-anomaly signals.

[0173] Therefore, the STEFRM algorithm exhibits more robust risk perception and judgment capabilities in multi-source heterogeneous sensing environments, significantly reducing the false alarm rate caused by single-point failures and maintaining high response sensitivity during periods of real risk. Figure 6 The simulation results shown verify the robustness and reliability improvement of the proposed STEFRM algorithm in risk state identification in complex laboratory scenarios.

[0174] Figure 7 The results show the performance comparison of different algorithms under the main risk assessment indicators. The horizontal axis represents different evaluation indicators, including FPR (false positive rate), FNR (false negative rate), and ACC (accuracy rate), and the vertical axis represents the numerical value of the corresponding indicators.

[0175] like Figure 7 As shown, the STEFRM algorithm significantly improves accuracy compared to the baseline algorithm, with the average accuracy increasing from 0.70 to 0.90, and the overall recognition accuracy improving by about 20%. The false alarm rate is reduced from 0.30 to 0.10, a reduction of more than two-thirds. The false negative rate is reduced from 0.25 to 0.15, demonstrating higher event detection sensitivity.

[0176] Specifically, STEFRM introduces a sensor reliability weighting mechanism and a spatiotemporal energy aggregation mechanism in risk modeling. These mechanisms dynamically adjust the contribution weights of each sensor based on historical stability and baseline deviation, thereby suppressing false alarms caused by single-point drift. Furthermore, the time integration and spatial diffusion processes strengthen multi-source consistency judgment, enabling the algorithm to remain sensitive to real-world risk events. Therefore, as... Figure 7 The results verify that the STEFRM algorithm has higher accuracy and robustness than the linear weighted model in a multi-source heterogeneous sensing environment, and can effectively reduce the false alarm rate while ensuring the sensitivity of risk detection.

[0177] For laboratory safety monitoring scenarios, this paper proposes a highly innovative multi-source data fusion assessment method. By introducing grey relational analysis, risk energy, and adaptive reliability weights, a laboratory risk field energy model is constructed, which can comprehensively consider the relationships between various sensor data and their reliability differences. Experimental results show that this method can effectively reduce false alarms caused by sensor failures and provide risk source interpretation capabilities compared to the linearly weighted baseline method.

[0178] Future work can be carried out in the following directions: 1) Collect long-term, multi-source data to train and validate the model; 2) Explore more accurate reliability assessment strategies, such as combining sensor health diagnosis models and cross-validation mechanisms; 3) Combine the proposed risk field model with machine learning methods, and use deep learning to extract more complex features to improve anomaly detection capabilities; 4) Considering data privacy and security, design a system to achieve cross-laboratory data sharing and federated model learning while ensuring that sensitive laboratory information is not leaked.

[0179] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of combined actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Secondly, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application. In the above embodiments, the descriptions of each embodiment have their own emphasis; for parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0180] Figure 8 A structural block diagram of an example of a laboratory safety status assessment system based on multi-source data fusion according to an embodiment of this application is shown.

[0181] like Figure 8 As shown, the laboratory safety status assessment system 800 based on multi-source data fusion includes a data acquisition unit 810, an associated risk analysis unit 820, a sensor reliability analysis unit 830, a sensor risk energy analysis unit 840, and a laboratory risk field energy generation unit 850.

[0182] The data acquisition unit 810 is used to acquire multi-source sensor data corresponding to the laboratory environment based on the sensor module, and to construct a laboratory sensor data matrix with a unified time axis based on the multi-source sensor data.

[0183] The associated risk analysis unit 820 is used to process the laboratory sensor data matrix using grey relational analysis to calculate the risk correlation degree corresponding to each of the sensors; the risk correlation degree is used to indicate the degree of proximity between the current sensor sampling characteristics and the preset hazard mode characteristics.

[0184] The sensor reliability analysis unit 830 is used to acquire the baseline deviation information corresponding to each of the sensors, and determine the corresponding reliability weight based on the historical stability, baseline deviation information and calibration cycle of each sensor; the historical stability is defined as the standard deviation of the data collected by the sensor under historical normal conditions, and the baseline deviation information is defined as the degree of deviation of the current measurement value of the sensor from the baseline mean of the time period.

[0185] The sensor risk energy analysis unit 840 is used to multiply the risk correlation degree of each sensor by the reliability weight and the risk severity coefficient to obtain the weighted risk energy of each sensor; each sensor has a unique corresponding risk severity coefficient.

[0186] The laboratory risk field energy generation unit 850 is used to spatiotemporally couple and superimpose the weighted risk energies of all sensors to generate laboratory risk field energy, and calculate the corresponding laboratory risk status level based on the laboratory risk field energy.

[0187] In some embodiments, this application provides a non-volatile computer-readable storage medium storing one or more programs including execution instructions, which can be read and executed by electronic devices (including but not limited to computers, servers, or network devices) to perform the steps of any of the above-described laboratory safety status assessment methods based on multi-source data fusion.

[0188] In some embodiments, this application also provides a computer program product, the computer program product including a computer program stored on a non-volatile computer-readable storage medium, the computer program including program instructions, which, when executed by a computer, cause the computer to perform the steps of any of the above-described laboratory safety status assessment methods based on multi-source data fusion.

[0189] In some embodiments, this application also provides an electronic device comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform steps of a laboratory safety status assessment method based on multi-source data fusion.

[0190] The above-described product can perform the methods provided in the embodiments of this application, and has the corresponding functional modules and beneficial effects for performing the methods. Technical details not described in detail in this embodiment can be found in the methods provided in the embodiments of this application.

[0191] The electronic devices in this application can exist in various forms, including but not limited to: mobile communication devices, ultra-mobile personal computer devices, portable entertainment devices, or other airborne electronic devices with data interaction functions.

[0192] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0193] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented using software plus a general-purpose hardware platform, or of course, using hardware. Based on this understanding, the above technical solutions, in essence or the parts that contribute to the related technology, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0194] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A laboratory safety status assessment method based on multi-source data fusion, characterized in that, The method includes: The sensor module collects multi-source sensor data corresponding to the laboratory environment, and constructs a laboratory sensor data matrix with a unified time axis based on the multi-source sensor data. The laboratory sensor data matrix is ​​processed using grey relational analysis to calculate the risk correlation degree corresponding to each sensor; the risk correlation degree is used to indicate the degree of proximity between the current sensor sampling characteristics and the preset hazard mode characteristics. The baseline deviation information corresponding to each sensor is obtained, and the corresponding reliability weight is determined based on the historical stability, baseline deviation information and calibration cycle of each sensor. The historical stability is defined as the standard deviation of the data collected by the sensor under historical normal conditions, and the baseline deviation information is defined as the degree of deviation of the current measurement value of the sensor from the baseline mean of the time period. The risk correlation of each sensor is multiplied by the reliability weight and risk severity coefficient to obtain the weighted risk energy of each sensor; each sensor has a unique corresponding risk severity coefficient. The weighted risk energies of all sensors are spatiotemporally coupled and superimposed to generate the laboratory risk field energy, and the corresponding laboratory risk status level is calculated based on the laboratory risk field energy. The step of determining the corresponding reliability weight based on the historical stability, baseline deviation information, and calibration cycle of each sensor includes: Calculate the standardized deviation of each sensor: ， In the formula, For the first The current measurement value of each sensor, For the first The baseline mean of each sensor during its respective time period. For the first The baseline standard deviation of each sensor over its respective time period; For the first The standardized deviation of each sensor represents baseline deviation information; Calculate the historical stability of each sensor: ， In the formula, For the first The basic reliability of a sensor represents its historical stability. For the first The stability and sensitivity coefficient of each sensor For reference stability constants; Calculate the calibration cycle attenuation factor for each sensor: ， ， In the formula, Indicates the current assessment time. For the first The calibration cycle of each sensor, For the first Calibration overdue ratio of individual sensors For the first The most recent calibration completion time for each sensor To calibrate the overdue sensitivity coefficient, Indicates the first Calibration cycle attenuation factor for each sensor; The sensor reliability weights for each sensor are determined by combining the deviation threshold and the modulation coefficient. ， In the formula, For deviation from the threshold, For attenuation modulation coefficients, For the first Reliability weights for each sensor; The step of spatiotemporally coupling and superimposing the weighted risk energies of all sensors to generate the laboratory risk field energy includes: For each sensor in the time window Integrating the weighted risk energy within the time frame yields the single-sensor time aggregated energy: ， ， In the formula, Represents the time integral variable; Indicates the first Each sensor at time Instantaneous weighted risk energy, Indicates the first Each sensor at time The degree of risk correlation, No. Each sensor at time Reliability weights, Indicates the first Risk severity coefficient of each sensor Indicates the first Each sensor at time Time-based energy aggregation; Based on the spatial relationship of the sensors, nuclear diffusion is performed with each sensor location as the center to construct a spatial energy field with the sensors as field points: ， Spatial kernel function Choose the Gaussian kernel function: ， In the formula, Sensor indexes for spatial field sampling points. ; For sensors The spatial neighborhood set centered on, This represents the sensor index within the spatial neighborhood set. Indicates the first Each sensor at time Time-based energy aggregation and They represent the first The spatial coordinates of the first sensor and the first The spatial coordinates of each sensor; For Gaussian kernel bandwidth, For distance Let Gaussian kernel function be the independent variable. For in the sensor The spatial energy field value at the field point; Summing the spatial energies at all field points yields the laboratory risk field energy: ， In the formula, Indicates time Laboratory risk field energy.

2. The method according to claim 1, characterized in that, Before constructing a laboratory sensor data matrix with a unified time axis based on the multi-source sensor data, the method further includes data preprocessing, specifically... Includes the following operations: Time synchronization and interpolation are performed on sensor data with different sampling frequencies; A sliding window filtering algorithm is used to smooth high-frequency sensor data in order to remove transient noise interference; When missing sensor data is detected, statistical reconstruction is performed using observations from adjacent time periods and associated sensors to fill in the missing samples, thereby ensuring the continuity and integrity of the laboratory sensor data matrix under a unified time axis.

3. The method according to claim 1, characterized in that, The process of processing the laboratory sensor data matrix using grey relational analysis to calculate the risk correlation degree corresponding to each of the sensors includes: For each sensor, define a hazard reference feature sequence corresponding to the hazard mode; Calculate the grey correlation coefficient between the current sensor sampling sequence indicated by the laboratory sensor data matrix and the hazard reference feature sequence: ， Calculate the mean of the grey relational coefficients within the time window to obtain the risk correlation degree: ， In the formula, and For sensor indexing, , This represents the total number of sensors in the sensor module. and For discrete-time indexes within a time window, where and , The window length is the time window length. and These are the reference sequences for the dangerous mode at time [time]. and time The value, For the first Each sensor at time The sampled values, For the resolution coefficient, satisfying ; For the first Each sensor at time The sampled values; For the first Each sensor at time The grey relational coefficient; For the first Risk correlation of individual sensors.

4. The method according to claim 1, characterized in that, The calculation of the corresponding laboratory risk status level based on the laboratory risk field energy includes: Based on historical data of laboratory risk field energy within a preset historical time period, a baseline model of laboratory activities is constructed. The laboratory activity baseline model is divided into several activity clusters using an unsupervised clustering method, with each activity cluster corresponding to a typical experimental runtime period. The target activity cluster to which the current activity cluster belongs is determined based on the similarity between the current multi-source sensor state vector and the center vector of each activity cluster. Calculate the risk field energy distribution parameters of the target active cluster. To dynamically set risk level thresholds: ， ， ， In the formula, , and These represent the safety threshold, warning threshold, and danger threshold, respectively. These are preset coefficients; and Let these represent the mean risk energy and standard deviation of the current activity cluster, respectively. The current risk field energy is compared with the dynamically set risk level threshold, and the corresponding laboratory risk status level is output based on the comparison result. The laboratory risk status level includes any one of the following: safe status, low risk status, alert status, and dangerous status.

5. The method according to claim 1, characterized in that, After spatiotemporally coupling and superimposing the weighted risk energies of all sensors to generate the laboratory risk field energy, the method further includes: Calculate the contribution of each sensor to the total risk field energy: ， in, and For sensor indexing, For the first Risk energy contribution of each sensor and They represent the first Each sensor at time Time-based energy aggregation and the first Each sensor at time Time-based energy aggregation; Constructing a laboratory risk thermal field based on the spatial coordinates of each sensor: ， in, Represents the laboratory plane coordinates The risk intensity value at the location, Indicates the first The spatial coordinates of each sensor within the laboratory plane; This is the thermal diffusion radius coefficient, used to control the diffusion range of the two-dimensional Gaussian kernel function in space; The laboratory risk thermal field is displayed to indicate the main risk areas within the laboratory.

6. A laboratory safety status assessment system based on multi-source data fusion, characterized in that, The system is used to implement the method as described in any one of claims 1-5; the system comprises: The data acquisition unit is used to acquire multi-source sensor data corresponding to the laboratory environment based on the sensor module, and to construct a laboratory sensor data matrix under a unified time axis based on the multi-source sensor data. The associated risk analysis unit is used to process the laboratory sensor data matrix using grey relational analysis to calculate the risk correlation degree corresponding to each of the sensors; the risk correlation degree is used to indicate the degree of proximity between the current sensor sampling characteristics and the preset hazard mode characteristics. The sensor reliability analysis unit is used to acquire baseline deviation information corresponding to each of the sensors, and determine the corresponding reliability weight based on the historical stability, baseline deviation information and calibration cycle of each sensor; the historical stability is defined as the standard deviation of the data collected by the sensor under historical normal conditions, and the baseline deviation information is defined as the degree of deviation of the current measurement value of the sensor from the baseline mean of the time period. The sensor risk energy analysis unit is used to multiply the risk correlation degree of each sensor by the reliability weight and the risk severity coefficient to obtain the weighted risk energy of each sensor; each sensor has a unique corresponding risk severity coefficient. The laboratory risk field energy generation unit is used to spatiotemporally couple and superimpose the weighted risk energies of all sensors to generate laboratory risk field energy, and calculate the corresponding laboratory risk status level based on the laboratory risk field energy.

Images